Methods and materials for assessing homologous recombination deficiencies in breast cancer subtypes

Abstract

Provided herein are methods and materials involved in assessing a sample (e.g., cancer cells) for the presence of homologous recombination deficiency (HRD) or HRD signature. For example, provided are methods and materials for determining whether a cell (e.g., cancer cell) contains an HRD signature. Also provided are materials and methods for identifying cells (e.g., cancer cells) with homology directed repair (HDR) deficiency, as well as materials and methods for identifying cancer patients who are likely to respond to a particular cancer treatment regimen.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit under 35 U.S.C. §119(e) of U.S. Provisional Patent Application No. 63 / 287,374, filed December 8, 2021, the contents of which are incorporated herein by reference in their entirety. [Background technology]

[0002] Cancer is a serious public health problem, with 562,340 cancer deaths in the United States in 2009 alone. American Cancer Society, Cancer Facts & Figures 2009 (available on the American Cancer Society website). One of the major challenges in cancer treatment is discovering relevant, clinically useful characteristics of a patient's cancer and, based on these characteristics, administering an optimal treatment plan for the patient's cancer. Although progress has been made in this area of ​​personalized medicine, there remains a significant need for better molecular diagnostic tools to characterize a patient's cancer. Summary of the Invention

[0003] The present specification relates to methods and materials involved in evaluating a sample (e.g., a cancer cell or nucleic acid derived therefrom) for homologous recombination deficiency (HRD) (e.g., HRD signature) based on the detection of specific chromosomal abnormalities ("CA"). For example, the present specification provides methods and materials for detecting CA regions to determine whether a cell (e.g., a cancer cell) has HRD (e.g., exhibits an HRD signature). The present specification also provides materials and methods for identifying cancer patients who are likely to respond to a particular cancer treatment regimen based on the presence, absence, or severity of HRD. Throughout the specification, unless otherwise indicated, HRD and homology-dependent repair (HDR) deficiency are used interchangeably.

[0004] In general, one aspect of the invention features a method for assessing HRD in cancer cells or DNA (e.g., genomic DNA) derived therefrom. In some embodiments, the method comprises or consists essentially of (a) detecting CA regions (as defined herein) in a sample or DNA derived therefrom in at least one pair of human chromosomes (e.g., any pair of human chromosomes other than the human X / Y sex chromosome pair) of the sample or DNA derived therefrom, and (b) determining the number, size (e.g., length), and / or characteristics of the CA regions. In some embodiments, the CA regions are analyzed in several chromosome pairs representative of the entire genome (e.g., sufficient chromosomes are analyzed such that the number and size of CA regions are expected to represent the number and size of CA regions across the genome).

[0005] Various aspects of the present invention include the use of combined analysis of two or more CA regions to assess (e.g., detect) HRD in a sample. Three types of CA regions useful in such methods include (1) chromosomal regions that exhibit loss of heterozygosity ("LOH regions" as defined herein), (2) chromosomal regions that exhibit telomeric allelic imbalance ("TAI regions" as defined herein), and (3) chromosomal regions that exhibit large-scale transitions ("LST regions" as defined herein). CA regions of certain sizes, chromosomal locations, or characteristics (e.g., "indicator CA regions" as defined herein) may be particularly useful in various aspects of the present invention described herein.

[0006] Thus, in one aspect, the invention provides a method of assessing (e.g., detecting) HRD in a sample, comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); and (3) assessing HRD in the sample based at least in part on the determinations made in (1) and (2). In another aspect, the invention provides a method of assessing (e.g., detecting) HRD in a sample, comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); and (3) assessing HRD in the sample based at least in part on the determinations made in (1) and (2). In another aspect, the present invention provides a method of assessing (e.g., detecting) HRD in a sample, the method comprising: (1) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (2) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); and (3) assessing HRD in the sample based at least in part on the determinations made in (1) and (2).In another aspect, the present invention provides a method of assessing (e.g., detecting) HRD in a sample, the method comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (3) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); and (4) assessing (e.g., detecting) HRD in the sample based at least in part on the determinations made in (1), (2), and (3).

[0007] In one aspect, the invention provides a method of diagnosing the presence or absence of an HRD in a patient sample, the method comprising: (1) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) a total number of LOH regions of a certain size or characteristic in the samples (e.g., "indicator LOH regions" as defined herein); (2) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) a total number of TAI regions of a certain size or characteristic in the samples (e.g., "indicator TAI regions" as defined herein); and (3) either (a) diagnosing the presence of an HRD in the patient sample if the number from (1) and / or the number from (2) exceeds a certain reference, or (3)(b) diagnosing the absence of an HRD in the patient sample if neither the number from (1) nor the number from (2) exceeds a certain reference. In another aspect, the invention provides methods of diagnosing the presence or absence of an HRD in a patient sample, the method comprising: (1) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) a total number of LOH regions of a certain size or characteristic in the samples (e.g., "indicator LOH regions" as defined herein); (2) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) a total number of LST regions of a certain size or characteristic in the samples (e.g., "indicator LST regions" as defined herein); and (3) either (a) diagnosing the presence of an HRD in the patient sample if the number from (1) and / or the number from (2) exceeds some reference, or (3)(b) diagnosing the absence of an HRD in the patient sample if neither the number from (1) nor the number from (2) exceeds some reference.In another aspect, the invention provides methods of diagnosing the presence or absence of an HRD in a patient sample, the method comprising: (1) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) the total number of TAI regions of a certain size or characteristic in the samples (e.g., "indicator TAI regions" as defined herein); (2) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) the total number of LST regions of a certain size or characteristic in the samples (e.g., "indicator LST regions" as defined herein); and (3) either (a) diagnosing the presence of an HRD in the patient sample if the number from (1) and / or the number from (2) exceeds a certain reference, or (3)(b) diagnosing the absence of an HRD in the patient sample if neither the number from (1) nor the number from (2) exceeds a certain reference. In another aspect, the invention provides a method of diagnosing the presence or absence of HRD in a patient sample, the method comprising: (1) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) the total number of LOH regions of a certain size or characteristic in the samples (e.g., "indicator LOH regions" as defined herein); (2) analyzing (e.g., assaying) one or more patient samples to determine (e.g., detect) the total number of TAI regions of a certain size or characteristic in the samples (e.g., "indicator TAI regions" as defined herein); and (3) detecting one or more TAI regions of a certain size or characteristic in the samples. (3)(a)(1) is a sequence that identifies a patient sample that is a subset of LST regions of a certain size or characteristic (e.g., "indicator LST regions" as defined herein), and (b) is a sequence that identifies a patient sample that is a subset of LST regions of a certain size or characteristic (e.g., "indicator LST regions" as defined herein).

[0008] Various embodiments of the present invention include using the average (e.g., arithmetic mean) of three CA regions to assess (e.g., detect) HRD in a sample. Three CA regions useful in such methods include (1) chromosomal regions showing loss of heterozygosity ("LOH regions" as defined herein), (2) chromosomal regions showing telomeric allelic imbalance ("TAI regions" as defined herein), and (3) chromosomal regions showing large-scale transitions ("LST regions" as defined herein). CA regions of certain sizes or characteristics (e.g., "indicator CA regions" as defined herein) may be particularly useful in various embodiments of the present invention described herein. Thus, in one aspect, the invention provides a method of assessing (e.g., detecting) HRD in a sample, the method comprising: (1) determining a total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining a total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (3) determining a total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); (4) calculating an average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) assessing HRD in the sample based at least in part on the calculated average (e.g., arithmetic mean) made in (4).

[0009] In some embodiments, assessing (e.g., detecting) HRD is based on a score ("CA region score" as defined herein) derived or calculated from (e.g., representing or corresponding to) the detected CA region. Scores are described in more detail herein. In some embodiments, HRD is detected if the CA region score of the sample exceeds some threshold (e.g., a reference or index CA region score), and optionally, HRD is not detected if the CA region score of the sample does not exceed some threshold (e.g., a reference or index CA region score, which in some embodiments may be the same threshold for positive detection). Those skilled in the art will readily understand that scores can be devised in the opposite direction within the present disclosure (e.g., HRD is detected if the CA region score is below a certain threshold and not detected if the score is above a certain threshold).

[0010] In some embodiments, the CA region score is a combination of scores derived or calculated from (e.g., representative of or corresponding to) two or more of: (1) the detected LOH regions ("LOH region score" as defined herein), (2) the detected TAI regions ("TAI region score" as defined herein), and / or (3) the detected LST regions ("LST region score" as defined herein). In some embodiments, the LOH region score and the TAI region score are combined to obtain the CA region score as follows: CA region score = A*(LOH region score) + B*(TAI region score)

[0011] In some embodiments, the LOH region score and the TAI region score are combined to obtain a CA region score as follows: CA region score = 0.32 * (LOH region score) + 0.68 * (TAI region score)

[0012] In some embodiments, the LOH and LST region scores are combined to obtain a CA region score as follows: CA region score = A*(LOH region score) + B*(LST region score)

[0013] In some embodiments, the TAI and LST domain scores are combined to obtain a CA domain score as follows: CA domain score = A*(TAI domain score) + B*(LST domain score)

[0014] In some embodiments, the LOH region score, the TAI region score, and the LST region score are combined to obtain a CA region score as follows: CA region score = A*(LOH region score) + B*(TAI region score) + C*(LST region score)

[0015] In some embodiments, the LOH region score, the TAI region score, and the LST region score are combined to obtain a CA region score as follows: CA region score = 0.21 * (LOH region score) + 0.67 * (TAI region score) + 0.12 * (LST region score)

[0016] In some embodiments, the CA region score is a combination of scores derived or calculated from (e.g., representative of or corresponding to) the average (e.g., arithmetic mean) of (1) the detected LOH regions ("LOH region score" as defined herein), (2) the detected TAI regions ("TAI region score" as defined herein), and / or (3) the detected LST regions ("LST region score" as defined herein) to obtain the CA region score. CA region score = A*(LOH region score) + B*(TAI region score) + C*(LST region score) 3

[0017] In another aspect, the present invention provides a method for predicting the status of BRCA1 and BRCA2 genes in a sample. Such methods are similar to the methods described above, except that the determination of CA regions, LOH regions, TAI regions, LST regions, or scores incorporating these are used to assess (e.g., detect) BRCA1 and / or BRCA2 defects in a sample. In another aspect, the present invention provides a method for predicting the response of a cancer patient to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor. Such methods are similar to the methods described above, except that the determination of CA regions, LOH regions, TAI regions, LST regions, or scores incorporating these are used to predict the likelihood that a cancer patient will respond to a cancer treatment regimen. In some embodiments, the patient is a treatment-naive patient. In another aspect, the present invention provides a method for treating cancer. Such methods are similar to the methods described above, differing in that a particular treatment regimen is administered (recommended, prescribed, etc.) based at least in part on the determination of CA regions, LOH regions, TAI regions, LST regions, or a score incorporating the same. In another aspect, the present invention features the use of one or more drugs selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors in the manufacture of a medicament useful for treating cancer in a patient identified as having (or having had) cancer cells determined to have HRD (e.g., HRD signature) as described herein. In another aspect, the present invention features a method for evaluating a sample for the presence of a mutation in a gene from an HDR pathway. Such methods are similar to the methods described above, differing in that the determination of CA regions, LOH regions, TAI regions, LST regions, or a score incorporating the same is used to detect (or not detect) the presence of a mutation in a gene from an HDR pathway.

[0018] In another aspect, the present invention provides a method for evaluating a patient, the method comprising or consisting essentially of: (a) determining whether the patient has (or has had) cancer cells with CA regions (e.g., CA region scores above a reference CA region score) above a reference number; and (b)(1) diagnosing the patient as having cancer cells with HRD if it is determined that the patient has (or has had) cancer cells with CA regions (e.g., CA region scores above a reference CA region score) above a reference number, or (b)(2) diagnosing the patient as not having cancer cells with HRD if it is determined that the patient does not have (or has never had) cancer cells with CA regions above a reference number (e.g., the patient does not have (or has never had) cancer cells with CA regions above a reference number).

[0019] In another aspect, the invention features the use of a plurality of oligonucleotides capable of hybridizing to a plurality of polymorphic regions of human genomic DNA in the manufacture of a diagnostic kit useful for determining the total number or combined length of CA regions in at least a chromosome pair (or DNA derived therefrom) in a sample obtained from a cancer patient, and for detecting (a) HRD or the likelihood of HRD in the sample (e.g., an HRD signature), (b) a defect (or the likelihood of a defect) in the BRCA1 or BRCA2 gene in the sample, or (c) an increased likelihood that a cancer patient will respond to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor.

[0020] In another aspect, the invention features a system for detecting HRD (e.g., HRD signature) in a sample. The system comprises or consists essentially of (a) a sample analyzer configured to generate a plurality of signals for genomic DNA (or DNA derived therefrom) of at least one pair of human chromosomes in the sample, and (b) a computer subsystem programmed to calculate the number or combined length of CA regions in at least one pair of human chromosomes based on the plurality of signals. The computer subsystem can be programmed to compare the number or combined length of CA regions to a reference number to detect (a) HRD or a likelihood of HRD (e.g., HRD signature) in the sample, (b) a defect (or a likelihood of a defect) in the BRCA1 or BRCA2 gene in the sample, or (c) an increased likelihood that a cancer patient will respond to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor. The system can comprise an output module configured to display (a), (b), or (c). The system may include an output module configured to display a recommendation for use of a cancer treatment regimen.

[0021] In another aspect, the present invention provides a computer program product embodied in a computer readable medium that, when executed on a computer, provides instructions for detecting the presence or absence of any CA region (wherein the CA region is optionally an indicator CA region) along one or more of human chromosomes other than human X and Y sex chromosomes, and for determining the total number or combined length of CA regions in one or more chromosome pairs. The computer program product may include other instructions.

[0022] In another aspect, the present invention provides a diagnostic kit. The kit comprises or consists essentially of at least 500 oligonucleotides that can hybridize to multiple polymorphic regions of human genomic DNA (or DNA derived therefrom) and a computer program product provided herein. The computer program product can be embodied in a computer readable medium that, when executed on a computer, provides instructions for detecting the presence or absence of any CA region (wherein the CA region is optionally an indicator CA region) along one or more of human chromosomes other than human X and Y sex chromosomes, and for determining the total number or combined length of CA regions in one or more chromosome pairs. The computer program product can include other instructions.

[0023] In some embodiments of any one or more of the aspects of the invention described in the previous paragraph, any one or more of the following may be appropriately applied. The CA region may be determined in at least 2, 5, 10, or 21 pairs of human chromosomes. The cancer cells may be ovarian cancer cells, breast cancer cells, lung cancer cells, or esophageal cancer cells. The reference may be 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, or 20 or more. The at least one pair of human chromosomes may exclude human chromosome 17. The DNA damaging agent can be cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline can be epirubincin or doxorubicin, the topoisomerase I inhibitor can be campothecin, topotecan, or irinotecan, or the PARP inhibitor can be iniparib, olaparib, or velapirib. The patient can be a treatment-naive patient.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention pertains. The present invention can be carried out using methods and materials similar or equivalent to those described herein, but suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will take precedence. In addition, the materials, methods, and examples are merely illustrative and are not intended to be limiting.

[0025] The details of one or more embodiments of the invention are set forth in the following description and the accompanying drawings. The materials, methods, and examples are illustrative only and are not intended to be limiting. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims. [Brief description of the drawings]

[0026] [Figure 1] Graphs are shown plotting allele dosages of breast cancer cells from fresh frozen samples from breast cancer patients along chromosomes as determined using SNP arrays (top) and high-throughput sequencing (bottom). [Diagram 2] Graphs are shown plotting the allele dosage of breast cancer cells from FFPE samples from breast cancer patients along chromosomes, as determined using SNP arrays (top) and high-throughput sequencing (bottom). [Diagram 3] 1 is a flow chart of an exemplary process for evaluating the genome of a cell (e.g., a cancer cell) for an HRD signature. [Figure 4] 1A-1C are diagrams of example computing devices and mobile computing devices that can be used to implement the techniques described herein. [Figure 5A] Shown are LOH region scores across breast cancer IHC subtypes. The top three panels are BRCA1 / 2-deficient samples. The bottom panel is BRCA1 / 2-intact samples. [Figure 5B] TAI region scores across breast cancer IHC subtypes are shown. The top three panels are BRCA1 / 2-deficient samples. The bottom panel is BRCA1 / 2-intact samples. [Figure 6] Correlation between LOH region score and TAI region score is shown. Correlation coefficient=0.69. X-axis: LOH score, Y-axis: TAI score, red dots: intact samples, blue dots (with superimposed "X"): BRCA1 / 2 deletion samples. The area under the dots is proportional to the number of samples with the combination of LOH score and TAI score. p=10-39. [Figure 7A] Figure 1 shows the LOH locus scores of patients analyzed in Example 2 herein. The top three panels are BRCA1 / 2-deficient samples. The bottom panel is a BRCA1 / 2-intact sample. [Figure 7B] Figure 1 shows the TAI region scores of patients analyzed in Example 2 herein. The top three panels are BRCA1 / 2-deficient samples. The bottom panel is a BRCA1 / 2-intact sample. [Figure 7C] Figure 1 shows the LST region scores of patients analyzed in Example 2 herein. The top three panels are BRCA1 / 2 defective samples. The bottom panel is BRCA1 / 2 intact samples. [Figure 7D] Shows LOH vs. TAI for patients analyzed in Example 2 herein. X-axis: LOH score, Y-axis: TAI score, red dots: intact samples, blue dots (with overlaid "X"): BRCA1 / 2-deficient samples. The area under the dots is proportional to the number of samples with the combination of LOH and TAI scores. [Figure 7E] Shows LOH vs. LST for patients analyzed in Example 2 herein. X-axis: LOH score, Y-axis: LST score, red dots: intact samples, blue dots (with superimposed "X"): BRCA1 / 2-deficient samples. The area under the dots is proportional to the number of samples with the combination of LOH and LST scores. [Figure 7F]Shows TAI vs. LST for patients analyzed in Example 2 herein. X-axis: TAI score, Y-axis: LST score, red dots: intact samples, blue dots (with superimposed "X"): BRCA1 / 2-deficient samples. The area under the dots is proportional to the number of samples with the combination of TAI and LST scores. [Figure 8] 1 is a graph plotting the number of LOH regions longer than 15 Mb and shorter than an entire chromosome for ovarian cancer cell samples with somatic BRCA mutations, germline BRCA mutations, low BRCA1 expression, or intact BRCA (BRCA normal). The size of the circle is proportional to the number of samples with such a number of LOH regions. [Figure 9A] Illustrates HRD-LOH scores in BRCA1 / 2 null (mutated or methylated) samples (upper panel) and intact samples (lower panel) in the all-participant breast cohort. [Figure 9B] Illustrates HRD-TAI scores in BRCA1 / 2 null (mutated or methylated) samples (upper panel) and intact samples (lower panel) in the all-participant breast cohort. [Figure 9C] Illustrates HRD-LST scores in BRCA1 / 2 null (mutated or methylated) samples (upper panel) and intact samples (lower panel) in the all-participant breast cohort. [Figure 10] Illustrates mean (eg, arithmetic mean) HRD combined score (Y-axis) stratified by Miller-Payne score (horizontal axis) in the combined cisplatin-1 and cisplatin-2 cohorts. [Figure 11] FIG. 1 illustrates the Spearman correlation of three different measures of HR deficit. The panels above the diagonal show the correlations. The diagonal panels show density plots. [Figure 12] Illustrates the association of clinical variables with the HRD combined score. [Figure 13]Illustrates the association of clinical variables with BRCA1 / 2 loss. The top and bottom left panels show the proportion of BRCA1 / 2 loss patients within each category of grade, stage, and type of breast cancer. The width of each bar is proportional to the number of patients in each category. The bottom right panel shows the conditional density estimates of BRCA1 / 2 loss according to age. [Figure 14] Illustrates the determination of high HRD with a reference score of 42 or greater. [Figure 15] Illustrates a histogram showing the distribution of HRD scores in the cisplatin cohort: the four columns on the left represent low HRD, and the five columns on the right, with a reference score above 42, represent high HRD. [Figure 16] Figure 1 illustrates the distribution of HRD scores within the response classes of pCR, RCB-I, RCB-II, and RCB-III. Boxes represent the interquartile range (IQR) of scores with a horizontal line at the median. The dotted line at 42 represents the HRD threshold between low and high scores. [Figure 17] Illustrates the response curve of the quantitative HRD score. The curve is modeled by generalized logistic regression. The shaded boxes indicate the probability of response in HR-deficient versus non-deficient samples. [Figure 18] Illustrates the HRD scores of the individual HRD components (LOH, TAI, and LST). [Figure 19] 1 is a graph showing a comparison of the distribution of BRCA1 deletion samples, according to certain exemplary embodiments. [Figure 20] 1 is a graph illustrating ER+BC thresholds, according to certain exemplary embodiments. [Figure 21A] 21A-B are graphs illustrating thresholds applied to TNBC and ER+BC, according to certain example embodiments. [Figure 21B] This is a continuation of Figure 21A. [Figure 22]A-B show the distribution of genomic instability scores (GIS) by cancer type and BRCA status. (A) Distribution of GIS for BRCA-deficient and BRCAwt tumors in ovarian cancer, TNBC, and ER+ breast cancer. (B) Distribution of GIS for BRCA-deficient tumors fitted to a normal distribution for ovarian cancer, TNBC, and ER+ breast cancer. [Diagram 23] A-B show the distribution of genomic instability scores (GIS) by pathological complete response (pCR) status in triple-negative breast cancer (TNBC). Distribution of GIS for (A) the full clinical validation cohort and (B) the BRCAwt clinical validation cohort. Samples are stratified based on whether pCR was achieved ('pCR' vs 'no pCR'). [Figure 24] Figure 1 shows the probability of pathological complete response (pCR) by genomic instability score (GIS) in triple negative breast cancer (TNBC). Likelihood of pCR versus range of GIS from a three-parameter logistic regression model fitted to the full clinical validation cohort (N=211, solid line) and the BRCAwt clinical validation cohort (N=171, dashed line). Vertical grey dashed lines represent potential thresholds of 33 or higher and 42 or higher. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] In general, one aspect of the invention features a method for assessing HRD in cancer cells or DNA (e.g., genomic DNA) derived therefrom. In some embodiments, the method includes, or consists essentially of, (a) detecting CA regions in at least one pair of human chromosomes or DNA derived therefrom in a sample or DNA derived therefrom, and (b) determining the number, size (e.g., length), and / or characteristics of the CA regions.

[0028] As used herein, "chromosomal abnormality" or "CA" refers to a somatic change in the chromosomal DNA of a cell, which is classified into at least one of three overlapping categories: LOH, TAI, or LST. Polymorphic loci (e.g., single nucleotide polymorphisms (SNPs)) in the human genome are generally heterozygous in an individual's germline because an individual typically receives one copy from the biological father and one copy from the biological mother. However, somatically, this heterozygosity can change (through mutation) to homozygosity. This change from heterozygosity to homozygosity is called loss of heterozygosity (LOH). LOH can result from several mechanisms. For example, in some cases, a locus on one chromosome can be deleted in somatic cells. A locus that remains present on the other chromosome (the other non-sex chromosome in the case of males) is an LOH locus because there is only one copy (rather than two copies) of that locus present in the genome of the affected cell. This type of LOH event results in a reduction in copy number. In other cases, a locus of one chromosome in a somatic cell (e.g., one non-sex chromosome in the case of a male) may be replaced with a copy of that locus from the other chromosome, thereby eliminating any heterozygosity that may exist within the replaced locus. In such cases, the locus that remains present on each chromosome is the LOH locus, and may be referred to as a copy-neutral LOH locus. LOH and its use in determining HRD are described in detail in International Patent Application No. PCT / US2011 / 040953 (published as WO / 2011 / 160063), the entire contents of which are incorporated herein by reference.

[0029] A broader class of chromosomal abnormalities that encompasses LOH is allelic imbalance. Allelic imbalance occurs when the relative copy number (i.e., copy ratio) at a particular locus in somatic cells differs from that of the germline. For example, if the germline has one copy of allele A and one copy of allele B at a particular locus, and the somatic cell has two copies of A and one copy of B, there is an allelic imbalance at the locus because the copy ratio (2:1) in the somatic cell is different from that of the germline (1:1). LOH is an example of allelic imbalance because the somatic cell has a different copy ratio (1:0 or 2:0) than the germline (1:1). However, allelic imbalance encompasses many more types of chromosomal abnormalities, such as 2:1 germline becoming 1:1 somatic, 1:0 germline becoming 1:1 somatic, 1:1 germline becoming 2:1 somatic, etc. Analysis of regions of allelic imbalance that encompass the telomeres of chromosomes is particularly useful in the present invention. Thus, a "telomeric allelic imbalance region" or "TAI region" is defined as a region having allelic imbalance that (a) extends to one of the subtelomeres and (b) does not cross the centromere. TAI and its use in determining HRD are described in detail in U.S. Patent Application Nos. 13 / 818,425 (published as US2013 / 0281312A1) and 14 / 466,208 (published as US2015 / 0038340A1), the entire contents of each of which are incorporated herein by reference.

[0030] An even broader class of chromosomal abnormalities, including LOH and TAI, is referred to herein as large scale transitions ("LSTs"). LSTs refer to any somatic copy number transitions (i.e., breakpoints) along the length of a chromosome that lie between two regions of at least some minimum length (e.g., at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 megabases or more) after filtering out regions shorter than some maximum length (e.g., 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4 megabases or more). For example, if after filtering regions shorter than 3 megabases, a somatic cell has, for example, a copy number of 1:1 for at least 10 megabases, and then has a breakpoint transition to a region of at least 10 megabases with, for example, a copy number of 2:2, then it is an LST. An alternative way of defining the same phenomenon as an LST region is a genomic region that has stable copy number over at least some minimum length (e.g., at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 megabases) bounded by a breakpoint (i.e., transition), and also has a change in copy number for another region of at least this minimum length. For example, after filtering regions shorter than 3 megabases, if a somatic cell has a region of at least 10 megabases with a copy number of 1:1, bounded on one side by a breakpoint transition to a region of at least 10 megabases with a copy number of, say, 2:2, and on the other side by a breakpoint transition to a region of at least 10 megabases with a copy number of, say, 1:2, then this is two LSTs. Note that this is more extensive than allelic imbalance, since such copy number changes are not considered allelic imbalance (because the copy ratios of 1:1 and 2:2 are the same, i.e., there is no change in the copy ratio).LST and its use in determining HRD are described in detail in U.S. Patent Application No. 14 / 402,254 (published as US2015 / 0140122A1), the entire contents of which are incorporated herein by reference.

[0031] Different cutoffs of LST score may be used for "near diploid" and "near tetraploid" tumors to separate BRCA1 / 2 intact and defective samples. LST score may increase with ploidy in both intact and defective samples. As an alternative to using ploidy specific cutoffs, some embodiments may use a modified LST score that adjusts it by ploidy: LSTm=LST-kP, where P is ploidy and k is a constant. Based on multivariate logistic regression analysis with defective as outcome and LST and P as predictors, k=15.5 gives the best separation between intact and defective samples (although other values ​​of k can be envisaged by those skilled in the art).

[0032] Chromosomal abnormalities may be spread across multiple loci to define chromosomal abnormality regions, referred to herein as "CA regions". Such CA regions may be of any length (e.g., from less than about 1.5 Mb long to as long as the entire length of the chromosome). The abundance of large CA regions ("indicator CA regions") indicates a deficiency in the cellular homology-dependent repair (HDR) mechanism. The definition of a CA region, and therefore what constitutes an "indicator" region for each type of CA (e.g., LOH, TAI, LST), depends on the specific characteristics of the CA. For example, a "LOH region" refers to at least some minimal number of contiguous loci that exhibit LOH, or some minimal stretch of genomic DNA with contiguous loci that exhibit LOH. On the other hand, a "TAI region" refers to at least some minimal number of contiguous loci that exhibit allelic imbalance extending from the telomere to the remainder of the chromosome (or some minimal stretch of genomic DNA with contiguous loci that exhibit allelic imbalance). Because LSTs have already been defined in terms of regions of genomic DNA of at least some minimal size, "LST" and "LST region" are used interchangeably herein to refer to the minimum number of contiguous loci (or some minimal stretches of genomic DNA) that have the same copy number or a transition from that copy number to a different copy number bounded by breakpoints.

[0033] In some embodiments, a CA region (either an LOH region, a TAI region, or an LST region) is an indicator CA region (either an indicator LOH region, an indicator TAI region, or an indicator LST region) if it is at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, or 100 megabases or more in length. In some embodiments, an indicator LOH region is an LOH region that is greater than about 1.5, 5, 12, 13, 14, 15, 16, 17 megabases or more (preferably 14, 15, 16 megabases or more, more preferably 15 megabases or more) in length, but less than the full length of the respective chromosome in which the LOH region is located. Alternatively or additionally, the total combined length of such indicator LOH regions may be determined. In some embodiments, the indicator TAI region is a TAI region with allelic imbalance that (a) extends to one of the subtelomeres, (b) does not cross the centromere, and (c) is longer than 1.5, 5, 12, 13, 14, 15, 16, 17 megabases or more (preferably 10, 11, 12 megabases or more, more preferably 11 megabases or more). Alternatively or additionally, the total combined length of such indicator TAI regions can be determined. Since the concept of LST already includes a region with some minimum size (such minimum size is determined based on its ability to distinguish HRD from HDR intact samples), the indicator LST region as used herein is the same as the LST region. Furthermore, the LST region score can be derived from either the number of regions exhibiting LST as described above or the number of LST breakpoints.In some embodiments, the minimum length of the stable copy number region bounded by the LST breakpoints is at least 3, 4, 5, 6, 7, 8, 9, 10, 11 12, 13, 14, 15, 16, 17, 18, 19, or 20 megabases (preferably 8, 9, 10, 11 megabases or more, more preferably 10 megabases) and the maximum region that remains unfiltered is less than 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4 megabases or less (preferably 2, 2.5, 3, 3.5, or 4 megabases or less, more preferably less than 3 megabases).

[0034] As used herein, if such a sample has a number of indicator CA regions (as described herein) or a CA region score (as described herein) that exceeds a reference as described herein, then the sample has an "HRD signature", with a number or score exceeding such a reference being indicative of a homologous recombination deficiency.

[0035] Thus, the invention generally involves detecting and quantifying indicator CA regions in a sample to determine whether cells in the sample (or cells from which DNA in the sample is derived) have an HRD signature. Often this involves comparing the number of indicator CA regions (or a test value or score derived or calculated therefrom and corresponding to such a number) to a reference or index number (or score).

[0036] Various aspects of the invention include using a combined analysis of two or more CA regions (including two or more indicator CA regions) to assess (e.g., detect, diagnose) HRD in a sample. Thus, in one aspect, the invention provides a method of assessing (e.g., detect, diagnose) HRD in a sample, comprising: (1) determining the total number (or combined length) of indicator LOH regions in the sample; (2) determining the total number (or combined length) of indicator TAI regions in the sample; and (3) determining (e.g., detecting, diagnosing) the presence or absence of HRD in the sample based at least in part on the determinations made in (1) and (2). In another aspect, the invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining the total number (or combined length) of indicator LOH regions in the sample, (2) determining the total number (or combined length) of indicator LST regions in the sample, and (3) determining the presence or absence (e.g., detecting, diagnosing) of HRD in the sample based at least in part on the determinations made in (1) and (2). In another aspect, the invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining the total number (or combined length) of indicator TAI regions in the sample, (2) determining the total number (or combined length) of indicator LST regions in the sample, and (3) determining the presence or absence (e.g., detecting, diagnosing) of HRD in the sample based at least in part on the determinations made in (1) and (2).In another aspect, the invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, the method comprising: (1) determining the total number (or combined length) of indicator LOH regions in the sample; (2) determining the total number (or combined length) of indicator TAI regions in the sample; (3) determining the total number (or combined length) of indicator LST regions in the sample; and (4) determining (e.g., detecting, diagnosing) the presence or absence of HRD in the sample based at least in part on the determinations made in (1), (2), and (3).

[0037] Various aspects of the invention include assessing (e.g., detecting, diagnosing) HRD in a sample using a combined analysis of the average of three different CA regions. Thus, in one aspect, the invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (3) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); (4) calculating the average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) assessing HRD in the sample based at least in part on the calculated average (e.g., arithmetic mean) made in (4).

[0038] As used herein, "CA region score" refers to a test value or score derived or calculated (e.g., representative of or corresponding to) the indicator CA regions detected in a sample (e.g., a score or test value derived or calculated from the number of indicator CA regions detected in a sample). Similarly, as used herein, "LOH region score" refers to a test value or score that is a subset of the CA region score and is derived or calculated (e.g., representative of or corresponding to) the indicator LOH regions detected in a sample (e.g., a score or test value derived or calculated from the number of indicator LOH regions detected in a sample), and similarly for the TAI region score and the LST region score. Such a score may, in some embodiments, simply be the number of indicator CA regions detected in a sample. In some embodiments, the score is more complex and takes into account the length of each indicator CA region or a subset of indicator CA regions detected.

[0039] As discussed above, the present invention generally involves combining the analysis of two or more CA region scores (which may include the number of such regions). Thus, in one aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining the LOH region score of the sample; (2) determining the TAI region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on either the LOH region score above the reference or the TAI region score above the reference, or optionally (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on both the LOH region score not above the reference and the TAI region score not above the reference. In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining an LOH region score of the sample; (2) determining an LST region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on either the LOH region score above the reference or the LST region score above the reference, or optionally (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on both the LOH region score not above the reference and the LST region score not above the reference. In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining a TAI region score of the sample; (2) determining an LST region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on either a TAI region score above a reference or an LST region score above a reference, or optionally, (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on both a TAI region score not above the reference and an LST region score not above the reference.In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining an LOH region score of the sample; (2) determining a TAI region score of the sample; (3) determining an LST region score of the sample; and (4)(a) detecting (or diagnosing) HRD in the sample based at least in part on either the LOH region score above the reference, the TAI region score above the reference, or the LST region score above the reference, or optionally (4)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on the LOH region score not above the reference, the TAI region score not above the reference, and the LST region score not above the reference.

[0040] In some embodiments, the CA region score is a combination of scores derived or calculated from (e.g., representative of or corresponding to) two or more of: (1) the detected LOH regions ("LOH region score" as defined herein), (2) the detected TAI regions ("TAI region score" as defined herein), and / or (3) the detected LST regions ("LST region score" as defined herein). In some embodiments, the LOH region score and the TAI region score are combined to obtain the CA region score as follows: CA region score = A*(LOH region score) + B*(TAI region score)

[0041] In some embodiments, the LOH region score and the TAI region score are combined to obtain a CA region score as follows: CA region score = 0.32 * (LOH region score) + 0.68 * (TAI region score) or CA region score = 0.34 * (LOH region score) 0.66 * (TAI region score)

[0042] In some embodiments, the LOH and LST region scores are combined to obtain a CA region score as follows: CA region score = A*(LOH region score) + B*(LST region score)

[0043] In some embodiments, the LOH region score of a sample and the LST region score of a sample are combined to obtain a CA region score as follows: CA region score = 0.85 * (LOH region score) 0.15 * (LST region score)

[0044] In some embodiments, the TAI and LST domain scores are combined to obtain a CA domain score as follows: CA domain score = A*(TAI domain score) + B*(LST domain score)

[0045] In some embodiments, the LOH region score, the TAI region score, and the LST region score are combined to obtain a CA region score as follows: CA region score = A*(LOH region score) + B*(TAI region score) + C*(LST region score)

[0046] In some embodiments, the LOH region score, the TAI region score, and the LST region score are combined to obtain a CA region score as follows: CA region score = 0.21 * (LOH region score) + 0.67 * (TAI region score) + 0.12 * (LST region score) or CA region score = [0.24] * (LOH region score) + [0.65] * (TAI region score) + [0.11] * (LST region score) or CA region score = [0.11] * (LOH region score) + [0.25] * (TAI region score) + [0.12] * (LST region score)

[0047] In some embodiments, the CA region score is a combination of scores derived or calculated from (e.g., representative of or corresponding to) the average (e.g., arithmetic mean) of (1) the detected LOH regions ("LOH region score" as defined herein), (2) the detected TAI regions ("TAI region score" as defined herein), and / or (3) the detected LST regions ("LST region score" as defined herein) to obtain a CA region score calculated from one of the following formulas:

number

[0048] In some embodiments, including some specifically exemplified herein, one or more of these coefficients (i.e., A, B, or C, or any combination thereof) is 1, and in some embodiments, all three coefficients (i.e., A, B, and C) are 1. Thus, in some embodiments, CA region score=(LOH region score)+(TAI region score)+(LST region score), where LOH region score is the number of indicator LOH regions (or total length of LOH), TAI region score is the number of indicator TAI regions (or total length of TAI), and LST region score is the number of indicator LST regions (or total length of LST).

[0049] In some cases, a formula may not have all of the specified coefficients (and thus does not incorporate the corresponding variable(s)). For example, the above embodiment may be applied to formula (2) where A in formula (2) is 0.95 and B in formula (2) is 0.61. C and D do not apply because these coefficients and their corresponding variables are not found in formula (2) (although the clinical variables are incorporated into the clinical score found in formula (2)). In some embodiments, A is 0.9 to 1, 0.9 to 0.99, 0.9 to 0.95, 0.85 to 0.95, 0.86 to 0.94, 0.87 to 0.93, 0.88 to 0.92, 0.89 to 0.91, 0.85 to 0.9, 0.8 to 0.95, 0.8 to 0.9, 0.8 to 0.85, 0.75 to 0.99, 0.75 to 0.95, 0.75 to 0.9, 0.75 to 0.85, or 0.75 to 0.8. In some embodiments, B is between 0.40 and 1, between 0.45 and 0.99, between 0.45 and 0.95, between 0.55 and 0.8, between 0.55 and 0.7, between 0.55 and 0.65, between 0.59 and 0.63, or between 0.6 and 0.62. In some embodiments, C is between 0.9 and 1, between 0.9 and 0.99, between 0.9 and 0.95, between 0.85 and 0.95, between 0.86 and 0.94, between 0.87 and 0.93, between 0.88 and 0.92, between 0.89 and 0.91, between 0.85 and 0.9, between 0.8 and 0.95, between 0.8 and 0.9, between 0.8 and 0.85, between 0.75 and 0.99, between 0.75 and 0.95, between 0.75 and 0.95, between 0.75 and 0.9, between 0.75 and 0.85, or between 0.75 and 0.8, as applicable. In some embodiments, D is between 0.9 and 1, between 0.9 and 0.99, between 0.9 and 0.95, between 0.85 and 0.95, between 0.86 and 0.94, between 0.87 and 0.93, between 0.88 and 0.92, between 0.89 and 0.91, between 0.85 and 0.9, between 0.8 and 0.95, between 0.8 and 0.9, between 0.8 and 0.85, between 0.75 and 0.99, between 0.75 and 0.95, between 0.75 and 0.95, between 0.75 and 0.9, between 0.75 and 0.85, or between 0.75 and 0.8, as applicable.

[0050] In some embodiments, A is 0.1 to 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.2 to 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.3 to 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.4 to 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.5 to 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.6 to 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3 0.7 to 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.8 to 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.9 to 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20 or 20, or 1 to 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1.5 to 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2 to 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2.5 to 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3 to 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3.5 to 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4 to 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4.5 to 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 5 to 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 6 to 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 7 to 8, 9, 10 , 11, 12, 13, 14, 15, or 20, or 8 to 9, 10, 11, 12, 13, 14, 15, or 20, or 9 to 10, 11, 12, 13, 14, 15, or 20, or 10 to 11, 12, 13, 14, 15, or 20, or 11 to 12, 13, 14, 15, or 20, or 12 to 13, 14, 15, or 20, or 13 to 14, 15, or 20, or 14 to 15 or 20, or 15 to 20, and B is 0.1 to 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3 , 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.2 to 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.3 to 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.4 to 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.5 to 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.6 to 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.7 to 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.8 to 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.9 to 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1 to 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1.5 to 2, 2.5 , 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2 to 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2.5 to 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3 to 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3.5 to 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 2 0, or 4 to 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4.5 to 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 5 to 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 6 to 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 7 to 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 8 to 9, 10, 11, 12, 13, 14, 15, or 20, or 9 to 10, 11, 12, 13, 14, 15, or 2 0, or 10 to 11, 12, 13, 14, 15, or 20, or 11 to 12, 13, 14, 15, or 20, or 12 to 13, 14, 15, or 20, or 13 to 14, 15, or 20, or 14 to 15 or 20, or 15 to 20, and C is 0.1 to 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.2 to 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, as applicable.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.3-0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.4-0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20,. or 0.5 to 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.6 to 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.7 to 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.8 to 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.9 to 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1 ~1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1.5 to 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2 to 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2.5 to 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3 to 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3.5 to 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4 to 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4.5 to 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 5 to 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 6 to 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 7 to 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 8 to 9, 10, 11, 12, 13, 14, 15, or 20, or 9 to 10, 11, 12, 13, 14, 15, or 20, or 10 to 11, 12, 13, 14, 15, or 20, or 11 to 12, 13, 14, 15, or is 20, or 12 to 13, 14, 15, or 20, or 13 to 14, 15, or 20, or 14 to 15 or 20, or 15 to 20, and D is 0.1 to 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.2 to 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.3 to 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.4 to 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.5 to 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 1 1, 12, 13, 14, 15, or 20, or 0.6 to 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.7 to 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.8 to 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 0.9 to 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1 to 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 1.5 to 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2 to 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 2.5 to 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3 to 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 3.5 to 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 , or 20, or 4 to 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 4.5 to 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 5 to 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 6 to 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20, or 7 to 8, 9, 10, 11, 12, 13, 14 , 15, or 20, or 8 to 9, 10, 11, 12, 13, 14, 15, or 20, or 9 to 10, 11, 12, 13, 14, 15, or 20, or 10 to 11, 12, 13, 14, 15, or 20, or 11 to 12, 13, 14, 15, or 20, or 12 to 13, 14, 15, or 20, or 13 to 14, 15, or 20, or 14 to 15 or 20, or 15 to 20. In some embodiments, A, B, and / or C are within the nearest round of any of these values ​​(e.g., A is between 0.45 and 0.54, etc.).

[0051] Thus, in one aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining an LOH region score of the sample; (2) determining a TAI region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on a combination of the LOH region score and the TAI region score (e.g., a combined CA region score) that exceeds a reference, or optionally (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on a combination of the LOH region score and the TAI region score (e.g., a combined CA region score) that does not exceed the reference. In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining an LOH region score of the sample; (2) determining an LST region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on a combination of the LOH region score and the LST region score (e.g., a combined CA region score) that exceeds a reference; or, optionally, (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on a combination of the LOH region score and the LST region score (e.g., a combined CA region score) that does not exceed the reference. In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining a TAI region score of the sample; (2) determining an LST region score of the sample; and (3)(a) detecting (or diagnosing) HRD in the sample based at least in part on a combination of the TAI region score and the LST region score (e.g., a combined CA region score) that exceeds a reference, or optionally (3)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on a combination of the TAI region score and the LST region score (e.g., a combined CA region score) that does not exceed a reference.In another aspect, the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining an LOH region score of the sample; (2) determining a TAI region score of the sample; (3) determining an LST region score of the sample; and (4)(a) detecting (or diagnosing) HRD in the sample based at least in part on a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined CA region score) that exceeds a reference; or, optionally, (4)(b) detecting (or diagnosing) the absence of HRD in the sample based at least in part on a LOH region score, the TAI region score, and the LST region score (e.g., a combined CA region score) that does not exceed the reference.

[0052] Accordingly, another aspect of the present invention provides a method of assessing (e.g., detecting, diagnosing) HRD in a sample, comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (3) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); (4) calculating an average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) assessing HRD in the sample based at least in part on the calculated average (e.g., arithmetic mean) made in (4).

[0053] In some embodiments, the reference (or index) discussed above for the CA region score (e.g., number of indicator CA regions) can be 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, 20 or more, preferably 5, preferably 8, more preferably 9 or 10, and most preferably 10. Reference to the length of all (e.g., combined) of the indicator CA regions can be about 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325 350, 375, 400, 425, 450, 475, 500 megabases or more, preferably about 75 megabases or more, preferably about 90 or 105 megabases or more, more preferably about 120 or 130 megabases or more, more preferably about 135 megabases or more, and most preferably about 150 megabases or more. In some embodiments, the references discussed above for the combined CA region score (e.g., the combined number of indicator LOH regions, indicator, TAI regions, and / or indicator LST regions) are 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50 or more, preferably 5, preferably 10, preferably 15, preferably 20, preferably 25, preferably 30, preferably 35, preferably 40-44, most preferably 42 or more. References to the total (e.g., combined) length of the indicator LOH region, indicator TAI region, and / or indicator LST region can be about 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325 350, 375, 400, 425, 450, 475, 500 megabases or more, preferably about 75 megabases or more, preferably about 90 or 105 megabases or more, more preferably about 120 or 130 megabases or more, more preferably about 135 megabases or more, and most preferably about 150 megabases or more.

[0054] In some embodiments, the present invention provides a method for detecting an HRD signature in a sample. Accordingly, another aspect of the present invention provides a method for detecting an HRD signature in a sample, comprising: (1) determining the total number of LOH regions of a certain size or characteristic in the sample (e.g., "indicator LOH regions" as defined herein); (2) determining the total number of TAI regions of a certain size or characteristic in the sample (e.g., "indicator TAI regions" as defined herein); (3) determining the total number of LST regions of a certain size or characteristic in the sample (e.g., "indicator LST regions" as defined herein); (4) combining the determinations made in (1), (2), and (3) (e.g., calculating or deriving a combined CA region score); and (5) characterizing a sample whose combined CA region score is greater than a reference value as having an HRD signature. In some embodiments, the reference value is 42. Thus, in some embodiments, a sample is characterized as having an HRD signature if the reference value is 42. In some embodiments, the references discussed above for the combined CA region score (e.g., the combined number of indicator LOH regions, indicator, TAI regions, and / or indicator LST regions) are 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50 or more, preferably 5, preferably 10, preferably 15, preferably 20, preferably 25, preferably 30, preferably 35, preferably 40-44, most preferably 42 or more.

[0055] In some embodiments, the number of indicator CA regions (or combined length, CA region score, or combined CA region score) in a sample is considered "greater" than the reference if it is at least 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater than the reference, while in some embodiments, it is considered "greater" if it is at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference. Conversely, in some embodiments, the number of indicator CA regions (or combined length, CA region score, or combined CA region score) in a sample is considered "not greater" than the reference if it is no more than 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater than the reference, while in some embodiments, it is considered "not greater" if it is no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference.

[0056] In some embodiments, the reference number (or length, value, or score) is derived from a relevant reference population. Such a reference population may include: (a) patients with the same cancer as the patient being tested, (b) patients with the same cancer subtype, (c) patients with cancers with similar genetic or other clinical or molecular characteristics, (d) patients who have responded to a particular treatment, (e) patients who have not responded to a particular treatment, (f) patients who are apparently healthy (e.g., free of any cancer, or at least free of the cancer of the patient being tested), etc. The reference numbers (or lengths, values, or scores) may be selected such that they (a) represent the numbers (or lengths, values, or scores) found in a reference population as a whole; (b) represent the average (mean, median, etc.) of the numbers (or lengths, values, or scores) found in a reference population as a whole or as a particular subpopulation; (c) represent the numbers (or lengths, values, or scores) (e.g., an average such as the mean or median) found in tertiles, quartiles, quintiles, etc. of a reference population ranked by (i) their respective numbers (or lengths, values, or scores) or (ii) the clinical characteristics they are found to have (e.g., strength of response, prognosis (including time to cancer-specific death), etc.); or (d) have high sensitivity for detecting HRDs to predict response to a particular treatment (e.g., platinum, PARP inhibitors, etc.).

[0057] In some embodiments, the reference or index indicates that the HRD is the same as the reference if the test value or score from the sample is exceeded, and indicates the absence of HRD (or functional HDR) if the test value or score from the sample is not exceeded. In some embodiments, they are different.

[0058] In another aspect, the present invention provides a method for predicting the status of BRCA1 and BRCA2 genes in a sample. Such methods are similar to the methods described above, except that the determination of CA regions, LOH regions, TAI regions, LST regions, or scores incorporating the same are used to assess (e.g., detect) BRCA1 and / or BRCA2 defects in the sample.

[0059] In another aspect, the present invention provides a method for predicting the response of a cancer patient to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor. Such a method is similar to the above-described method, except that the determination of CA regions, LOH regions, TAI regions, LST regions, or scores incorporating these, comprising a high HRD score (e.g., HRD signature or a high combined CA region score) is used to predict the likelihood that a cancer patient will respond to a cancer treatment regimen.

[0060] In some embodiments, the patient is a treatment-naive patient. In another aspect, the present invention provides a method of treating cancer. Such a method is similar to the above-mentioned method, except that a specific treatment regimen is administered (recommended, prescribed, etc.) based at least in part on the determination of CA region, LOH region, TAI region, LST region, or score incorporating them.

[0061] In another aspect, the invention features the use of one or more drugs selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors in the manufacture of a medicament useful for treating cancer in a patient identified as having (or having) cancer cells determined to have high levels of HRD (e.g., an HRD signature) as described herein.

[0062] In another aspect, the present specification features a method for evaluating a sample for the presence of mutations in genes from HDR pathways.Such a method is similar to the above-mentioned method, and differs in that CA region, LOH region, TAI region, LST region, or the score determination incorporating them is used to detect (or not detect) the presence of mutations in genes from HDR pathways.

[0063] In another aspect, the document features a method for evaluating a patient's cancer cells for the presence of an HRD signature. The method comprises, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes in a cancer patient's cancer cells that exceed a reference number, and (b) identifying the patient as having cancer cells with an HRD signature. In another aspect, the document features a method for evaluating a patient's cancer cells for the presence of an HDR-deficient state. The method comprises, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes in a cancer patient's cancer cells that exceed a reference number, and (b) identifying the patient as having cancer cells with an HDR-deficient state. In another aspect, the document features a method for evaluating a patient's cancer cells for the presence of an HRD signature. The method comprises, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes in a cancer patient's cancer cells that exceed a reference number, and (b) identifying the patient as having cancer cells with an HRD signature. In another aspect, the present specification features a method for evaluating a patient's cancer cells for the presence of a genetic mutation in a gene from a HDR pathway. The method includes or consists essentially of: (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes in a cancer patient's cancer cells that exceed a reference number; and (b) identifying the patient as having cancer cells that have a genetic mutation.

[0064] In another aspect, the document features a method for determining whether a patient is likely to respond to a cancer treatment regimen, comprising administering radiation or a drug selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor. The method comprises, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes in cancer cells of a cancer patient that exceeds a reference number, and (b) identifying the patient as likely to respond to a cancer treatment regimen. In another aspect, the document features a method for evaluating a patient. The method comprises, or consists essentially of, (a) determining that the patient contains cancer cells with an HRD signature, where the presence of indicator CA regions in at least one pair of human chromosomes in cancer cells of the cancer patient that exceeds a reference number indicates that the cancer cells have an HRD signature, and (b) diagnosing the patient as having cancer cells with an HRD signature. In another aspect, the document features a method for evaluating a patient. The method includes or consists essentially of: (a) determining that the patient contains cancer cells with HDR-deficient status, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient's cancer cells above a reference number indicates that the cancer cells have HDR-deficient status; and (b) diagnosing the patient as having cancer cells with HDR-deficient status. In another aspect, the document features a method for evaluating a patient. The method includes or consists essentially of: (a) determining that the patient contains cancer cells with HDR-deficient status, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient's cancer cells above a reference number indicates that the cancer cells have high HDR; and (b) diagnosing the patient as having cancer cells with HDR-deficient status. In another aspect, the document features a method for evaluating a patient.The method comprises, or consists essentially of: (a) determining that a patient contains cancer cells that have a genetic mutation in a gene from the HDR pathway, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient's cancer cells above a reference number indicates that the cancer cells have a genetic mutation; and (b) diagnosing the patient as having cancer cells that have a genetic mutation. In another aspect, the document features a method for assessing a patient's likelihood of responding to a cancer treatment regimen, comprising administering radiation or a drug selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor. The method includes, or consists essentially of, (a) determining that the patient contains cancer cells having an HRD signature, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient that exceeds a reference number indicates that the cancer cells have an HRD signature, and (b) diagnosing the patient as likely to respond to a cancer treatment regimen based at least in part on the presence of the HRD signature. In another aspect, the document features a method for assessing a patient for likelihood of responding to a cancer treatment regimen, the method including administering radiation or a drug selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor. The method comprises, or consists essentially of, (a) determining that the patient contains cancer cells having an HRD signature, wherein the presence of indicator CA regions in excess of a reference number in at least one pair of human chromosomes in the cancer patient's cancer cells indicates that the cancer cells have an HRD signature, and (b) diagnosing the patient as likely to respond to a cancer treatment regimen based at least in part on the presence of the HRD signature.

[0065] In another aspect, the document features a method for performing a diagnostic analysis of cancer cells of a patient. The method includes, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of the cancer cells in excess of a reference number, and (b) identifying or classifying the patient as having cancer cells with an HRD signature. In another aspect, the document features a method for performing a diagnostic analysis of cancer cells of a patient. The method includes, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of the cancer cells in excess of a reference number, and (b) identifying or classifying the patient as having cancer cells with an HDR-deficient state. In another aspect, the document features a method for performing a diagnostic analysis of cancer cells of a patient. The method includes, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of the cancer cells in excess of a reference number, and (b) identifying or classifying the patient as having cancer cells with an HDR-deficient state. In another aspect, the document features a method for performing a diagnostic analysis of cancer cells of a patient. The method includes, or consists essentially of, (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of the cancer cells in excess of a reference number, and (b) identifying or classifying the patient as having cancer cells with an HDR-deficient state. The method comprises or consists essentially of: (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of a cancer cell that is longer than the reference number; and (b) identifying or classifying the patient as having cancer cells with genetic mutations in genes from the HDR pathway. In another aspect, the present specification features a method for performing a diagnostic analysis of a cancer cell of a patient to determine whether the cancer patient is likely to respond to a cancer treatment regimen, comprising administering radiation or administering a drug selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors. The method comprises or consists essentially of: (a) detecting the presence of indicator CA regions in at least one pair of human chromosomes of a cancer cell that is longer than the reference number; and (b) identifying or classifying the patient as having a high likelihood of responding to a cancer treatment regimen.

[0066] In another aspect, the description features a method for diagnosing a patient as having cancer cells with an HRD signature. The method comprises, or consists essentially of: (a) determining that the patient comprises cancer cells with an HRD signature, where the presence of indicator CA regions in at least one pair of human chromosomes in the cancer patient's cancer cells beyond a reference number indicates that the cancer cells have an HRD signature; and (b) diagnosing the patient as having cancer cells with an HRD signature. In another aspect, the description features a method for diagnosing a patient as having cancer cells with an HDR-deficient state. The method comprises, or consists essentially of: (a) determining that the patient comprises cancer cells with an HDR-deficient state, where the presence of indicator CA regions in at least one pair of human chromosomes in the cancer patient's cancer cells beyond a reference number indicates that the cancer cells have an HDR-deficient state; and (b) diagnosing the patient as having cancer cells with an HDR-deficient state. In another aspect, the description features a method for diagnosing a patient as having cancer cells with an HDR-deficient state. The method comprises, or consists essentially of: (a) determining that the patient comprises cancer cells with an HDR-deficient state, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient's cancer cells beyond a reference number indicates that the cancer cells have an HDR-deficient state; and (b) diagnosing the patient as having cancer cells with an HDR-deficient state. In another aspect, the description features a method for diagnosing a patient as having cancer cells with a genetic mutation in a gene from the HDR pathway. The method comprises, or consists essentially of: (a) determining that the patient comprises cancer cells with a genetic mutation, where the presence of indicator CA regions in at least one pair of human chromosomes of the cancer patient's cancer cells beyond a reference number indicates that the cancer cells have an HDR-deficient state; and (b) diagnosing the patient as having cancer cells with an genetic mutation.In another aspect, the description features a method for diagnosing a patient as a candidate for a cancer treatment regimen, comprising administering radiation or administering a drug selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors. The method comprises, or consists essentially of, (a) determining that the patient contains cancer cells with an HRD signature, where the presence of indicator CA regions in at least one pair of human chromosomes in the cancer patient's cancer cells above a reference number indicates that the cancer cells have an HRD signature, and (b) diagnosing the patient as likely to respond to the cancer treatment regimen based at least in part on the presence of the HRD signature. In another aspect, the description features a method for diagnosing a patient as a candidate for a cancer treatment regimen, comprising administering radiation or administering a drug selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors. The method comprises, or consists essentially of, (a) determining that the patient contains cancer cells having an elevated HRD signature, wherein the presence of indicator CA regions in excess of a reference number in at least one pair of human chromosomes in the cancer patient's cancer cells indicates that the cancer cells have an HRD signature; and (b) diagnosing the patient as likely to respond to a cancer treatment regimen based at least in part on the presence of the HRD signature.

[0067] In another aspect, the invention provides a method for evaluating a patient, the method comprising, or consisting essentially of: (a) determining whether the patient has (or has had) cancer cells with indicator CA regions (or, e.g., a CA region score above a reference CA region score) above a reference number; and (b)(1) diagnosing the patient as having cancer cells with HRD if it is determined that the patient has (or has had) cancer cells with CA regions (or, e.g., a CA region score above a reference CA region score) above the reference number, or (b)(2) diagnosing the patient as not having cancer cells with HRD if it is determined that the patient does not have (or has never had) cancer cells with CA regions above the reference number (e.g., the patient does not have (or has never had) cancer cells with CA regions above the reference number).

[0068] In another aspect, the invention features the use of a plurality of oligonucleotides capable of hybridizing to a plurality of polymorphic regions of human genomic DNA in the manufacture of a diagnostic kit useful for determining the total number or combined length of CA regions in at least a chromosome pair (or DNA derived therefrom) in a sample obtained from a cancer patient, and for detecting (a) HRD, high HRD, or likelihood of HRD (e.g., an HRD signature, respectively) in the sample, (b) a defect (or likelihood of a defect) in the BRCA1 or BRCA2 gene in the sample, or (c) an increased likelihood that a cancer patient will respond to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor.

[0069] In another aspect, the invention features a system for detecting HRD (e.g., an HRD signature) in a sample. The system comprises or consists essentially of: (a) a sample analyzer configured to generate a plurality of signals for genomic DNA (or DNA derived therefrom) of at least one pair of human chromosomes in the sample; and (b) a computer subsystem programmed to calculate the number or combined length of CA regions in at least one pair of human chromosomes based on the plurality of signals. The computer subsystem can be programmed to compare the number or combined length of CA regions to a reference number to detect (a) HRD, high HRD, or a likelihood of HRD (respectively, e.g., an HRD signature) in the sample, (b) a defect (or a likelihood of a defect) in the BRCA1 or BRCA2 gene in the sample, or (c) an increased likelihood that a cancer patient will respond to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor. The system can comprise an output module configured to display (a), (b), or (c). The system may include an output module configured to display a recommendation for use of a cancer treatment regimen.

[0070] In another aspect, the present invention provides a computer program product embodied in a computer readable medium that, when executed on a computer, provides instructions for detecting the presence or absence of any CA region (wherein the CA region is optionally an indicator CA region) along one or more of human chromosomes other than human X and Y sex chromosomes, and for determining the total number or combined length of CA regions in one or more chromosome pairs. The computer program product may include other instructions.

[0071] In another aspect, the present invention provides a diagnostic kit. The kit comprises or consists essentially of at least 500 oligonucleotides that can hybridize to multiple polymorphic regions of human genomic DNA (or DNA derived therefrom) and a computer program product provided herein. The computer program product can be embodied in a computer readable medium that, when executed on a computer, provides instructions for detecting the presence or absence of any CA region (wherein the CA region is optionally an indicator CA region) along one or more of human chromosomes other than human X and Y sex chromosomes, and for determining the total number or combined length of CA regions in one or more chromosome pairs. The computer program product can include other instructions.

[0072] In some embodiments of any one or more of the aspects of the invention described in the preceding paragraph, any one or more of the following may be suitably applied. The CA region may be determined in at least 2, 5, 10, or 21 pairs of human chromosomes. The cancer cells may be ovarian cancer cells, breast cancer cells, lung cancer cells, or esophageal cancer cells. The reference may be 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, or 20 or more. The at least one pair of human chromosomes may exclude human chromosome 17. The DNA damaging agent may be cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline may be epirubicin or doxorubicin, the topoisomerase I inhibitor may be campothecin, topotecan, or irinotecan, or the PARP inhibitor may be iniparib, olaparib, or verapirib. The patient may be a treatment-naive patient.

[0073] As described herein, a sample (e.g., a cancer cell sample or a sample containing DNA from one or more cancer cells) can be identified as having an "HRD signature" (or alternatively referred to as an "HDR-deficient signature") if the genome of the cell being evaluated contains (a) any of a LOH region score, a TAI region score, or a LST region score that exceeds the reference, or (b) a combined CA region score that exceeds the reference. Conversely, a sample (e.g., a cancer cell sample or a sample containing DNA from one or more cancer cells) can be identified as lacking an "HRD signature" (or alternatively referred to as an "HDR-deficient signature") if the genome of the cell being evaluated contains (a) a LOH region score, a TAI region score, and a LST region score that does not exceed the reference, or (b) a combined CA region score that does not exceed the reference.

[0074] A cell (e.g., a cancer cell) identified as having an HRD signature may be classified as having an increased likelihood of having HDR deficiency and / or having an increased likelihood of having a defective state in one or more genes in an HDR pathway. For example, a cancer cell identified as having an HRD signature may be classified as having an increased likelihood of having an HDR defective state. In some cases, a cancer cell identified as having an HRD signature may be classified as having an increased likelihood of having a defective state for one or more genes in an HDR pathway. As used herein, a defective state of a gene means that the sequence, structure, expression, and / or activity of the gene or its product is defective compared to normal. Examples include, but are not limited to, low or absent expression of mRNA or protein, deleterious mutations, hypermethylation, attenuated activity (e.g., enzymatic activity, ability to bind to another biomolecule), and the like. As used herein, a defective state of a pathway (e.g., an HDR pathway) means that at least one gene (e.g., BRCA1) of the pathway is defective. Examples of highly harmful mutations include frameshift mutations, stop codon mutations, and mutations that lead to changes in RNA splicing.The defective state of genes in HDR pathway can lead to defective or reduced activity in homology-directed repair in cancer cells.Examples of genes in HDR pathway include, but are not limited to, the genes listed in Table 1. [Table 1]

[0075] As described herein, identifying CA loci (as well as the size and number of CA regions) may include, first, genotyping the sample at various genomic loci (e.g., SNP loci, individual bases in large-scale sequencing), and, second, determining whether the loci exhibit LOH, TAI, or LST. Any suitable technique may be used to determine the genotype at loci of interest in the genome of a cell. For example, single nucleotide polymorphism (SNP) arrays (e.g., SNP arrays across the human genome), targeted sequencing of loci of interest (e.g., sequencing of SNP loci and their surrounding sequences), and even large-scale sequencing (e.g., whole exome, transcriptome, or genome sequencing) may be used to identify loci that are homozygous or heterozygous. Typically, an analysis of the homozygous or heterozygous nature of loci across the length of a chromosome may be performed to determine the length of the CA region. For example, stretches of SNP locations spaced along a chromosome (e.g., spaced about 25 kb to about 100 kb apart) may be evaluated using the results of an SNP array to determine not only the presence of regions of homozygosity (e.g., LOH) along the chromosome, but also the length of the regions. Results from an SNP array may be used to generate graphs plotting allele dosage along the chromosome. Allele dosage of SNPi i The two alleles (A i and B. i ) can be calculated from the adjusted signal intensity: i =A i / (A i +B i ). Examples of such graphs are shown in Figures 1 and 2, showing the differences between fresh frozen and FFPE samples, and between SNP microarray and SNP sequencing analysis. Numerous variations of nucleic acid arrays useful in the present invention are known in the art. These include the arrays used in the various examples below (e.g., the Affymetrix 500K GeneChip array in Example 3, the Affymetrix OncoScan™ FFPE Express 2.0 service (formerly MIP CN service) in Example 4).

[0076] Once a sample has been genotyped for multiple loci (e.g., SNPs), common techniques may be used to identify loci and regions of LOH, TAI, and LST (including those described in International Application No. PCT / US2011 / 040953 (published as WO / 2011 / 160063), International Application No. PCT / US2011 / 048427 (published as WO / 2012 / 027224), Popova et al., Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1 / 2 inactivation, CANCER RES. (2012) 72:5454-5462). In some embodiments, determining whether a chromosomal imbalance or large-scale transition comprises determining whether these are somatic or germline abnormalities. One way of determining this is to compare the somatic genotype to the germline. For example, the genotypes of multiple loci (e.g., SNPs) can be determined in both a germline (e.g., blood) sample and a somatic (e.g., tumor) sample. The genotypes of each sample can be compared (typically computationally) to determine where the genome of the germline cells is heterozygous and the genome of the somatic cells is homozygous. Such loci are LOH loci, and regions of such loci are LOH regions.

[0077] Computational techniques can also be used to determine whether an abnormality is germline or somatic. Such techniques are particularly useful when germline samples are not available for analysis and comparison. For example, algorithms such as those described elsewhere can be used to detect LOH regions using information from SNP arrays (Nannya et al., Cancer Res. (2005) 65:6071-6079 (2005)). Typically, these algorithms do not explicitly take into account contamination of tumor samples with benign tissue. See Abkevich et al., International Application No. PCT / US2011 / 026098; Goransson et al., PLoS One (2009) 4(6):e6057. This contamination is often high enough to make detection of LOH regions difficult. Despite contamination, improved analytical methods according to the present invention for identifying LOH, TAI, and LST include those embodied in computer software products as described below.

[0078] The following is an example. If the observed ratio of signals of two alleles, A and B, is 2 to 1, there are two possibilities. The first possibility is that the cancer cells have LOH with a deletion of allele B in a sample with 50% contamination by normal cells. The second possibility is that there is no LOH, but allele A is replicated in a sample without contamination with normal cells. The algorithm can be implemented as a computer program described herein to reconstruct LOH regions based on genotype (e.g., SNP genotype) data. One point of the algorithm is to first reconstruct the allele-specific copy number (ASCN) at each locus (e.g., SNP). The ASCN is the copy number of both the paternal and maternal alleles. The LOH regions are then determined as the stretches of SNPs where one of the ASCNs (paternal or maternal) is zero. The algorithm can be based on maximizing the likelihood function and can be conceptually similar to the previously described algorithm designed to reconstruct the total copy number (rather than the ASCN) at each locus (e.g., SNP). See International Application No. PCT / US2011 / 026098 to Abkevich et al. The likelihood function can be maximized over the ASCN of all loci, the level of contamination with benign tissue, the total copy number averaged over the entire genome, and the sample-specific noise level. The input data for the algorithm can include or consist of (1) sample-specific normalized signal intensities for both alleles of each locus, and (2) an assay-specific (specific for different SNP arrays and sequence-based approaches) set of parameters defined based on the analysis of a large number of samples with known ASCN profiles.

[0079] In some cases, nucleic acid sequencing techniques may be used to genotype loci. For example, genomic DNA from a cell sample (e.g., a cancer cell sample) may be extracted and fragmented. Any suitable method may be used to extract and fragment genomic nucleic acid, including, but not limited to, commercially available kits such as QIAamp™ DNA Mini Kit (Qiagen™), MagNA™ Pure DNA Isolation Kit (Roche Applied Science™), and GenElute™ Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich™). Once extracted and fragmented, either targeted or untargeted sequencing may be performed to genotype the sample at the loci. For example, whole genome, whole transcriptome, or whole exome sequencing may be performed to genotype millions or even billions of base pairs (i.e., a base pair may be the "loci" being evaluated).

[0080] In some cases, targeted sequencing of known polymorphic loci (e.g., SNPs and surrounding sequences) can be performed as an alternative to microarray analysis. For example, genomic DNA can be enriched for those fragments containing the loci (e.g., SNP positions) to be analyzed using kits designed for this purpose (e.g., Agilent SureSelect™, Illumina TruSeq Capture™, and Nimblegen SeqCap EZ Choice™). For example, genomic DNA containing the loci to be analyzed can be hybridized to biotinylated capture RNA fragments to form biotinylated RNA / genomic DNA complexes. Alternatively, DNA capture probes can be utilized to result in the formation of biotinylated DNA / genomic DNA hybrids. Streptavidin-coated magnetic beads and magnetic forces can be used to separate biotinylated RNA / genomic DNA complexes from those genomic DNA fragments that are not present in the biotinylated RNA / genomic DNA complexes. The resulting biotinylated RNA / genomic DNA complex can be treated to remove the captured RNA from the magnetic beads, thereby leaving intact genomic DNA fragments containing the loci to be analyzed. These intact genomic DNA fragments containing the loci to be analyzed can be amplified, for example, using PCR technology. The amplified genomic DNA fragments can be sequenced using high-throughput sequencing technology or next-generation sequencing technology such as Illumina HiSeq™, Illumina MiSeq™, Life Technologies SoLID™ or Ion Torrent™, or Roche 454™.

[0081] Sequencing results from genomic DNA fragments may be used to identify loci that do or do not exhibit CA, similar to the microarray analysis described herein. In some cases, analysis of the genotypes of loci across the length of a chromosome may be performed to determine the length of the CA region. For example, stretches of SNP locations spaced apart along a chromosome (e.g., spaced apart from about 25 kb to about 100 kb) may be assessed by sequencing and the sequencing results used to determine not only the presence of a CA region, but also the length of that CA region. The resulting sequencing results may be used to generate a graph plotting allele dosage along the chromosome. Allele dosage of SNPi d i The two alleles (A i and B. i ) can be calculated from the adjusted number of captured probes for: i =A i / (A i +B i ). Examples of such graphs are shown in Figures 1 and 2. Determining whether an abnormality is germline or somatic can be performed as described herein.

[0082] In some cases, a selection process may be used to select loci (e.g., SNP loci) to be evaluated using assays configured to genotype the loci (e.g., SNP array-based assays and sequencing-based assays). For example, any human SNP location may be selected for inclusion in a SNP array-based assay or a sequencing-based assay configured to genotype the locus. In some cases, 0.5, 1.0, 1.5, 2.0, 25 million or more SNP locations present in the human genome may be evaluated to identify those SNPs that (a) are not present on the Y chromosome, (b) are not mitochondrial SNPs, (c) have a minor allele frequency of at least about 5 percent in Caucasians, (d) have a minor allele frequency of at least about 1 percent in three races other than Caucasians (e.g., Chinese, Japanese, and Yoruba), and / or (e) do not have significant deviations from Hardy-Weinberg equilibrium in any of the four races. In some cases, more than 100,000, 150,000, or 200,000 human SNPs that meet criteria (a)-(e) may be selected. Of the human SNPs that meet criteria (a)-(e), a group of SNPs (e.g., the top 110,000 SNPs) may be selected such that the SNPs have a high degree of allele frequency in Caucasians, cover the human genome in a somewhat evenly spaced manner (e.g., at least one SNP every about 25 kb to about 500 kb), and are not in linkage disequilibrium with another selected SNP in any of the four races. In some cases, about 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, 110,000, 120,000, 130,000 or more SNPs may be selected as meeting each of these criteria and included in an assay configured to identify CA regions across the human genome. For example, about 70,000 to about 90,000 (e.g., about 80,000) SNPs may be selected for analysis in a SNP array-based assay, and about 45,000 to about 55,000 (e.g., about 54,000) SNPs may be selected for analysis in a sequencing-based assay.

[0083] As described herein, any suitable type of sample may be evaluated. For example, a sample containing cancer cells may be evaluated to determine whether the genome of the cancer cells includes an HRD signature, lacks an HRD signature, has an increased number of indicator CA regions, or has an increased CA region score. Examples of samples containing cancer cells that may be evaluated as described herein include, but are not limited to, tumor biopsy samples (e.g., breast tumor biopsy samples), formalin-fixed paraffin-embedded tissue samples containing cancer cells, core needle biopsies, fine needle aspirates, and samples containing cancer cells shed from tumors (e.g., blood, urine, or other bodily fluids). In the case of formalin-fixed paraffin-embedded tissue samples, the sample may be prepared by DNA extraction using a genomic DNA extraction kit optimized for FFPE tissue, including but not limited to those described above (e.g., QuickExtract™ FFPE DNA Extraction Kit (Epicentre™) and QIAamp™ DNA FFPE Tissue Kit (Qiagen™)).

[0084] In some cases, laser dissection techniques may be performed on tissue samples to minimize the number of non-cancer cells in the cancer cell sample being evaluated. In some cases, antibody-based purification methods may be used to enrich for cancer cells and / or deplete non-cancer cells. Examples of antibodies that may be used for cancer cell enrichment include, but are not limited to, anti-EpCAM, anti-TROP-2, anti-c-Met, anti-folate binding protein, anti-N-cadherin, anti-CD318, anti-antimesencymal stem cell antigen, anti-Her2, anti-MUC1, anti-EGFR, anti-cytokeratin (e.g., cytokeratin 7, cytokeratin 20, etc.), anti-caveolin-1, anti-PSA, anti-CA125, and anti-surfactant protein antibodies.

[0085] Any type of cancer cell can be evaluated using the methods and materials described herein. For example, breast cancer cells, ovarian cancer cells, liver cancer cells, esophageal cancer cells, lung cancer cells, head and neck cancer cells, prostate cancer cells, colon, rectum or colorectal cancer cells, and pancreatic cancer cells can be evaluated to determine whether the genome of the cancer cell contains an HRD signature, lacks an HRD signature, has an increased number of indicator CA regions, or has an increased CA region score. In some embodiments, the cancer cell is a primary or metastatic cancer cell of ovarian cancer, breast cancer, lung cancer, or esophageal cancer.

[0086] When evaluating the genome of a cancer cell for the presence or absence of an HRD signature, one or more pairs of chromosomes (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, or 23 pairs) can be evaluated. In some cases, the genome of a cancer cell is evaluated for the presence or absence of an HRD signature using one or more pairs of chromosomes (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 pairs).

[0087] In some cases, it may be beneficial to exclude certain chromosomes from this analysis. For example, for females, the pair evaluated may include the pair of X sex chromosomes, while for males, any autosomal pair (i.e., any pair other than the pair of X and Y sex chromosomes) may be evaluated. As another example, in some cases, the pair of chromosome number 17 may be excluded from analysis. It has been determined that certain chromosomes have abnormally high levels of CA in certain cancers, and therefore, when analyzing the samples described herein from patients with these cancers, it may be beneficial to exclude such chromosomes. In some cases, the sample is from a patient with ovarian cancer, and the chromosome excluded is chromosome 17.

[0088] Thus, a given number of chromosomes may be analyzed to determine the number of indicator CA regions (or CA region scores or combined CA region scores), preferably the number of CA regions greater than 9 megabases, 10 megabases, 12 megabases, 14 megabases, more preferably greater than 15 megabases in length. Alternatively, or in addition, the sizes of all identified indicator CA regions may be summed to obtain the total length of the indicator CA regions.

[0089] As described herein, a patient having cancer cells (or a sample derived therefrom) identified as having an HRD signature status may be classified as likely to respond to a particular cancer treatment regimen based at least in part on such HRD signature. For example, a patient having cancer cells with an HRD signature may be classified as likely to respond to a cancer treatment regimen that includes the use of a DNA damaging agent, a synthetic lethal agent (e.g., a PARP inhibitor), radiation, or a combination thereof based at least in part on such HRD signature. In some embodiments, the patient is a treatment-naive patient. Examples of DNA damaging agents include, but are not limited to, platinum-based chemotherapeutic agents (e.g., cisplatin, carboplatin, oxaliplatin, and picoplatin), anthracyclines (e.g., epirubicin and doxorubicin), topoisomerase I inhibitors (e.g., campothecin, topotecan, and irinotecan), DNA crosslinkers such as mitomycin C, and triazene compounds (e.g., dacarbazine and temozolomide). Synthetic lethal therapeutic approaches typically involve administering an agent that inhibits at least one critical component of a biological pathway that is particularly important for the survival of a particular tumor cell. For example, if a tumor cell has a defective homology repair pathway (e.g., as determined according to the present invention), an inhibitor of poly ADP-ribose polymerase (or a platinum drug, a double-strand break repair inhibitor, etc.) may be particularly potent against such tumors, since two pathways important for survival are blocked (one biologically, e.g., by BRCA1 mutation, and the other synthetically, e.g., by administering a pathway drug). Synthetic lethal approaches to cancer therapy are described, for example, in O'Brien et al., Converting cancer mutations into therapeutic opportunities, EMBO MOL.MED.(2009)1:297-299. Examples of synthetic lethal agents include, but are not limited to, PARP inhibitors or double-strand break repair inhibitors in homology repair-deficient tumor cells, PARP inhibitors in PTEN-deficient tumor cells, methotrexate in MSH2-deficient tumor cells, and the like.Examples of PARP inhibitors include, but are not limited to, olaparib, iniparib, and veliparib. Examples of double-stranded break repair inhibitors include, but are not limited to, KU55933 (ATM inhibitor) and NU7441 (DNA-PKcs inhibitor). Examples of information that can be used in addition to the presence of HRD signatures that are the basis for classification as likely to respond to a particular cancer treatment regimen include, but are not limited to, previous treatment results, germline or somatic DNA mutations, gene or protein expression profiling (e.g., ER / PR / HER2 status, PSA level), tumor histology (e.g., adenocarcinoma, squamous cell carcinoma, serous papillary carcinoma, mucinous carcinoma, invasive ductal carcinoma, ductal carcinoma in situ (non-invasive), etc.), disease stage, tumor or cancer grade (e.g., well, moderate, or poorly differentiated (e.g., Gleason, modified Bloom Richardson), etc.), number of previous treatment courses, etc.

[0090] Once classified as likely to respond to a particular cancer treatment regimen (e.g., a cancer treatment regimen that includes the use of a DNA damaging agent, a PARP inhibitor, radiation, or a combination thereof), the cancer patient may be treated with such a cancer treatment regimen. In some embodiments, the patient is a treatment naive patient. Thus, the present invention provides a method of treating a patient, comprising detecting an HRD signature as described herein and administering (or recommending or prescribing) a treatment regimen that includes the use of a DNA damaging agent, a PARP inhibitor, radiation, or a combination thereof. Any suitable method for treating the cancer in question may be used to treat a cancer patient identified as having cancer cells with an HRD signature. For example, a platinum-based chemotherapy drug or a combination of platinum-based chemotherapy drugs may be used to treat cancers described elsewhere (see, e.g., U.S. Pat. Nos. 3,892,790, 3,904,663, 7,759,510, 7,759,488, and 7,754,684). In some cases, an anthracycline or a combination of anthracyclines may be used to treat cancers described elsewhere (see, e.g., U.S. Pat. Nos. 3,590,028, 4,138,480, 4,950,738, 6,087,340, 7,868,040, and 7,485,707). In some cases, a topoisomerase I inhibitor or a combination of topoisomerase I inhibitors can be used to treat cancers described elsewhere (see, for example, U.S. Patent Nos. 5,633,016 and 6,403,563). In some cases, a PARP inhibitor or a combination of PARP inhibitors can be used to treat cancers described elsewhere (see, for example, U.S. Patent Nos. 5,177,075, 7,915,280, and 7,351,701). In some cases, radiation can be used to treat cancers described elsewhere (see, for example, U.S. Patent No. 5,295,944).In some cases, combinations including different agents (e.g., combinations including any of platinum-based chemotherapy agents, anthracyclines, topoisomerase I inhibitors, and / or PARP inhibitors) with or without radiation therapy may be used to treat cancer. In some cases, combination treatments may combine any of the above agents or treatments (e.g., DNA damaging agents, PARP inhibitors, radiation, or combinations thereof) with another agent or treatment (e.g., taxane agents (e.g., doxetaxel, paclitaxel, abraxane), growth factor or growth factor receptor inhibitors (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and / or antimetabolites (e.g., 5-flourouracil, methotrexate).

[0091] In some cases, patients identified as having cancer cells lacking HRD signatures may be classified as unlikely to respond to a treatment regimen comprising a DNA damaging agent, a PARP inhibitor, radiation, or a combination thereof, based at least in part on the sample lacking HRD signatures. Such patients may then be classified as likely to respond to a cancer treatment regimen comprising the use of one or more cancer treatment agents not associated with HDR, such as taxanes (e.g., doxetaxel, paclitaxel, abraxane), growth factors or growth factor receptor inhibitors (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumab), and / or antimetabolites (e.g., 5-fluorouracil, methotrexate). In some embodiments, the patient is a treatment-naive patient. Once classified as likely to respond to a particular cancer treatment regimen (e.g., a cancer treatment regimen that includes the use of a cancer treatment agent not associated with HDR), the cancer patient may be treated with such a cancer treatment regimen. Thus, the present invention provides a method of treating a patient, comprising detecting the absence of an HRD signature as described herein and administering (or recommending or prescribing) a treatment regimen that does not include the use of a DNA damaging agent, a PARP inhibitor, radiation, or a combination thereof. In some embodiments, the treatment regimen includes one or more of a taxane agent (e.g., doxetaxel, paclitaxel, abraxane), a growth factor or growth factor receptor inhibitor (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and / or antimetabolite (e.g., 5-flurouracil, methotrexate). Any suitable method for treating cancer may be used to treat cancer patients identified as having cancer cells that lack the HRD signature.Examples of information that may be used in addition to the absence of an HRD signature upon which a classification as likely to respond to a particular cancer treatment regimen can be based include, but are not limited to, previous treatment results, germline or somatic DNA mutations, gene or protein expression profiling (e.g., ER / PR / HER2 status, PSA level), tumor histology (e.g., adenocarcinoma, squamous cell carcinoma, serous papillary carcinoma, mucinous carcinoma, invasive ductal carcinoma, ductal carcinoma in situ (non-invasive), etc.), disease stage, tumor or cancer grade (e.g., well, moderate, or poorly differentiated (e.g., Gleason, modified Bloom Richardson), etc.), number of previous courses of treatment, etc.

[0092] Upon treatment for a certain period of time (e.g., 1-6 months), the patient may be evaluated to determine whether the treatment regimen is effective. If a beneficial effect is detected, the patient may continue the same or similar cancer treatment regimen. If a beneficial effect is minimal or not detected at all, adjustments to the cancer treatment regimen may be made accordingly. For example, the dose, frequency of administration, or duration of treatment may be increased. In some cases, additional anti-cancer drugs may be added to the treatment regimen, or a particular anti-cancer drug may be replaced with one or more different anti-cancer drugs. The patient being treated may continue to be monitored as necessary, and changes may be made to the cancer treatment regimen as needed.

[0093] In addition to predicting likely therapeutic responses or selecting desirable therapeutic regimens, the HRD signature may be used to determine a patient's prognosis. Thus, in one aspect, the present specification features a method for determining a patient's prognosis based at least in part on detecting the presence or absence of an HRD signature in a sample from a patient. The method includes, or essentially includes: (a) determining whether a sample from a patient contains cancer cells (or whether the sample contains DNA derived from such cells) that have an HRD signature described herein (e.g., the presence of more indicator CA regions than a reference, or a high CA region score or a combined CA region score) (sometimes referred to herein as having high HRD); and (b)(1) determining that the patient has a relatively good prognosis based at least in part on the presence of the HRD signature or the presence of high HRD, or (b)(2) determining that the patient has a relatively poor prognosis based at least in part on the absence of the HRD signature. Prognosis may include the patient's chance of survival (e.g., progression-free survival, overall survival), with a relatively good prognosis including an increased chance of survival compared to some reference population (e.g., the average patient with this patient's cancer type / subtype, the average patient without the HRD signature, etc.). Conversely, a relatively poor prognosis in terms of survival includes a decreased chance of survival compared to some reference population (e.g., the average patient with this patient's cancer type / subtype, the average patient with the HRD signature, etc.).

[0094] As described herein, the present disclosure provides methods for evaluating a patient for cells (e.g., cancer cells) having an HRD signature. In some embodiments, one or more clinicians or medical professionals may determine whether a sample from a patient contains cancer cells (or whether the sample contains DNA from such cells) having an HRD signature. In some cases, one or more clinicians or medical professionals may determine whether a patient contains cancer cells having an HRD signature by obtaining a cancer cell sample from a patient and evaluating the DNA of the cancer cells of the cancer cell sample to determine the presence or absence of an HRD signature as described herein.

[0095] In some cases, one or more clinicians or medical professionals may obtain a cancer cell sample from a patient and provide the sample to a testing laboratory capable of evaluating the DNA of the cancer cells of the cancer cell sample to provide an indication of the presence or absence of an HRD signature described herein. In some embodiments, the patient is a treatment-naive patient. In such cases, one or more clinicians or medical professionals may determine whether a sample from a patient contains cancer cells having an HRD signature (or whether the sample contains DNA derived from such cells) by receiving information directly or indirectly from a testing laboratory about the presence or absence of an HRD signature described herein. For example, after evaluating the DNA of the cancer cells for the presence or absence of an HRD signature described herein, the testing laboratory may provide the clinician or medical professional with, or access to, a written, electronic, or verbal report or medical record that provides an indication of the presence or absence of an HRD signature of the particular patient (or patient sample) being evaluated. Such written, electronic, or verbal reports, or medical records, may enable one or more clinicians or medical professionals to determine whether a particular patient being evaluated contains cancer cells having an HRD signature.

[0096] Once a clinician or medical professional, or a group of clinicians or medical professionals, has determined that a particular patient being evaluated contains cancer cells with an HRD signature, the clinician or medical professional (or group) may classify the patient as having cancer cells with a genome containing the presence of an HRD signature. In some embodiments, the patient is a treatment-naive patient. In some cases, the clinician or medical professional, or a group of clinicians or medical professionals, may diagnose a patient whose genome has been determined to have cancer cells that contain the presence of an HRD signature as having cancer cells that are HDR deficient (or likely to be deficient). Such a diagnosis may be based solely on a determination that a sample from the patient contains cancer cells (or whether the sample contains DNA derived from such cells) with an HRD signature, or may be based at least in part on a determination that a sample from the patient contains cancer cells (or whether the sample contains DNA derived from such cells) with an HRD signature. For example, a patient determined to have cancer cells with an HRD signature may be diagnosed as likely to be HDR deficient based on the presence of the HRD signature in combination with a deletion state in one or more tumor suppressor genes (e.g., BRCA1 / 2, RAD51C), a family history of cancer, or the presence of behavioral risk factors (e.g., smoking).

[0097] In some cases, a clinician or medical professional, or a group of clinicians or medical professionals, may diagnose a patient whose genome is determined to have cancer cells that include the presence of an HRD signature as having cancer cells that are likely to contain genetic mutations in one or more genes in HDR pathways. In some embodiments, the patient is a treatment-naive patient. Such a diagnosis may be based solely on the determination that the particular patient being evaluated contains cancer cells that have a genome that contains an HRD signature, or may be based at least in part on the determination that the particular patient being evaluated contains cancer cells that have a genome that contains an HRD signature. For example, a patient whose genome is determined to have cancer cells that contain the presence of an HRD signature may be diagnosed as having cancer cells that are likely to contain genetic mutations in one or more genes in HDR pathways based on the combination of the presence of the HRD signature and a family history of cancer, or the presence of behavioral risk factors (e.g., smoking).

[0098] In some cases, a clinician or medical professional or a group of clinicians or medical professionals may diagnose a patient whose genome is determined to have cancer cells that include the presence of an HRD signature as having cancer cells that are likely to respond to a particular cancer treatment regimen. In some embodiments, the patient is a treatment-naive patient. Such a diagnosis may be based solely on a determination that a sample from the patient contains cancer cells that have an HRD signature (or whether the sample contains DNA derived from such cells), or may be based at least in part on a determination that a sample from the patient contains cancer cells that have an HRD signature (or whether the sample contains DNA derived from such cells). For example, a patient whose cancer cells are determined to have an HRD signature may be diagnosed as likely to respond to a particular cancer treatment regimen based on the presence of the HRD signature in combination with a deletion state in one or more tumor suppressor genes (e.g., BRCA1 / 2, RAD51), a family history of cancer, or the presence of behavioral risk factors (e.g., smoking). As described herein, patients determined to have cancer cells with an HRD signature may be diagnosed as likely to respond to a cancer treatment regimen that includes the use of platinum-based chemotherapy drugs such as cisplatin, carboplatin, oxaliplatin, or picoplatin, anthracyclines such as epirubicin or doxorubicin, topoisomerase I inhibitors such as campothecin, topotecan, or irinotecan, PARP inhibitors, radiation, combinations thereof, or combinations of any of the foregoing with other anti-cancer drugs. In some embodiments, the patient is a treatment-naive patient.

[0099] When a clinician or medical professional, or a group of clinicians or medical professionals, determines that a sample from a patient contains cancer cells with a genome lacking an HRD signature (or if the sample contains DNA from such cells), the clinician or medical professional (or group) may classify the patient as having cancer cells whose genome lacks an HRD signature. In some embodiments, the patient is a treatment-naive patient. In some cases, the clinician or medical professional, or a group of clinicians or medical professionals, may diagnose a patient determined to have cancer cells containing a genome lacking an HRD signature as having cancer cells that are likely to have functional HDR. In some cases, the clinician or medical professional, or a group of clinicians or medical professionals, may diagnose a patient determined to have cancer cells containing a genome lacking an HRD signature as having cancer cells that are likely not to contain genetic mutations in one or more genes in the HDR pathway. In some cases, a clinician or medical professional or group of clinicians or medical professionals may diagnose a patient determined to have cancer cells that contain a genome lacking an HRD signature or that contain an increased number of CA regions covering an entire chromosome as having cancer cells that are less likely to respond to platinum-based chemotherapy drugs such as cisplatin, carboplatin, oxalaplatin, or picoplatin, anthracyclines such as epirubincin or doxorubicin, topoisomerase I inhibitors such as campothecin, topotecan, or irinotecan, PARP inhibitors, or radiation, and / or are more likely to respond to a cancer treatment regimen that includes the use of cancer therapeutic agents not associated with HDR, such as one or more taxanes, growth factor or growth factor receptor inhibitors, antimetabolites, etc. In some embodiments, the patient is a treatment-naive patient.

[0100] As described herein, the present specification also provides a method for performing diagnostic analysis of a nucleic acid sample (e.g., a genomic nucleic acid sample or amplified nucleic acid therefrom) of a cancer patient to determine whether the sample from the patient contains cancer cells (or whether the sample contains DNA derived from such cells) that contain an HRD signature and / or an increased number of CA regions covering an entire chromosome. In some embodiments, the patient is a treatment-naive patient. For example, one or more laboratory technicians or laboratory professionals may detect the presence or absence of an HRD signature in the genome (or DNA derived therefrom) of the patient's cancer cells, or the presence or absence of an increased number of CA regions covering an entire chromosome in the genome of the patient's cancer cells. In some cases, one or more laboratory technicians or laboratory professionals may (a) receive a cancer cell sample obtained from a patient, receive a genomic nucleic acid sample obtained from a cancer cell obtained from a patient, or receive a sample containing enriched and / or amplified nucleic acid from such a genomic nucleic acid sample obtained from a cancer cell obtained from a patient, and (b) perform an analysis (e.g., an SNP array-based assay or a sequencing-based assay) using the received material to detect the presence or absence of an HRD signature covering all chromosomes or the presence or absence of an increased number of CA regions in the genome of the cancer cell of the patient by detecting the presence or absence of an HRD signature covering all chromosomes as described herein, or the presence or absence of an increased number of CA regions. In some cases, one or more laboratory technicians or laboratory professionals may receive the sample to be analyzed (e.g., a cancer cell sample obtained from a patient, a genomic nucleic acid sample obtained from a cancer cell obtained from a patient, or a sample containing enriched and / or amplified nucleic acid from such a genomic nucleic acid sample obtained from a cancer cell obtained from a patient) directly or indirectly from a clinician or medical professional. In some embodiments, the patient is a treatment-naive patient.

[0101] When a laboratory technician or a laboratory professional, or a group of laboratory technicians or laboratory professionals, detects the presence of an HRD signature described herein, the laboratory technician or laboratory professional (or group) can associate the HRD signature or the results (or results or summary of results) of the diagnostic analysis performed with the corresponding patient's name, medical record, symbol / numeric identifier, or combinations thereof. Such identification can be based solely on detecting the presence of the HRD signature, or can be based at least in part on detecting the presence of the HRD signature. For example, a laboratory technician or a laboratory professional can identify a patient having cancer cells detected as having an HRD signature as having cancer cells that are potentially HDR deficient (or as having an increased likelihood of responding to a particular treatment as described at length herein) based on a combination of the HRD signature and the results of other genetic and biochemical tests performed in the testing laboratory. In some embodiments, the patient is a treatment-naive patient.

[0102] The reverse of the above is also true. That is, when a laboratory technician or a laboratory professional, or a group of laboratory technicians or laboratory professionals, detects the absence of an HRD signature, the laboratory technician or laboratory professional (or group) can associate the absence of the HRD signature or the results of the diagnostic analysis performed (or the results or a summary of the results) with the corresponding patient's name, medical record, symbol / numeric identifier, or a combination thereof. In some cases, the laboratory technician or laboratory professional, or a group of laboratory technicians or laboratory professionals, can identify a patient having cancer cells detected as lacking an HRD signature as having cancer cells with potentially intact HDR (or having a reduced likelihood of responding to a particular treatment as described at length herein), either based on the absence of the HRD signature alone, or based on the absence of the HRD signature in combination with the results of other genetic and biochemical tests performed in the testing laboratory. In some embodiments, the patient is a treatment-naive patient.

[0103] The results of any analysis according to the present invention are often communicated to a physician, genetic counselor, and / or patient (or other interested parties such as researchers) in a transmittable form that can be communicated or transmitted to any of the above parties. Such forms may vary and may be tangible or intangible. The results may be embodied in explanatory descriptions, diagrams, photographs, charts, images, or other visual forms. For example, graphs or diagrams showing genotype or LOH (or HRD status) information may be used to illustrate the results. The descriptions and visual forms may be recorded on tangible media such as paper, computer readable media such as floppy disks, compact disks, flash memory, or intangible media, e.g., electronic media in the form of e-mail or a website on the Internet or an intranet. In addition, the results may also be recorded in audio format and transmitted via telephone, facsimile, wireless mobile phone, Internet telephone, etc., via any suitable medium, e.g., analog or digital cable lines, fiber optic cables, etc.

[0104] In this way, information and data about test results can be generated anywhere in the world and transmitted to different locations. As an illustrative example, if an assay is performed outside the United States, information and data about test results can be generated, posted in the above-mentioned transmittable form, and then imported into the United States. Thus, the present invention also encompasses a method for generating information in transmittable form about an HRD signature for at least one patient sample. The method includes the steps of (1) determining an HRD signature by the method of the present invention, and (2) embodying the results of the determining step in a transmittable form. The transmittable form is a product of such a method.

[0105] Some embodiments of the invention described herein include a step of correlating the presence of an HRD signature according to the invention (e.g., a total number of indicator CA regions greater than a reference, or a CA region score or a combined CA region score) to a particular clinical feature (e.g., an increased likelihood of a defect in the BRCA1 or BRCA2 gene, an increased likelihood of HDR deficiency, an increased likelihood of response to a treatment regimen including a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, etc.), and optionally a step of correlating the absence of the HRD signature to one or more other clinical features. Wherever such an embodiment is described throughout this specification, another embodiment of the invention may include one or both of the following steps in addition to or instead of the correlating step: (a) concluding that the patient has the clinical feature based at least in part on the presence or absence of the HRD signature, or (b) communicating that the patient has the clinical feature based at least in part on the presence or absence of the HRD signature.

[0106] By way of illustration and not limitation, one embodiment described herein is a method of predicting a cancer patient's response to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: (1) determining in a sample two or more of: (a) the sample's LOH region score; (b) the sample's TAI region score; or (c) the sample's LST region score; and (2) determining in a sample two or more of (a) the LOH region score, the TAI region score, and the LST region score that are greater than the reference. Correlating the above combination (e.g., the combined CA region score) to an increased likelihood of responding to the treatment regimen, or optionally (2)(b) Correlating a combination of two or more of the LOH region score, TAI region score, and LST region score that is not greater than the reference (e.g., the combined CA region score) to no increased likelihood of responding to the treatment regimen, or optionally (2)(c) Correlating the average (e.g., arithmetic mean) of the LOH region score, TAI region score, and LST region score. According to the previous paragraph, this description of this embodiment is understood to include the description of two alternative related embodiments.One such embodiment provides a method of predicting a cancer patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: (1) determining in a sample two or more of (a) the sample's LOH region score, (b) the sample's TAI region score, or (c) the sample's LST region score, or an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score; and (2) determining in a sample a combination of two or more of (a) the LOH region score, the TAI region score, and the LST region score that are greater than the reference. (2)(b) concluding that the patient has an increased likelihood of responding to the cancer treatment regimen based at least in part on a combination (e.g., a combined CA region score) of two or more of the LOH region score, TAI region score, and LST region score not exceeding a reference (e.g., a combined CA region score), or optionally concluding that the patient has a non-increased likelihood of responding to the cancer treatment regimen based at least in part on an average (e.g., an arithmetic mean) of the LOH region score, TAI region score, and LST region score.Another such embodiment provides a method of predicting a cancer patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: (1) determining in a sample two or more of (a) the sample's LOH region score, (b) the sample's TAI region score, or (c) the sample's LST region score, or an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score; and (2) determining a combination of two or more of (a) the LOH region score, the TAI region score, and the LST region score (e.g., a combined CA region score) that is greater than a reference. or, optionally, (2)(b) communicating that the patient has an increased likelihood of responding to the cancer treatment regimen based at least in part on an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score not exceeding a reference (e.g., a combined CA region score), or that the patient has a non-increased likelihood of responding to the cancer treatment regimen based at least in part on an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score.

[0107] In each embodiment described herein that includes correlating a particular assay or analytical output (e.g., total number of indicator CA regions greater than a reference number, presence of an HRD signature, etc.) to some probability (e.g., increased, not increased, decreased, etc.) of some clinical feature (e.g., response to a particular treatment, cancer-specific mortality, etc.), or additionally or alternatively concluding or communicating such clinical feature based at least in part on such particular assay or analytical output, such correlating, concluding, or communicating may include assigning a risk or likelihood of the clinical feature occurring based at least in part on the particular assay or analytical output. In some embodiments, such risk is a percentage probability of an event or outcome occurring. In some embodiments, patients are assigned to a risk group (e.g., low risk, intermediate risk, high risk, etc.). In some embodiments, "low risk" is any percentage probability below 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, or 50%. In some embodiments, "intermediate risk" is any percentage probability greater than 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, or 50% and less than 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, or 75%. In some embodiments, "high risk" is any percentage probability greater than 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 99%.

[0108] As used herein, "communicating" certain information means making such information known to another person or transmitting such information to a thing (e.g., a computer). In some methods of the invention, a patient's prognosis or likelihood of response to a particular treatment is communicated. In some embodiments, information used to arrive at such a prognosis or response prediction (e.g., an HRD signature according to the invention, etc.) is communicated. This communication can be audible (e.g., verbal), visual (e.g., written), electronic (e.g., data transmitted from one computer system to another computer system), etc. In some embodiments, communicating the cancer classification (e.g., prognosis, likelihood of response, appropriate treatment, etc.) comprises generating a report communicating the cancer classification. In some embodiments, the report is a paper report, an audible report, or an electronic record. In some embodiments, the report is displayed and / or stored on a computing device (e.g., a handheld device, a desktop computer, a smart device, a website, etc.). In some embodiments, the cancer classification is communicated to a physician (e.g., a report communicating the classification is provided to a physician). In some embodiments, the cancer classification is communicated to the patient (e.g., a report communicating the classification is provided to the patient). Communicating the cancer classification may also be accomplished by transmitting information (e.g., data) embodying the classification to a server computer and allowing an intermediary or end user to access such information (e.g., by viewing the information displayed from the server, by downloading the information in the form of one or more files transmitted from the server to the intermediary or end user's device, etc.).

[0109] Whenever an embodiment of the present invention involves concluding some fact (e.g., a patient's prognosis or the patient's likelihood of response to a particular treatment regimen), this may in some embodiments include a computer program that concludes such a fact, typically after implementing an algorithm that applies information about the CA region according to the present invention.

[0110] In each embodiment described herein that includes a number of CA regions (e.g., indicator CA regions), or the combined length of such CA regions in total, or the average (e.g., arithmetic mean) of the combined CAR region scores, the invention encompasses related embodiments that include test values ​​or scores (e.g., CA region scores, LOH region scores, etc.) derived from, incorporated into, and / or at least to some extent reflective of such numbers or lengths. In other words, the number or length of bare CA regions need not be used in the various methods, systems, etc. of the invention, but test values ​​or scores derived from such numbers or lengths may be used.For example, one embodiment of the present invention is a method of treating cancer in a patient, comprising: (1) determining in a sample from the patient two or more of: (a) a number of indicator LOH regions, (b) a number of indicator TAI regions, or (c) a number of indicator LST regions, or an average (e.g., arithmetic mean) thereof; (2) providing one or more test values ​​derived from the numbers of indicator LOH regions, indicator TAI regions, and / or indicator LST regions; (3) comparing the test value(s) to one or more reference values ​​(e.g., reference values ​​(e.g., mean, median, tertile, quartile, quintile, etc.) derived from the numbers of indicator LOH regions, indicator TAI regions, and / or indicator LST regions in a reference population); and (4) determining whether one or more of (a) the test values ​​are greater than (e.g., less than) at least one of the reference values. or optionally, (4)(b) administering an anti-cancer drug to the patient or recommending, prescribing, or initiating a treatment regimen that includes chemotherapy and / or synthetic lethal agents based, at least in part, on the comparing step that reveals one or more of the test values ​​is not greater than at least one of the reference values ​​(e.g., no more than 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater). The invention encompasses corresponding embodiments in which, mutatis mutandis, the test value or score is used to determine a patient's prognosis, the likelihood of a patient's response to a particular therapeutic regimen, the likelihood of a patient, or a patient sample, having a BRCA1, BRCA2, RAD51C or HDR deficiency, and the like.

[0111] FIG. 8 illustrates an exemplary process by which a computing system (or a computer program (e.g., software) containing computer executable instructions) can identify LOH loci or regions from genotype data as described herein. This process can be adapted for use in determining TAI and LST, as would be apparent to one of skill in the art. When the observed ratio of signals for two alleles, A and B, is 2 to 1, there are two possibilities. The first possibility is that the cancer cells have LOH with a deletion of allele B in a sample with 50% contamination with normal cells. The second possibility is that there is no LOH, but allele A is replicated in a sample without contamination with normal cells. The process begins in box 1500, where the following data is collected by the computing system: (1) sample-specific normalized signal intensities for both alleles of each locus, and (2) a set of assay-specific (specific to different SNP arrays and sequence-based approaches) parameters defined based on the analysis of a large number of samples with known ASCN profiles. As described herein, loci along chromosomes may be evaluated for homozygosity or heterozygosity using any suitable assay, such as an SNP array-based assay or a sequencing-based assay. In some cases, a system including a signal detector and a computer may be used to collect data (e.g., fluorescent signals or sequencing results) regarding the homozygosity or heterozygosity nature of multiple loci (e.g., sample-specific normalized signal intensities for both alleles at each locus). At box 1510, an allele-specific copy number (ASCN) is reconstructed at each locus (e.g., each SNP). The ASCN is the copy number of both the paternal and maternal alleles. At box 1530, a likelihood function is used to determine whether a homozygous locus or region of a homozygous locus is due to LOH. This may be conceptually similar to the previously described algorithm designed to reconstruct the total copy number (rather than the ASCN) at each locus (e.g., SNP).See International Application No. PCT / US2011 / 026098 to Abkevich et al. The likelihood function may be maximized over the ASCN of all loci, the level of contamination with benign tissue, the total copy number averaged over the entire genome, and the sample-specific noise level. At box 1540, LOH regions are determined as stretches of SNPs where one of the ASCNs (paternal or maternal) is zero. In some embodiments, the computer process further comprises the step of querying or determining whether the patient is treatment naive.

[0112] FIG. 3 shows an exemplary process by which a computing system can determine the presence or absence of an LOH signature, and is included to illustrate how this process can be applied to TAI and LST, as would be apparent to one of skill in the art. The process begins at box 300, where data regarding the homozygous or heterozygous nature of a plurality of loci along a chromosome is collected by a computing system. As described herein, any suitable assay, such as an SNP array-based assay or a sequencing-based assay, may be used to evaluate the loci along a chromosome for homozygous or heterozygous nature. In some cases, a system including a signal detector and a computer may be used to collect data regarding the homozygous or heterozygous nature of the plurality of loci (e.g., fluorescent signals or sequencing results). At box 310, data regarding the homozygous or heterozygous nature of the plurality of loci and the location or spatial relationship of each locus is evaluated by the computing system to determine the length of any LOH regions present along the chromosome. At box 320, the data regarding the number of LOH regions detected and the length of each detected LOH region is evaluated by a computing system to determine the number of LOH regions having a length (a) equal to or greater than a preset number of Mb (e.g., 15 Mb) and (b) less than the full length of the chromosome containing the LOH region. Alternatively, the computing system may determine the sum or combined LOH lengths of the above. At box 330, the computing system formats an output that provides an indication of the presence or absence of the HRD signature. Once formatted, the computing system may provide the output to a user (e.g., a laboratory technician, clinician, or medical professional). As described herein, the presence or absence of the HRD signature may be used to provide an indication of the likely HDR status of the patient, an indication of the possible presence or absence of a genetic mutation in a gene of an HDR pathway, and / or an indication of a possible cancer treatment regimen.

[0113] 4 is a diagram of an example of a computing device 1400 and a mobile computing device 1450 that may be used with the techniques described herein. The computing device 1400 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The computing device 1450 is intended to represent various forms of mobile devices, such as personal digital assistants, mobile phones, smartphones, and other similar computing devices. The components shown here, their connections and relationships, and their functions are merely exemplary and are not meant to limit the implementation of the invention(s) described and / or claimed in this document.

[0114] The computing device 1400 includes a processor 1402, a memory 1404, a storage device 1406, a high-speed interface 1408 connected to the memory 1404 and a high-speed expansion port 1410, and a low-speed bus 1414 and a low-speed interface 1415 connected to the storage device 1406. Each of the components 1402, 1404, 1406, 1408, 1410, and 1415 may be interconnected using various buses and mounted on a common motherboard or otherwise as desired. The processor 1402 may process instructions for execution within the computing device 1400, including instructions stored in the memory 1404 or on the storage device 1406, to display graphical information for a GUI on an external input / output device, such as a display 1416 coupled to the high-speed interface 1408. In other implementations, multiple processors and / or multiple buses may be used, along with multiple memories and memory types, as desired. Additionally, multiple computing devices 1400 may be connected, with each device providing a portion of the required operations (eg, a bank of servers, a group of blade servers, or a multi-processor system).

[0115] The memory 1404 stores information within the computing device 1400. In one implementation, the memory 1404 is a volatile memory unit(s). In another implementation, the memory 1404 is a non-volatile memory unit(s). The memory 1404 may also be another form of computer-readable medium, such as a magnetic disk or an optical disk.

[0116] The storage device 1406 can provide mass storage for the computing device 1400. In one implementation, the storage device 1406 can be or contain a computer-readable medium such as a floppy disk device, a hard disk device, an optical disk device, or an array of devices including a tape device, a flash memory or other similar solid-state memory device, or a device in a storage area network or other configuration. The computer program product can be tangibly embodied in an information carrier. The computer program product can also contain instructions that, when executed, perform one or more methods, such as those described above. The information carrier is a computer-readable medium or machine-readable medium, such as the memory 1404, the storage device 1406, a memory on the processor 1402, or a propagating signal.

[0117] The high-speed controller 1408 manages bandwidth-intensive operations for the computing device 1400, while the low-speed controller 1415 manages low-bandwidth-intensive operations. Such an allocation of functions is merely exemplary. In one implementation, the high-speed controller 1408 is coupled to the memory 1404, the display 1416 (e.g., via a graphics processor or accelerator), and a high-speed expansion port 1410 that may accept various expansion cards (not shown). In an implementation, the low-speed controller 1415 is coupled to the storage device 1406 and the low-speed expansion port 1414. The low-speed expansion port, which may include various communication ports (e.g., USB, Bluetooth, Ethernet, wireless Ethernet), may be coupled, for example, via a network adapter, to one or more input / output devices, such as a keyboard, a pointing device, a scanner, an optical reader, a fluorescent signal detector, or a networking device such as a switch or router.

[0118] The computing device 1400 may be implemented in a number of different forms, as shown in the figure. For example, it may be implemented as a standard server 1420, or multiple times within a group of such servers. It may also be implemented as part of a rack server system 1424. In addition, it may be implemented in a personal computer, such as a laptop computer 1422. Alternatively, components from the computing device 1400 may be combined with other components in a mobile device (not shown), such as device 1450. Each such device may contain one or more of the computing devices 1400, 1450, and the entire system may be made up of multiple computing devices 1400, 1450 in communication with each other.

[0119] Computing device 1450 includes, among other components, a processor 1452, memory 1464, input / output devices such as a display 1454, a communication interface 1466, and a transceiver 1468. Device 1450 may also be provided with a storage device, such as a microdrive or other device, to provide additional storage. Each of the components 1450, 1452, 1464, 1454, 1466, and 1468 are interconnected using various buses, and some components may be mounted on a common motherboard or otherwise as desired.

[0120] The processor 1452 may execute instructions within the computing device 1450, including instructions stored in the memory 1464. The processor may be implemented as a chipset of chips including separate and multiple analog and digital processors. The processor may provide coordination of other components of the device 1450, such as control of a user interface, applications run by the device 1450, and wireless communication by the device 1450.

[0121] The processor 1452 may communicate with a user via a control interface 1458 and a display interface 1456 coupled to a display 1454. The display 1454 may be, for example, a TFT LCD (thin film transistor liquid crystal display) or an OLED (organic light emitting diode) display, or other suitable display technology. The display interface 1456 may comprise appropriate circuitry for driving the display 1454 to present graphical and other information to the user. The control interface 1458 may receive commands from the user and convert them for transmission to the processor 1452. Additionally, an external interface 1462 may be provided in communication with the processor 1452 to enable short-range communication of the device 1450 with other devices. The external interface 1462 may provide, for example, for wired communication in some implementations or for wireless communication in other implementations, and multiple interfaces may also be used.

[0122] The memory 1464 stores information within the remote device 1450. The memory 1464 may be implemented as one or more of a computer-readable medium, a volatile memory unit, or a non-volatile memory unit. An expansion memory 1474 may also be provided to and connected to the device 1450 via an expansion interface 1472, which may include, for example, a SIMM (single in-line memory module) card interface. Such expansion memory 1474 may provide additional storage space to the device 1450 or may also store applications or other information in the device 1450. For example, the expansion memory 1474 may include instructions that perform or complement the processes described herein and may include secure information. Thus, for example, the expansion memory 1474 may be provided as a security module for the device 1450 and may be programmed with instructions that enable secure use of the device 1450. In addition, secure applications may be provided via a SIMM card along with additional information, such as placing identifying information on the SIMM card in an unhackable manner.

[0123] The memory may include, for example, flash memory and / or NVRAM memory, as discussed below. In one implementation, a computer program product is tangibly embodied in an information carrier. The computer program product includes instructions that, when executed, perform one or more methods, such as those described herein. The information carrier is a computer-readable or machine-readable medium, such as memory 1464, expansion memory 1474, memory on processor 1452, or a propagated signal that may be received, for example, via transceiver 1468 or external interface 1462.

[0124] Device 1450 may communicate wirelessly via communication interface 1466, which may include digital signal processing circuitry, as appropriate. Communication interface 1466 may provide for communication under various modes or protocols, such as GSM voice calls, SMS, EMS, or MMS messaging, CDMA, TDMA, PDC, WCDMA, CDMA2000, or GPRS, among others. Such communication may occur, for example, via radio frequency transceiver 1468. In addition, short range communication may occur, such as using Bluetooth, WiFi, or other such transceivers (not shown). In addition, GPS (Global Positioning System) receiver module 1470 may provide additional navigation-related and location-related wireless data to device 1450, which may be used as needed by applications executing on device 1450.

[0125] Device 1450 may also communicate audibly using audio codec 1460, which may receive spoken information from a user and convert it into usable digital information. Audio codec 1460 may also generate audible sounds for the user, such as through a speaker, for example, in a handset of device 1450. Such sounds may include sounds from voice telephone calls, may include recorded sounds (e.g., voice messages, music files, etc.), and may also include sounds generated by applications running on device 1450.

[0126] The computing device 1450 may be implemented in a number of different forms, as shown in the figure, such as as a mobile phone 1480, or as part of a smartphone 1482, personal digital assistant, or other similar mobile device.

[0127] Various implementations of the systems and techniques described herein may be realized in digital electronic circuitry, integrated circuits, specially designed ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementation in one or more computer programs executable and / or interpretable on a programmable system including at least one programmable processor, which may be special-purpose or general-purpose, coupled to receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0128] These computer programs (also known as programs, software, software applications or codes) include machine instructions for a programmable processor and may be implemented in high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus and / or device used to provide machine instructions and / or data to a programmable processor (e.g., magnetic disks, optical disks, memory, and programmable logic devices (PLDs)), including machine-readable media that receive machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0129] To provide for interaction with a user, the systems and techniques described herein may be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user, and a keyboard and pointing device (e.g., a mouse or trackball) by which the user can provide input to the computer. Other types of devices may also be used to provide interaction with a user, for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or haptic feedback), and input from the user may be received in any form, including acoustic, velocity, or tactile input.

[0130] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., data servers), middleware components (e.g., application servers), or front-end components (e.g., client computers having a graphical user interface or web browser through which a user can interact with an implementation of the systems and techniques described herein), or any combination of such back-end, middleware, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communications network). Examples of communications networks include local area networks ("LANs") and wide area networks ("WANs"), and the Internet.

[0131] A computing system may include clients and servers. Clients and servers are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0132] In some cases, the computing system provided herein may be configured to include one or more sample analyzers. The sample analyzer may be configured to generate a plurality of signals for genomic DNA of at least one pair of human chromosomes of the cancer cell. For example, the sample analyzer may generate a signal that can be interpreted in a manner that identifies the genotype of the locus along the chromosome. In some cases, the sample analyzer may be configured to perform one or more steps of a SNP array-based assay or a sequencing-based assay, and may be configured to generate and / or capture a signal from such an assay. In some cases, the computing system provided herein may be configured to include a computing device. In such cases, the computing device may be configured to receive a signal from the sample analyzer. The computing device may include computer executable instructions or a computer program (e.g., software) containing computer executable instructions for performing one or more of the methods or steps described herein. In some cases, such computer executable instructions may instruct the computing device to analyze a signal from a sample analyzer, from another computing device, from a SNP array-based assay, or from a sequencing-based assay. Analysis of such signals may be performed to determine genotype, homozygosity, or other chromosomal abnormality at a particular locus, region of CA, number of CA regions, determine the size of the CA region, determine the number of CA regions having a particular size or range of sizes, determine whether the sample is positive for an HRD signature, determine the number of indicator CA regions in at least one pair of human chromosomes, determine the likelihood of defects in the BRCA1 and / or BRCA2 genes, determine the likelihood of HDR defects, determine the likelihood that a cancer patient will respond to a particular cancer treatment regimen (e.g., a regimen including a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, a PARP inhibitor, or a combination thereof), or determine a combination of these items.

[0133] In some cases, the computing system provided herein may include computer executable instructions or computer programs (e.g., software) containing computer executable instructions for formatting an output that provides an indication of the number of CA regions, the size of CA regions, the number of CA regions having a particular size or size range, whether the sample is positive for an HRD signature, the number of indicator CA regions in at least one pair of human chromosomes, the likelihood of a defect in the BRCA1 and / or BRCA2 genes, for determining the likelihood of HDR defects, the likelihood that a cancer patient will respond to a particular cancer treatment regimen (e.g., a regimen including a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, a PARP inhibitor, or a combination thereof), or a combination of these items. In some cases, the computing system provided herein may include computer executable instructions or computer programs (e.g., software) containing computer executable instructions for determining a desired cancer treatment regimen for a particular patient based at least in part on the presence or absence of an HRD signature or the number of indicator CA regions.

[0134] In some cases, the computing systems provided herein may include a pre-processing device configured to process a sample (e.g., cancer cells) so that an SNP array-based assay or a sequencing-based assay can be performed. Examples of pre-processing devices include, but are not limited to, devices configured to enrich a cell population of cancer cells as opposed to non-cancerous cells, devices configured to lyse cells and / or extract genomic nucleic acid, and devices configured to enrich a sample for specific genomic DNA fragments.

[0135] The present specification also provides kits for evaluating samples (e.g., cancer cells) described herein. For example, the present specification provides kits for evaluating cancer cells for the presence of HRD signatures or for determining the number of indicator CA regions in at least one pair of human chromosomes. The kits provided herein may include either SNP probes (e.g., an array of SNP probes for performing an SNP array-based assay described herein) or primers (e.g., primers designed to sequence SNP regions via a sequencing-based assay) in combination with a computer program product containing computer executable instructions (e.g., computer executable instructions for determining the number of indicator CA regions) for performing one or more of the methods or steps described herein. In some cases, the kits provided herein may include at least 500, 1000, 10,000, 25,000, or 50,000 SNP probes that can hybridize to polymorphic regions of human genomic DNA. In some cases, the kits provided herein may include at least 500, 1000, 10,000, 25,000, or 50,000 primers capable of sequencing polymorphic regions of human genomic DNA. In some cases, the kits provided herein may include one or more other components for performing an SNP array-based assay or a sequencing-based assay. Examples of such other components include, but are not limited to, buffers, sequencing nucleotides, enzymes (e.g., polymerases), and the like. The present specification also provides for the use of any suitable number of materials provided herein in the manufacture of a kit for performing one or more of the methods or steps described herein. For example, the present specification provides for the use of a collection of SNP probes (e.g., a collection of 10,000 to 100,000 SNP probes) and a computer program product provided herein in the manufacture of a kit for evaluating cancer cells for the presence of an HRD signature.As another example, the present specification provides the use of a collection of primers (e.g., a collection of 10,000 to 100,000 primers for sequencing SNP regions) and a computer program product provided herein in the manufacture of a kit for evaluating cancer cells for the presence of an HRD signature.

[0136] Specific Embodiments Below are illustrative, but non-limiting details of particular embodiments of the present disclosure, namely, methods and systems in accordance with the more general description above.

[0137] In some embodiments, the sample used is a frozen tumor sample. In some embodiments, the sample is from a particular breast cancer subtype selected from triple negative, ER+ / HER2-, ER- / HER2+, or ER+ / HER2+. In some embodiments, the laboratory assay portion of the method, system, etc., comprises assaying the sample to sequence the BRCA1 and / or BRCA2 genes (as well as any other gene(s) in Table 1). In some embodiments, the laboratory assay portion of the method, system, etc., comprises assaying the sample to determine the allele dosage (e.g., genotype, copy number, etc.) of at least 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000 or more selected SNPs across the complete genome. In some embodiments, the SNP analysis is performed using oligonucleotide microarrays as discussed above. In some embodiments, BRCA sequence analysis, SNP analysis, or both are performed using probe capture (e.g., a probe for each SNP analyzed and / or a probe capturing the entire coding region of BRCA1 and / or BRCA2) with subsequent PCR enrichment technology (e.g., Agilent™ SuerSelect XT). In some embodiments, BRCA sequence analysis, SNP analysis, or both are performed by processing the output from the enrichment technology using a "next generation" sequencing platform (e.g., Illumina™ HiSeq2500). In some embodiments, samples are analyzed for BRCA1 / 2 somatic and / or germline mutations, which may include large rearrangements. In some embodiments, samples are analyzed for BRCA1 promoter methylation (e.g., by qPCR assay (e.g., SA Biosciences)). In some embodiments, a sample is determined to have high methylation (or is "methylated") if it has more than 10% (or 5%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%) methylation (e.g., % of BRCA1 or BRCA2 promoter CpGs that are methylated).In some embodiments, DNA from the patient's matched normal (non-tumor) tissue can be analyzed to determine, for example, whether the BRCA1 or BRCA2 mutation is germline or somatic.

[0138] In some embodiments, the LOH region score may be calculated by counting the number of LOH regions that are longer than 15 Mb but shorter than the length of a complete chromosome. In some embodiments, the TAI region score may be calculated by counting the number of telomeric regions that are longer than 11 Mb and have allelic imbalance that extends to one of the subtelomeres but does not cross the centromere. In some embodiments, the LST region score may be calculated by counting the number of breakpoints between regions longer than 10 megabases with stable copy number after filtering regions shorter than 3 megabases. In some embodiments, the LST region score may be corrected by adjusting for ploidy: LSTm=LST-kP, where P is ploidy and k is a constant (in some embodiments, k=15.5). In some embodiments, BRCA1 / 2 deficiency may be defined as loss of function due to BRCA1 or BRCA2 mutations or methylation of the BRCA1 or BRCA2 promoter region, with LOH in the affected gene. In some embodiments, a response to treatment may be a partial complete response ("pCR"), which in some embodiments may be defined as Miller-Payne 5 status after treatment (e.g., neoadjuvant).

[0139] In some embodiments, the claimed method comprises at least 8*10 -12 , 6*10 -6 , 0.0009, 0.01, 0.03, 2*10 -16 , 3*10 -6 , 10 -6 , 0.0009, 8*10 -12 , 2*10 -16 , 8*10 -8 , 6*10 -6 , 3*10 -6, or predicts BRCA deficiency with a p-value of 0.0002 (e.g., each CA region score is predefined, and optionally multiple scores are combined in a manner to obtain these p-values). In some embodiments, the p-value is calculated according to the Kolmogorov-Smirnov test. In some embodiments, the HRD score and age at diagnosis may be coded as numerical (e.g., integer) variables, breast cancer stage and subtype may be coded as categorical variables, and grade may be analyzed as either a numerical or categorical variable, or both.

[0140] In some embodiments, p-values ​​are two-sided. In some embodiments, logistic regression analysis may be used to predict BRCA1 / 2 deficiency based on the HRD scores disclosed herein, including the HRD combined score. In some embodiments, the various CA region scores are correlated (e.g., defined to achieve) according to the following correlation coefficient: LOH region score and TAI region score = 0.69 (p = 10 -39 ), between LOH and LST = 0.55 (p = 2*10 -19 ), and between TAI and LST = 0.39 (p = 10 -9 ).

[0141] In some embodiments, the method detects BRCA1 / 2 defects and / or predicts treatment response (e.g., platinum therapy response, e.g., cisplatin) by combining the LOH region score and the TAI region score as follows: Combined CA region score = 0.32 * LOH region score + 0.68 * TAI region score. In some embodiments, the method detects BRCA1 / 2 defects and / or predicts treatment response (e.g., platinum therapy response, e.g., cisplatin) by combining the LOH region score, the TAI region score, and the LST region score as follows: Combined CA region score = 0.21 * LOH region score + 0.67 * TAI region score + 0.12 * LST region score. In some embodiments, the method detects BRCA1 / 2 defects and / or predicts treatment response (e.g., platinum therapy response, e.g., cisplatin) by combining the LOH region score, the TAI region score, and the LST region score as follows: Combined CA region score = 0.11 * LOH region score 0.25 * TAI region score + 0.12 * LST region score. In some embodiments, the method detects BRCA1 / 2 defects and / or predicts treatment response (e.g., platinum therapy response, e.g., cisplatin) by combining the LOH region score, the TAI region score, and the LST region score as follows: Combined CA region score = arithmetic mean of the LOH region score, the TAI region score, and the LST region score.

[0142] In some embodiments, BRCA deficiency status and HRD status may be combined to predict treatment response. For example, the disclosure may include a method of predicting a patient's (e.g., a triple-negative breast cancer patient's) response to a cancer treatment regimen that includes a DNA damaging agent (e.g., a platinum agent, e.g., cisplatin), an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: determining, in a cancer cell from a patient sample, the number of indicator CA regions (e.g., indicator LOH regions, indicator TAI regions, indicator LST regions, or any combination thereof) in at least one pair of human chromosomes in the cancer cell of the cancer patient; determining whether cancer cells from a patient sample are defective in BRCA1 or BRCA2 (e.g., deleterious mutations, hyper-promoter methylation); and diagnosing a patient in a sample where either (a) the number of indicator CA regions is greater than a reference number, or (b) there is a BRCA1 or BRCA2 deficiency, or both (a) and (b), as having an increased likelihood of responding to the cancer treatment regimen.

[0143] Additional Specific Embodiments Embodiment 1. An in vitro method for predicting a patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor, comprising: (1) determining the number of indicator CA regions, including at least two types selected from indicator LOH regions, indicator TI regions, and indicator LST regions, in at least one pair of human chromosomes in the cancer cells of a cancer patient in a sample containing cancer cells; (2) diagnosing a patient in which the number of indicator LOH regions, indicator TAI regions, or indicator LST regions in the sample is greater than the reference number as having an increased likelihood of responding to the cancer treatment regimen.

[0144] Embodiment 2. The method of embodiment 1, wherein said at least one pair of human chromosomes represents the entire genome.

[0145] Embodiment 3. The method of embodiment 1 or embodiment 2, wherein the indicator CA regions are determined in at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 pairs of human chromosomes.

[0146] Embodiment 4. The method of any one of embodiments 1-3, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0147] Embodiment 5. The number of references in the indicator LOH region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more, and the number of references in the indicator TAI region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more. , 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more and the number of references in the indicator LST region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more.

[0148] Embodiment 6. The indicator LOH region is defined as an LOH region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, and the indicator TAI region is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, 6. The method of any one of embodiments 1-5, wherein the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length but does not extend across the centromere, and the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length.

[0149] Embodiment 7. The method of any one of embodiments 1-6, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0150] Embodiment 8. The method of any one of embodiments 1-7, further comprising administering said cancer treatment regimen to said patient diagnosed with an increased likelihood of responding to said cancer treatment regimen.

[0151] Embodiment 9. An in vitro method for predicting a patient's response to a cancer treatment regimen that includes a platinum agent, comprising: (1) determining the number of indicator CA regions, including at least two types selected from indicator LOH regions, indicator TI regions, and indicator LST regions, in at least one pair of human chromosomes in the cancer cells of a cancer patient in a sample containing cancer cells; (2) determining whether a sample containing cancer cells is defective in BRCA1 or BRCA2; and (3) diagnosing a patient in a sample in which either (a) the number of indicator LOH regions, indicator TAI regions, or indicator LST regions is greater than a reference number, or (b) the patient has a BRCA1 or BRCA2 deficiency, or both (a) and (b), as having an increased likelihood of responding to the cancer treatment regimen.

[0152] Embodiment 10. The method of embodiment 9, wherein said at least one pair of human chromosomes represents the entire genome.

[0153] Embodiment 11. The method of embodiment 9 or embodiment 10, wherein the indicator CA regions are determined in at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 pairs of human chromosomes.

[0154] Embodiment 12. The method of any one of embodiments 9 to 11, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0155] Embodiment 13. The number of references in the indicator LOH region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more, and the number of references in the indicator TAI region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more. , 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more and the number of references in the indicator LST region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more.

[0156] Embodiment 14. The indicator LOH region is defined as an LOH region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, and the indicator TAI region is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, 14. The method of any one of embodiments 9-13, wherein the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length but does not extend across the centromere, and the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length.

[0157] Embodiment 15. The method of any one of embodiments 9-14, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0158] Embodiment 16. The method of any one of embodiments 9-15, wherein the sample is BRCA1 or BRCA2 deficient if a deleterious mutation, loss of heterozygosity, or hypermethylation is detected in either BRCA1 or BRCA2 in the sample.

[0159] Embodiment 17. The method of embodiment 16, wherein high methylation is detected if methylation is detected in at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, or 50% or more of the BRCA1 or BRCA2 promoter CpGs analyzed.

[0160] Embodiment 18. An in vitro method for predicting a patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor, comprising: (1) determining the number of indicator CA regions, including at least two types selected from indicator LOH regions, indicator TI regions, and indicator LST regions, in at least one pair of human chromosomes in the cancer cells of a cancer patient in a sample containing cancer cells; (2) providing a test value derived from the number of such indicator CA regions; and (3) comparing the test value to one or more reference values ​​derived from the number of indicator CA regions in a reference population; and (4) diagnosing a patient in a sample in which the test value is greater than the one or more reference values ​​as having an increased likelihood of responding to the cancer treatment regimen.

[0161] Embodiment 19. The method of embodiment 18, wherein said at least one pair of human chromosomes represents the entire genome.

[0162] Embodiment 20 The method of embodiment 18 or embodiment 19, wherein the indicator CA regions are determined in at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 pairs of human chromosomes.

[0163] Embodiment 21. The method of any one of embodiments 18-20, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0164] 22. The number of references in the indicator LOH region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more, and the number of references in the indicator TAI region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more. , 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more and the number of references in the indicator LST region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more.

[0165] Embodiment 23. The indicator LOH region is defined as an LOH region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, and the indicator TAI region is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 23. The method of any one of embodiments 18-22, wherein the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length but does not extend across the centromere, and the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length.

[0166] Embodiment 24. The method of any one of embodiments 18-23, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0167] Embodiment 25. The method of any one of embodiments 18-24, further comprising diagnosing a patient in which the test value in the sample is not greater than the one or more reference numbers as not having an increased likelihood of responding to the cancer treatment regimen, and either (5)(a) recommending, prescribing, initiating or continuing a treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor in the patient diagnosed as having an increased likelihood of responding to the cancer treatment regimen, or (5)(b) recommending, prescribing, initiating or continuing a treatment regimen not comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor in the patient diagnosed as not having an increased likelihood of responding to the cancer treatment regimen.

[0168] Embodiment 26. The test value is derived by calculating the arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in the sample as follows: Test Value = (Number of Indicator LOH Regions) + (Number of Indicator TAI Regions) + (Number of Indicator LST Regions) 3 26. The method of any one of embodiments 18-25, wherein the one or more reference values ​​are derived by calculating an arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in samples from the reference population as follows: Test Value = (Number of Indicator LOH Regions) + (Number of Indicator TAI Regions) + (Number of Indicator LST Regions) 3

[0169] Embodiment 27. The method of any one of embodiments 18-26, comprising diagnosing a patient in a sample in which the test value is at least 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, or at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater, or at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% greater than the one or more reference numbers as having an increased likelihood of responding to the cancer treatment regimen.

[0170] Embodiment 28. A method of treating a cancer patient, comprising: (1) determining the number of indicator LOH regions, indicator TI regions, and indicator CA regions including indicator LST regions in at least one pair of human chromosomes in the cancer cells of the cancer patient in a sample containing cancer cells; (2) providing a test value derived from the number of such indicator CA regions; and (3) comparing the test value to one or more reference values ​​derived from the number of indicator CA regions in a reference population; and (4)(a) recommending, prescribing, initiating, or continuing a therapeutic regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor in a patient whose sample has a test value that is greater than at least one such reference value; or (4)(b) in a patient whose sample has a test value not greater than at least one of the reference values, either recommending, prescribing, initiating, or continuing a treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor.

[0171] Embodiment 29. The method of embodiment 28, wherein said at least one pair of human chromosomes represents the entire genome.

[0172] Embodiment 30 The method of embodiment 28 or embodiment 29, wherein the indicator CA region is determined in at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 pairs of human chromosomes.

[0173] Embodiment 31. The method of any one of embodiments 28-30, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0174] 32. The number of references in the indicator LOH region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more, and the number of references in the indicator TAI region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more. , 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more and the number of references in the indicator LST region is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more.

[0175] Embodiment 33. The indicator LOH region is defined as an LOH region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length, but less than either an entire chromosome or an entire chromosome arm, and the indicator TAI region is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 33. The method of any one of embodiments 28-32, wherein the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length but does not extend across the centromere, and the indicator LST region is defined as an LST region that is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 30, 35, 40, 45, 50 or more megabases in length.

[0176] Embodiment 34. The method of any one of embodiments 28-33, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0177] Embodiment 35. The test value is derived by calculating the arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in the sample as follows: Test Value = (Number of Indicator LOH Regions) + (Number of Indicator TAI Regions) + (Number of Indicator LST Regions) 3 The method of any one of embodiments 28-34, wherein the one or more reference values ​​are derived by calculating an arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in samples from the reference population as follows: Test Value = (Number of Indicator LOH Regions) + (Number of Indicator TAI Regions) + (Number of Indicator LST Regions) 3

[0178] Embodiment 36. The method of any one of embodiments 28-35, comprising diagnosing a patient in a sample in which the test value is at least 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, or at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater, or at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% greater than the one or more reference numbers as having an increased likelihood of responding to the cancer treatment regimen.

[0179] Embodiment 37. A method for assessing HRD in a cancer cell or its genomic DNA, the method comprising: (a) detecting, in a cancer cell or genomic DNA derived therefrom, an indicator CA region in at least one pair of human chromosomes of the cancer cell, wherein the at least one pair of human chromosomes is not the human X / Y sex chromosome pair; (b) determining the total number of indicator CA regions in the at least one pair of human chromosomes.

[0180] Embodiment 38. A method for predicting the status of BRCA1 and BRCA2 genes in cancer cells, comprising: determining, in a cancer cell, the total number of indicator CA regions in at least one pair of human chromosomes in the cancer cell; and diagnosing a patient in which the total number in cancer cells is greater than a reference number as having an increased likelihood of a defect in the BRCA1 or BRCA2 gene.

[0181] Embodiment 39. A method for predicting HDR status in cancer cells, comprising: determining, in a cancer cell, the total number of indicator CA regions in at least one pair of human chromosomes in the cancer cell; and diagnosing a patient in which the total number is greater than the reference number in cancer cells as having an increased likelihood of HDR deficiency.

[0182] Embodiment 40. A method of predicting a cancer patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: determining, in a cancer cell from the cancer patient, the number of indicator CA regions in at least one pair of human chromosomes in the cancer cell of the cancer patient; and diagnosing a patient in which the total number of cancer cells is greater than a reference number as having an increased likelihood of responding to the cancer treatment regimen.

[0183] Embodiment 41. A method for predicting a cancer patient's response to a treatment regimen, comprising: determining, in a cancer cell from the cancer patient, the total number of indicator CA regions in at least one pair of human chromosomes in the cancer cell from the cancer patient; and diagnosing a patient in which the count in cancer cells is greater than a reference count as having an increased likelihood of not responding to a treatment regimen comprising paclitaxel or docetaxel.

[0184] Embodiment 42. A method of treating cancer, comprising: (a) determining, in a cancer cell from a cancer patient or in genomic DNA obtained therefrom, the total number of indicator CA regions in at least one pair of human chromosomes in the cancer cell; (b) administering to the cancer patient a cancer treatment regimen comprising one or more drugs selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor if the total number of indicator CA regions is greater than a reference number.

[0185] Embodiment 43. Use of one or more drugs selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors for the manufacture of a medicament useful for treating cancer in a patient identified as having cancer cells determined to have a total of five or more indicator CA regions.

[0186] Embodiment 44. A system for determining the LOH status of cancer cells in a cancer patient, comprising: (a) a sample analyzer configured to generate a plurality of signals for genomic DNA of at least one pair of human chromosomes of the cancer cell; (b) a computer subsystem programmed to calculate a number of indicator CA regions in the at least one pair of human chromosomes based on the plurality of signals.

[0187]

[0036] Embodiment 45. The computer subsystem compares the number of indicator CA regions to a reference number. (a) the possibility of a defect in the BRCA1 and / or BRCA2 genes in the cancer cells; (b) the possibility of HDR deficiency in the cancer cells; or (c) The system of embodiment 8, wherein the system is programmed to determine the likelihood that the cancer patient will respond to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor.

[0188] Embodiment 46. When executed on a computer, detecting the presence or absence of any indicator CA region along one or more of the human chromosomes; determining a total number of said indicator CA regions in said one or more chromosome pairs.

[0189] Embodiment 47. A diagnostic kit comprising: at least 500 oligonucleotides capable of hybridizing to multiple polymorphic regions of human genomic DNA; A diagnostic kit comprising the computer program product of embodiment 10.

[0190] Embodiment 48. The use of a plurality of oligonucleotides capable of hybridizing to a plurality of polymorphic regions of human genomic DNA to produce a diagnostic kit useful for determining the total number of indicator CA regions in at least one chromosome pair of human cancer cells obtained from a cancer patient; and (a) an increased likelihood of a defect in the BRCA1 or BRCA2 gene in the cancer cells; (b) an increased likelihood of HDR deficiency in the cancer cells; or (c) Use to detect an increased likelihood that a cancer patient will respond to a cancer treatment regimen that includes a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor.

[0191] Embodiment 49. The method of any one of embodiments 37 to 42, wherein the indicator CA regions are indicator LOH regions, indicator TAI regions, and indicator LST regions, optionally determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0192] 50. The method of any one of embodiments 36-42, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0193] Embodiment 51 The method of any one of embodiments 36-42, wherein the total number of indicator LOH regions, indicator TI regions, or indicator LST regions is 9, 15, 20, or more.

[0194] Embodiment 52 The method of any one of embodiments 36-42, wherein the indicator LOH region, indicator TAI region, or indicator LST region is defined as having a length of about 6, 12, or 15 megabases or more.

[0195] Embodiment 53. The method of any one of embodiments 36 to 42, wherein the reference number is 6, 7, 8, 9, 10, 11, 12, or 13 or more.

[0196] Embodiment 54. The use of embodiment 43 or 48, wherein the indicator CA regions are indicator LOH regions, indicator TAI regions, and indicator LST regions, optionally determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0197] Embodiment 55. The use of embodiment 43 or 48, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0198] Embodiment 56. The use of embodiment 43 or 48, wherein the total number of indicator LOH regions, indicator TI regions, or indicator LST regions is 9, 15, 20 or more.

[0199] Embodiment 57. The use of embodiment 43 or 48, wherein the indicator LOH region, indicator TAI region, or indicator LST region is defined as having a length of about 6, 12, or 15 megabases or more.

[0200] Embodiment 58. The system of embodiment 44 or 45, wherein the indicator CA regions are indicator LOH regions, indicator TAI regions, and indicator LST regions, optionally determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0201] Embodiment 59. The system of embodiment 44 or 45, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0202] Embodiment 60. The system of embodiment 44 or 45, wherein the total number of indicator LOH regions, indicator TI regions, or indicator LST regions is 9, 15, 20, or more.

[0203] Embodiment 61. The system of embodiment 44 or 45, wherein the indicator LOH region, indicator TAI region, or indicator LST region is defined as having a length of about 6, 12, or 15 megabases or more.

[0204] Embodiment 62. The computer program product of embodiment 46, wherein the indicator CA regions are indicator LOH regions, indicator TAI regions, and indicator LST regions, optionally determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0205] Embodiment 63. The computer program product of embodiment 46, wherein the cancer cells are ovarian cancer cells, breast cancer cells, or esophageal cancer cells.

[0206] Embodiment 64. The computer program product of embodiment 46, wherein the total number of indicator LOH regions, indicator TI regions, or indicator LST regions is 9, 15, 20, or more.

[0207] Embodiment 65. The computer program product of embodiment 46, wherein the indicator LOH region, indicator TAI region, or indicator LST region is defined as having a length of about 6, 12, or 15 megabases or more.

[0208] Embodiment 66 The method of any one of embodiments 36-42, wherein the at least one pair of human chromosomes is not human chromosome 17.

[0209] Embodiment 67. The use of embodiment 43 or 48, wherein the indicator CA region is not in human chromosome 17.

[0210] Embodiment 68. The system of embodiment 44 or 45, wherein the indicator CA region is not on human chromosome 17.

[0211] Embodiment 69. The computer program product of embodiment 46, wherein the indicator CA region is not in human chromosome 17.

[0212] Embodiment 70. The method of embodiment 40 or 42, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0213] Embodiment 71. The use of embodiment 48, wherein the DNA damaging agent is a platinum-based chemotherapy drug, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0214] Embodiment 72. The system of embodiment 45, wherein the DNA damaging agent is a platinum-based chemotherapy drug, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0215] Embodiment 73. The computer program product of embodiment 46, wherein the DNA damaging agent is a platinum-based chemotherapy drug, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0216] Embodiment 74. A method comprising: (a) detecting indicator CA regions in cancer cells or genomic DNA derived therefrom, the indicator CA regions including at least two types selected from indicator LOH regions, indicator TI regions, and indicator LST regions in a representative number of pairs of human chromosomes in the cancer cells; (b) determining the number and size of the indicator CA regions.

[0217] Embodiment 75. The method of embodiment 74, wherein said representative number of pairs of human chromosomes represents the entire genome.

[0218] Embodiment 76 The method of embodiment 74, further comprising correlating an increase in the number of indicator CA regions of a particular size with an increased likelihood of HDR loss.

[0219] Embodiment 77. The method of embodiment 76, wherein the specified size is greater than about 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 75, or 100 megabases and less than the length of the entire chromosome containing the indicator CA region.

[0220] Embodiment 78. The method of any of embodiments 76 or 77, wherein 6, 7, 8, 9, 10, 11, 12, or 13 or more indicator CA regions of said particular size correlate with an increased likelihood of HDR loss.

[0221] Embodiment 79. A method for determining the prognosis of a cancer patient, comprising: (a) determining whether a sample comprising cancer cells has an HRD signature, wherein the presence of indicator CA regions in excess of a reference number, including at least two selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes of the cancer cells of the cancer patient indicates that the cancer cells have an HRD signature; (b)(1) diagnosing a patient in whom an HRD signature is detected in a sample as having a relatively good prognosis; or (b)(2) diagnosing a patient in whom the HRD signature is not detected in the sample as having a relatively poor prognosis.

[0222] Embodiment 80. A composition comprising a therapeutic agent selected from the group consisting of DNA damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors for use in treating a disease cancer selected from the group consisting of breast cancer, ovarian cancer, liver cancer, esophageal cancer, lung cancer, head and neck cancer, prostate cancer, colon cancer, rectal cancer, colorectal cancer, and pancreatic cancer in a patient having indicator CA regions in excess of the reference number in at least one pair of human chromosomes in the patient's cancer cells.

[0223] Embodiment 81 The composition of embodiment 80, wherein the indicator CA regions are determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0224] Embodiment 82. The composition of embodiment 80, wherein the total number of indicator CA regions is 9, 15, 20 or more.

[0225] Embodiment 83 The composition of embodiment 80, wherein said first length is greater than or equal to about 6, 12, or 15 megabases.

[0226] Embodiment 84. The composition of embodiment 80, wherein the reference number is 6, 7, 8, 9, 10, 11, 12, or 13 or more.

[0227] Embodiment 85. A method of treating cancer in a patient, comprising: determining the number of indicator CA regions, including at least two of indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in the cancer cells of the cancer patient, which indicates that the cancer cells have an HRD signature; providing a test value derived from the number of said indicator CA regions; comparing the test value to one or more reference values ​​(e.g., mean, median, tertile, quartile, quintile, etc.) derived from the number of indicator CA regions in a reference population; administering an anti-cancer drug to the patient or recommending, prescribing, or initiating a treatment regimen including chemotherapy and / or a synthetic lethal agent based at least in part on the comparing step revealing that the test value is greater than at least one of the reference values ​​(e.g., at least 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater); or and recommending, prescribing, or initiating a treatment regimen that does not include chemotherapy and / or synthetic lethal agents based at least in part on the comparing step revealing that the test value is not greater than at least one of the reference values ​​(e.g., no more than 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater).

[0228] Embodiment 86 The method of embodiment 85, wherein the indicator CA regions are determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0229] Embodiment 87. The method of embodiment 85, wherein the total number of indicator CA regions is 9, 15, 20 or more.

[0230] Embodiment 88 The method of embodiment 85, wherein said first length is greater than or equal to about 6, 12, or 15 megabases.

[0231] Embodiment 89. The method of embodiment 85, wherein the reference number is 6, 7, 8, 9, 10, 11, 12, or 13 or more.

[0232] Embodiment 90. The method of embodiment 85, wherein said chemotherapy is selected from the group consisting of DNA damaging agents, anthracyclines, and topoisomerase I inhibitors, and / or said synthetic lethal agent is a PARP inhibitor drug.

[0233] Embodiment 91. The method of embodiment 85, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, and / or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0234] Embodiment 92. A method for assessing HRD in a cancer cell or its genomic DNA, the method comprising: (a) detecting an indicator CA region in a cancer cell or genomic DNA derived therefrom, the indicator CA region including at least two of an indicator LOH region, an indicator TAI region, or an indicator LST region in at least one pair of human chromosomes in the cancer cell, the at least one pair of human chromosomes being other than the human X / Y sex chromosome pair; (b) determining an average (e.g., arithmetic mean) over the total number of indicator CA regions by calculating the average of the number of each type of indicator CA region detected in the at least one pair of human chromosomes (e.g., for 16 indicator LOH regions and 18 indicator LST regions, the arithmetic mean is calculated to be 17).

[0235] Embodiment 93. A method for predicting the status of BRCA1 and BRCA2 genes in cancer cells, comprising: determining an average (e.g., arithmetic mean) of the total number of indicator CA regions of each type, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in the cancer cell; and correlating the average (e.g., the arithmetic mean) across total numbers greater than the reference number with an increased likelihood of a defect in the BRCA1 or BRCA2 gene.

[0236] Embodiment 94. A method for predicting HDR status in cancer cells, comprising: determining an average (e.g., arithmetic mean) of the total number of indicator CA regions of each type, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in the cancer cell; and correlating the average (e.g., arithmetic mean) over a total number greater than the reference number with an increased likelihood of a defect in the HDR.

[0237] Embodiment 95. A method for predicting a cancer patient response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and / or a PARP inhibitor, the method comprising: determining an average (e.g., arithmetic mean) of the total number of each type of indicator CA region, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in the sample including cancer cells (e.g., for 16 indicator LOH regions and 18 indicator LST regions, the arithmetic mean is determined to be 17); and diagnosing patients in which the average (e.g., arithmetic mean) across total numbers in the samples is greater than a reference number as having an increased likelihood of responding to the cancer treatment regimen.

[0238] Embodiment 96. A method for predicting a cancer patient response to a treatment regimen, comprising: determining an average (e.g., arithmetic mean) over a total number of indicator CA regions, including at least two selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in a patient sample comprising cancer cells; and diagnosing patients in which the average (e.g., arithmetic mean) across the total numbers in the samples is greater than a reference number as having an increased likelihood of not responding to a treatment regimen comprising paclitaxel or docetaxel.

[0239] Embodiment 97. A method for treating cancer, comprising: (a) determining an average (e.g., arithmetic mean) of the total number of indicator CA regions of each type, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes of the cancer cells or a patient sample containing genomic DNA obtained therefrom; (b) administering to patients in whom the total number of indicator CA regions in the sample is greater than the reference number a cancer treatment regimen comprising one or more drugs selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor.

[0240] Embodiment 98. The method of embodiment 95 or 97, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0241] Embodiment 99. A composition comprising a therapeutic agent selected from the group consisting of a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, and a PARP inhibitor for use in treating a disease cancer selected from the group consisting of breast cancer, ovarian cancer, liver cancer, esophageal cancer, lung cancer, head and neck cancer, prostate cancer, colon cancer, rectal cancer, colorectal cancer, and pancreatic cancer in a patient having an average (e.g., arithmetic mean) that exceeds a reference number across types of indicator CA regions, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions, in at least one pair of human chromosomes in the patient's cancer cells.

[0242] Embodiment 100. A method of treating cancer in a patient, comprising: determining the average (e.g., arithmetic mean) of the total number of indicator CA regions in at least one pair of human chromosomes of the cancer cells of the cancer patient, indicative of the cancer cells having an HRD signature, in a sample from the patient; providing a test value derived from an average (e.g., an arithmetic mean) over a number of types of indicator CA regions, including at least two types selected from indicator LOH regions, indicator TAI regions, or indicator LST regions; comparing the test value to one or more reference values ​​(e.g., mean, median, tertile, quartile, quintile, etc.) derived from the average (e.g., arithmetic mean) count across indicator CA region types in a reference population; administering an anti-cancer drug to the patient or recommending, prescribing, or initiating a treatment regimen including chemotherapy and / or a synthetic lethal agent based at least in part on the comparing step revealing that the test value is greater than at least one of the reference values ​​(e.g., at least 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater); or and recommending, prescribing, or initiating a treatment regimen that does not include chemotherapy and / or synthetic lethal agents based at least in part on the comparing step revealing that the test value is not greater than at least one of the reference values ​​(e.g., no more than 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-, or 10-fold greater, no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater).

[0243] Embodiment 101. The method of embodiment 100, wherein the average (e.g., arithmetic mean) across types of indicator CA regions is determined in at least 2, 5, 10, or 21 pairs of human chromosomes.

[0244] Embodiment 102. The method of embodiment 100, wherein the chemotherapy is selected from the group consisting of DNA damaging agents, anthracyclines, and topoisomerase I inhibitors, and / or the synthetic lethal agent is a PARP inhibitor drug.

[0245] Embodiment 103. The method of embodiment 100, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, and / or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0246] Embodiment 104 The method of embodiment 1, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0247] Embodiment 105. The method of embodiment 104, wherein the reference number is 42.

[0248] Embodiment 106 The method of embodiment 9, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0249] Embodiment 107. The method of embodiment 106, wherein the reference number is 42.

[0250] Embodiment 108 The method of embodiment 18, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0251] Embodiment 109. The method of embodiment 108, wherein the reference number is 42.

[0252] Embodiment 110. The method of embodiment 28, wherein the reference number is 42.

[0253] Embodiment 111 The method of embodiment 37, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0254] Embodiment 112. The method of embodiment 111, wherein the reference number is 42.

[0255] Embodiment 113 The method of embodiment 38, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0256] Embodiment 114. The method of embodiment 113, wherein the reference number is 42.

[0257] Embodiment 115 The method of embodiment 39, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0258] Embodiment 116. The method of embodiment 115, wherein the reference number is 42.

[0259] Embodiment 117 The method of embodiment 40, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0260] Embodiment 118. The method of embodiment 117, wherein the reference number is 42.

[0261] Embodiment 119 The method of embodiment 41, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0262] Embodiment 120. The method of embodiment 119, wherein the reference number is 42.

[0263] Embodiment 121 The method of embodiment 42, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0264] Embodiment 122. The method of embodiment 121, wherein the reference number is 42.

[0265] Embodiment 123 The method of embodiment 79, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0266] Embodiment 124. The method of embodiment 123, wherein the reference number is 42.

[0267] Embodiment 125 The method of embodiment 85, wherein the indicator CA region is a combination of an indicator LOH region, an indicator TAI region, and an indicator LST region.

[0268] Embodiment 126. The method of embodiment 125, wherein the reference number is 42.

[0269] Embodiment 127. An in vitro method for predicting a patient's response to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor, comprising: (1) determining the number of indicator LOH regions, indicator TI regions, and indicator CA regions including indicator LST regions in at least one pair of human chromosomes in the cancer cells of the cancer patient in a sample containing cancer cells; (2) combining the indicator CA regions to provide a test value as follows: Test Value = (number of indicator LOH regions) + (number of indicator TAI regions) + (number of indicator LST regions); (3) providing a reference value for comparison with the test value.

[0270] Embodiment 128. The method of embodiment 127, wherein the reference value represents the 5th percentile of indicator CA region scores in a training cohort of HDR-deficient patients.

[0271] Embodiment 129. The method of embodiment 127 or embodiment 128, wherein the reference value is 42.

[0272] Embodiment 130. The method of any one of embodiments 127 to 129, further comprising comparing the test value with the reference value.

[0273] Embodiment 131. The method of any one of embodiments 127 to 130, further comprising diagnosing a patient in which the test value in the sample is greater than the reference value as having an increased likelihood of responding to the cancer treatment regimen.

[0274] Embodiment 132. The determining step comprises assaying the sample to determine at least 150, 200, 250, 300, 350, 400, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 1900, 2100, 2200, 2300, 2400, 2500, 3000, 3500, 4000, 4500, 5000, 6000, 7000, 8000, 9000, 10000, 11000, 12000, 13000, 14000, 15000, 16000, 17000, 18000, 19000, 21000, 22000, 23000, 24000, 25000, 26000, 27000, 28000, 29000, 30000, 31000, 32000, 33000, 34000, 35000, 36000, 37000, 38000, 39000, 40000, 41000, 42000, 43000, 44000, 45000, 46000, 47000, 48000, 49000, 50000, 51000, 52000, 450, 500, 600, 700, 800, 900, 1,000, 1,500, 2,000, 2,500, 3,000, 3,500, 4,000, 4,500, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000, 11,000, 12,000, 13,000, 14,000, 15,000, 16,000, 17,000, 18,000, 19,000, 20,000, 25,000, 30,000, 35,000, 40,000, 45,000, 50,000, 60,000, 70,000 132. The method of any one of embodiments 127-131, comprising determining the copy number of each allele for 80,000, 90,000, 100,000, 125,000, 150,000, 175,000, 200,000, 250,000, 300,000, 400,000, 500,000, 600,000, 700,000, 800,000, 900,000, 1,000,000 or more polymorphic genomic loci.

[0275] Embodiment 133 The method of embodiment 132, wherein said determining step comprises assaying said polymorphic genomic loci in at least 10 autosomal pairs.

[0276] Embodiment 134. The method of embodiment 133, wherein said polymorphic genomic loci in 22 autosomal pairs.

[0277] Embodiment 135. The method of any one of embodiments 132-134, wherein said determining step comprises assaying said sample to measure the copy number of each allele for at least 5,000 polymorphic genomic loci in said autosomal pair.

[0278] Embodiment 136. The method of embodiment 135, wherein said determining step comprises assaying said sample to measure the copy number of each allele for at least 10,000 polymorphic genomic loci in said autosomal pair.

[0279] Embodiment 137. The method of embodiment 136, wherein said determining step comprises assaying said sample to measure the copy number of each allele for at least 50,000 polymorphic genomic loci in said autosomal pair.

[0280] Embodiment 138. A method for determining a homologous recombination (HR) deficiency status of triple negative breast cancer (TNBC) cells of a patient, comprising: (1) determining the combined number of loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large state transition (LST) regions in at least one pair of human chromosomes in a sample comprising the patient's TNBC cells; and (2) identifying ER+ BC cancer cells as likely HR deficient if the combined number of LOH, TAI, and LST regions is greater than 32.

[0281] Embodiment 139. The method of embodiment 138, wherein the indicator LOH region is greater than 1.5 megabases in length but less than the full length of the respective chromosome in which the LOH region is located.

[0282] Embodiment 140 The method of embodiment 139, wherein the indicator LOH region is at least 10 megabases in length.

[0283] Embodiment 141 The method of embodiment 139, wherein the indicator LOH region is at least 15 megabases in length.

[0284] Embodiment 142. The method of any one of embodiments 138 to 141, wherein the indicator TAI region is a region that (i) extends to one of the subtelomeres, (ii) does not cross the centromere, and (iii) has an allelic imbalance greater than 1.5 megabases in length.

[0285] Embodiment 143 The method of any one of embodiments 142, wherein the indicator TAI region is at least 10 megabases in length.

[0286] Embodiment 144. The method of any one of embodiments 138 to 143, wherein the indicator LST region is a region that contains a somatic copy number breakpoint along the length of the chromosome that is between two regions that are at least 10 megabases in length after filtering out regions shorter than 3 megabases in length.

[0287] Embodiment 145. The method of any one of embodiments 138-144, wherein the cancer cells are identified as HR deficient if the combined number is 38 or greater.

[0288] Embodiment 146. The method of any one of embodiments 138-144, wherein the cancer cells are identified as HR deficient if the combined number is 42 or greater.

[0289] Embodiment 147. The method of any one of embodiments 138 to 146, wherein at least one pair of human chromosomes is an autosome.

[0290] Embodiment 148. The method of any one of embodiments 138 to 146, wherein the human chromosomes are autosomes and the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 10 pairs of autosomes.

[0291] Embodiment 149. The method of any one of embodiments 138 to 146, wherein the human chromosomes are autosomes and the number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 15 pairs of autosomes.

[0292] Embodiment 150 The method of any one of embodiments 147-149, further comprising assaying at least 150 polymorphic genomic loci in each autosomal pair.

[0293] Embodiment 151. The method of any one of embodiments 138-146, further comprising assaying at least 5,000 polymorphic genomic loci in at least 20 human chromosomes, wherein the chromosomes are autosomes.

[0294] Embodiment 152. The method of any one of embodiments 138 to 151, further comprising calculating a test value derived from the combined numbers of indicator LOH regions, indicator TAI regions, and indicator LST regions, and identifying the cancer cell as HR deficient if the test value exceeds a reference value, wherein the reference value is derived from a reference number greater than or equal to 33.

[0295] Embodiment 153. The method of embodiment 152, wherein the test value is the arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and the reference value is 8 or greater.

[0296] Embodiment 154. The method of embodiment 152 or 153, wherein the test value is derived by calculating the arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in the sample as follows: Test value = (number of indicator LOH regions) + (number of indicator TAI regions) + (number of indicator LST regions) ÷ 3.

[0297] Embodiment 155. The method of any one of embodiments 138-154, further comprising identifying the patient as likely to respond to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor based on identifying the cancer cells as likely to be HR deficient.

[0298] Embodiment 156. The method of embodiment 155, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubicin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

[0299] Embodiment 157. The method of embodiment 155 or 156, further comprising administering, recommending or prescribing a treatment regimen.

[0300] Embodiment 158 ​​The method of any one of embodiments 138-157, wherein the breast cancer cells are BRAC1 / 2 deficient.

[0301] Embodiment 159. The method of any one of embodiments 138 to 158, wherein the combined number consists of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions.

[0302] The invention is further described in the following examples, which do not limit the scope of the invention described in the claims. EXAMPLES

[0303] Example 1 - LOH and TAI region scores across breast cancer subtypes and association with BRCA1 / 2 loss LOH signatures based on genome-wide tumor LOH profiles have been developed that are highly correlated with defects in BRCA1 / 2 and other HDR pathway genes in ovarian cancer (Abkevich, et al., Patterns of Genomic Loss of Heterozygosity Predict Homologous Recombination Repair Defects, BR. J. CANCER (2012)), which predict response to DNA damaging agents (e.g., platinum-based neoadjuvant) therapy in breast cancer (Telli et al., Homologous Recombination Deficiency (HRD) score predicts response following neoadjuvant platinum-based therapy in triple-negative and BRCA1 / 2 mutation-associated breast cancer (BC), CANCER RES. (2012)). A second score based on the TAI score also showed a strong correlation with BRCA1 / 2 defects and predicted response to platinum therapy in triple-negative breast cancer (Birkbak et al.,Telomeric allelic imbalance indicates defective DNA repair and sensitivity to DNA-damaging agents,CANCER DISCOV.(2012)). This study examined the frequency of BRCA1 / 2 defects and elevated LOH or TAI region scores across breast cancer subtypes, defined by ER / PR / HER2 status.

[0304] Frozen tumors were purchased from three commercial tissue biobanks. Approximately 50 randomly confirmed tumors from each of four breast cancer subtypes (triple negative, ER+ / HER2-, ER- / HER2+, ER+ / HER2+) were selected for analysis. A targeted custom hybridization panel was developed targeting BRCA1, BRCA2, and 50,000 selected SNPs across the entire genome. This panel was used in combination with sequencing on an Illumina HiSeq2500 to analyze tumors for BRCA1 / 2 somatic and germline mutations, including large rearrangements, and SNP allele dosage. BRCA1 promoter methylation was determined by qPCR assay (SA Biosciences). When available, DNA from normal tissue was used to determine whether deleterious mutations were germline or somatic.

[0305] SNP data was analyzed using an algorithm that determines the most likely allele-specific copy number at each SNP position. LOH region scores were calculated by counting the number of LOH regions that were longer than 15 Mb but shorter than the length of a complete chromosome. TAI region scores were calculated by counting the number of telomeric regions with allelic imbalance that were longer than 11 Mb but did not cross the centromere. Samples with low quality SNP data and / or high contamination by normal DNA were excluded. 191 of 213 samples obtained a robust score. [Table 2] [Table 3] [Table 4]

[0306] Figure 5 shows LOH and TAI region scores across breast cancer IHC subtypes. 5A: LOH score, 5B: TAI score. Blue bars: BRCA1 / 2-deficient samples. Red bars: BRCA1 / 2-intact samples. Figure 6 shows the correlation between LOH region scores and TAI region scores (correlation coefficient = 0.69). X-axis: LOH score, Y-axis: TAI score, red dots: intact samples, blue dots: BRCA1 / 2-deficient samples. The area under the dots is proportional to the number of samples with the LOH and TAI score combination (p = 10 -39 ).

[0307] Logistic regression analysis was used to predict BRCA1 / 2 loss based on LOH and TAI scores. Both scores were significant in multivariate analysis (chi-square for LOH was 10.8 and for TAI was 44.7, p=0.001 and 2.3*10 -11 The best model for discriminating between BRCA1 / 2-deficient and intact samples was 0.32*LOH region score + 0.68*TAI region score (p=9*10 -18 ).

[0308] Conclusions: Elevated LOH and TAI region scores are each highly associated with BRCA1 / 2 loss in all subtypes of breast cancer, LOH and TAI region scores are highly significantly correlated, and the combined CA region score (i.e., LOH and TAI combined) shows the best correlation with BRCA1 / 2 loss in this dataset. Based on the present disclosure, the combination of LOH-HRD and TAI-HRD scores can predict response to DNA damaging agents and other agents (e.g., platinum therapy) in triple-negative breast cancer, allowing the expansion of platinum use to other breast cancer subtypes.

[0309] Example 2 - LOH, TAI, and LST region scores across breast cancer subtypes and association with BRCA1 / 2 loss SNP allele frequency ratios were obtained and used to calculate LOH, TAI and LST region scores as described in Example 1. LST score was defined as the number of breakpoints between regions longer than 10 megabases with stable copy number after filtering regions shorter than 3 megabases. We observed that LST score increased with ploidy in both intact and defective samples. Therefore, instead of using a ploidy-specific cutoff in this Example 2, we modified the LST region score by adjusting for ploidy. LSTm=LST-kP, where P is ploidy and k is a constant. Based on multivariate logistic regression analysis with defect as outcome and LST and P as predictors, k=15.5.

[0310] 191 of the 214 samples gave scores that passed the QC criteria used. Thirty-eight of these samples were BRCA1 / 2 defective. The corresponding p-value from the Kolmogorov-Smirnov test of the LOH region scores was 8*10 -12 and the TAI domain score is 2*10 -16 and the LST domain score is 8*10 -8 53 / 191 samples were triple-negative breast cancers, including 22 that were BRCA1 / 2 defective. The corresponding p-values ​​were 6*10 for LOH, TAI, and LST region scores, respectively. -6 , 3*10 -6 , and 0.0002. When performing the same analysis for each individual breast cancer subtype, significant p-values ​​are also seen for all subtypes with at least one of the scores (Table 5). The distribution of scores is shown for BRCA1 / 2-deficient versus BRCA1 / 2-intact samples in Figures 7A-7C.

[0311] The scores were then analyzed to determine whether they correlated (Figures 2D-2F). The correlation coefficient between the LOH region score and the TAI region score was 0.69 (p=10 -39 ), and the relationship between LOH and LST was 0.55 (p=2*10 -19 ), and the correlation between TAI and LST was 0.39 (p=10-9 ) was.

[0312] Logistic regression analysis was used to predict BRCA1 / 2 loss based on LOH, TAI, and LST region scores. All three scores were significant in multivariate analysis (chi-square for LOH was 5.1 (p=0.02) and for TAI was 44.7 (p=2*10 -11 ) and for LST, 5.4 (p=0.02). The best model for distinguishing between BRCA1 / 2-deficient and intact samples in this dataset was 0.21*LOH+0.67*TAI+0.12*LST (p=10 -18 ). This Example 2 extends the conclusions from Example 1 (i.e., a model combining LOH and TAI region scores) to a model combining LOH, TAI, and LST region scores.

[0313] Other clinical data that were available for many samples included stage, grade, and age at diagnosis. Stage information was available for 64 / 191 samples. The correlation coefficients between stage and LOH region score (0.07) and TAI region score (0.1) were not significant. Grade information was available for 164 / 191 samples. The correlation coefficients between grade and LOH region score (0.33) and TAI region score (0.23) were significant (p=2*10 respectively). -5 and 0.004). Age at diagnosis was known for 184 / 191 samples. The correlation coefficient between age and LOH region score (-0.13) was not significant. The correlation coefficient between age and TAI region score (-0.25) was significant (p=0.0009). [Table 5]

[0314] Example 3 - Arithmetic mean LOH, TAI, and LST region scores across breast cancer subtypes and association with BRCA1 / 2 loss The following study shows how the HRD score described herein can predict the efficacy of drugs targeting BRCA1 / 2 deficiency and HR deficiency in triple-negative breast cancer (TNBC). To investigate the rate of BRCA1 / 2 deficiency across breast cancer subtypes, breast tumor samples were assayed for BRCA1 / 2 mutations and promoter methylation. The three HRD scores described in Example 2 were determined for the samples, and then the arithmetic mean of LOH / TAI / LST scores was used to examine the association with BRCA1 / 2 deficiency. Analysis of a neoadjuvant TNBC cohort treated with cisplatin was further examined for the relationship between all three HRD scores and response.

[0315] Invasive breast tumor samples and matched normal tissues were obtained from three commercial suppliers. Samples were selected to represent approximately equal representation of all subtypes of breast cancer, as defined by IHC analysis of ER, PR, and HER2. BRCA1 promoter methylation analysis was performed by qPCR. BRCA1 / 2 mutation screening and genome-wide SNP profiles were generated using custom Agilent SureSelect XT capture and subsequently sequenced on an Illumina HiSeq2500. These data were used to calculate HRD-LOH, HRD-TAI, and HRD-LST scores.

[0316] SNP microarray and clinical data were downloaded from public repositories for the Cisplatin-1 and Cisplatin-2 study cohorts. BRCA1 / 2 mutation data were not available for one of these cohorts. All three HRD scores were calculated using publicly available data and analyzed for association with response to cisplatin. The two cohorts were combined to improve power.

[0317] To calculate the HRD score, the SNP data were analyzed using an algorithm that determines the most likely allele-specific copy number at each SNP position. LOH-LOH was calculated by counting the number of LOH regions that were longer than 15 Mb but shorter than the length of a complete chromosome. HRD-TAI score was calculated by counting the number of regions longer than 11 Mb that had allelic imbalance that extended to one of the subtelomeres but did not cross the centromere. HRD-LST score was the number of breakpoints between regions longer than 10 Mb after filtering out regions shorter than 3 Mb.

[0318] The combined score was the arithmetic mean of the LOH / TAI / LST scores. All p-values ​​were from logistic regression models with BRCA deficiency or response to cisplatin as dependent variables.

[0319] Table 6 shows BRCA1 / 2 mutation and BRCA1 promoter methylation frequencies across four breast cancer subtypes. BRCA1 / 2 variant analysis was successful in 100% of samples, while large-scale rearrangement analysis was less robust with 198 / 214 samples generating data that passed QC criteria. Deleterious mutations were observed in 24 / 214 individuals (one had a somatic mutation in BRCA1 and a germline mutation in BRCA2). Matched normal DNA was available for 23 / 24 variants and was used to determine whether the identified mutation was germline or somatic. BRCA1 promoter methylation analysis was successful in 100% of samples. Figure 9 illustrates the HRD scores in BRCA1 / 2-deficient samples. [Table 6]

[0320] Table 7 shows the association between the three HRD scores and BRCA1 / 2 defects in the all-participants breast cohort. The combined score was the arithmetic mean of the three HRD scores. [Table 7]

[0321] Table 8 shows the association between HRD score and pCR (Miller-Payne 5) in TNBC treated with cisplatin in the neoadjuvant setting. Data were available from samples from Cisplatin-1 (Silver et al., Efficacy of neoadjuvant Cisplatin in triple-negative breast cancer. J.CLIN. ONCOL.28:1145-53(2010)) and Cisplatin-2 (Birkbak et al., (2012)) studies, and pCR was defined as patients with Miller-Payne 5 status after neoadjuvant treatment. HRD combined was the arithmetic mean of the three HRD scores. [Table 8]

[0322] Conclusions: BRCA1 / 2 deficiency and elevated HRD scores were observed in all breast subtypes, and the HRD score detected BRCA1 / 2 deficiency. All three HRD scores predicted / detected response to cisplatin treatment in TNBC. The average (arithmetic mean) of the three HRD scores detected BRCA1 / 2 status in the all-participants breast cohort and detected cisplatin response in a second independent TNBC cohort. The combined HRD arithmetic mean was a stronger predictor / detector of BRCA1 / 2 deficiency or treatment response than the individual HRD scores.

[0323] Example 4 - Multivariate analysis of BRCA1 / 2 status and DNA-based assays for homologous recombination deficiency The previous examples describe DNA-based scores that measure homologous recombination deficiency (HRD), which show that each score is significantly associated with BRCA1 / 2 deficiency, as well as the HRD combined score, defined as the arithmetic mean of three different HRD scores. This example extends the results of the previous examples by examining (1) the association between each of the three scores and the HRD combined score, (2) the association between clinical variables and the HRD combined score, and (3) the association between clinical variables and the HRD combined score with BRCA1 / 2 deficiency.

[0324] Methods: The analysis in this Example 4 includes samples from the same 197 patients as described in the previous Example. Briefly, 215 breast tumor samples were purchased as fresh frozen samples from three commercial suppliers. Samples were selected to give approximately equal representation of breast cancer subtypes according to IHC analysis of ER, PR, and HER2. 198 samples produced reliable HRD scores according to the Kolmogorov-Smirnov quality index. One patient who passed the HRD score was excluded from the analysis due to an abnormal breast cancer subtype (ER / PR+HER2-). The tumor and clinical characteristics of the patients are detailed in Table 9.

[0325] Patient clinical data were provided for 91 variables, but data for most variables were too sparse to be included in the analysis. Breast cancer subtype (TNBC, ER+ / HER2-, ER- / HER2+, ER+ / HER2+) was available for all patients. Other variables considered were age at diagnosis (provided for 196 / 197 patients), stage (provided for 191 / 197 patients), and grade (provided for 190 / 197 patients). [Table 9-1] [Table 9-2]

[0326] BRCA1 / 2 mutation screening and genome-wide SNP profiles were generated using custom Agilent SureSelect XT capture and subsequently sequenced on an Illumina HiSeq2500. Methylation of the BRCA-1 promoter region was determined by qPCR. Samples with >10% methylation were classified as methylated.

[0327] The HRD score was calculated from the genome-wide tumor loss of heterozygosity (LOH) profile (HRD-LOH), telomeric allelic imbalance (HRD-TAI), and large-scale state transitions (HRD-LST), which are three HRD scores combined in the "HRD combined score" discussed in this Example 4.

[0328] BRCA1 / 2 deficiency was defined as loss of function due to BRCA-1 or BRCA-2 mutations, or methylation of the BRCA-1 promoter region, accompanied by loss of heterozygosity (LOH) in the affected genes.

[0329] All statistical analyses were performed using R version 3.0.2. All reported p-values ​​are two-sided. Statistical tools used included Spearman rank-sum correlation, Kruskal-Wallis one-way analysis of variance, and logistic regression.

[0330] For logistic regression modeling, HRD score and age at diagnosis were coded as numerical variables. Breast cancer stage and subtype were coded as categorical variables. Grade was analyzed as both a numerical and categorical variable, but was categorical unless otherwise noted. Coding grade as a numerical value is not appropriate unless the increased odds of BRCA1 / 2 deletion are the same when comparing grade 2 with grade 1 as when comparing grade 3 with grade 2.

[0331] P values ​​reported for univariate logistic regression models are based on partial likelihood ratios. Multivariate p values ​​are based on partial likelihood ratios of the change in deviation from the full model (including all relevant predictors) versus the reduced model (including all predictors except the predictor being evaluated and any interaction terms involving the predictor being evaluated). Odds ratios for HRD scores are reported by interquartile range.

[0332] Results: Pairwise correlations of HRD-LOH, HRD-TAI, and HRD-LST scores were examined graphically (Figure 1) and quantified with Spearman rank-sum correlations. Because right skewness and outliers were observed in the HRD score distributions, Spearman rank-sum correlations were preferred over the more commonly used Pearson product-moment correlations. All pairwise comparisons of scores showed positive correlations that were significantly different from zero (p<10 -16 ).

[0333] The degree of independent BRCA1 / 2 deletion information captured by each of the HRD-LOH, HRD-TAI, and HRD-LST scores was measured by examining a multivariate logistic regression model with all three scores included as predictors of BRCA1 / 2 deletion status (Table 10). The HRD-TAI score captured significant BRCA1 / 2 deletion information independently of that provided by the other two scores (p=0.00016), as did the HRD-LST score (p=0.00014). At the 5% significance level, the HRD-LOH score did not add significant independent BRCA1 / 2 deletion information (p=0.069). [Table 10]

[0334] Table 10 illustrates the results from a three-term multivariate logistic regression model using HRD-LOH, HRD-TAI, and HRD-LST as predictors of BRCA1 / 2 loss.

[0335] To evaluate whether the HRD combined score adequately captured the BRCA1 / 2 deficiency information of its three components, three bivariate logistic regression models were tested. Each model included the HRD combined score and one of the HRD-LOH, HRD-TAI, or HRD-LST scores. None of the component scores added significantly to the HRD combined score at the 5% significance level (HRD-LOH p=0.89, HRD-TAI p=0.090, HRD-LST p=0.28). This suggests that the HRD combined score adequately captured the BRCA1 / 2 deficiency information of the HRD-LOH, HRD TAI, and HRD-LST scores.

[0336] The HRD combined score was finally compared to a model-based combined score optimized to predict BRCA1 / 2 loss in this patient set. The HRD combined score weights each of the HRD-LOH, HRD-TAI, and HRD-LST scores equally, while the model-based score assigns the HRD-TAI score approximately twice the weight of the HRD-LOH or HRD-LST scores. The formula for the model-based score is given by HRD-model = 0.11 × (HRD-LOH) + 0.25 × (HRD-TAI) + 0.12 × (HRD-LST).

[0337] The results of the univariate analysis (Table 11) show that the HRD model score is approximately one order of magnitude better than the HRD combined score (HRD model p=2.5×10 -25 , HRD combination p=1.1×10 -24 ). [Table 11]

[0338] Table 11 shows the results from the univariate logistic regression. Odds ratios for HRD score are reported per IQR of score. Odds ratios for age are reported per year. Odds ratios for grade (numeric) are per unit.

[0339] In a bivariate logistic regression model, the HRD model score did not add significant independent BRCA1 / 2 deficiency information to the HRD combined score (p=0.089), further suggesting that the HRD combined score adequately captures the BRCA1 / 2 deficiency information of the HRD-LOH, HRD-TAI, and HRD-LST scores.

[0340] Associations between clinical variables and the HRD combined score are shown in Figure 12. The HRD combined score correlated significantly with tumor grade (Spearman correlation 0.23, p=0.0017). Correlations with breast cancer stage and age at diagnosis were not significantly different from zero at the 5% level. According to the Kruskal-Wallis one-way ANOVA test, the mean HRD combined score differed significantly between breast cancer subtypes (p=1.6×10 -5 ).

[0341] Heterogeneity of the HRD combined score among clinical subpopulations was tested by examining the significance of interaction terms in multivariate logistic regression models. For each clinical variable, the interaction term between the HRD combined score and the HRD combined score was added to a model including all clinical variables. The interaction term did not reach significance at the 5% significance level. Thus, there is no evidence to suggest that the probability of BRCA1 / 2 deletion conferred by the HRD combined score differs among clinical subpopulations.

[0342] Similar tests for each of the HRD-LOH, HRD-TAI, and HRD-LST scores showed significant interactions of HRD-TAI score with age (p=0.0072) and grade (p=0.015), as well as a significant interaction of HRD-LST score with breast cancer subtype (p=0.021). After adjusting for multiple comparisons, only the interaction of HRD-TAI score with age remained significant at the 5% level (p=0.029). The significance of this interaction suggests that the increase in the odds of BRCA1 / 2 deletion per unit increase in HRD-TAI score decreases as age increases.

[0343] Associations between clinical variables and BRCA1 / 2 defects are displayed in Figure 13. Clinical variables and HRD combined scores were evaluated in univariate (Table 11) and multivariate (Table 12) logistic regression models. Odds ratios for HRD scores are reported by IQR. Odds ratios for age at diagnosis are reported per year. [Table 12]

[0344] Table 12 shows the results of the multivariate logistic regression. Odds ratios for HRD score are reported by the IQR of the score. Odds ratios for age are reported for each year.

[0345] In univariate analysis, each of the HRD scores (HRD-LOH, HRD-TAI, HRD-LST, HRD combined, and HRD-model) was significantly associated with BRCA1 / 2 loss. Higher scores indicated a higher likelihood of loss. Increasing age at diagnosis was significantly associated with a decreased risk of BRCA1 / 2 loss (p=0.0071). Univariate results for breast cancer subtype and tumor grade (both categorical and numerical) were also statistically significant. Cancer stage was not associated with BRCA1 / 2 status.

[0346] In multivariate analyses, we examined models based on the HRD combined score and all available clinical variables. The HRD combined score captured significant BRCA1 / 2 deletion information that was not captured by the clinical variables (p = 1.2 × 10 -16 ). Of the available clinical variables, only age at diagnosis maintained significance in the multivariate setting (p = 0.027). Grade was coded as a categorical variable and was not statistically significant (p = 0.40). Grade was also not significant when coded as a numerical variable (p = 0.28). Quadratic and cubic effects of the HRD combined score were tested in a multivariate model including all clinical variables and were not statistically significant.

[0347] Discussion In this Example 4, the frequency of BRCA1 / 2 deletions across the four subtypes of breast cancer defined by IHC subtyping ranged from about 9% to about 16%. Sequencing of matched tumor and normal DNA samples suggests that about 75% of the observed mutations were of germline origin. The primary method for loss of the second allele in breast cancer is via LOH, however, about 24% of tumors carried subsequent somatic deleterious mutations in the second allele. In addition, apparently sporadic breast tumors were found in one individual with a BRCA2 somatic deleterious mutation.

[0348] All three HRD scores show a strong correlation with BRCA1 / 2 deficiency regardless of subtype, and the frequency of elevated scores suggests that a significant proportion of all breast tumor subtypes have defects in the homologous recombination DNA repair pathway. These findings, especially when combined with the findings of Example 3 above, indicate that agents that target or exploit DNA damage repair (e.g., platinum agents) may prove effective across a subset of tumors from all subtypes of breast cancer (those with homologous recombination deficiencies as detected by the present disclosure).

[0349] Implementation of these HRD scores, alone or in combination, in a clinical setting is best achieved using assays that are compatible with formalin-fixed and paraffin-embedded ("FFPE") core needle biopsies. This type of sample yields very low amounts and poor quality DNA. DNA extracted from these FFPE-processed samples often does not perform well in SNP microarray analysis.

[0350] Liquid hybridization-based target enrichment techniques have been developed for the generation of libraries for next-generation sequencing. These methodologies allow targeted sequencing of regions of interest after genomic complexity reduction, resulting in reduced sequencing costs. Preliminary tests have shown that available assays are compatible with DNA derived from FFPEDNA. In this Example 4, we report the development of a capture panel that targets approximately 54,000 SNPs distributed across the genome. The allele counts from the sequencing information provided by this panel can be used to calculate copy number and LOH reconstruction, as well as all three HRD scores. In addition, as in this Example 4, BRCA1 and BRCA2 capture probes can be included in the panel, which allows high-quality mutation screening for deleterious variants in these genes in the same assay.

[0351] All three scores were significantly correlated with each other, suggesting that they all measure the same core genomic phenomenon, however, logistic regression analysis shows that the scores can be combined, resulting in a stronger association with BRCA1 / 2 defects in this dataset.

[0352] The combination of a robust score capable of identifying tumors with defects in homologous recombination DNA repair, and an assay compatible with formalin-fixed paraffin-embedded clinical pathology specimens, facilitates the diagnostic identification and classification of patients with a high likelihood of responding to agents targeting double-stranded DNA damage repair. In addition, such agents may be useful across all subtypes of breast cancer in which HRD is detected according to the present disclosure.

[0353] Example 5 - High HRD threshold (eg, one example of an HRD signature) This example illustrates the determination of high HRD. The threshold reference value was selected to have high sensitivity for detecting HRD in breast and ovarian tumors that were non-specific for treatment response or outcome. The total number of LOH, TAI, and LST regions was determined. To calculate the HRD score, the SNP data was analyzed using an algorithm that determines the most likely allele-specific copy number at each SNP position. LOH-LOH was calculated by counting the number of LOH regions that were longer than 15 Mb but shorter than the length of a full chromosome. The HRD-TAI score was calculated by counting the number of regions longer than 11 Mb with allelic imbalance that extended to one of the subtelomeres but did not cross the centromere. The HRD-LST score was the number of breakpoints between regions longer than 10 Mb after filtering out regions shorter than 3 Mb. The combined score (HRD score) was the sum of the LOH / TAI / LST scores.

[0354] The training set was assembled from four different cohorts (497 breast and 561 ovarian cases). Because the distribution of HRD scores in BRCA-deficient samples generally represents the distribution of scores in HRD samples, the set consisted of 78 breast and 190 ovarian tumors that lacked functional copies of BRCA1 or BRCA2. The threshold was set at the 5th percentile of the HRD scores in the training set, giving a sensitivity of >95% for detecting HR deficiency. High HRD (or HRD signature) was defined as having a reference score of 42 or higher (Figure 14).

[0355] Example 6 - HRD predicts cisplatin response in triple-negative breast cancer This example shows how the HRD scores described herein can predict the efficacy of drugs targeting HR defects in triple-negative breast cancer (TNBC) samples. Analysis of a neoadjuvant TNBC cohort treated with cisplatin was examined for the relationship between all three HRD scores and response. All p-values ​​were from a logistic regression model with response to cisplatin as the dependent variable.

[0356] HR-deficient status was determined for 62 of the 70 samples (70 individual patients) received from the cisplatin cohort (8 had insufficient tumors for analysis). Of these, 31 (50%) were HR-deficient, 22 (35%) were non-HR-deficient, and 9 (15%) were undetermined. Figure 15 provides a histogram showing the distribution of HRD scores in the cohort. A score of 42 or higher was considered to have high HRD (see also Example 5). The bimodality illustrated in Figure 15 illustrates that the HRD score effectively distinguished between HR-deficient and non-deficient tumor status. Pathological complete response (pCR), associated with long-term survival, defined as 0 residual cancer burden (RBC), was observed in 11 / 59 (19%) samples. Pathological response (PR), defined as 0 or 1 RBC, was observed in 22 / 59 (37%) samples. These overall response rates correlated with expectations for monotherapy.

[0357] Statistical analysis followed a predefined statistical analysis plan (SAP) including primary, secondary, and BRCA wild-type subset analyses.

[0358] The primary analysis used HR deficiency status to predict response in 50 samples. As shown in Table 13, HR deficiency samples provided a better predictor of both PR and pCR response. For example, 52% of HR deficiency samples had a pathological response, in contrast to 9.5% of non-deficient samples having a pathological response. Similarly, 28% of HR deficiency samples had a pathological complete response, in contrast to 0% of non-deficient samples having a pathological complete response. [Table 13]

[0359] A secondary analysis predicted response in 48 samples using the quantitative HRD score described in Example 5. As shown in Table 14, HRD scores were significantly higher in samples from responders, defined as either PR or pCR, than in non-responders. [Table 14]

[0360] The distribution of HRD scores within each class of response for the secondary analysis defined by BRCA mutation status is illustrated in Figure 16, with the dotted line at 42 representing the HRD threshold between low and high scores. The response curves or probabilities of PR associated with each value of the quantitative HRD score for the secondary analysis are illustrated in Figure 17. The curves shown in Figure 17 were modeled by generalized logistic regression, which estimates four parameters: shape, scale, and the lower and upper limits of the curve. The shaded boxes indicate the probability of response in HR-deficient versus non-deficient samples. Table 15 shows that in the secondary analysis, HR status remained significantly associated with pathological response. [Table 15]

[0361] The individual HRD component scores versus pathological response are shown in Table 16 and illustrated in Figure 18. Table 16 shows that each component score (i.e., LOH, TAI, and LST) predicted response and their sum (i.e., HRD score) was as significant or more significant than any of the individual components (HRD p-value=3.1×10−4). Figure 18 illustrates the strong pairwise correlations between the component scores. [Table 16]

[0362] Further examined in secondary analyses was the association of BRCA1 / 2 mutation status with response. Table 17 confirmed that BRCA mutation status was associated with response; however, in this cohort (n=51), the association was not significant and BRCA mutation status was not as predictive as HR deficiency. [Table 17]

[0363] Further, subset analysis was performed using HR deficiency status in 38 BRCA wild-type samples, showing that HR deficiency is predicted in samples without BRCA1 / 2 mutation.As shown in Table 18, HR deficiency samples provide better predictors of both PR and pCR response in BRCA wild-type samples.For example, 52.6% of HR deficiency samples have pathological response, in contrast to 10.5% of non-deficiency samples have pathological response.Similarly, 26.3% of HR deficiency samples have pathological complete response, in contrast to 0% of non-deficiency samples have pathological complete response. [Table 18]

[0364] Subset analysis was further performed using quantitative HRD scores in 38 BRCA wild-type samples. As shown in Table 19, samples with high HRD (score 42 or higher) provided better predictors of both PR and pCR response in BRCA wild-type samples. [Table 19]

[0365] In conclusion, this example shows that the sum of all three HRD scores significantly predicted response to cisplatin treatment in TNBC.

[0366] Example 7 - HRD determination in estrogen receptor positive breast cancer As described herein, the total number of loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large state transition (LST) regions in breast cancer (BC) and ovarian cancer (OC) tumor tissues can be used to determine whether a tumor is likely to be homologous recombination (HR) deficient. This determination is important because patients with homologous HR deficient tumors can benefit from treatment with agents that target the deficient HR pathway, such as DNA damaging agents, anthracyclines, topoisomerase I inhibitors, radiation, and / or PARP inhibitors. Conversely, patients whose tumors are identified as not HR deficient can benefit from treatment with agents that do not target the HR pathway, such as taxane agents or hormonal therapy.

[0367] For patients with ovarian cancer, the FDA-approved threshold for the combined LOH-TAI-LST region to identify HR deficiency is 42, which reflects the 5th percentile of BRCA-deficient tumors (see Example 5). Similarly, a lower threshold for the combined LOH-TAI-LST region can be used for OC. As shown in Figure 16, for example, patients in the complete response group (pCR) or beneficial RCB-I group have HRD scores with a low 1st percentile threshold (i.e., greater than 32) equal to or above a, which is significantly associated with improved outcomes after platinum-based therapy (see also Example 6 and Figure 16, Mol Cancer Res. 2018; 16(7): 1103-11 and Cancers. 2021; 13(5): 946).

[0368] The ability to determine the optimal threshold of combined LOH-TAI-LST regions for different tumor types is important because this threshold can vary between different cancers and even between different cancer subtypes. Triple-negative breast cancer (TNBC) and estrogen receptor positive breast cancer (ER+BC) are the main focus of most breast cancer clinical trials that evaluate outcomes based on HRD status. In this example, we use an exploratory threshold of 33 or more for OC as a comparison to identify separate thresholds for estrogen receptor positive breast cancer (ER+BC) subtypes.

[0369] Briefly, the combined LOH-TAI-LST regions in BRCA-deficient tumors (or "genomic instability scores" or "GIS" as referred to in this example) were determined for newly diagnosed patients at various stages of OC, TNBC, or ER+BC across five cohorts: Abkevich et al. (Br. J. Cancer. 2012;107(10):1776-82), TCGA (Nature. 2012;490(7418):61-70), Timms et al. (Breast Cancer Res. 2014;16(145):1-9), TBCRC008 (J. Nucl. Med. 2015;56(1):31-7), and OlympiAD studies (NEJM. 2017;377(17):1700) (Table 20). That is, GIS was determined as a combination of LOH, TAI, and LST identified via next-generation sequencing-based assays. BRCA deficiency was defined by loss of function due to pathogenic variants in BRCA1 or BRCA2, or methylation of the BRCA1 promoter region, with LOH in the affected gene. GIS distributions in different cancer types and subtypes were compared using the Kolmogorov-Smirnov test. A normal distribution was fitted to GIS in BRCA-deficient ER+BC tumors. The 1st percentile of the fitted distribution was selected as the threshold.

[0370] With reference to Table 20, in all cohorts, BRCA1 / 2 deficiency was defined as loss of function due to BRCA1 or BRCA2 mutations with LOH in the affected gene. In Abkevich et al., TCGA, and Timms et al., deficiency can also be caused by methylation of the BRCA1 promoter region with LOH of BRCA1. A total of 561 OC tumors (190 BRCA deficiency), 118 TNBC tumors (46 BRCA deficiency), and 406 ER+BC tumors (76 BRCA deficiency) were included across the five cohorts (Table 20). [Table 20]

[0371] When score distribution was evaluated for BRCA-deficient tumors, the GIS distribution within ER+BC was significantly higher than that of OC (p=9.6×10 -5 ) and TNBC (p = 2.1 × 10 -4 ) (Figure 19). The 1st percentile of the normal distribution fit in BRCA-negative ER+ BC tumors results in a threshold of 24 (Figure 20). Using a threshold of 24 or higher, for example, 45.1% of ER+ BC tumors (183 / 406, 75 / 76 BRCA-negative, 108 / 330 BRCA-intact) were GIS-positive (Figure 21A). In contrast, the GIS distribution of TNBC was not significantly different from that of OC (p=0.72) (Figure 21B). Using a search threshold of 33 or higher, 64.4% of TNBC tumors (76 / 118, 46 / 46 BRCA-negative, 30 / 72 BRCA-intact) were GIS-positive (Figure 21B).

[0372] When compared with OC, the distribution of GIS in BRCA-deficient tumors was different for ER+BC, but not for TNBC. This indicates that different GIS thresholds are appropriate for breast cancer subtypes and that the GIS thresholds developed for OC are distinguishable from ER+BC. These findings are also consistent with the fact that OC and TNBC are known to share similar cancer development mechanisms (Int. J. Mol Sci. 2016;17(5):759). These data further validate that the 1st percentile of TNBC (i.e., a threshold of 33 or more of combined LOH, TAI, or LST regions) is useful for identifying HRD in TNBC tumors. Similarly, these data indicate that the 1st percentile of ER+BC tumors (i.e., a threshold of 24 or more of combined LOH, TAI, or LST regions) is useful for identifying HRD in ER+BC tumors.

[0373] Example 8: Identifying homologous recombination deficiencies in breast cancer: The distribution of genomic instability scores differs between breast cancer subtypes The present disclosure further evaluated whether the cutoff for ovarian cancer may also be appropriate for the major breast cancer subtypes. To assess this, the genomic instability score (GIS) distributions of BRCA-deficient estrogen receptor positive breast cancer (ER+BC) and triple negative breast cancer (TNBC) were compared to the GIS distribution of BRCA-deficient ovarian cancer. For TNBC, a threshold was set and validated using clinical outcomes.

[0374] Methods: Briefly, ovarian and breast cancer (ER+BC and TNBC) tumors from 10 study cohorts were sequenced to identify BRCA1 / 2 mutations and GIS was calculated.Pathological complete response (pCR) to platinum therapy was assessed in a subset of TNBC samples.

[0375] Tumor Samples: The complete cohort consisted of ovarian and breast cancer tumors (TNBC and ER+) from 10 individual study cohorts (Hennessy et al., The Cancer Genome Atlas Network-Breast, The Cancer Genome Atlas Network-Ovarian, NCT01372579, NCT00148694 / NCT00580333, PrECOG 0105, Timms et al., TBCRC008, TBCRC030, and OlympiAD studies). All included samples had known GIS and were obtained under an Institutional Review Board approved protocol. MyChoice CDx (Myriad Genetics) testing was performed on all samples to determine somatic BRCA1 / BRCA2 status and GIS.

[0376] BRCA1 / BRCA2 Sequencing: BRCA1 and BRCA2 gene mutation detection and single nucleotide polymorphism (SNP) whole genome analysis were performed using custom hybridization capture methods as previously described. BRCA mutation status was defined as a deleterious or suspected deleterious mutation in BRCA1 or BRCA2, regardless of heterozygosity. BRCA wild type (BRCAwt) refers to samples without deleterious mutations or without suspected deleterious mutations in BRCA1 or BRCA2. BRCA deficiency was defined as loss of function resulting from germline or somatic deleterious or suspected deleterious variants in BRCA1 or BRCA2 with loss of heterozygosity in the affected gene, or from multiple deleterious or suspected deleterious mutations in the same BRCA gene. BRCA intact refers to samples that are not BRCA deficient, regardless of BRCA mutation status.

[0377] Genomic instability score: GIS was calculated using an algorithm that combines the LOH, TAI, and LST measurements described herein. Binary GIS status was determined based on whether the GIS score was above or below a threshold of 33 or above or 42 or above.

[0378] Pathological complete response: Pathological complete response (pCR) to neoadjuvant chemotherapy was available for TNBC samples from five cohorts (NCT01372579, NCT00148694 / NCT00580333, PrECOG 0105, TBCRC008, and TBCRC030). pCR status was not available for ER+ samples. In some studies, residual cancer burden (RCB) was used and pCR status was not available. Patients with data on residual cancer burden (RCB) after treatment with platinum therapy were dichotomized into those with pCR (RCB-0) and those with incomplete response (RCB-I / II / III). Patients with RCB-0 who did not receive crossover therapy before surgery and did not finish treatment due to progression or toxicity were considered to have achieved pCR.

[0379] Statistics: All p-values ​​were considered significant at the α=0.05 level. GIS distributions in subsets of samples were compared using the Kolmogorov-Smirnov test. Binary logistic regression was used to measure the ability of binary GIS status (i.e., scores above or below a threshold) to predict pCR status in TNBC tumors. Odds ratios (ORs) with 95% profile likelihood confidence intervals (CIs) and partial likelihood ratio test p-values ​​were reported. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated by comparing binary GIS status and binary pCR status, with pCR events above the threshold considered true positives. Univariate 3-parameter logistic regression models optimized for upper bound, slope, and midpoint were used to estimate the probability of pCR for each GIS value.

[0380] result Ovarian cancer tumors A total of 560 ovarian cancer tumors from two cohorts (Hennessy et al. and The Cancer Genome Atlas Network- Ovarian) were included, of which 20.1% were known to be BRCA-deficient (N=115 / 560, Table 21). Of the BRCA-deficient samples, 67.8% (N=78 / 115) had a pathogenic mutation in BRCA1, 31.3% (N=36 / 115) had a pathogenic mutation in BRCA2, and 0.9% (N=1 / 115) had a pathogenic mutation in both BRCA1 and BRCA2. The GIS distributions of BRCA-deficient and BRCA-intact tumors are shown in Figure 22A. In this analysis, the GIS distributions of BRCA-deficient ovarian cancer samples were used as a comparison to evaluate the GIS distributions in BRCA-deficient ER+ breast cancer and TNBC samples. [Table 21]

[0381] ER+ breast cancer tumors A total of 805 ER+ breast cancer tumors were included from five cohorts (The Cancer Genome Atlas Network-Breast, PrECOG 0105, Timms et al. (Breast Cancer Research. 2014;16(6):1-9), TBCRC008, and OlympiAD studies). Of these, 579 were ER+HER2-, 174 were ER+HER2+, and 52 were ER+ with unknown HER2 status. To determine whether it was appropriate to combine all ER+ breast cancer tumors, the GIS distribution of BRCA-negative tumors for ER+HER2- (N=60) and ER+HER2+ (N=10) was compared. No significant difference was observed between the GIS distribution of ER+HER2- and ER+HER2+BRCA-negative tumors (p=0.88).

[0382] Among ER+ breast cancer tumors, 8.8% (71 / 805) were BRCA-deficient, of which 40.8% (N=29 / 71) had a pathogenic mutation in BRCA1 and 59.2% (N=42 / 71) had a pathogenic mutation in BRCA2. The GIS distributions of BRCA-deficient and BRCA-intact tumors are shown in Figure 22A. A significant difference was observed between the GIS distributions for BRCA-deficient ER+ breast cancer tumors and ovarian cancer tumors (p=0.027, Figure 22B), indicating that a separate threshold should be established for ER+ breast cancer tumors. Potential GIS thresholds will be established in future studies when clinical outcomes of ER+ breast tumors treated with platinum or other DNA damaging agents are available.

[0383] TNBC Tumors A total of 443 TNBC tumors were included from seven cohorts (The Cancer Genome Atlas Network-Breast, NCT01372579, NCT00148694 / NCT00580333, PrECOG 0105, Timms et al. (Breast Cancer Research. 2014;16(6):1-9), TBCRC008, and TBCRC030). Of the 56 (12.6%) BRCA-deficient TNBC tumors, 47 (83.9%) harbored pathogenic mutations in BRCA1, 8 (14.3%) harbored pathogenic mutations in BRCA2, and 1 (1.8%) harbored pathogenic mutations in both BRCA1 and BRCA2. The GIS distribution of BRCA-deficient and BRCA-intact tumors is shown in Figure 22A. When comparing the GIS distribution of BRCA-deficient samples, TNBC tumors were significantly different from ER+ breast cancer tumors (p=0.002, FIG. 22B), but not from ovarian cancer tumors (p=0.49, FIG. 22B), indicating that the same thresholds used for ovarian cancer tumors may also be appropriate for TNBC tumors.

[0384] Clinical validation of thresholds in TNBC GIS thresholds of 42 or more and 33 or more have been previously validated in patients with ovarian cancer. Because the GIS distributions in ovarian and TNBC samples were similar, in this study, the thresholds used for ovarian cancer were applied to TNBC samples. The TNBC clinical validation cohort (following preoperative studies: NCT01372579, NCT00148694 / NCT00580333, PrECOG 0105, TBCRC008, and TBCRC030) included 211 platinum-treated samples (N=55 with pCR), of which 171 were BRCA wild-type (BRCAwt) tumors (N=39 with pCR). The GIS distributions for all TNBC clinical validation samples (full clinical validation cohort) and the subset of BRCAwt samples (BRCAwt clinical validation cohort) are summarized by binary pCR status (i.e., pCR vs. no pCR) in Figures 23A-23B.

[0385] Univariate logistic regression models were used to assess the ability of GIS thresholds of 33 or greater and 42 or greater to independently predict binary pCR status in both the full and BRCAwt clinical validation cohorts. GIS thresholds of 33 or greater and 42 or greater were significant independent predictors of pCR in both the full and BRCAwt clinical validation cohorts. Compared with a GIS threshold of 42 or more, a threshold of 33 or more resulted in larger effect sizes in both the full clinical validation cohort (GIS 33 or more: OR 11.1, 95% CI 3.9 to 47.1, p=2.2×10-7; GIS 42 or more: OR 8.2, 95% CI 3.5 to 22.3, p=5.6×10-8) and the BRCAwt clinical validation cohort (GIS 33 or more: OR 9.4, 95% CI 3.2 to 40.4, p=5.6×10-6; GIS 42 or more: OR 7.0, 95% CI 2.9 to 19.6, p=3.0×10-6).

[0386] A bivariate logistic regression model including both GIS thresholds (≥42 and ≥33) as binary variables was used to assess the ability of thresholds to predict pCR. In the full clinical validation cohort, a GIS threshold of ≥42 was significant (OR 3.6, 95% CI 1.1–15.8, p=0.03), whereas GIS status of ≥33 was not (OR 3.6, 95% CI 0.6–21.0, p=0.15). In the same model fitted to the BRCAwt clinical validation cohort, none of the GIS thresholds were significant (GIS ≥33: OR 3.6, 95% CI 0.6–21.3, p=0.15; GIS ≥42: OR 3.0, 95% CI 0.9–13.7, p=0.07).

[0387] Sensitivity, specificity, PPV, and NPV for given thresholds are reported in Table 22 for GIS thresholds of 33 or greater and 42 or greater. [Table 22]

[0388] A high proportion of samples with pCR events had a GIS of 33 or greater in both the full clinical validation cohort (94.5%, N=52 / 55) and the BRCAwt clinical validation cohort (92.3%, N=36 / 39). The proportion of pCR events captured by the threshold decreased at higher GIS thresholds of 42 or greater (full clinical validation cohort: 89.1%, N=49 / 55; BRCAwt clinical validation cohort: 84.6%, N=33 / 39). Of all samples with pCR events, 5.5% in the full clinical validation cohort and 7.7% in the BRCAwt subset had a GIS between 33 and 42.

[0389] The difference in utility between thresholds of 33 or greater and 42 or greater can also be characterized by the difference in probability of pCR calculated by 3-parameter logistic regression with a continuous GIS predicting binary pCR status (Figure 24). In both the full and BRCAwt clinical validation cohorts, patients with a GIS between 33 and 42 had an intermediate probability of pCR, and a GIS threshold of 33 or greater separated patients with a low probability of response from those with an intermediate to high probability of response. The reverse was true for a GIS threshold of 42 or greater, identifying only those patients with the highest likelihood of response.

[0390] In this study, the GIS distribution of BRCA-deficient tumors was evaluated for two different major breast cancer subtypes. The GIS distribution of BRCA-deficient tumors in ER+ breast cancer was significantly different from that in ovarian cancer, indicating that the GIS thresholds used for ovarian cancer may not be appropriate for ER+ breast cancer. The GIS distribution of BRCA-deficient TNBC tumors in this study was not statistically significantly different from ovarian cancer, and clinical validation analysis demonstrated the ability of GIS thresholds of 33 or higher and 42 or higher to predict platinum-based therapy pCR in a subset of TNBC samples. Taken together, these findings highlight the importance of determining individual thresholds for different cancer lineages and different cancer subtypes.

[0391] Compared with BRCA-deficient ovarian cancer tumors, GIS distribution was significantly different in BRCA-deficient ER+ breast tumors, but not in TNBC tumors. Differences in the underlying biology, and therefore GIS, between BRCA1- and BRCA2-mutated tumors may at least partially explain the observed differences between GIS distributions for TNBC and ER+ breast cancer.

[0392] GIS thresholds of ≥33 and ≥42, set at the 1st and 5th percentiles of BRCA-deficient tumors, respectively, have been previously validated in ovarian cancer. Therefore, both thresholds were evaluated in the TNBC clinical validation cohort. When evaluated in independent analyses, both GIS thresholds of ≥33 and ≥42 were found to significantly predict pCR to platinum therapy, but a larger effect size was observed for a GIS threshold of ≥33 compared to ≥42 (OR 11.1 vs. 8.2). In a bivariate model that evaluated the relationship between the two thresholds (i.e., whether one threshold added significant information to the other) in the full clinical validation cohort, a GIS threshold of ≥42 was significant, whereas a GIS threshold of ≥33 was not. In the BRCAwt clinical validation cohort, none of the GIS thresholds were found to be significant. Although analysis in the full clinical validation cohort showed that a threshold of 42 or greater added significant predictive information to a threshold of 33 or greater, null findings in the BRCAwt analysis suggested that the two GIS thresholds had similar predictive value for pCR. The clinical significance of these inconsistent findings was unclear. Therefore, additional criteria were evaluated to assess the clinical validity of the two thresholds.

[0393] In both the full clinical validation cohort and the BRCAwt clinical validation cohort, a GIS threshold of 42 or more had lower sensitivity but higher specificity than a threshold of 33 or more. When selecting a GIS threshold to identify patients who will benefit from DNA damaging agents (e.g., platinum, PARP inhibitors), it is important to consider the appropriate balance of sensitivity and specificity. A GIS threshold of 42 or more with higher specificity will result in fewer false positives (i.e., fewer patients who will not benefit from treatment who will be classified as HRD positive), but will also result in lower sensitivity and therefore fewer true positives (i.e., fewer patients who will benefit from treatment who will be classified as HRD positive). Of patients who achieved pCR to platinum therapy, 5.5% of patients in the full clinical validation cohort and 7.7% of patients in the BRCAwt cohort were not identified as eligible for treatment using a threshold of 42 or more. In clinical settings, when alternative treatment options are scarce, it may be beneficial to utilize a lower threshold of 33 or more to maximize the identification of eligible patients. The decision to pursue treatment with a DNA damaging agent may then be considered on an individual basis, which may depend on a number of clinical factors.

[0394] When selecting a GIS threshold for a clinical trial, the balance between sensitivity and specificity must also be considered. This is particularly relevant when the eligibility criteria of the study may affect the GIS distribution. For example, a clinical trial with enrollment criteria that enriches for patients with HR-deficient tumors (e.g., BRCA1 / 2 mutated tumors, high-grade and / or serous subtype, platinum-sensitive tumors) will shift the distribution toward higher GIS because patients with BRCA mutated tumors have higher GIS. A higher GIS threshold may seem appropriate based on high specificity alone (i.e., fewer patients benefiting from treatment are classified as HRD positive). However, whether it may be appropriate to prioritize specificity or sensitivity may depend on the study population, or other clinical factors (e.g., first-line treatment, metastatic disease).

[0395] While the present invention has been described in conjunction with its detailed description, it is understood that the above description is intended to be illustrative, and not limiting, of the scope of the invention as defined by the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

1. 1. An in vitro method for determining the homologous recombination (HR) deficiency status of estrogen receptor positive (ER+), BRCA-deficient breast cancer (BC) cells in a patient, comprising: (1) determining the combined number of loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large state transition (LST) regions in at least one pair of human chromosomes in a sample containing the ER+, BRCA-deficient BC cells of the patient; and (2) identifying the ER+, BRCA-deficient BC cancer cells as HR-deficient if the combined number of LOH, TAI, and LST regions is greater than 24; A method comprising:

2. 2. The method of claim 1, wherein the indicator LOH region is greater than 1.5 megabases in length but less than the entire length of the respective chromosome in which the LOH region is located.

3. 3. The method of claim 2, wherein the indicator LOH region is at least 10 megabases in length.

4. 3. The method of claim 2, wherein the indicator LOH region is at least 15 megabases in length.

5. 5. The method of any one of claims 1 to 4, wherein the indicator TAI region is a region that (i) extends to one of the subtelomeres, (ii) does not cross the centromere, and (iii) has an allelic imbalance greater than 1.5 megabases in length.

6. 6. The method of claim 5, wherein the indicator TAI region is at least 10 megabases in length.

7. 5. The method of any one of claims 1 to 4, wherein the indicator LST region is a region that comprises a somatic copy number breakpoint along the length of a chromosome that is between two regions that are at least 10 megabases in length after filtering out regions shorter than 3 megabases in length.

8. The method according to any one of claims 1 to 4, wherein the at least one pair of human chromosomes is an autosome.

9. 5. The method of any one of claims 1 to 4, wherein the human chromosomes are autosomes and the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 10 pairs of the autosomes.

10. 5. The method of any one of claims 1 to 4, wherein the human chromosomes are autosomes, and the number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 15 pairs of autosomes.

11. 11. The method of claim 10, further comprising assaying at least 150 polymorphic genomic loci in each autosome pair.

12. 5. The method of any one of claims 1 to 4, further comprising assaying at least 5,000 polymorphic genomic loci in at least 20 human chromosomes, wherein the chromosomes are autosomes.

13. The method of any one of claims 1 to 4, further comprising calculating a test value derived from the combined numbers of indicator LOH regions, indicator TAI regions, and indicator LST regions.

14. 14. The method of claim 13, wherein the test value is the arithmetic mean of the numbers in the indicator LOH region, indicator TAI region, and the reference value is 8 or greater.

15. 14. The method of claim 13, wherein the test value is derived by calculating the arithmetic mean of the number of Indicator LOH Regions, Indicator TAI Regions, and Indicator LST Regions in the sample as follows: Test value = (number of indicator LOH regions) + (number of indicator TAI regions) + (number of indicator LST regions) ÷ 3.

16. 5. The method of any one of claims 1 to 4, further comprising identifying the patient as responsive to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor based on identifying the cancer cells as HR-deficient.

17. 17. The method of claim 16, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin, the anthracycline is epirubincin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan, or the PARP inhibitor is iniparib, olaparib, or velapirib.

18. The method of any one of claims 1 to 4, wherein the combined number consists of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions.

19. 1. An in vitro method for determining the homologous recombination (HR) deficiency status of BRCA-deficient triple-negative breast cancer (TNBC) cells in a patient, comprising: (1) determining the combined number of loss of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large state transition (LST) regions in at least one pair of human chromosomes in a sample containing the TNBC cells of the patient; and (2) identifying the TNBC cells as HR-deficient if the combined number of LOH, TAI, and LST regions is greater than 33; A method comprising:

20. 20. The method of claim 19, wherein the indicator LOH region is greater than 1.5 megabases in length but less than the entire length of the respective chromosome in which the LOH region is located.

21. 21. The method of claim 20, wherein the indicator LOH region is at least 10 megabases in length.

22. 21. The method of claim 20, wherein the indicator LOH region is at least 15 megabases in length.

23. 23. The method of any one of claims 19 to 22, wherein the indicator TAI region is a region that (i) extends to one of the subtelomeres, (ii) does not cross the centromere, and (iii) has an allelic imbalance greater than 1.5 megabases in length.

24. 24. The method of claim 23, wherein the indicator TAI region is at least 10 megabases in length.

25. 23. The method of any one of claims 19 to 22, wherein the indicator LST region is a region that comprises a somatic copy number breakpoint along the length of a chromosome that is between two regions that are at least 10 megabases in length after filtering out regions shorter than 3 megabases in length.

26. 23. The method of any one of claims 19 to 22, wherein the cancer cells are identified as HR deficient if the combined number is 38 or greater.

27. 23. The method of any one of claims 19 to 22, wherein the cancer cells are identified as HR deficient if the combined number is 42 or greater.

28. The method according to any one of claims 19 to 22, wherein the at least one pair of human chromosomes are autosomes.

29. 23. The method of any one of claims 19 to 22, wherein the human chromosomes are autosomes and the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 10 pairs of the autosomes.

30. 23. The method of any one of claims 19 to 22, wherein the human chromosomes are autosomes and the number of indicator LOH regions, indicator TAI regions, and indicator LST regions is determined in at least 15 pairs of autosomes.

31. 23. The method of any one of claims 19 to 22, further comprising assaying at least 150 polymorphic genomic loci in each autosome pair.

32. 23. The method of any one of claims 19 to 22, further comprising assaying at least 5,000 polymorphic genomic loci in at least 20 human chromosomes, wherein the chromosomes are autosomes.

33. 23. The method of any one of claims 19 to 22, further comprising: calculating a test value derived from the combined numbers of indicator LOH regions, indicator TAI regions, and indicator LST regions; and identifying the TNBC cells as HR-deficient if the test value exceeds a reference value, wherein the reference value is derived from a reference number of 33 or greater.

34. 34. The method of claim 33, wherein the test value is the arithmetic mean of the numbers in the indicator LOH region, indicator TAI region, and the reference value is 8 or greater.

35. 34. The method of claim 33, wherein the test value is derived by calculating the arithmetic mean of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions in the sample as follows: Test value = (number of indicator LOH regions) + (number of indicator TAI regions) + (number of indicator LST regions) ÷ 3.

36. 23. The method of any one of claims 19 to 22, further comprising identifying the patient as responsive to a cancer treatment regimen comprising a DNA damaging agent, an anthracycline, a topoisomerase I inhibitor, or a PARP inhibitor based on identifying the cancer cells as HR-deficient.

37. 37. The method of claim 36, wherein the DNA damaging agent is cisplatin, carboplatin, oxalaplatin, or picoplatin; the anthracycline is epirubicin or doxorubicin; the topoisomerase I inhibitor is campothecin, topotecan, or irinotecan; or the PARP inhibitor is iniparib, olaparib, or velapirib.

38. 23. The method of any one of claims 19 to 22, wherein the combined number consists of the number of indicator LOH regions, indicator TAI regions, and indicator LST regions.

Images