Method and apparatus for evaluating a patient's response to therapy

Abstract

The present invention relates to monitoring patient response to therapy. To facilitate monitoring of patient response to therapy, methods are provided that use similar genetic changes over time to identify groups of trends in each of the genetic changes and analyze those groups to determine patient response to therapy.

Description

Technical Field

[0001] The present invention relates to monitoring a patient's response to cancer therapy, and more particularly to a method, apparatus, and system, computer program element, and computer-readable medium for monitoring a patient's response to cancer therapy.

Background Art

[0002] Oncologists typically use images created by computed tomography (CT) and / or magnetic resonance imaging (MRI) scans to guide treatment, compare the tumor to its pre-treatment state, and assist the physician in evaluating the patient's response to treatment. These imaging methods have the advantage of providing the location and shape of the tumor (e.g., with respect to other organs), but because the resolution of medical imaging is generally in the millimeter range, it is difficult to evaluate early changes in tumor volume. Monitoring of the therapy response of cancer patients is also performed by analyzing a patient's sample, such as a liquid biopsy including blood or other body fluids, to track the occurrence of specific mutations over time during treatment. Since tumors carry various mutations, a wide panel covering many different mutations is examined. As a result, it is very difficult to monitor the patient's response to therapy because all of those mutations need to be tracked and analyzed.

Summary of the Invention

Problems to be Solved by the Invention

[0003] It is necessary to facilitate monitoring of the patient's response to therapy.

[0004] The object of the present invention is solved by the subject matter of the independent claims, and further embodiments are incorporated in the dependent claims. It should be noted that the aspects described below of the present invention are further applicable to methods, apparatuses, systems, computer program elements, and computer-readable media.

Means for Solving the Problems

[0005] A first aspect of the present invention is a computer-implemented method for monitoring a patient's response to cancer therapy, comprising: a) receiving data indicating the occurrence of genetic changes over time in a patient, each genetic change being associated with a common disease or group of common diseases, the data being obtained based on the analysis of biomarkers in a patient sample; b) grouping the trends in each of the genetic changes into a plurality of trend groups according to the similarity between trends such that each trend group has a group-specific temporal behavior pattern, each trend group including trends representing similar temporal behavior of the allele frequency of the genetic change, i.e., AF, and a similarity measure quantifying the similarity between trends within the same trend group being within a predetermined range; c) analyzing the group-specific temporal behavior patterns of the plurality of trend groups to monitor the patient's response to cancer therapy. The invention relates to a computer-implemented method.

[0006] Since cancer is a highly heterogeneous disease, cancer patients carry many different mutations in their DNA. Thus, for example, to monitor responses to therapies across a wide patient population using liquid biopsies, a panel that covers many different mutations is required. As a result, it becomes very difficult to monitor a patient's response to therapy because all of those mutations must be tracked.

[0007] To address this problem, even when a complex panel with many different mutations (e.g., about 40 different mutations) is used, a simple model that reduces its complexity is proposed in the present disclosure. To monitor the therapy response of cancer patients to therapies such as drug therapy, radiation therapy, serum therapy, and / or surgical procedures, one or more biomarkers are extracted from a patient's sample such as blood or other body fluids. Examples of biomarkers include circulating tumor DNA (ct-DNA) generated from dead (e.g., necrotic or apoptotic) tumor cells and circulating tumor cells (CTCs) flowing in the bloodstream from primary tumors and metastases. Currently, ct-DNA and CTCs are the clinically optimal biomarkers used to monitor therapy response. Based on the analysis results of one or more biomarkers in a blood sample, the occurrence of genetic changes over time in an individual patient is determined and stored in a dataset. Each of the genetic changes is associated with a common disease or a group of common diseases. That is, various genetic changes associated with a disease such as breast cancer or a group of common diseases are tracked.

[0008] Genetic changes are also called mutations, i.e., changes in the gene sequence. Mutations include changes as small as the substitution of one nucleotide base for another single DNA building block or nucleotide base. On the other hand, larger mutations can affect many genes on a chromosome. Along with substitutions, mutations can also be caused by insertions, deletions, or amplifications of the DNA sequence. Some mutations are genetic because they are passed from parents carrying the mutation in the germline to offspring, meaning that the genetic changes of the present disclosure are related to non-genetic mutations that occur in cells outside the germline, called somatic mutations. For example, some powerful somatic mutations can cause cancers that affect the survival of a single organism.

[0009] When it is determined that genetic changes occur over time in individual patients, classification methods such as cluster analysis of genetic changes or self-organizing maps are used to identify groups of tendencies that tend to have similar genetic changes over time, i.e., group-specific temporal behaviors of genetic changes. Each tendency group includes a tendency representing a similar temporal behavior of the allele frequency of genetic changes. The allele frequency represents the occurrence of gene variants in a population. Alleles are diverse forms of genes located at the same position or locus on a chromosome. The allele frequency is calculated by dividing the number of times an allele is observed in a population by the total number of replicates of all alleles at a particular locus in that population. The allele frequency is expressed as a decimal value, a percentage value, or a fractional value. Changes in allele frequency over time may indicate that genetic drift has occurred or that a new mutation has been induced. Also, in relation to changes in population genetic germline variants, it should be noted that the term "allele frequency" is also referred to as, for example, the variant allele frequency (VAF) of variants that give rise to traits such as hair and eye color. In relation to tumor mutations, the term "allele frequency" is also referred to as the mutant allele frequency (MAF). Since the genetic changes in the present disclosure are related to tumor mutations, the allele frequency in the present disclosure is also referred to as MAF.

[0010] Thereafter, to monitor the patient's response to cancer therapy, the group-specific temporal behavior patterns of these identified tendency groups are further analyzed. That is, these identified tendency groups are associated with characteristics related to the patient's response to cancer therapy. Exemplary group-specific temporal behavior patterns are described below, particularly in relation to the embodiments shown in FIGS. 4A to 4C. That is, it is not necessary to analyze a large dataset of genetic changes being tracked. Rather, only the tendency groups are analyzed to determine the patient's response to cancer therapy. Since the number of tendency groups is significantly less than the number of genetic changes being monitored, the complexity is reduced, even when using a complex panel with many different mutations, for example, more than 40 mutations.

[0011] Also, it should be noted that this analysis yields technical results, or more narrowly, technical intermediate results that are useful to the physician when later reaching a diagnosis. That is, analysis step c) does not actually make a decision without the need for the physician's expertise.

[0012] According to one embodiment of the present invention, in step b), the trends are (i) a method of performing cluster analysis to identify different trend groups of trends, (ii) a method of using a self-organizing map to identify different trend groups of trends (iii) a method of comparing trends using a predetermined trend grouping rule to associate each trend with a respective trend group and are grouped using at least one of them.

[0013] This will be described below with respect to the exemplary embodiment particularly shown in FIG. 1.

[0014] According to a first aspect of the present invention, step c) further comprises a step of dividing the trend groups into a plurality of trend classes based on trend classification rules, wherein each trend class is associated with the respective patient's response to treatment such that each trend class includes a trend group having a trend representing the same patient's response to treatment, and the trend classification rules define and divide the trend classes with respect to the temporal behavior of the genetic changes, a dividing step; a step of applying a trend score assignment rule to assign a trend score to each of the trend groups based on the associated trend class, wherein the trend score assignment rule defines the trend score with respect to the trend class, an applying step; a step of determining the worst-case trend score of the genetic changes so as to be used to determine the response state of the cancer patient.

[0015] That is, it is proposed to classify the trends in each of the genetic changes over time and summarize them into a single measure of treatment response. More specifically, trend classification rules are used to identify trend classes of trend groups that have similar trend effects on the patient's response to treatment, such as trends representing positive or negative tendencies. Therefore, the number of trend classes representing trend effects is significantly less than the number of trend groups and, of course, the number of genetic changes being monitored for a disease, such as breast cancer. Therefore, according to the trend score assignment rules, those trend classes are assigned trend scores. For each individual patient, the worst-case trend score is calculated. The worst-case trend score corresponds to the trend score representing the most negative trend effect in the patient's response to treatment. Finally, the patient is evaluated or stratified according to the calculated worst-case trend score.

[0016] In this way, for each individual patient, a highly reliable and easily accessible distinction is achieved between different patient responses to treatment, such as complete remission, partial response, non-response (also called stable disease), or progressive disease. The calculation of the worst-case trend score for each individual patient from the monitored biomarker is performed without human supervision or intervention.

[0017] According to one embodiment of the present invention, the trend classification rules classify the trends into (i) a category that classifies an AF that has dropped to 0 and does not rise thereafter as the trend class of complete remission CR, (ii) a category that classifies an AF that has dropped below a predetermined percentage value of the initial value and maintains a value below that for a certain period of time as the trend class of partial response PR, (iii) a category that classifies an AF that has risen from the initial value to a predetermined threshold and maintains a state above the predetermined threshold towards the end of the time period as the trend class of progressive disease PD, (iv) a category that classifies the remaining trends in AF as the trend class of stable disease SD and classifies them.

[0018] For example, the predetermined percentage value of the initial value is in the range between 10% and 90%, such as 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%.

[0019] For example, the predetermined threshold value of AF is in the range between 10% and 90%, such as 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%.

[0020] For example, the predetermined percentage value and / or the predetermined threshold value of AF are defined by the user.

[0021] This will be described below, particularly in relation to the exemplary embodiment of FIG. 1.

[0022] According to an embodiment of the present invention, a tendency score assignment rule defines a correspondence between a tendency score and a tendency class such that (i) the first score is assigned to the tendency class CR, (ii) the second score is assigned to the tendency class PR, (iii) the third score is assigned to the tendency class SD, (iv) the fourth score is assigned to the tendency class PD.

[0023] The first, second, third, and fourth scores have values in ascending order. Or, the first, second, third, and fourth scores have values in descending order.

[0024] In one example, the first, second, third, and fourth scores are 0, 1, 2, and 3 respectively, and thus have values in ascending order.

[0025] In another example, the first, second, third, and fourth scores are 3, 2, 1, and 0 respectively, and thus have values in descending order.

[0026] According to an embodiment of the present invention, the worst-case tendency score is given by the extreme value of the tendency score of the genetic change, and the patient's response state is​ (i) When the extreme value of the propensity score is equal to the first score, the class in which the CR of the patient is determined, and (ii) When the extreme value of the propensity score is equal to the second score, the class in which the PR of the patient is determined, and (iii) When the extreme value of the propensity score is equal to the third score, the class in which the SD of the patient is determined, and (iv) When the extreme value of the propensity score is equal to the fourth score, the class in which the PD of the patient is determined, and are classified into.

[0027] The extreme value corresponds to the maximum value of the propensity score when the first, second, third, and fourth scores have values in ascending order. Or, the extreme value corresponds to the minimum value of the propensity score when the first, second, third, and fourth scores have values in descending order.

[0028] For example, when the first, second, third, and fourth scores are 0, 1, 2, and 3 respectively, and the maximum value of the propensity score is 2, the patient response to the treatment is determined to be stable disease.

[0029] On the other hand, when the first, second, third, and fourth scores are 3, 2, 1, and 0 respectively, and the minimum value of the propensity score is 2, the patient response to the treatment is determined to be partial response.

[0030] According to one embodiment of the present invention, step c) further has a step of counting the number of propensity groups in order to estimate the number of different tumor subclones.

[0031] That is, it is proposed to use a classification method such as cluster analysis or self-organizing map to identify different groups of propensities in order to determine the number of different time courses. This gives an estimate of the number of different tumor subclones in order to evaluate intratumoral heterogeneity when subclones respond differently to a particular treatment. This is beneficial in determining the treatment regimen, and changes in heterogeneity are triggers for adapting the therapy for optimal therapy selection.

[0032] According to one embodiment of the present invention, step c) further comprises, for each subclone at each time point, estimating the relative ratio of the tumor covered, based on two limits: the ratio of the tumors that are tumors divided by the number of subclones found, and the ratio of the sum of the MAFs of the subclones to the sum of the total MAFs for all variants. the ratio of the tumors that are tumors divided by the number of subclones found, and the ratio of the sum of the MAFs of the subclones to the sum of the total MAFs for all variants Based on these two limits, for each subclone at each time point, estimating the relative ratio of the tumor covered.

[0033] For example, the ratio of the tumor covered by a particular subclone is close to two limits: the ratio of the tumors that are tumors divided by the number of subclones found, and the ratio of the sum of the MAF (%) of the subclone to the sum of the MAF (%) for all variants. the ratio of the tumors that are tumors divided by the number of subclones found, and the ratio of the sum of the MAF (%) of the subclone to the sum of the MAF (%) for all variants close to these two limits.

[0034] According to one embodiment of the present invention, the biomarker comprises at least one of circulating tumor DNA (ct-DNA) and genetic analysis of circulating tumor cells CTC.

[0035] Currently, ct-DNA and CTC are clinically optimal biomarkers used to monitor therapy response. Among them, ct-DNA is the most promoted because the occurrence of CTC is low in most cancer patients unless they have metastatic disease.

[0036] According to one embodiment of the present invention, the method further comprises outputting at least one of the number of different tumor subclones, the relative ratio of the tumor covered at each time point for each tumor subclone, and / or the worst-case trend score to a clinical decision support system. the number of different tumor subclones, for each tumor subclone, the relative ratio of the tumor covered at each time point, and / or the worst-case trend score to a clinical decision support system.

[0037] The number of different tumor subclones and the relative ratio of the tumor covered at each time point are used to provide additional information to the clinician when evaluating tumor heterogeneity.

[0038] The worst-case trend score is a trigger to perform medical imaging, such as MRI or CT, to provide further information about a patient's response to therapy. In some examples, the output is combined with medical imaging to improve the efficiency and accuracy of monitoring a patient's response to therapy.

[0039] A second aspect of the present invention relates to a decision support apparatus for monitoring a patient's response to cancer therapy. The decision support apparatus includes an input unit and a processing unit. The input unit is configured to receive data indicating the occurrence of genetic changes over time in a patient, each of the genetic changes being associated with a common disease or group of common diseases. The data is obtained based on the analysis of biomarkers in a patient sample. The processing unit is configured to execute the methods as described above and below.

[0040] As used herein, the terms "part" and "unit" refer to one or more software or firmware programs, combinational logic circuits, and / or other suitable components that implement the described functionality, an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or group), and / or a memory (shared, dedicated, or group) that execute them, or a part thereof, or that includes them.

[0041] In some examples, the processing unit is a single unit that performs steps b) and c).

[0042] In some examples, the processing unit includes two sub-units, one for performing step b) and the other for performing step c). These sub-units can be integrated into a single processing unit. Alternatively, these sub-units can be located in different places to perform distributed computing.

[0043] The third aspect of the present invention relates to a system for monitoring a patient's response to cancer therapy. The system includes a decision support device as described above and below, and a sample analysis device configured to analyze a patient sample to obtain data output to the device and indicating the occurrence of genetic changes over time in the patient.

[0044] In some examples, the decision support device is integrated with the sample analysis device. For example, the computing unit of the sample analysis device is further configured to execute the steps of the method as described above and below.

[0045] In some examples, the decision support device is provided separately from the sample analysis device. For example, the decision support device is coupled to the sample analysis device via a physical cable or wirelessly to receive data.

[0046] Examples of the sample analysis device include, but are not limited to, sequencers such as Illumina's HiSeq or ThermoFisher's IonTorrent, mass spectrometry systems such as Agena Biosciences MassARRAY system, digital droplet polymerase chain reaction (PCR) systems such as Bio-Rad QX200 droplet digital PCR, and / or multiplex PCR systems such as Biocartis Idylla platform.

[0047] According to an embodiment of the present invention, the system further includes a medical imaging device for monitoring a patient's response to cancer therapy by imaging.

[0048] Examples of the medical imaging device include, but are not limited to, MRI, CT, and positron emission tomography (PET) imaging devices. The combination of imaging-based therapy response monitoring and therapy response monitoring using a patient sample, such as a liquid biopsy, improves the efficiency and accuracy of monitoring a patient's response to therapy.

[0049] According to another aspect of the present invention, there is provided a computer program element for controlling an apparatus as described above and below, adapted to execute the steps of the method as described above and below when executed by a processing unit.

[0050] According to a further aspect of the present invention, there is provided a computer-readable medium storing the program element.

[0051] The above and other aspects of the present invention will become apparent from the embodiments described below and will be described in detail with reference to the embodiments.

[0052] The above and other aspects of the present invention will become apparent from the embodiments described by way of example in the following description and with reference to the accompanying drawings, and will be described in further detail with reference to the embodiments.

Brief Description of the Drawings

[0053]

Figure 1

Figure 2

Figure 3

Figure 4A

Figure 4B

Figure 4C

Figure 5

Figure 6

Figure 7

Figure 8

[0054] Note that the figures are only schematic and are not drawn to scale. In the figures, elements corresponding to elements already described have the same reference numerals. Whether or not shown as non-limiting, examples, embodiments, or any features should not be understood to limit the present invention as recited in the claims.

[0055] Monitoring of the therapy response of cancer patients is performed using medical imaging devices such as MRI and CT. Although those methods have the advantage of providing the location of the tumor, since the resolution of medical imaging is generally in the range of millimeters, it is not always easy to evaluate early changes in tumor volume. Other methods for monitoring the therapy response of cancer patients are performed by analyzing biomarkers such as circulating tumor DNA in a patient's sample, such as blood or other body fluids, and tracking the occurrence of specific mutations over time during treatment. However, since tumors carry various mutations, a wide panel covering many different mutations showing different trends over time is generally examined. As a result, it is very difficult to monitor the patient's response to the administered therapy.

[0056] To facilitate monitoring of a patient's response to cancer therapy, the following disclosure describes methods, apparatus, systems, computer program elements, and computer-readable media related to the individual evaluation of therapy response using a dataset representing the dynamics of the occurrence of genetic changes in an individual patient and a statistical classification model such as cluster analysis or self-organizing maps. More specifically, the techniques described herein identify groups that tend to have similar genetic changes over time and use a statistical classification model to process the data to associate those groups with characteristics related to a patient's response to cancer therapy. Since the number of groups of tendencies is significantly less than the number of genetic changes being tracked, it becomes very easy to determine a patient's response.

[0057] FIG. 1 shows a flowchart of a computer-implemented method 100 for monitoring a patient's response to cancer therapy. Examples of cancer therapies include, but are not limited to, drug therapy, radiation therapy, serum therapy, and / or surgical procedures. This method is also applicable to non-invasive prenatal diagnosis or infectious disease monitoring.

[0058] The computer-implemented method 100 is implemented as a device, module, or related component in a set of logic instructions stored in a non-transitory machine or computer-readable storage medium, such as random access memory (RAM), read-only memory (ROM), programmable ROM (PROM), firmware, flash memory, etc., in configurable logic such as a programmable logic array (PLA), field programmable gate array (FPGA), combined programmable logic circuit (CPLD), etc., fixed functionality hardware logic such as an application-specific integrated circuit (ASIC), complementary MOS (CMOS), or transistor-transistor logic (TTL) technology, or any combination thereof. For example, the computer program code for executing the operations shown in method 100 is written in one or more programming languages including languages such as JAVA (registered trademark), SMALLTALK (registered trademark), C++, Python, TypeScript, Java (registered trademark)Script, and conventional procedural programming languages such as any combination of the "C" programming language or similar programming languages. As an example, the computer-implemented method 100 is executed by the exemplary decision support device 200 shown in FIG. 7.

[0059] In step 110, i.e., step a), data indicating the occurrence of genetic changes over time in a patient is received by a decision support device, such as that illustrated in FIG. 6. Each of those genetic changes is associated with a general disease or group of general diseases. That is, various genetic changes associated with diseases or groups of general diseases, such as breast cancer, are tracked. This data is obtained based on the analysis of biomarkers in a patient sample, such as a liquid biopsy, and stored in a dataset. For example, the biomarker is extracted from a blood sample.

[0060] Figure 2 is reproduced from a figure on the Wikipedia page "Circulating tumor DNA" showing some examples of biomarkers extracted from blood, including red blood cells (RBCs), phagocytes, circulating tumor DNA (ct-DNA), normal circulating free DNA (cf-DNA), circulating tumor cells (CTCs), normal cells, and tumor cells. Tumors release multiple biomarkers into the bloodstream, such as ct-DNA, CTCs, exosomes, and platelets. Ct-DNA is generated from dead tumor cells. CTCs are flushed into the bloodstream from primary tumors and metastases. Exosomes are cell-derived vesicles containing tumor messenger RNA (mRNA), microRNA (miRNA), proteins, and double-stranded DNA (dSDNA). Platelets have been reported to capture circulating tumor RNA. Currently, ct-DNA and CTCs are the clinically optimal biomarkers used to monitor therapy response. Among them, ct-DNA is commonly used because the occurrence of CTCs is low (e.g., less than 10 times) in most cancer patients unless they have metastatic disease.

[0061] Since different cancer patients harbor different mutations in their tumors, it is necessary to monitor a large panel of possible genetic changes, i.e., mutations and / or epigenetic changes, so that at least some of them can be reliably detected in biomarkers such as ct-DNA. For example, breast cancer is a very heterogeneous disease with a wide range of possible genetic changes, but all of these genetic changes occur at limited frequencies in the patient population, e.g., 15% or less. For example, the analysis of various genetic changes associated with breast cancer and their general occurrence in patients are shown in Table 1.

[0062]

Table 1

[0063] Furthermore, several hotspots within the above genes need to be monitored, and as a result, at least 44 locations in the biomarker need to be monitored. In certain cases, this is insufficient, and some additional 12 sites need to be monitored due to epigenetic changes. Furthermore, the occurrence of copy number variations (CNVs) such as the well-known HER2 CNV also needs to be examined.

[0064] In step 120, i.e., step b), the trends in the change of allele frequency (AF) over time for each of the genetic changes, i.e., each of the genetic changes, are grouped into a plurality of trend groups, also called TG, according to the similarity between those trends. Each trend group includes trends representing similar temporal behavior of the genetic changes. That is, this step identifies groups of trends with similar temporal behavior patterns, i.e., trends having similar changes in allele frequency over time. The similarity between trends is quantified by a similarity measure. Within the same trend group, the similarity measure is within a predetermined range.

[0065] Various methods are used to identify groups that tend to have group-specific temporal behavior patterns. For example, cluster analysis is performed to identify groups with different tendencies. Cluster analysis or clustering is an unsupervised learning technique aimed at grouping a set of tendencies into clusters, i.e., groups, such that tendencies within the same cluster should be as similar as possible, and tendencies in one cluster should be as different as possible from those in other clusters. Cluster analysis aims to group a set of patterns into clusters based on similarity. General clustering techniques use a similarity function to compare various data items. In particular, clustering is performed based on a similarity measure to group similar data objects together. This similarity measure is based on distance functions such as Euclidean distance, Manhattan distance, Minkowski distance, cosine similarity, etc. to group the tendencies in a cluster. Clusters are formed such that any two tendencies within a cluster have a minimum distance value, and any two tendencies spanning different clusters have a maximum distance value. That is, the similarity measure between tendencies within the same tendency group is within a predetermined range. In other examples, a self-organizing map is used to identify different tendency groups of tendencies. A self-organizing map or self-organizing feature map is a type of artificial neural network (ANN) that is trained using unsupervised learning to generate a low-dimensional discretized representation of the input space of training samples called a map, based on a similarity measure such as Euclidean distance. In a further example, tendencies are compared using a predetermined tendency grouping rule that associates each tendency with its respective tendency group. For example, a predetermined tendency grouping rule stipulates that tendencies with an AF that drops to 0 and then does not rise are grouped as the same group.

[0066] In step 130, i.e., step c), a plurality of trend groups are analyzed to monitor the patient's response to cancer therapy. That is, this step associates those trend groups with features related to the patient's response to cancer therapy. Grouping trends into trend groups compresses a broad panel of genetic changes to be tracked into simpler but representative groups of trends with group-specific temporal behavior patterns, thereby facilitating the analysis of trends for monitoring the patient's response to cancer therapy. It should also be noted that the analysis step c) does not actually make a decision without the need for a physician's expertise. That is, this analysis yields technical results, or more narrowly, technical intermediate results that will assist the physician in reaching a diagnosis later.

[0067] In one example, a plurality of trend groups are analyzed to determine whether a patient is responding to therapy. As described above, since it is not known in which patients such mutations as described above are occurring and new mutations may occur as a progression of the disease, such as resistance genes, in several patients, a very large number of genetic changes need to be measured in all patients. As a result, since all of those mutations should be tracked, it becomes very difficult to monitor the patient's response to therapy. As will be described below, particularly with reference to the exemplary embodiment of FIG. 3, it is proposed to further analyze a plurality of trend groups and aggregate the trends of a large set of genetic changes into a single response score. In this way, a simple model is proposed to reduce complexity even when a complex panel having many different mutations, for example, more than 40 mutations, has to be used, and a simple distance function is estimated to determine whether a patient is responding to therapy.

[0068] FIG. 3 is a flowchart of a method 100 for monitoring a patient's response to cancer therapy according to the above example. In the illustrated flowchart, step 130, i.e., step c), further has the following steps.

[0069] In step 132a, based on the trend classification rules, the trend group is divided into a plurality of trend classes. Each trend class is associated with the reaction of each patient to the treatment such that each trend class includes a trend group having a trend representing the same patient reaction to the treatment. The trend classification rules define the trend classes with respect to the temporal behavior of gene changes. That is, in this step, the trend classification rules are used to further group the trend group into trend classes and associate those trend classes with features related to different patient reactions to the treatment. For example, the trend classification rules classify the trend into categories of (i) AF that does not rise after dropping to 0 as the trend class of complete remission (CR), (ii) AF that drops below a predetermined percentage value of the initial value and maintains a value below that at a certain time elapse as the trend class of partial response (PR), (iii) AF that rises from the initial value to a predetermined threshold and maintains a state above the predetermined threshold towards the end of the time elapse as the trend class of progressive disease (PD), and (iv) the remaining trends in AF as the trend class of stable disease (SD).

[0070] Figures 4A, 4B, and 4C are diagrams showing examples from the literature that represent the dynamics of various types of ct-DNA in the form of the ratio of variants to normal cells, such as normal wild-type cells, in a patient's blood. The data was adapted from a poster presentation by S. Toomey et al. titled "Non-invasive genotyping of locally advanced rectal cancer patients using circulating tumour DNA", presented at the International Symposium on Minimal Residual Cancer (ISMRC) held in Dublin, Ireland from May 3 - 5, 2018. Depending on the sensitivity of the technology used, it is measured between 0.001 and 2 or 3%, and may reach 10% in exceptional cases. The above-described trend classification rules are used to identify to which set of trend classes the above example belongs. For example, in Figure 4A, the curve shows that the AF does not rise after dropping in the third week of treatment. The term "Bx" represents the mutation identified in the pre-treatment biopsy. Note that this is the same mutation that is usually tracked using liquid biopsy. Therefore, the trend in Figure 4A is classified as CR. In Figure 4B, the curve shows that the AF drops below a predetermined percentage value and maintains a value below that for a certain period of time. For example, the predetermined percentage value of its initial value is in the range between 10% and 90%, such as 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%. The user defines the predetermined percentage value. Therefore, the trend in Figure 4B is classified as PR. In Figure 4C, the upper curve shows that the AF rises from its initial value to a predetermined threshold and maintains a state above the predetermined threshold towards the end of the time period. Therefore, the trend represented by the upper curve is classified as PD. On the other hand, the trend represented by the lower curve indicates that since the AF does not follow any of the above patterns, it is classified as SD. For example, the predetermined threshold of AF is in the range between 10% and 90%, such as 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%.A predetermined threshold value of AF is defined by the user.

[0071] In step 134a, the tendency score assignment rule is applied to assign a tendency score, also referred to as TS, to each of the tendency groups based on the associated tendency class. The tendency score assignment rule defines the tendency score for the tendency class. This step links the tendency class to the tendency score. That is, for each tendency in the allele frequency, i.e., for the occurrence of a mutant allele relative to the normal or wild-type allele, the corresponding score is incorporated.

[0072] Referring to FIGS. 4A, 4B, and 4C, the tendency score assignment rule defines a correspondence between the tendency score and the tendency class such that (i) the first score is assigned to the tendency class CR, (ii) the second score is assigned to the tendency class PR, (iii) the third score is assigned to the tendency class SD, and (iv) the fourth score is assigned to the tendency class PD. The first, second, third, and fourth scores have values in ascending order, such as 0, 1, 2, 3, as illustrated in FIGS. 4A to 4C. That is, specifically, it is shown as 0 for a reaction, 1 for a partial reaction, 2 for a stable disease, and 4 for a progressive disease. Alternatively, the first, second, third, and fourth scores have values in descending order, such as 3, 2, 1, and 0. That is, specifically, it is shown as 3 for a reaction, 2 for a partial reaction, 1 for a stable disease, and 0 for a progressive disease.

[0073] In step 136a, a worst-case trend score of genetic changes is determined for use in judging the response state of a cancer patient. For example, the worst-case trend score is given by the extreme value of the trend score of genetic changes. The response state of the patient is classified into (i) a class in which the patient's CR is judged when the extreme value of the trend score is equal to the first score, (ii) a class in which the patient's PR is judged when the extreme value of the trend score is equal to the second score, (iii) a class in which the patient's SD is judged when the extreme value of the trend score is equal to the third score, and (iv) a class in which the patient's PD is judged when the extreme value of the trend score is equal to the fourth score.

[0074] In one example, when the first, second, third, and fourth scores have values in ascending order, such as 0, 1, 2, and 3, the extreme value corresponds to the maximum value of the trend score. That is, the final score classifies the patient into one of the four specified classes: (i) 0 = complete remission of the patient, (ii) 1 = partial response of the patient, (iii) 2 = stable disease of the patient, and (iv) 3 = progressive disease of the patient, for example, to obtain a value between 0 and 3.

[0075] In another example, when the first, second, third, and fourth scores have values in descending order, such as 3, 2, 1, and 0, the extreme value corresponds to the minimum value of the trend score. That is, the final score classifies the patient into one of the four specified classes: (i) 3 = complete remission of the patient, (ii) 2 = partial response of the patient, (iii) 1 = stable disease of the patient, and (iv) 0 = progressive disease of the patient, for example, to obtain a value between 0 and 3.

[0076] In the examples of FIGS. 4A, 4B, and 4C, the final score is 3. Accordingly, the patient's PD is judged.

[0077] Thus, the above approach compresses the data space of a large dataset of monitored genetic changes into simple yet representative response scores. In this way, the number of trend scores representing trend effects is significantly less than the number of genetic changes monitored for a disease such as breast cancer. Therefore, this method can easily and with high confidence distinguish clearly between complete remission, partial response, non-response (i.e., stable disease), or progressive disease for each individual patient. Furthermore, the calculation of worst-case scores, such as extreme trend scores, for individual patients from the monitored biomarkers is performed without human supervision or intervention.

[0078] In other examples, multiple trend groups are analyzed to assess intratumoral heterogeneity. Intratumoral heterogeneity suggests that a cancer patient's tumor is not composed of a single type of cancer cell, but rather multiple subclones exist, each with its own characteristics and reacting differently to a specific treatment. Therefore, assessing tumor heterogeneity is important for determining a treatment regimen, and changes in heterogeneity are triggers for adapting therapy. For this purpose, analysis of the heterogeneity of the primary tumor alone is not sufficient, nor is analysis of biomarkers such as ct-DNA at a single point in time sufficient to provide sufficient insight. As will be described below, with particular reference to the exemplary embodiment of FIG. 5, it is proposed to further analyze trend groups to assess tumor heterogeneity, which is relevant to optimal therapy selection.

[0079] FIG. 5 is a flowchart of a method 100 for monitoring a patient's response to cancer therapy according to the above example. In the shown flowchart, step 130, i.e., step c), further has the following steps.

[0080] In step 132b, the number of trend groups is counted to estimate the number of different tumor subclones. For example, FIG. 6 is a diagram showing an example of five trends in genetic changes A to E tracked over five time points. Basically, there are three different trend groups. Trends B and E have only a scale factor difference and belong to one trend group. Trends C and D belong to another trend group, and only trend A belongs to a further trend group. Therefore, it can be concluded that there are three subclones, two of which exist at time points 1 to 3 and three exist at time points 4 and 5.

[0081] Optionally, step 130 further has step 134b, and at each time point, for each of the subclones, the relative ratio of the tumor covering it is estimated. This is estimated using two limits: the ratio of the tumor that is the tumor divided by the number of subclones found, and the ratio of the sum of the MAF (%) of that subclone to the sum of the total MAF (%) for all variants.

[0082] FIG. 7 is a diagram showing a decision support device 200 for monitoring a patient's response to cancer therapy according to some embodiments of the present disclosure. The system 200 includes an input unit 210 and a processing unit 220.

[0083] The input unit 210 is configured to receive data indicating the occurrence of genetic changes over time in a patient. Each of the genetic changes is associated with a general disease or a group of general diseases. The data is obtained based on the analysis of biomarkers in a patient sample.

[0084] The processing unit 220 is configured to execute any one of the steps of the method described above. For example, the processing unit 220 is configured to execute any one of the methods shown in FIGS. 1, 3, and 5.

[0085] Optionally, as shown in FIG. 7, the decision support device further includes an output unit 230 for outputting results. In one example, the output unit 230 includes a display.

[0086] FIG. 8 is a diagram showing a system 300 for monitoring a patient's response to cancer therapy according to some embodiments of the present disclosure. The system 300 includes the decision support device 200 as described above and a sample analysis device 310.

[0087] The sample analysis device 310 is configured to analyze a patient sample to obtain data indicating the occurrence of genetic changes over time in the patient, which is output to the device. In one example, the sample analysis device 310 is a sequencer such as Illumina's HiSeq or ThermoFisher's IonTorrent. In another example, the sample analysis device 310 is a mass spectrometry system such as the Agena Biosciences MassARRAY system. In a further example, the sample analysis device 310 is a digital droplet polymerase chain reaction (PCR) system such as the Bio-Rad QX200 droplet digital PCR. In a further example, the sample analysis device 310 is a multiplex PCR system such as the Biocartis Idylla platform.

[0088] Optionally, the system further comprises a medical imaging device 320 for monitoring a patient's response to cancer therapy by imaging. That is, monitoring the therapy response using a patient sample is combined with monitoring the therapy response in cancer patients by imaging. Examples of medical imaging devices include, but are not limited to, magnetic resonance imaging (MRI) devices, X-ray imaging devices, computed tomography (CT) imaging devices, ultrasound (US) imaging devices, and positron emission tomography (PET) imaging devices. Medical imaging methods have the advantage of providing the location of tumors, but since the resolution of medical imaging is generally in the millimeter range, it is not always easy to evaluate changes in tumor volume. On the other hand, biomarkers such as ct-DNA provide molecular information regarding which mutations in a patient's DNA are causing the cancer and thus which drugs should be administered. Further, since biomarkers are measured at the molecular level, responses can be detected much earlier than by imaging-based detection. Further, biomarkers provide additional information such as tumor genetic analysis, tumor heterogeneity, and resistance to therapy. Thus, the combination of both methods for monitoring a patient's response to cancer therapy provides a more efficient and accurate approach for monitoring progression, response, and / or tumor heterogeneity over time.

[0089] In another exemplary embodiment of the invention, there is provided a computer program or computer program element having the feature of being adapted to perform the steps of a method according to one of the above-described embodiments on a suitable system.

[0090] Accordingly, the computer program elements are stored on the computing unit which is also part of an embodiment of the invention. The computing unit is adapted to execute or cause the execution of the steps of the method described above. Further, it is adapted to operate the components of the device described above. The computing unit may be adapted to operate automatically and / or execute the instructions of a user. The computer program is loaded into the working memory of the data processor. Thereby, the data processor is equipped to execute the method of the invention.

[0091] Exemplary embodiments of the invention cover both a computer program for using the invention from the start and a computer program for converting an existing program into a program using the invention by means of an update.

[0092] Furthermore, the computer program elements can provide all the necessary steps for fulfilling the procedure of an exemplary embodiment of the method as described above.

[0093] According to a further exemplary embodiment of the invention, a computer-readable medium such as a CD-ROM is presented, on which computer program elements are stored, which computer program elements are those described by the foregoing part.

[0094] The computer program is stored and / or distributed on a suitable medium such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless communication systems.

[0095] However, the computer program may be presented via a network such as the World Wide Web and may be downloaded from such a network into the working memory of the data processor. According to a further exemplary embodiment of the present invention, there is provided a medium making a computer program element available for download, the computer program element being configured to execute a method according to one of the above-described embodiments of the present invention.

[0096] It should be noted that the embodiments of the present invention are described with reference to different subject matters. In particular, some embodiments are described with reference to method claims and other embodiments are described with reference to apparatus claims. However, those skilled in the art will infer from the above and the following descriptions that, unless otherwise specified, in addition to the combinations of features belonging to one type of subject matter, combinations between features related to different subject matters are also disclosed in relation to this application. However, not all features are combined to achieve a synergistic effect exceeding the simple sum of the features.

[0097] Although the present invention has been illustrated and described in detail in the drawings and the above description, such illustrations and descriptions should be considered to be illustrative or exemplary and not restrictive. The present invention is not limited to the disclosed embodiments. Other variations of the disclosed embodiments can be understood and implemented by those skilled in the art in practicing the invention as claimed, upon consideration of the drawings, the present disclosure, and the dependent claims.

[0098] In the claims, the terms "comprising" and "having" do not exclude other elements or steps, and the singular form does not exclude the plural. A single processor or other unit may perform the functions of several elements recited in the claims. The fact that certain means are recited in mutually different dependent claims does not indicate that a combination of these means cannot be used effectively. Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A computer-implemented method for monitoring a patient's response to cancer therapy, the computer-implemented method comprising: a) receiving data indicative of the occurrence of genetic changes over time in a patient, each of the genetic changes being associated with a common disease or group of common diseases, the data being obtained based on the analysis of biomarkers in a patient sample; b) grouping the trends in each of the genetic changes into a plurality of trend groups according to the similarity between trends such that each trend group has a group-specific temporal behavior pattern, each trend group including trends representing a similar temporal behavior of the allele frequencies of the genetic changes, and a similarity measure quantifying the similarity between trends within the same trend group being within a predetermined range; c) analyzing the group-specific temporal behavior patterns of the plurality of trend groups to monitor the patient's response to the cancer therapy. Step c) further comprises: dividing the trend groups into a plurality of trend classes based on trend classification rules, each trend class being associated with the response of a respective patient to treatment and including trend groups having trends representing the same patient response to treatment, the trend classification rules defining the trend classes with respect to the temporal behavior of the genetic changes; applying a trend score assignment rule to assign a trend score to each of the trend groups based on its associated trend class, the trend score assignment rule defining the trend score with respect to the trend class; determining the worst-case trend score of the genetic changes for use in determining the response status of a cancer patient. A computer-implemented method.

2. In step b), the trends are grouped using at least one of: a method of performing a cluster analysis to identify different trend groups of the trends; a method of using a self-organizing map to identify different trend groups of the trends; a method of comparing the trends using a predetermined trend grouping rule to associate each of the trends with a respective trend group. The computer-implemented method according to claim 1.

3. The trend classification rules classify the trends ​ (i)A category that classifies the allele frequency that does not rise after dropping to 0 as a complete remission, which is a trend class, (ii)A category that classifies the allele frequency that drops below a predetermined percentage value of the initial value and maintains a value below that for a certain period of time as a partial response, which is a trend class, (iii)A category that classifies the allele frequency that rises from the initial value to a predetermined threshold value and maintains a state above the predetermined threshold value towards the end of the time course as a progressive disease, which is a trend class, (iv)A category that classifies the remaining trends in the allele frequency as a stable disease, which is a trend class into which The computer-implemented method according to claim 1.

4. The tendency score assignment rule is between the tendency score and the tendency class, (i)The first score is assigned to a complete remission, which is the tendency class, (ii)The second score is assigned to a partial response, which is the tendency class, (iii)The third score is assigned to a stable disease, which is the tendency class, (iv)The fourth score is assigned to a progressive disease, which is the tendency class, defines the correspondence of the first, second, third, and fourth scores have values in ascending or descending order, The computer-implemented method according to claim 3.

5. The worst-case tendency score is given by the extreme value of the tendency score of the gene change, The reaction state of the patient is (i)When the extreme value of the tendency score is equal to the first score, a class in which a complete remission of the patient is determined, (ii)When the extreme value of the tendency score is equal to the second score, a class in which a partial response of the patient is determined, (iii)When the extreme value of the tendency score is equal to the third score, a class in which a stable disease of the patient is determined, (iv)When the extreme value of the tendency score is equal to the fourth score, a class in which a progressive disease of the patient is determined, and is classified into the extreme value is When the first, second, third, and fourth scores have values in ascending order, the maximum value of the tendency score, or When the first, second, third, and fourth scores have values in descending order, the minimum value of the tendency score corresponding to The computer-implemented method according to claim 4.

6. Step c) further has a step of counting the number of the tendency groups to estimate the number of different tumor subclones, The computer-implemented method according to any one of claims 1 to 5.

7. The ratio of tumors that are tumors divided by the number of discovered subclones, and The ratio of the sum of the mutant allele frequencies of the subclones to the sum of the mutant allele frequencies of all alleles with respect to all variants Based on these two limits, at each time point, further comprising the step of estimating the relative ratio of the tumors covered for each of the subclones, The computer-implemented method according to claim 6.

8. The biomarker includes at least one of circulating tumor DNA and genetic analysis of circulating tumor cells, The computer-implemented method according to any one of claims 1 to 7.

9. The number of different tumor subclones, For each of the tumor subclones, the relative ratio of the tumors covered at each time point, and / or The worst-case trend score Further comprising the step of outputting at least one of to a clinical decision support system, The computer-implemented method according to any one of claims 1 to 8.

10. A decision support device for monitoring a patient's response to cancer therapy, the decision support device comprising: An input unit configured to receive data indicating the occurrence of genetic changes over time in a patient, each of the genetic changes being associated with a general disease or a group of general diseases, the data being obtained based on the analysis of biomarkers in a patient sample, the input unit; A processing unit for executing the method according to any one of claims 1 to 9 A decision support device comprising.

11. A system for monitoring a patient's response to cancer therapy, The decision support device according to claim 10, and A sample analysis device configured to analyze a patient sample in order to obtain data indicating the occurrence of genetic changes over time in a patient output to the device A system comprising.

12. The system according to claim 11, further comprising a medical imaging device for monitoring a patient's response to cancer therapy by imaging.

13. A computer program for controlling the device according to any one of claims 10 to 12, which executes the steps of the method according to any one of claims 1 to 9 when executed by a processing unit.

Images