Biomarkers for the treatment of peripheral T-cell lymphoma and methods of use thereof

Abstract

The present disclosure relates to methods for genetically subtyping peripheral T-cell lymphoma. TIFF2025506567000021.tif56128

Description

[Technical field]

[0001] This application is an international application claiming the benefit of priority from U.S. Provisional Patent Application No. 63 / 307,905, filed February 8, 2022, the entire contents of which are incorporated herein by reference in their entirety.

[0002] All patents, patent applications, and publications cited herein are hereby incorporated by reference in their entireties. The disclosures of these publications are incorporated by reference in their entireties into this application in order to more fully describe the state of the art known to those skilled in the art as of the date of the invention described and claimed herein.

[0003] This patent disclosure contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure as it appears in the U.S. Patent and Trademark Office patent file or records, but otherwise reserves all and any copyright rights whatsoever.

[0004] Government Interests This invention was made with Government support under U01 CA157581 and UH3 CA206127 awarded by the National Institutes of Health (NIH). The Government has certain rights in this invention.

[0005] Technical Field The subject matter disclosed herein relates to methods for the diagnosis, prognosis, and / or treatment of peripheral T-cell lymphoma (PTCL). In addition, the subject matter disclosed herein relates to kits and reagents for the diagnosis, prognosis, and / or treatment of PTCL. Furthermore, the subject matter disclosed herein relates to methods for identifying a set of biomarkers for the diagnosis, prognosis, and / or treatment of PTCL using software and analysis systems. [Background technology]

[0006] Introduction Malignancies derived from T-cell and natural killer (NK) cell lineages constitute 10% of all non-Hodgkin's lymphomas (NHL) in Western countries and are more prevalent in Asia. Peripheral T-cell lymphomas (PTCL) are heterogeneous cancers that constitute up to 20% of all non-Hodgkin's lymphomas (NHL). The World Health Organization (WHO) classification recognizes several distinct subtypes of peripheral T-cell lymphomas (PTCL), including angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma (ALCL), adult T-cell leukemia / lymphoma (ATLL), and entities that are mostly derived from NK cells, including extranodal NK / T-cell lymphoma, nasal type (ENKTL). There are even rarer PTCLs, most of which are extranodal tumors. The diverse morphology and lack of definitive diagnostic markers for most subtypes make the diagnosis and classification of these diseases difficult and hinder their investigation. With current immunophenotypic and molecular markers, approximately 30%-50% of PTCL cases are unclassifiable and classified as PTCL not otherwise specified (PTCL-NOS). Although several recurrent genetic abnormalities have been reported in PTCL, no specific genetic abnormality is diagnostic of a particular PTCL subtype, except for t(2;5)(p23;q35) in ALK(1)ALCL. Patients with PTCL generally have a poor prognosis with current standard of care, and no improvement in outcomes for PTCL patients has been achieved in the past two decades. Thus, accurate diagnosis of PTCL and its subtypes is required to ensure the selection of the most appropriate and effective treatment for each individual patient. In particular, the creation of a commercial assay for routine clinical application would improve the diagnosis, treatment, and outcome of PTCL patients. Summary of the Invention

[0007] In a first aspect, the invention includes a method of distinguishing between subtypes of peripheral T-cell lymphoma (PTCL). In an embodiment, the method includes subjecting a sample from a subject to nucleic acid isolation, obtaining a gene expression profile from the sample, and identifying the subtype of PTCL based on the presence of specific genes in the gene expression profile. In certain embodiments, the PTCL subtypes are angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not specified (PTCL-NOS), or indeterminate, where indeterminate indicates that the sample contains characteristics of at least two PTCL subtypes.

[0008] If the gene expression profile includes any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2, the PTCL subtype may be identified as AITL. If the gene expression profile includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101, the PTCL subtype may be identified as ALK- ALCL. In embodiments, the PTCL subtype may be identified as ENKTL if the gene expression profile includes any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25. The PTCL subtype may be identified as ATLL if the gene expression profile includes any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, HBZ, and ZCCHC12.

[0009] 3. The method of claim 2, wherein the PTCL subtype is identified as ALK+ ALCL if the gene expression profile includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2. In certain embodiments, the PTCL subtype is identified as PTCL-NOS if the gene expression profile includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, WARS, TBX21, CXCR3, GATA3, and CCR4.

[0010] In embodiments, the sample is a biopsy specimen from a subject. The sample may include formalin-fixed paraffin-embedded tissue. In certain embodiments, the sample includes fresh frozen tissue.

[0011] The method may further include providing to the subject an effective amount of a subtype-specific therapy. In embodiments, the subtype-specific therapy is a histone deacetylase (HDAC) inhibitor, an antifolate, an acylating agent, a proteosome inhibitor, an antibody-drug conjugate, a phosphoinositide 3-kinase (PI3K) inhibitor, a Janus kinase (JAK) inhibitor, a signal transducer and activator of transcription (STAT) 3 inhibitor, a STAT5 inhibitor, an anaplastic lymphoma kinase (ALK) inhibitor, a hepatocyte growth factor (HGF) inhibitor, a cMET inhibitor, a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor, a platelet-derived growth factor receptor beta (PDGFRβ) inhibitor, or a rapamycin. and / or combinations thereof.In embodiments, the HDAC inhibitor comprises romidepsin, belinostat, panobinostat, or a combination thereof, the antifolate comprises pralatrexate, the akylating agent comprises bendamustine, the proteosome inhibitor comprises bortezomib, the antibody-drug conjugate comprises brentuximab vedotin, the PI3K inhibitor comprises duvelisib, tenalisib, or a combination thereof, the JAK inhibitor comprises ruxolitinib, the ALK inhibitor comprises crizotinib, the mTOR pathway inhibitor comprises everolimus, and the hypomethylating agent comprises , 5-azacytidine, anti-CD52 antibodies include alemtuzumab, ImiDs include lenalidomide, CCR4 inhibitors include mogamulizumab, IDH inhibitors include enasidenib, BCL2 inhibitors include venetoclax, anti-CD25 antibodies include camidinelumabtesilin, calcineurin inhibitors include cyclosporine A, SYK inhibitors include celdulatinib, bispecific antibodies include AFM13, and chimeric antigen receptor T (CAR-T) cells include CD30, CD7, or both, or a combination thereof.

[0012] Subtype-specific treatments for PTCL-NOS may include romidepsin, belinostat, brentuximab vedotin, duvelisib, or combinations thereof. Subtype-specific treatments for AITL may include romidepsin, 5-Aza, isocitrate dehydrogenase (IDH) inhibitors, calcineurin inhibitors, or combinations thereof. Subtype-specific treatments for ALK- ALCL may include brentuximab vedotin. Subtype-specific treatments for ALK+ ALCL may include ALK inhibitors, platelet-derived growth factor receptor beta (PDGFRβ) inhibitors, or combinations thereof. Subtype-specific treatments for ATLL may include NOTCH inhibitors, hepatocyte growth factor (HGF) inhibitors, cMET inhibitors, mogamulizumab, or combinations thereof. Subtype-specific treatments for ENKTL may include platelet-derived growth factor receptor alpha (PDGFRα) inhibitors.

[0013] Another aspect of the invention includes a diagnostic kit for identifying a subtype of peripheral T-cell lymphoma (PTCL) in a sample from a subject, the kit comprising at least one of means for detecting the presence of one or a combination of genes representing a gene signature indicative of a particular PTCL subtype, and instructions for use. In embodiments, the PTCL subtypes are angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, where indeterminate indicates that the sample contains characteristics of at least two PTCL subtypes. If the gene signature includes any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2, the PTCL subtype may be indicative of AITL. If the gene signature includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101, the PTCL subtype may be indicative of ALK- ALCL. In embodiments, the gene signature may indicate ENKTL if it includes any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25. The gene signature may indicate ATLL if it includes any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, TAX, and ZCCHC12.If the gene signature includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2, the PTCL subtype may be indicative of ALK+ ALCL. If the gene signature includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, and WARS, the PTCL subtype may be indicative of PTCL-NOS.

[0014] Another embodiment includes a method of identifying a particular subtype of peripheral T-cell lymphoma (PTCL) in a sample, comprising: obtaining a gene expression profile from the sample; comparing the gene expression profile to gene signatures associated with particular PTCL subtypes, wherein each subtype of PTCL comprises a unique gene signature; and identifying the subtype of PTCL in the sample as either angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, wherein indeterminate indicates that the gene expression profile of the sample comprises genes from the unique gene signatures of at least two of the PTCL subtypes.

[0015] In embodiments, the gene signature of AITL includes any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2; The gene signature of ALCL includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101, the gene signature of ENKTL includes any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25, the gene signature of ATLL includes any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, TAX, and ZCCHC12, and is ALK+. The gene signature of ALCL includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2, and the gene signature of PTCL-NOS includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, and WARS, or any combination thereof.

[0016] In embodiments, identifying the subtype of PTCL comprises: A set of at least four binary predictors, the first binary predictor comprises determining whether a gene expression profile of the sample is more consistent with a gene signature of AITL or PTCL-NOS; a second binary predictor determining whether the gene expression profile of the sample is more consistent with a gene signature of ALCL or PTCL-NOS; a third binary predictor determining whether the gene expression profile of the sample is more consistent with a gene signature of ATLL or PTCL-NOS; and a fourth binary predictor comprising determining whether the gene expression profile of the sample is more consistent with a gene signature of ENKTL or PTCL-NOS. A set of at least four binary predictors, as well as and assigning a PTCL subtype of the specimen based on the results from the binary predictors. If the gene expression profile of the sample is more consistent with the gene signature of AITL in the first binary predictor, and if the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS in the second, third, and fourth binary predictors, the PTCL subtype may be identified as AITL. If the gene expression profile of the sample is more consistent with the gene signature of ALCL in the second binary predictor, and if the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS in the first, third, and fourth binary predictors, the PTCL subtype may be identified as ALCL. In one embodiment, the method further comprises: a fifth binary predictor comprising determining whether the gene expression profile of the sample is more consistent with a gene signature of ALK+ ALCL or ALK- ALCL; and identifying the PTCL subtype as ALK+ ALCL if the gene expression profile of the sample is more consistent with a gene signature of ALK+ ALCL; or Identifying the PTCL subtype as ALK- ALCL if the gene expression profile of the sample is more consistent with the gene signature of ALK- ALCL.

[0017] In one embodiment, the PTCL subtype is identified as ATLL if the gene expression profile of the sample is more consistent with the gene signature of ATLL in the third binary predictor, and if the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS in the first, second, and fourth binary predictors. The PTCL subtype may be identified as ENKTL if the gene expression profile of the sample is more consistent with the gene signature of ENKTL in the fourth binary predictor, and if the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS in the first, second, third, and fourth binary predictors. The PTCL subtype may be identified as PTCL-NOS if the gene expression profile of the sample is more consistent with PTCL-NOS in the first, second, third, and fourth binary predictors.

[0018] If at least two of the binary predictors are less consistent with PTCL-NOS, the PTCL subtype may be identified as indeterminate. In one embodiment, the sample identified as indeterminate is subjected to an additional binary predictor, the additional binary predictor including a first potential subtype including one of the PTCL subtypes less consistent with PTCL-NOS and a second potential subtype including at least one other PTCL subtype less consistent with PTCL-NOS. In such an embodiment, the method may further include determining whether the gene expression profile of the sample is more consistent with the gene expression profile of the first potential subtype or the second potential subtype, and identifying the PTCL subtype as the first potential subtype if the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype, or identifying the PTCL subtype as the second potential subtype if the gene expression profile of the sample is more consistent with the gene signature of the second potential subtype. If the PTCL subtype is identified as ALCL, the method may further include determining the ALCL subtype as either ALK+ ALCL or ALK- ALCL using a fifth binary predictor disclosed herein.

[0019] In embodiments, samples identified as PTCL-NOS cases are further separated into GATA3 high and TBX21 high subgroups. Samples identified as ENKTL may be subdivided into NK or gamma / delta T cell lineages.

[0020] Another embodiment includes a method of treating peripheral T-cell lymphoma. In embodiments, the method includes obtaining a gene expression profile from a sample of the subject; comparing the gene expression profile to a gene signature associated with a particular PTCL subtype, wherein each subtype of PTCL comprises a unique gene signature; identifying the subtype of PTCL in the sample as either angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, wherein indeterminate indicates that the gene expression profile of the sample comprises genes from the unique gene signatures of at least two PTCL subtypes; and administering to the subject an effective amount of a therapeutic agent, wherein the therapeutic agent is configured to treat the identified PTCL subtype. In embodiments, the therapeutic agent is a histone deacetylase (HDAC) inhibitor, an antifolate, an acylating agent, a proteosome inhibitor, an antibody-drug conjugate, a phosphoinositide 3-kinase (PI3K) inhibitor, a Janus kinase (JAK) inhibitor, a signal transducer and activator of transcription (STAT) 3 inhibitor, a STAT5 inhibitor, an anaplastic lymphoma kinase (ALK) inhibitor, a hepatocyte growth factor (HGF) inhibitor, a cMET inhibitor, a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor, a platelet-derived growth factor receptor beta (PDGFRβ) inhibitor, a mammalian and / or combinations thereof.

[0021] In one embodiment, the HDAC inhibitor comprises romidepsin, belinstat, panobinostat, or a combination thereof, the antifolate comprises pralatrexate, the akylating agent comprises bendamustine, the proteosome inhibitor comprises bortezomib, the antibody-drug conjugate comprises brentuximab vedotin, the PI3K inhibitor comprises duvelisib, tenalisib, or a combination thereof, the JAK inhibitor comprises ruxolitinib, the ALK inhibitor comprises crizotinib, the mTOR pathway inhibitor comprises everolimus, and the hypomethylating agent comprises 5- include azacitidine (5-Aza), anti-CD52 antibodies include alemtuzumab, ImiDs include lenalidomide, CCR4 inhibitors include mogamulizumab, IDH inhibitors include enasidenib, BCL2 inhibitors include venetoclax, anti-CD25 antibodies include camidinelumabtesilin, calcineurin inhibitors include cyclosporine A, SYK inhibitors include celdulatinib, bispecific antibodies include AFM13, chimeric antigen receptor T (CAR-T) cells include CD30, CD7, or both, or combinations thereof.

[0022] In embodiments, when the PTCL subtype is identified as PTCL-NOS, the therapeutic agent includes romidepsin, belinostat, brentuximab vedotin, duvelisib, or a combination thereof. When the PTCL subtype is identified as AITL, the therapeutic agent may include romidepsin, 5-Aza, isocitrate dehydrogenase (IDH) inhibitors, calcineurin inhibitors, or a combination thereof. When the PTCL subtype is identified as ATLL, the therapeutic agent may include NOTCH inhibitors, hepatocyte growth factor (HGF) inhibitors, cMET inhibitors, mogamulizumab. In embodiments, when the PTCL subtype is identified as ALK- ALCL, the therapeutic agent includes brentuximab vedotin. When the PTCL subtype is identified as ALK+ ALCL, the therapeutic agent may include ALK inhibitors, platelet-derived growth factor receptor beta (PDGFRβ) inhibitors, or a combination thereof. In one embodiment, when the PTCL subtype is identified as ENKTL, the therapeutic agent comprises a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor. [Brief description of the drawings]

[0023] The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Patent Office upon request and payment of the necessary fee. The drawings described herein correspond to Amador C, Bouska A, Wright G, et al. Gene Expression Signatures for the Accurate Diagnosis of Peripheral T-Cell Lymphoma Entities in the Routine Clinical Practice. J Clin Oncol. 2022; 40(36): 4261-4275. doi: 10.1200 / JCO.21.02707, which is incorporated by reference in its entirety.

[0024] Certain features of the invention are set forth with particularity in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by reference to the following detailed description that sets forth illustrative embodiments, in which the principles of the invention are utilized, and the accompanying drawings. [Figure 1] Molecular diagnostic signatures of PTCL subgroups are shown. Using a multi-covariate predictive model, unique gene expression signatures were identified for the major PTCL entities (for details, see references 10 and 11). Each column represents a PTCL patient and each row represents a unique gene of the classifier. [Diagram 2] The two main molecular subgroups within PTCL-NOS are shown. (A) Bayesian predictors for GATA3 and TBX21 subtypes were derived. Leave-one-out cross-validation was used for classification accuracy. Approximately 18% of cases were not clearly defined (UC, unclassifiable). (B) Overall survival (OS) analysis of GATA3 and TBX21 subgroups showed significant differences in clinical outcomes (P=.01) (Ref. 10). [Diagram 3] Figure 1 shows the error rate versus number of genes in the model. The Y-axis shows the average cross-validated error rate for 500 simulated datasets, with noise added as a function of the X-axis to approximate the effect of platform variation in the number of available candidate genes. Each colored line represents a different diagnostic distinction (see figure legend). There is a steep drop in the curves from left to right, with each curve flattening out at 15-20 genes. These results demonstrate that this number of genes is sufficient to make each diagnostic distinction, and that additional genes do not further decrease the error rate. [Figure 4]Performance of the Lymph2Cx assay in an independent validation cohort. (A) The Lymph2Cx assay is shown in the form of a gene expression heatmap of 67 patients. The 20 genes contributing to the model, including 5 housekeeping genes, are shown on the left. Cell-of-origin assignment for the assay is shown and compared to the gold standard method, using a previously published algorithm for gene expression from fresh frozen tissue and three immunohistochemistry-based algorithms. (B) Comparison of Lymph2Cx scores in a validation cohort from two independent laboratories (MoCha, CLC). The horizontal and vertical light grey lines represent the thresholds between GCB, unclassified, and ABC subtypes. The R2 is 0.996 and the slope of the best fit line is 1.015. MoCha, Molecular Characterization Laboratory, Frederick National Laboratory for Cancer Research (Frederick, MD), CLC, Centre for Lymphoid Cancer, BC Cancer Agency (Vancouver, BC). [Diagram 5] (A) Histogram showing the distribution of correlation between Nanostring and Affymetrix data for 667 diagnostic and prognostic genes in DLBCL. (B) Genes present in the "academic" PTCL signature. The majority of the 40 genes are well correlated, with a median correlation of 0.76. [Figure 6] Assay development is shown using U133 array data from FF samples independent of the validation set to develop a linear predictor of subtypes including cutpoints. In the training set, replicate samples are used to adjust gene weights and additive shifts to account for platform differences and mimic FF predictor scores as closely as possible to FFPE specimens / NanoString analysis, establishing a locking aspect for validation. A Phase II trial following this proposal is also shown. [Figure 7]Variability (Y-axis) versus signal intensity (X-axis) is shown. Each colored line represents data from a different Lymphoma / Leukemia Molecular Profiling Project (LLMPP) center. Overlapping lines indicate excellent reproducibility between centers, with some variability at minimal signal intensity. [Figure 8] Exemplary immunohistochemical stains for subclassification of ALCL and PTCL-NOS are shown. ALCL can be subclassified as CD30+ or ​​ALK+, and GATA3 and TBX21 represent two subgroups of PTCL-NOS. [Figure 9] A histogram comparison of RNA yields from FFPE tissues collected at various times using different protocols for DNA / RNA isolation is provided. RNA isolation was performed using either the STORM protocol, the modified STORM protocol, or the Quiagen protocol. RNA recovery from older blocks was generally lower, but this did not differ using the newer kits. [Figure 10A] Representative gel images of RNA isolated from FFPE and fresh frozen samples using various protocols are included. Tissues were collected in the indicated years. [Figure 10B] A histogram comparing the effectiveness of different protocols in isolating larger RNA molecules (greater than 200 base pairs in length) from samples of various ages is provided. [Figure 11] Compared to Figure 1 , the refined gene signature is shown (modified from ref. 10 ), with additional genes added to the diagnostic signature list. [Figure 12]PTCL molecular classifier is provided. Figure 12A provides the study design and schematic of molecular diagnosis of PTCL. The molecular classifier for PTCL subclassification was derived using HG-U133plus2.0 array data from PTCL with fresh frozen (FF) tissue (n=109) designated as training cohort. This molecular classifier had 442 distinct genes, including housekeeping genes and other genes involved in T cell biology. This transcriptome signature was considered for nCounter® analysis (NanoString, Inc.) using corresponding matched FFPE samples. Transcripts that showed high correlation between FF and FFPE data were selected. The algorithm was further refined to have a minimum number of transcripts for subclassification and to mimic the FF predictor score using the NanoString platform (see M&M for details). The final diagnostic model yielded 153 transcripts (99 diagnostic, 5 viral and 16 housekeeping, and 33 T-cell biology related) and was validated in an independent cohort of PTCL cases rigorously characterized by pathology and other ancillary methods. The classification algorithm was based on a series of several binary predictors to distinguish one entity from another, as detailed in Figure 17. Figure 12B provides heat maps of the final nCounter® classifier in the training and validation cohorts. EBV and HTLV-1 viral transcripts commonly expressed in specific PTCL subtypes are shown and used for normalization, and housekeeping genes that do not vary between PTCL subtypes are shown for comparison. Figure 12C shows Kaplan-Meier curves of overall survival for 89 of the 105 training cohort cases and 66 of the 140 validation cases, with available outcome data. Figure 12D shows Kaplan-Meier curves of OS for PTCL subtypes included in the training cohort (molecular classification by HG-U133plus2.0 array). Figure 12E shows Kaplan-Meier curves of OS for PTCL entities included in the validation cohort (pathological classification). [Figure 13]AITL classification. Figure 13A provides a scatter plot of AITL diagnostic score against the mean expression of five TFH-related genes in training and validation AITL. Figure 13B provides a box plot of EBER and EBNA1 transcript expression in AITL (training / validation cohort). Figure 13C provides Kaplan-Meier curves of OS of AITL in the combined cohort by CD20 mRNA expression. Solid lines are cases classified as AITL by molecular classification (p=0.02). Dotted lines are AITL by pathology (p=0.06). CD20 expression (400x) in representative low and high AITL cases. Figure 13D shows the mutation status of cases with available sequencing data. Cases that were AITL-PTCL-NOS intermediate or not classified as AITL by the NanoString classifier are indicated with an asterisk. FIG. 13E provides violin and dot plots of AITL classification diagnostic scores in AITL and PTCL-NOS cases profiled with nCounter®. Cases that were discordant between AITL and PTCL-NOS are colored red, and intermediate AITL cases are colored gray. FIG. 13F provides a heat map of AITL showing discordance by nCounter® classification in the validation cohort. The average signature of the concordant cases is shown. For cases labeled as intermediate, the case diagnosed as AITL by consensus pathology is on the left, and the intermediate case diagnosed as PTCL-NOS is on the right. Two PTCL-NOS classified as AITL by nCounter®. Figures G-I provide an example of focal expression of BCL6 and ICOS (400x; G, H&I; B, BCL-6 and C, ICOS) seen in a PTCL-NOS case classified as AITL by the nCounter® platform. [Figure 14]ALCL classification. FIG. 14A provides a violin plot and scatter plot of ALCL classification score vs. PTCL-NOS. FIG. 14B shows ALK-positive vs. ALK-negative ALCL on the nCounter® platform. Cases that were discordant between ALCL and PTCL-NOS or between ALK-negative and ALK-positive are colored red. FIG. 14C provides a heat map of ALK-negative and ALK-positive ALCL cases showing discordance by nCounter® classification in the validation cohort. H and E and CD30 and ALK immunostaining of a representative ALK-negative ALCL case classified as ALK-positive ALCL by the NanoString assay are shown (400x). FIG. 14D provides Kaplan-Meier curves of OS of ALCL in the training and validation cohorts by NanoString classification. FIG. 14E provides a heat map of CD30 and cytotoxic transcript expression in ALCL and PTCL-NOS cases. Discordant cases are indicated by red lines. [Figure 15]ATLL and ENKTCL classification. Figure 15A provides violin and dot plots of ATLL classification scores in ATLL and PTCL-NOS cases profiled with nCounter®. Figure 15B provides a heat map of discordant ATLL cases by nCounter® classification in the validation cohort. H&E and CD4 staining (lower panel) for cases diagnosed as PTCL-NOS but classified as ATLL by nCounter®. Figure 15C provides a scatter plot of expression of the HTLV-1 specific transcript HBZ versus ATLL score in training and validation ATLL cases. The solid fitted line represents the training data and the dashed line represents the validation data. Figure 15D provides a heat map of HBZ and TAX expression in the ATLL and PTCL-NOS validation cohort. Discordant cases are indicated by red asterisks. Figure 15E provides a scatter plot of HBZ expression measured by qPCR versus nCounter®. Figures 15F-G provide violin and dot plots of ENKTCL classification scores (F) or EBER scores (G) in ENKTCL and PTCL-NOS cases profiled in training and validation cohorts in nCounter®. Mismatched cases are in red. Figure 15H provides heat maps of CD3 gamma and delta and EBV transcript expression in ENKTCL and PTCL-NOS cases in training and validation cohorts. Figure 15I provides heat maps of association signature expression in ENKTCL mismatched cases compared to the average signature in the validation cohort. [Figure 16]PTCL-NOS subclassification is shown. Figure 16A provides violin and dot plots of PTCL-GATA3 classification scores in PTCL-NOS cases profiled in training and validation cohorts in nCounter®. Figure 16B shows the mutation status of the PTCL-NOS cohort with available sequencing data. Figure 16C provides heat maps of expression of CD4, CD8, CD20, and cytotoxic genes in training (top) and validation (bottom) cases. Figure 16D provides a scatter plot of the mean expression of cytotoxic genes versus CD20 in cases classified as PTCL-TBX21 by NanoString. Figure 16E provides H&E and IHC staining for one representative PTCL-GATA3 case (left) showing GATA3, and one PTCL-TBX21 case (right) showing TBX21 and CD8 expression. FIG. 16F shows the KM curves for overall survival of PTCL-NOS cases with available outcome data in the combined training and validation cohorts classified as PTCL-GATA3 or PTCL-TBX21 NanoString. [Figure 17] A schematic diagram of the algorithm design for PTCL subclassification is provided. The six pairwise submodels are combined to generate the final predictor for classification. Stage 1: First, four categories (AITL, ALCL, ENKTCL, or ATLL) are differentiated from PTCL-NOS. Then, EBV-viral transcripts (EBER) are added to the ENKTCL classifier (vs. PTCL-NOS), and finally, ALCL is subclassified into ALK- ALCL vs. ALK+ ALCL. Stage 2: In the second stage, samples considered as PTCL-NOS are differentiated into PTCL-GATA3 or PTCL-TBX21 subtypes. If more than one of the four predictors results in a non-PTCL-NOS diagnosis (e.g., ALCL and AITL), an additional pairwise predictor between the types is applied to break ties (e.g., ALCL vs. AITL) using all genes characteristic of either type selected to be included in the model. [Figure 18A]For the same RNA extracted from fresh frozen tissue, a scatter plot of log2(counts) measured by nCounter442 gene assay (NanoString, Inc) versus log2(probe intensity) values ​​for the corresponding transcripts measured by HG-U133 plus2 (Affymetrix, Inc) is provided. [Figure 18B] A histogram of correlation values ​​for each mRNA transcript for 10 cases profiled by two platforms (nCounter, NanoString, Inc vs. HG-U133 plus 2 (Affymetrix, Inc) is provided. [Figure 18C] Gel images comparing RNA extracted by Qiagen (Q) or Storm kits (S) are provided. [Figure 18D] A bar graph comparing the % of RNA greater than 200 bases in RNA extracted by Qiagen (Q) or Storm kits (S) is provided. [Figure 18E] Scatter plots and histograms are provided comparing the log2(counts) generated by nCounter analysis of RNA from fresh frozen (FF) or FFPE samples extracted using Qiagen or Storm kits from corresponding biopsies. Numbers in boxes represent correlation coefficients. [Figure 18F] 13 provides a scatter plot of log2(counts) generated by nCounter analysis of RNA extracted from FF or FFPE tissue from corresponding biopsies. [Figure 19A] A histogram of correlation values ​​for each mRNA transcript for FFPE RNA analyzed by nCounter (NanoString, Inc) and FF RNA analyzed by HG-U133 plus 2 (Affymetrix, Inc) is provided for 100 cases profiled on both platforms. [Figure 19B]A histogram of correlation values ​​for each mRNA transcript in the classifier for FFPE RNA analyzed by nCounter and FF RNA analyzed by HG-U133 plus 2 is provided for 100 cases profiled on both platforms. [Figure 19C] Heatmaps of Affymetrix U133 plus2 data from FF RNA from cases included in the training cohort showing the original but expanded (C) and reduced (D) classifier gene lists that were ultimately used in the nCounter platform classification model. Housekeeping genes that do not vary between PTCL subtypes are shown for comparison. [Figure 19D] Heatmaps of Affymetrix U133 plus2 data from FF RNA from cases included in the training cohort showing the original but expanded (C) and reduced (D) classifier gene lists that were ultimately used in the nCounter platform classification model. Housekeeping genes that do not vary between PTCL subtypes are shown for comparison. [Figure 19E] 13 provides box plots and dot plots of the variance of housekeeping and model genes analyzed in FFPE tissues by nCounter in a run of training and validation cases. [Figure 19F] Kaplan-Meier curves of overall survival for PTCL-NOS cases classified as GATA3 or TBX21 using FF RNA analyzed by HG-U133 plus2 Affymetrix arrays (top) or FFPE RNA analyzed by nCounter (bottom). [Figure 20]Comparison of model scores (FFPE RNA) generated at UNMC (initial design) vs. model scores in the final diagnostic model run at three different locations (UNMC, Insight.Inc, and COH). The threshold for classifying a case as a PTCL entity is labeled in green. The light green line indicates the mean AITL / PTCL-NOS score threshold. The red line shows the results of the orthogonal regression, which can be compared to the grey line showing the diagonal of equal model. *nCounter panel using a reduced gene design containing 153 CodeSets. **nCounter panel using the initial design containing 442 CodeSets. All cases were run against the original 442 CodeSet design at UNMC and against the final 153 CodeSet design elsewhere, while 7 samples were also repeated at UNMC using the 153 final design CodeSet panel. All scores were generated using the final model. [Figure 21] Figure 1 provides violin and dot plots of classification scores for all PTCL cases profiled in the training and validation cohorts with nCounter. The x-axis is the known diagnosis. Cases where the nCounter classification did not match the expected diagnosis are colored according to how they were classified, while matched cases are gray. [Figure 22] Shown are H&E and IHC staining for one nodular ENKTL case with strong EBER and cytotoxicity (TIA-1 and perforin) marker expression. [Figure 23A] 1 provides heat maps of TFH, AITL, GATA3 and TBX21 signature expression in AITL, PTCL-TFH, PTCL-TBX21, PTCL-GATA3, and reactive hyperplasia. [Figure 23B] Box plots of the mean expression of TFH-specific genes in the indicated entities are provided. [Figure 24]Concordance between gold standard diagnosis and refined diagnostic signature in training cohort. AITL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ATLL, adult T-cell leukemia / lymphoma; ENKTCL, extranodal NK / T-cell lymphoma; PTCL-NOS, peripheral T-cell lymphoma not otherwise specified; FFPE, formalin fixed paraffin embedded tissue; FF, fresh frozen tissue. *Borderline cases were not considered discordant. #Combined pathologic and molecular diagnosis based on fresh frozen GEP. [Diagram 25] Concordance between the gold standard diagnosis and the refined diagnostic signature in the validation cohort is shown. The transcriptional classifier was concordant with the pathology diagnosis given by three expert hematopathologists in 85% (119 / 140) of cases and showed a “borderline association” with the molecular signature in 6% (8 / 140). Overall, 127 / 140 cases (green) are considered concordant between the consensus pathology test and the molecular diagnosis. The classifier improved the pathology diagnosis in two PTCL-NOS cases molecularly classified as ATLL (light blue) that were concordant with subsequent clinicopathological information on retesting. Of the remaining 11 cases that were “discordant,” four had molecular classifications (yellow) that could provide an improvement over the pathology diagnosis based on the overall transcriptomic signature and morphological features. AITL, angioimmunoblastic T-cell lymphoma, ALCL, anaplastic large cell lymphoma, ATLL, adult T-cell leukemia / lymphoma, ENKTCL, extranodal NK / T-cell lymphoma, PTCL-NOS, peripheral T-cell lymphoma not otherwise specified. *Borderline cases were not considered discordant. +PTCL-NOS cases molecularly classified as ATLL that were concordant with subsequent clinicopathologic information on restaging. ++PTCL-NOS cases molecularly classified as AITL that showed TFH marker expression on restaging. +++ALK-ALCL cases that may represent ALK-positive-like ALCL based on global transcriptomic signature. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Description of exemplary embodiments Details of one or more embodiments of the subject matter disclosed herein are provided herein. However, it should be understood that the subject matter disclosed herein can be embodied in various forms. Modifications to the embodiments described herein and other embodiments will be apparent to those skilled in the art after reviewing the information provided herein. The information provided herein, particularly the specific details of the exemplary embodiments described, are provided primarily for clarity of understanding and should not be construed as limiting. In case of conflict, the present specification, including definitions, shall prevail.

[0026] Although the terms used herein are believed to be well understood by those skilled in the art, certain definitions are provided to facilitate the description of the subject matter disclosed herein. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the subject matter disclosed herein belongs.

[0027] Although any methods, devices, and materials similar or equivalent to those described herein can be used in the practice or testing of the subject matter disclosed herein, representative methods, devices, and materials are described below.

[0028] In certain cases, the nucleotides and polypeptides disclosed herein are contained in publicly available databases, such as GENBANK® and SWISSPROT. Information, including sequences and other information related to such nucleotides and polypeptides, contained in such publicly available databases is expressly incorporated by reference. Unless otherwise indicated or apparent, references to such publicly available databases are to the most recent version of the database as of the filing date of this application.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this subject matter belongs. All patents, patent applications, published applications and publications, GenBank sequences, databases, websites and other published materials mentioned throughout this disclosure are incorporated by reference in their entirety unless otherwise noted. In the event that there are multiple definitions for terms herein, the definitions in this section prevail. When reference is made to a URL or other such identifier or address, it is understood that such identifiers may change and specific information on the Internet may come and go, but equivalent information may be found by searching the Internet. Reference thereto evidences the availability and public dissemination of such information.

[0030] Although any methods, devices, and materials similar or equivalent to those described herein can be used in the practice or testing of the subject matter disclosed herein, representative methods, devices, and materials are described below.

[0031] Following long-standing patent law convention, the terms "a," "an," and "the" when used in this application, including the claims, refer to "one or more." Thus, for example, a reference to "a cell" includes a plurality of such cells, and so forth.

[0032] The term "substantially" permits deviations from the descriptive language that do not adversely affect the intended purpose. It is understood that the descriptive language is modified by the term "substantially" even if the word "substantially" is not explicitly recited. Thus, for example, the phrase "wherein the lever extends vertically" means "wherein the lever extends substantially vertically," unless precise vertical positioning is necessary for the lever to perform its function.

[0033] Terms such as "comprising," "including," "having," and "involving" (and similarly "comprises," "includes," "has," and "involves") are used interchangeably and have the same meaning. Specifically, each of the terms is defined consistent with the general U.S. patent law definition of "comprising," and therefore is to be construed as an open term meaning "at least the following," and not excluding additional features, limitations, aspects, etc. Thus, for example, "a process comprising steps a, b, and c" means that the process includes at least steps a, b, and c. Whenever the terms "a" or "an" are used, they are to be understood as "one or more," unless such an interpretation is insignificant in the context.

[0034] Unless otherwise indicated, all numbers expressing properties such as quantities of ingredients, reaction conditions, and the like used in the specification and claims are to be understood as being modified in all instances by the term "about." Accordingly, unless otherwise indicated, the numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the subject matter disclosed herein.

[0035] As used herein, the term "about" when referring to a value or amount of mass, weight, time, volume, concentration, or percentage is meant to encompass variations from the particular amount of, in some embodiments, ±20%, in some embodiments, ±10%, in some embodiments, ±5%, in some embodiments, ±1%, in some embodiments, ±0.5%, and in some embodiments, ±0.1%, where such variations are appropriate to perform the disclosed methods.

[0036] As used herein, a range can be expressed as "about" one particular value and / or to "about" another particular value. It is also understood that there are many values ​​disclosed herein, and each value is also disclosed herein as "about" that particular value in addition to the value itself. For example, if the value "10" is disclosed, then "about 10" is also disclosed. It is also understood that each unit between two particular units is also disclosed. For example, if 10 and 15 are disclosed, then 11, 12, 13, and 14 are also disclosed.

[0037] As used herein, a "biological sample" refers to a sample of biological material obtained from a subject. In embodiments, the subject includes a human subject. Examples of biological samples include tissues, tissue samples, cell samples, fluid samples, or combinations thereof. Biological samples can be obtained, for example, in the form of tissue biopsies, such as suction biopsies, brush biopsies, surface biopsies, needle biopsies, punch biopsies, excision biopsies, open biopsies, and endoscopic biopsies. A "biological sample" can refer to a sample of tissue or fluid isolated from a subject, including, but not limited to, blood, plasma, serum, tumor biopsy, urine, feces, sputum, spinal fluid, pleural fluid, nipple aspirate, lymphatic fluid, external sections of skin, respiratory, intestinal, and genitourinary tracts, tears, saliva, milk, cells, tumors, organs, and also samples of components of in vitro cell cultures. Non-limiting examples of biological samples include formalin-fixed paraffin-embedded (FFPE) tissue, fresh frozen (FF) samples, or other prepared sample types.

[0038] Those skilled in the art will recognize that a biological sample from a subject can be obtained by a variety of conventional techniques. The methods described herein typically involve obtaining a biological sample from a subject, such as a subject suspected of having or afflicted with cancer. As used herein, the phrase "obtaining a biological sample" refers to any process for directly or indirectly acquiring a biological sample from a subject. For example, a biological sample can be obtained (e.g., at a point-of-care facility, e.g., a clinic, hospital, laboratory facility) by obtaining a tissue or fluid sample (e.g., tissue biopsy, blood draw, bone marrow sample, spinal tap) from a subject. Alternatively, a biological sample can be obtained by receiving a biological sample (e.g., at a laboratory facility) from one or more persons who have obtained the sample directly from the subject. A biological sample can be, for example, tissue (e.g., tissue biopsy, blood), cells (e.g., hematopoietic cells such as hematopoietic stem cells, white blood cells, or reticulocytes, stem cells, or plasma cells, cancer cells), vesicles, biomolecular aggregates, or platelets from a subject.

[0039] In some embodiments, the sample is from a resection, biopsy, or core needle biopsy of a primary or metastatic tumor. In addition, fine needle aspirate samples are used. The sample can be either paraffin-embedded or frozen tissue.

[0040] In embodiments, the biological sample may be a biopsy tissue, which contains at least one cell that is or is suspected to be a cancer cell. The term "biopsy tissue" may refer to a sample of tissue that is removed from a subject for the purpose of determining whether the sample contains cancerous tissue or testing the tissue for cancer. In some embodiments, the biopsy tissue is obtained because a subject is suspected of having cancer. The biopsy tissue is then tested for the presence or absence of cancer. In other embodiments, the biopsy tissue is obtained because a subject is known to have cancer, and the biopsy tissue is then tested for markers that indicate the stage and / or treatment options of the cancer.

[0041] As used herein, "diagnosing" refers to detecting and identifying a disease in a subject. The term can also encompass assessing or evaluating disease status (progression, regression, stabilization, response to treatment, etc.) in a patient known to have the disease (e.g., PTCL).

[0042] As used herein, the term "prognosis" refers to providing information regarding the impact of the presence of cancer (e.g., as determined by the diagnostic method of the present invention) on the future health of a subject (e.g., predicted morbidity or mortality, likelihood of developing cancer, and risk of metastasis). In other words, the term "prognosis" refers to providing a prediction of the expected course and outcome of cancer, or a prediction of the likelihood of recovery from cancer. The term "prognosis" is recognized in the art and includes predictions about the likely course of disease or disease progression, particularly with respect to the likelihood of disease remission, disease relapse, tumor recurrence, metastasis, and death. A "good prognosis" may refer to the likelihood that a patient suffering from cancer will remain cancer-free after therapy. A "poor prognosis" may refer to the likelihood of relapse or recurrence of the underlying cancer after treatment, the likelihood of developing metastasis, and / or the likelihood of death. In certain embodiments, the time frame for evaluating prognosis is, for example, less than 1 year, 1 year, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, 15 years, 20 years, or longer.

[0043] As used herein, the terms "treat," "treatment," or "treating" refer to any type of treatment that benefits a patient suffering from a disease, including improving the patient's condition (e.g., one or more symptoms), slowing the progression of a condition, and the like.

[0044] An aspect of the present invention relates to a method for the treatment of cancer in a patient in need thereof. Treating cancer, as used herein, may refer to partially or completely inhibiting, delaying, or preventing the progression of cancer, including cancer metastasis, inhibiting, delaying, or preventing the recurrence of cancer, or preventing the onset or development of cancer in a mammal, e.g., a human (chemoprevention). Treating cancer may be indicated by stopping or reducing tumor initiation, tumor growth, proliferation, angiogenesis, and / or metastasis.

[0045] Aspects of the invention also include providing diagnostic information about cancer in a subject. For example, aspects of the invention may include providing diagnostic information to a subject suspected of having cancer or at risk of having cancer. The term "subject suspected of having cancer" may refer to a subject who presents with one or more symptoms indicative of cancer (e.g., a noticeable lump or mass) or is being screened for cancer (e.g., during a routine physical exam). A subject suspected of having cancer may also have one or more risk factors. A subject suspected of having cancer generally has not been tested for cancer. However, "subject suspected of having cancer" encompasses individuals who have received an initial diagnosis but whose stage of cancer is unknown. The term further includes people who have previously had cancer (e.g., individuals in remission). As used herein, the term "subject at risk of cancer" refers to a subject who has one or more risk factors for developing a particular cancer. Risk factors include, but are not limited to, gender, age, genetic predisposition, environmental exposure, previous cancer occurrence, existing non-cancer disease, and lifestyle.

[0046] The phrase "effective amount" refers to an amount of a therapeutic agent that results in an improvement in a patient's condition. In certain embodiments, an effective amount increases survival rate for at least one year. An effective amount can be measured by the number of years of life after initiation of treatment without substantial disease progression. In embodiments, an effective amount can be an amount sufficient to cause a progression-free survival (PFS) of at least one year. An effective amount can be sufficient to induce a PFS of greater than one year. In embodiments, an effective amount is a dosage sufficient to cause a 1-year PFS, a 2-year PFS, a 3-year PFS, a 4-year PFS, or a 5-year PFS. In embodiments, the therapeutic agent is an anti-cancer agent.

[0047] As used herein, the term "subject" refers to the target of administration. The subject of the methods disclosed herein may be a vertebrate, such as a mammal, fish, bird, reptile, or amphibian. Thus, the subject of the methods disclosed herein may be a human or non-human. Thus, veterinary therapeutic applications are provided in accordance with the subject matter disclosed herein.

[0048] "Subject" may refer to mammals, such as humans and non-human primates, and mammals important due to being endangered, such as the Amur tiger; animals of economic importance, such as animals raised on farms for human consumption; and / or animals of social importance to humans, such as animals kept as pets or in zoos. Examples of such animals include, but are not limited to, carnivores, such as cats and dogs; swine, including pigs, hogs, and wild boars; ruminants and / or ungulates, such as cows, oxen, sheep, giraffes, deer, goats, bison, and camels; rabbits, guinea pigs, and rodents. Treatment of birds is also provided, including treatment of endangered and / or zoo-raised bird species, as well as poultry, more specifically domesticated poultry, i.e., turkeys, chickens, ducks, geese, guinea fowl, and the like, which are also economically important to humans. Thus, treatment of livestock is provided, including, but not limited to, domesticated swine, ruminants, ungulates, horses (including race horses), poultry, and the like.

[0049] In an embodiment, the subject has previously been diagnosed with cancer and perhaps has already been treated for cancer. In an alternative embodiment, the subject has not previously been diagnosed with cancer. The present invention may be useful for all patients at risk for cancer. Each type of cancer has its own set of risk factors, but the risk of developing cancer increases with age, sex, race, and personal and family medical history. Other risk factors are largely related to lifestyle choices, while certain infections, occupational exposures, and some environmental factors may also be associated with the development of cancer.

[0050] As used herein, the terms "administering" and "administration" refer to any method of providing a pharmaceutical preparation to a subject. Such methods are well known to those skilled in the art and include, but are not limited to, oral administration, transdermal administration, administration by inhalation, intranasal administration, topical administration, intravaginal administration, eye drop administration, buccal administration, intracerebral administration, rectal administration, and parenteral administration, including injections such as intravenous administration, intraarterial administration, intramuscular administration, and subcutaneous administration. Administration can be continuous or intermittent. The preparation can be administered therapeutically, i.e., administered to treat an existing condition of interest. The preparation can be administered prophylactically, i.e., administered for the prevention of a condition of interest.

[0051] The subject matter disclosed herein relates to a method for diagnosing, prognosing, or treating cancer in a subject by determining the presence or amount of one or more biomarkers in a biological sample from the subject. In an embodiment, the cancer is non-Hodgkin's lymphoma. The cancer can be PTCL.

[0052] In one embodiment, the method includes obtaining a biological sample from a subject, the biological sample comprising at least one cell that is or is suspected to be a cancer cell; measuring or determining an expression level of a gene from the biological sample; comparing the expression level in the biological sample to a control sample; and providing diagnostic, prognostic, or predictive information based on the measuring step.

[0053] In embodiments, measurements such as gene nucleotide or protein levels can be compared to a control, and if different from that of the control, the subject may have an increased likelihood of having and / or developing cancer, a decreased likelihood of having or developing cancer, an increased likelihood of responding to a given treatment, or a decreased likelihood of responding to a given treatment. A sample or subject that is "different from that of the control" is understood to have a level of an analyte or diagnostic or therapeutic indicator (e.g., marker) that is detected at a statistically different level than a sample from a control sample that is normal, untreated, or abnormal. Determining statistical significance is within the capabilities of one of ordinary skill in the art, for example, the number of standard deviations from the mean that constitute a positive or negative result. For example, the control sample may include one or more non-cancerous cells, or a biological sample from a patient not diagnosed with cancer.

[0054] In embodiments, measurements such as mRNA and / or protein levels can be compared to thresholds to determine whether a subject has cancer, diagnose a type or subtype of cancer, provide a patient prognosis, determine a particular treatment regimen, or combinations thereof. The term "threshold" can refer to a value derived from multiple biological samples for a biomarker, exceeding which threshold is associated with an increased likelihood of having and / or developing cancer, or an increased likelihood of responding to a given treatment.

[0055] According to certain embodiments of the method of the present invention, the presence or amount of a gene product (e.g., a polypeptide or a nucleic acid) encoded by a gene is detected in a sample derived from a subject (e.g., a tissue or cell sample obtained from a tumor or a blood sample obtained from a subject). The sample can be subjected to various processing steps prior to or during the detection sequence.

[0056] For purposes of the present invention, the term "gene" has the same meaning as understood in the art. However, it will be understood by those skilled in the art that the term "gene" has various meanings in the art, some of which include gene regulatory sequences (e.g., promoters, enhancers, etc.) and / or intron sequences, while others are limited to coding sequences. It will be further understood that the definition of "gene" includes reference to nucleic acids that do not code for proteins, but rather code for functional RNA molecules such as tRNAs. For clarity, the inventors note that the term "gene" as used in this application generally refers to a portion of a nucleic acid that codes for a protein. The term can optionally include regulatory sequences. This definition is not intended to exclude the application of the term "gene" to non-protein-coding expression units, but rather to clarify that in most cases, the term as used herein refers to a protein-coding nucleic acid.

[0057] A gene product or expression product is generally RNA transcribed from a gene or a polypeptide encoded by RNA transcribed from a gene.

[0058] Expression of a gene can be measured by a variety of techniques known in the art. Certain techniques may utilize a polynucleotide corresponding to part or all of a gene, rather than an antibody that binds to a polypeptide encoded by the gene. Suitable techniques include, but are not limited to, in situ hybridization, Northern blots, and various nucleic acid amplification techniques (e.g., PCR, quantitative PCR, and ligase chain reaction).

[0059] In some embodiments, the present application provides a method for identifying different subgroups of PTCL based on the presence or amount of one or more biomarkers in a biological sample. Non-limiting subgroups of PTCL include PTCL-NOS, AITL, ALCL, ALK(-)ALCL, ENKTL, NK and γδ-T-PTCL. In embodiments, PTCL-NOS can be further subdivided into one of two major molecular subtypes characterized by high expression of transcription factors GATA3 or TBX21 and either of their target genes.

[0060] In certain embodiments, the biomarkers include gene signatures that are unique to different subtypes of PTCL. The gene signature for each subtype of PTCL may include one or more of the genes listed in Table 1 below.

[0061] Table 1. Exemplary PTCL subtype gene signatures TIFF2025506567000002.tif108169

[0062] In some embodiments, the one or more biomarkers include GATA, TBX21, or their associated target genes. The one or more biomarkers may include NK or gamma / delta T cell lineage. In certain embodiments, the NK lineage includes EBV viral transcription (EBER and LMP1). In embodiments, the gene signature for ATLL subtypes further includes HTLV-1 viral transcript expression (TAX, HBZ, or a combination thereof).

[0063] In some embodiments, the present application provides methods for diagnosing, prognosing, and / or treating cancer in a subject by determining the expression profile of a specific panel of genes.

[0064] In some embodiments, the present application provides a method for predicting the survival of a subject diagnosed with a particular subgroup of PTCL. In some embodiments, the method involves detecting the expression of a profile of a particular panel of genes.

[0065] Further, in some embodiments of the subject matter disclosed herein, a method for treatment for a particular subgroup of PTCL in a subject is provided. In embodiments, the treatment method includes administering to the subject an effective amount of an anti-cancer agent. The anti-cancer agent may vary depending on the PTCL subtype of the subject. In embodiments, the treatment is optimized to specifically target the PTCL subtype of the subject. Non-limiting examples of such agents include chemotherapy, immunotherapy, toxin therapy, radiation therapy, or combinations thereof. The anti-cancer agent may include pralatrexate, romidepsin, belinostat, brentuximab vedotin, duvelisib, decitabine and related compounds, EZH1 / 2 inhibitors, other PTCL subtype-specific treatments now known or subsequently discovered, or combinations thereof.

[0066] In addition, the subject matter disclosed herein relates to kits and reagents for the diagnosis, prognosis, and / or treatment of PTCL. In some embodiments, the kits include a means for detecting the expression profile of a particular panel of genes. In some embodiments, the kits include one or more probes.

[0067] Furthermore, the subject matter disclosed herein relates to methods for identifying a set of biomarkers for the diagnosis, prognosis, and / or treatment of PTCL using software and analysis systems. In some embodiments, the subject matter disclosed herein provides a method for subtyping PTCL by comparing the expression levels of several genes.

[0068] The embodiments use methodologies for analyzing multiple nucleic acids in a single reaction. In embodiments, the methods or kits use high-throughput nucleic acid sequencing technologies, such as next-generation sequencing (NGS). In embodiments, targeted RNA sequence panels can be provided on NGS platforms. Non-limiting examples of such NGS technologies include instruments and protocols from Illumina, Inc. (San Diego, CA, USA), Thermo Fischer Scientific (Waltham, MA, USA), and Qiagen (Venlo, Netherlands). In still other embodiments, specific multiplex PCR-based platforms are used. The embodiments use qPCR-based platforms configured to accommodate a significant number of analytes, such as Qiagen Modaplex (Venlo, Netherlands) or similar platforms. The embodiments can use conventional real-time PCR platforms. Alternative embodiments utilize any of a variety of expression microarray-based platforms, including but not limited to Affymetrix (Santa Clara, CA, USA), Aknonni Biosystems (Frederick, MD, USA), Biofire Diagnostics (Salt Lake City, UT, USA) or similar platforms. Some embodiments use non-array platforms and protocols. In certain non-array embodiments, quantitative nuclease protection assays are utilized. Other non-array embodiments use platforms and protocols from NanoString Technologies, Inc. (Seattle, WA, USA). The examples provided herein are merely illustrative and should not be considered limiting in any way. Embodiments use any of a variety of molecular diagnostic platforms.

[0069] In a specific embodiment, a DNA chip is used, which is a convenient device for comparing the expression levels of a large number of genes simultaneously. DNA chip-based expression profiling can be carried out, for example, by the method disclosed in "Microarray Biochip Technology" (Mark Schena, Eaton Publishing, 2000).

[0070] A DNA chip contains immobilized high-density probes for detecting several genes. Therefore, the expression levels of many genes can be estimated simultaneously in one analysis. That is, the DNA chip can be used to determine the expression profile of a specimen.

[0071] In some embodiments, the subject matter disclosed herein relates to a method for identifying biomarkers specific to different subtypes of cancer, such as PTCL. An embodiment includes performing a diagnostic algorithm based on GEP analysis of a biological sample. Certain embodiments use a stepwise binary model algorithm for subgroup classification. An embodiment provides a method for substantially reducing the number of genes in a cancer subgroup classifier without significantly compromising diagnostic accuracy. In an embodiment, the total number of genes is reduced to less than 50. In certain embodiments, the total number of genes is less than 25. The total number of genes can be reduced to 15-20 (inclusive). In some embodiments, the total number of genes is reduced to about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, to about 30. The methods further include identifying distinct oncogenic pathways that contain potential therapeutic targets in these various molecular PTCL entities.

[0072] In some embodiments, the presently disclosed subject matter provides methods of distinguishing PTCL subtypes using refined diagnostic and prognostic biomarkers in a subject.

[0073] The methods described herein are also useful for selecting a treatment for cancer. For example, an embodiment may include providing a cancer sample or obtaining (e.g., isolating) a cancer sample from a subject, measuring at least one of the expression, activity, or product of a gene in a PTCL signature, and selecting a treatment based on the measuring step. An embodiment may further include comparing the measured level with a control sample or a threshold level.

[0074] Aspects of the invention also include kits, eg, kits for treating and / or diagnosing cancer.

[0075] In certain embodiments, the kits can include one or more compositions described herein, such as, for example, an inhibitor.

[0076] For example, the kit may include one or more reagents useful for detecting the level of genes in PTCL signatures or other proteins in a sample. One or more reagents may be immobilized on a solid support. Non-limiting examples of compositions of the solid support structure include plastic, cardboard, glass, plexiglass, tin, paper, or combinations thereof. The solid support may also include a dipstick, spoon, spatula, filter paper, or swab.

[0077] The reagents may include a labeled compound or agent capable of detecting cancer or tumor cells (e.g., scFv or monoclonal antibody) in a biological sample; a means for determining the amount of gene expression in the sample; and a means for comparing the amount of gene expression in the sample to a standard, such as a control sample or a threshold value. The standard, in some embodiments, is a non-cancerous cell or a cellular extract thereof. The compound or agent may be packaged in a suitable container. The kit may further include instructions for using the kit to detect cancer in a sample.

[0078] In embodiments, the kit may also include primers for amplifying mRNA transcribed from a gene encoding a polypeptide and / or a control sample for testing the primers. For example, the control sample may include a nucleic acid that hybridizes to the primers.

[0079] The kit may also include a sample collection apparatus, such as a device used to collect biological fluids or tumor biopsies.

[0080] In some embodiments, the kits may include containers containing one or more reagents, and, optionally, (b) informational material. The informational material may be descriptive, instructional, marketing, or other material relating to the methods described herein and / or use of the agents for diagnostic purposes. In some embodiments, the kits also include one or more anti-cancer therapeutic agents.

[0081] The informational material of the kit is not limited in its form. In one embodiment, the informational material may include information about the manufacture of the components of the kit, such as molecular weight, concentration, expiration date, batch or manufacturing site information, etc. In one embodiment, the informational material relates to methods of using the components of the kit. The information can be provided in a variety of formats, including printed text, computer readable materials, video or audio recordings, or information that provides a link or address to the substantive material.

[0082] The kit may include other components such as solvents or buffers, stabilizers, or preservatives. Optionally, the kit may include a therapeutic agent that can be provided in any form, for example, liquid, dry, or lyophilized, preferably substantially pure and / or sterile. If the agent is provided in a solution, the liquid solution is preferably an aqueous solution. If the agent is provided as a dry form, reconstitution is generally by the addition of a suitable solvent. A solvent, for example, sterile water or buffer, can be optionally provided in the kit.

[0083] Other embodiments Although the present invention has been described in conjunction with its detailed description, the foregoing description is intended to be illustrative, but not limiting, of the scope of the invention, which is defined by the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

[0084] The invention is further described in the following examples, which do not limit the scope of the invention described in the claims. EXAMPLES

[0085] Examples are provided below to facilitate a more complete understanding of the present invention. The following examples illustrate exemplary modes of making and practicing the present invention. However, the scope of the present invention is not limited to the specific embodiments disclosed in these examples, and are for illustrative purposes only, since alternative methods can be used to obtain similar results.

[0086] Example 1 The overarching goal of this study is to transition the "academic" PTCL diagnostic and prognostic gene expression (GE) signatures originally developed using fresh frozen specimens analyzed by expression microarrays into an optimized lockdown signature applicable to formalin-fixed paraffin-embedded (FFPE) specimens analyzed on the Nanostring nCounter platform. The two objectives of this proposal (Objective 1. To identify a set of reduced transcripts and generate a new model capable of replicating the existing "academic" diagnostic and prognostic PTCL signatures, Objective 2. To confirm the applicability of the refined diagnostic and prognostic PTCL gene signature set on the NanoString nCounter platform for the analysis of FFPE specimens) will be performed in a Phase II study immediately following this proposal to determine the feasibility of the optimized FFPE / nCounter assay as a candidate for further commercial test development, including a full evaluation and validation of the analytical and clinical performance and clinical utility of the assay.

[0087] Specific Purpose Phase I: Optimization of "academic" gene expression profiling (GEP) signatures for peripheral T-cell lymphoma (PTCL) diagnosis and prognosis and initial assessment of the feasibility of the optimized signatures for commercial assay development. Objectives. The overarching goal of this study is to transition the "academic" PTCL diagnostic / prognostic GEP signature, originally developed using fresh frozen (FF) specimens analyzed by expression microarray, into an optimized lockdown signature applicable to formalin-fixed paraffin-embedded (FFPE) specimens analyzed on the Nanostring nCounter platform. This effort is divided into two objectives, the specific activities of each of which are detailed below. These Phase I studies will help determine the feasibility of the optimized FFPE specimens / nCounter assay as a candidate for further commercial test development (including its offering in CLIA-certified laboratories and its further development as an in vitro diagnostic (IVD) kit and companion diagnostic (CDx) assay for new or existing PTCL therapies), which, if the Phase I objectives are met, will be followed up in a Phase II follow-up study that includes a full evaluation and validation of the analytical and clinical performance of the assay, as well as clinical utility.

[0088] Aim 1: To identify sets of reduced transcripts and generate new models capable of replicating existing "academic" diagnostic and prognostic PTCL signatures. 1a. The reduced set of transcripts required for optimal discrimination of each PTCL entity is identified. Deliverables: Identification of those gene transcripts from the original PTCL subtyping signature that can be transferred from fresh frozen (FF) specimens / expression array platform to FFPE specimens / nCounter platform. These studies will determine the performance of the codesets in RNA from approximately 60 FFPE specimens on the nCounter platform compared to data from FF samples from the same cases previously characterized using arrays.

[0089] 1b. To refine diagnostic and prognostic signatures that can be applied to FFPE tissues to differentiate PTCL entities.Deliverables: Identification of a single working minimal gene set capable of reproducing the original "academic" PTCL signature with as close a match as possible to 1. Using the transcripts represented in Objective 1a on a training set of approximately 120 PTCL cases sufficient to transfer from FF / microarray to the FFPE / nCounter platform, develop a classification model and algorithm comprised of the minimal number of genes capable of distinguishing clinically relevant diagnostic and prognostic PTCL subtypes.

[0090] Aim 2: To validate the applicability of the refined diagnostic and prognostic PTCL gene signature set on the NanoString nCounter platform for the analysis of FFPE specimens. 2a. A lockdown diagnostic and prognostic PTCL gene signature will be developed and its inter-laboratory reproducibility will be assessed. Deliverables: Lockdown of an optimized PTCL signature based on initial validation cohort and inter-laboratory study results. Approximately 100 PTCL FFPE cases will be subtyped using the refined diagnostic and prognostic gene signature independently in molecular diagnostic laboratories at UNMC, City of Hope, and HealthChart, Insight Molecular Labs along with a consulting commercial CLIA lab, and results will be compared to previous FF / microarray gold standard subtyping from matched specimens. Inter-laboratory assay reproducibility will be assessed, assay algorithms finalized, and final assay SOPs will be derived.

[0091] 2b. Validate "locked" diagnostic / prognostic signatures for Phase II assay transition. Deliverables: Validation cohort results confirming suitability of the lockdown signature for advancement into Phase II development. Analysis of a PTCL validation cohort of approximately 30 FFPE cases will be performed on the nCounter platform in multiple independent laboratories worldwide to confirm suitability of the optimized subtyping signature, algorithm, and SOP for advancement into Phase II assay development.

[0092] A. Chemical and Clinical Significance Peripheral T-cell lymphomas (PTCLs) are heterogeneous entities with therapeutic challenges: PTCLs constitute up to 20% of non-Hodgkin lymphomas 1、2 The World Health Organization (WHO) recognizes several PTCL subtypes, including angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma (ALCL), adult T-cell leukemia / lymphoma (ATLL), and extranodal NK / T-cell lymphoma (ENKTL). 3 The diagnosis of PTCL is difficult even for expert hematopathologists. 4 Using current immunophenotypic and molecular markers, 30-50% of PTCL cases cannot be classified and are classified as PTCL, not otherwise specified (PTCL-NOS). 3 ALK(+)ALCL (characterized by co-PI, Dr. Steve Morris) 5 With the exception of ALK(-)ALCL with DUSP22 rearrangements, patients with PTCL have a poor prognosis with CHOP-like chemotherapy or more intensive regimens. 4、6、7 In fact, the 5-year progression-free survival (PFS) rate has remained at only 20-25% over the past 20 years. 8~12 However, new therapies targeted at specific patient populations are beginning to be tested, and some have shown impressive results (see Table 2). 13~16 Accurate PTCL classification is therefore essential both in therapeutic trials and in selecting the best treatment for patients in the clinic. 17 (See, e.g., letters of support from Celgene and Seattle Genetics).

[0093] Table 2: Phase II trials of novel therapeutic agents with responses by subtype (adapted from ref. 13). TIFF2025506567000003.tif27166

[0094] New approaches are needed to improve PTCL diagnosis: We performed extensive gene expression profiling (GEP) studies to define robust molecular signatures capable of classifying the major subtypes of PTCL. 10~12 We were also able to reclassify some PTCL-NOS into separate entities instead (15% as AITL, 10% as ALCL, and 9% as γδ-PTCL). Importantly, we also identified two new major subgroups within PTCL-NOS that exhibit distinct enriched signaling pathways and distinct prognosis. Furthermore, we identified an expression signature that predicts survival in AITL patients. 10、11、18 .

[0095] Affymetrix 19 , NanoString 20、21 , and quantitative nuclease protection assay (qNPA; HTG Molecular Diagnostics, Inc.) 22 Among several platforms tested by the LLMPP for FFPE specimens, including IgG, the FDA-cleared NanoString nCounter system was superior in terms of specimen handling, lack of need for enzymatic reactions, and direct quantification of mRNA expression, while also providing high accuracy and sensitivity. 23 Of note, the LLMPP has utilized a similar approach to that described herein to establish a Lymph2Cx expression-based assay for DLBCL classification on the nCounter platform. 20、21 The assay was subsequently licensed exclusively by NanoString and is currently being developed in collaboration with Celgene (worth up to $45 million in upfront, developmental and regulatory milestone, and commercial payments to NanoString) as an in vitro diagnostic (IVD) companion diagnostic (CDx) test to support the development of lenalidomide as a treatment for patients with DLBCL (http: / / www.nanostring.com / company / corp_press_release?id=116).

[0096] innovation This approach will change practice for the diagnosis, prognosis and management of PTCL through the creation of high-content, quantitative and reproducible molecular assays that provide greater precision and accuracy in the diagnosis of currently defined PTCL subtypes as well as new entities not previously recognizable.

[0097] The inventors propose to translate the research diagnostic and prognostic signatures into a platform applicable to formalin-fixed paraffin-embedded (FFPE) tissues and validate the signatures initially in a CLIA laboratory setting and later as IVD kits, enabling widespread clinical use.

[0098] Because novel agents in clinical trials show differential response rates in different PTCL subtypes (see, e.g., Table 2), it is important to evaluate molecular signatures in experimental trials using standardized and validated assays. The molecular assays developed herein will make the implementation of precision medicine for PTCL much easier, just as the Lymph2Cx assay helps guide personalized therapy for DLBCL. 20、23 .

[0099] In summary, we develop and validate a molecular diagnostic / prognostic assay that reliably identifies clinically relevant PTCL subtypes in routinely available samples, enabling optimal personalized therapy and facilitating the testing of novel agents with precise patient stratification.

[0100] B. Preliminary Study B.1. Specimen cohort for assay development: The inventors have developed a cohort of over 1,000 FFPE cases with pathological and clinical data. 4 and has access to over 300 fresh frozen (FF) archived PTCL tissues. 10~12FFPE blocks corresponding to the FF samples used in our original GEP studies will be utilized for the refinement and optimization of the initial assay performed herein. We will obtain additional retrospectively collected PTCL samples as needed from multiple European and Asian collaborators (see supporting bios / letters) for the initial validation studies performed as part of this Phase I proposal, and during the subsequent Phase II studies. These tissue and data resources are unprecedented and, in combination with cutting-edge technology and sophisticated bioinformatics analysis, will enable us to further optimize the signature and validate the PTCL assay for clinical application.

[0101] B.2. A Robust Diagnostic and Prognostic Signature for PTCL: During development of the academic version of the PTCL assay, Drs. Chan and Iqbal, as well as extended LLMPP investigators, completed gene expression profiling (GEP) on 160 fresh frozen PTCL tissue biopsies using HG U133 plus 2 (U133) microarrays (Affymetrix). 11、12 These findings were then refined using a much larger series of cases (n=372). 10 Figure 1 shows the characteristic gene expression profiles of the major classes of PTCL. This study demonstrates: (a) the successful performance of a diagnostic algorithm based on GEP analysis of fresh frozen tissues from several centers using standardized operating procedures, (b) that a stepwise binary model algorithm can be used for subgroup classification in complex entities, and (c) that the number of genes in the subgroup classifier can be substantially reduced (e.g., from over 200 genes for AITL to 22 genes) without significantly compromising the diagnostic accuracy. 10 , and (d) clearly demonstrated that distinct oncogenic pathways, including potential therapeutic targets, can be identified in these various molecular PTCL entities.

[0102] Some additional notable findings included: (i) ALK(-) ALCL is a distinct entity, a finding that was later independently supported by both genomic and mutational data; 24~27 Thus, 11% of PTCL-NOS can be reclassified as this entity, which is highly responsive to the anti-CD30 therapy brentuximab vedotin (Table 2). (ii) ENKTL can be subdivided into NK and γδ-T cell subtypes. 10、12 Approximately 9% of PTCL-NOS cases are classified as γδ-PTCL, a subtype containing activating STAT5B mutations and sensitivity to JAK inhibitors. 28 ) can be reclassified as 10 Equally importantly, either Aurora kinase A or NOTCH1 inhibition was shown to induce significant cytotoxicity in NK-lymphoma cell lines. 12 (iii) AITL prognosis was highly dependent on the tumor microenvironment (i.e., B cell, dendritic cell, and monocyte cell content), with the patient quartile with the most favorable signature having a 5-year overall survival (OS) of 55%, whereas the most unfavorable quartile had an OS of only 15%. 10、11 .

[0103] B.3. Identification of distinct subgroups within PTCL-NOS: PTCL-NOS represents the largest portion of PTCL diagnosed with current histopathological methods. 4 Currently, high expression of either the transcription factors GATA3 (approximately 33%) or TBX21 (approximately 49%) and many of their target genes are 10 It can be divided into two main molecular subtypes characterized by high expression of GATA3 (Figure 2A). Cases with high expression of GATA3 show poorer clinical outcomes (Figure 2B). 10 .

[0104] B.4. Mutational analysis cross-validates molecular classification. Our mutational analysis of over 90 PTCL cases demonstrated that IDH2 R172 Exclusive presence of mutations 29、30, and CD28 and RHOA G17V showed preferential occurrence of mutations 31 In contrast, JAK1 and STAT3 mutations are more prevalent in ALK(-) ALCL. 26 STAT5B mutations are frequent in gamma delta-PTCL 28 These observations support the validity of molecular classification, and these mutations provide independent markers for evaluating the performance of molecular classifiers.

[0105] B.5. GEP Signatures in FFPE Tissues: Concerns have been raised that FFPE tissues may not produce RNA of sufficient quality for GEP assays. We have shown that GEP signatures obtained from FFPE tissues using expression arrays can robustly subclassify DLBCL. 19 In addition, the inventors have tested two non-array platforms, the quantitative nuclease protection assay (HTG Molecular Diagnostics, Inc.). 22、32 and nCounter technology 20、23、33 (NanoString Technologies) and demonstrated the feasibility of measuring the signature in archival DLBCL FFPE specimens. Additionally, industrial co-PI Dr. Morris and Brian Z. Ring, PhD, the lead industrial bioinformatician on the current proposal, performed GEP of triple-negative breast cancer FFPE specimens using RNA-seq and developed the commercially available Insight TNBCtype™ 101-gene assay to identify six molecular TNBC subtypes, each relevant for clinical management. 34 .

[0106] B.6. Gene signatures can be refined into smaller panels: Drs. Chan, Iqbal and colleagues selected 345 genes from their previous array studies as the most powerful predictors of PTCL diagnosis and prognosis. 10~12Here, we further refine this gene list to build a classifier with a smaller number of genes. Figure 3 shows calculations showing that using a smaller panel of 15-20 genes, there is no significant change in the error rate in distinguishing the two PTCL subtypes.

[0107] B.7. NanoString Technologies Platform for Profiling FFPE Tissue: The inventors were able to reduce the signature from FF specimens to a 15-gene set for FFPE tissue for DLBCL classification using the NanoString nCounter system: the "Lymph2Cx" signature. 20、23 (Figure 4A), demonstrating that the assay could be performed with high reproducibility across multiple laboratories (Figure 4B). The Lymph2Cx assay was recently selected as the CDx for FDA evaluation in the current Phase III ROBUST trial (Celgene; ClinicalTrials.gov ID: NCT02285062) testing R-CHOP + lenalidomide for ABC-DLBCL. The development of Lymph2Cx by Drs. Chan, Iqbal and his LLMPP colleagues will be replicated for the assay proposed herein to ensure the creation of a standardized and validated clinical diagnostic and prognostic signature. The likelihood of success is further enhanced by leveraging the core capabilities and strengths of our molecular diagnostic consulting company, Insight Genetics, Inc. (which will participate in the validation studies conducted in Objective 2a, as described below), as well as our corporate partner, HealthChart LLC.

[0108] C. Approach / Specific Objectives A note on technical terminology / statistical considerations: Throughout the text, transcripts are referred to as "probe sets" for the Affymetrix U133 array and "code sets" for the NanoString platform. The "training set" consists of cases for which corresponding GEP data from FF specimens analyzed on the U133 array are available. The "validation set" consists of non-duplicate cases against which the locked-down algorithm is evaluated. The final number of cases included in these sets, as well as all other statistical planning and analyses in this project, will be performed by Wu Consulting, Inc., a biostatistics company that has worked extensively in the pharmaceutical and biotechnology sectors since 1993 (www.wuconsulting.com; see letter).

[0109] Aim 1: To identify sets of reduced transcripts and generate new models capable of replicating existing "academic" diagnostic and prognostic PTCL signatures. Overview: The overarching goal of this Phase I STTR proposal is to transition academic PTCL diagnostic / prognostic GEP signatures developed using fresh frozen specimens analyzed by expression arrays into optimized lockdown signatures applicable to FFPE specimens analyzed by the Nanostring nCounter platform. Current signature 10 is larger than is needed and would be impractical and uneconomical for clinical assay development. Therefore, our initial work (Objective 1) will be to reduce the diagnostic set to the smallest possible number of genes using the same linear model methodology that was successfully used to develop the Lymph2Cx assay for DLBCL. 20 In addition to this, however, we will explore complementary approaches (e.g., centroid-based models as we have used previously) to help ensure success in creating robust models using minimal genes. 34 .

[0110] Deliverables: Creation of a reduced gene set used for the derivation of an optimized, commercially applicable PTCL classification model.

[0111] Objective 1a: Identify the reduced set of transcripts required for optimal differentiation of each PTCL entity. Pre-analytical considerations for FFPE specimens: We previously compared the performance of 667 diagnostic and prognostic genes identified using the U133 array for DLBCL against data obtained using the NanoString code set to analyze FFPE tissues (n=88) from the same cases. Figure 5A shows the correlation between code sets vs. probe sets for these transcripts. Most genes correlate very well with a median correlation of 0.75, but some perform very poorly. We specifically examined the correlation of 40 genes present in both this DLBCL gene set and our academic PTCL signature, observing a correlation of 0.76 between the two platforms (Figure 5B). We expect a similar correlation to be true for the remaining genes in our PTCL signature, and assuming that only about 15-20 genes are sufficient to distinguish each subtype in a binary model (Figure 3), our microarray study 10~12 A sufficient number of transcripts from the 345 genes identified in should be obtained to construct a model with a significantly reduced number of genes.

[0112] Power Calculation:The required number of matched FF and FFPE PTCL cases for Objective 1a is relatively small, since these studies only seek to determine the performance of CodeSets in RNA from FFPE specimens on the nCounter platform compared to data from corresponding FF samples previously characterized using arrays. We expect that 60 samples will be sufficient to constrain the correlation of well-performing transcripts, with a correlation between specimen and platform type of 0.80, between 0.68 and 0.88 with a 95% probability, and to constrain the correlation of poorly performing transcripts, with a correlation of 0.4, between 0.155 and 0.60, allowing easy discrimination between good and poorly performing transcripts. This sample size also allows for appropriate scaling and additional adjustments, such that for a typical sample, the error due to inadequate adjustments is less than one-fifth the error due to noise between FF and FFPE specimen modeling.

[0113] Aim 1b: Refining diagnostic and prognostic signatures that can be applied to FFPE tissue to distinguish PTCL entities. Method for iterative generation of gene signature predictors in FFPE tissue / NanoString assays: Our original FF RNA PTCL signature consists of a model score composed of a linear combination of expression values ​​with individual weights based on the discriminatory power of each gene, along with a set of cut points that divide samples into PTCL molecular subtypes. 10~12We generate a "training set" using FF PTCL specimens analyzed on the U133 array and matched FFPE tissues analyzed on the NanoString platform. This training set accurately estimates the correlation between any given Affymetrix probe set used in the FF tissue-based model and the corresponding NanoString code set on the FFPE specimen. This analysis indicates the extent to which platform changes introduce noise into the expression estimates, and therefore the extent to which certain genes should be downweighted or completely eliminated to show reduced predictive power in the model. Second, once the weighted scores are generated, the training set is used to estimate the centering and dynamic range changes of the final model scores to determine appropriate cut points. This should be easily accomplished by estimating the mean and SD of the model on the FFPE samples / NanoString assays and comparing the results to the mean and SD of the FF specimen / array model for the matched samples. To meet these goals, the training set required is estimated to be 120 cases (Figure 6).

[0114] Development of an algorithm for PTCL subtype classification and prognosis assessment: A detailed methodology for subtype classification is available in B-cell lymphomas. 20、23、35~40 and PTCL 10~12 has been previously described by our team. Briefly, in the current study, a series of four binary predictors are used to distinguish four individual subtypes AITL, ALCL, ATLL, and ENKTL from PTCL-NOS. Based on the results for these predictors, we classify specimens into one of six possible categories (AITL, ALCL, ATLL, ENKTL, PTCL-NOS, or indeterminate) (Table 3). "Indeterminate" samples have characteristics of two or more of the well-defined PTCL subtypes and, based on our previous work, are classified as "indeterminate" or "indeterminate". 10~12, indeterminate cases should represent a very small percentage (1-2%) of samples. PTCL-NOS cases are then separated into GATA3-high and TBX21-high subgroups, and ENKTL are also subdivided into NK and gamma / delta T cell lineages. Previously defined AITL prognostic subgroups 10、11 is transformed using the methodology described for diagnostic signatures. We identify well-performing code sets and divide patients into subgroups that show the greatest survival differentiation, adjusting their weights to best reproduce the model scores. To ensure that the prognostic model is robust, we proceed to lockdown and validation only if we can show that it differentiates survival with a one-sided p-value of less than 0.05.

[0115] (Table 3) Stepwise binary classification scheme for subtype analysis. TIFF2025506567000004.tif45139

[0116] -Sample size considerations:For this phase of the project, we propose to collect additional new samples to bring the total number of paired FF and FFPE specimens up to 120 cases (20 AITL, 10 ALK(+)ALCL, 10 ALK(-)ALCL, 20 ATLL, 20 ENKTL, 20 GATA3 PTCL, and 20 TBX21 PTCL). With this number of patients, we should be able to estimate the noise introduced by platform changes in well-measured genes (r>0.8) to within 12% of their true values ​​with 95% confidence. Estimates of the scaling and mean shift required to normalize our FFPE model scores to their frozen counterparts with 95% confidence suggest that the difference between the estimated model thresholds and their optimized values ​​is less than 20% of the predicted noise of individual samples due to platform changes. Furthermore, the 20 samples of each subtype for each binary distinction also allow us to normalize our model scores to their frozen counterparts, such that the difference between the estimated model cutpoints and their optimized values ​​is, with 95% confidence, only 12% of the variability of individual samples. Simulations suggest that this degree of imprecision would result in up to 4% of samples being predicted differently from how they would be predicted if, with 95% confidence, a similar NanoString model were used based on an infinite number of samples.

[0117] Potential issues and alternative approaches: Reclassification: Re-examine discordant cases for morphological differences between FF and FFPE tissues that could explain the discordance 10、20、23 Molecular classifications are cross-validated by analysis of somatic mutations and / or DNA copy number alterations that are unique or enriched in specific subtypes. 10~12、24~31 Clinical outcomes serve as validation of the prognostic signature. Classification models: There are many additional classification models that could be considered (e.g., the centroid-based model we have used previously while successfully developing a commercially available diagnostic assay based on similar expression). 34、41However, our PTCL signature has been tested and validated in FF tissues using a linear model, and we have not found any significant differences from other studies. 18、42 As shown in, the robustness of the same model in FFPE tissue is predicted. Number of genes in the model: The current "academic" signature contains 345 genes. Our preliminary bioinformatics work suggests that a total of 50-60 transcripts, or perhaps fewer, for the entire signature could be readily reduced. Notably, the NanoString platform can assay up to 800 transcripts in a single test.

[0118] Specific Aim 2: Verify the applicability of the refined diagnostic and prognostic PTCL gene signature set on the NanoString nCounter platform for analysis of FFPE specimens. Overview: The overarching goal of Objective 2 is to demonstrate that the refined diagnostic and prognostic signatures are robust, reproducible, and ready for transition to Phase II commercial development. The LLMPP consortium has demonstrated low inter-laboratory variability during sample rotation of frozen material trialed on the Affymetrix platform (Figure 7), demonstrating the ability to develop SOPs and obtain reproducible results across testing sites. Here, we will perform a similar study to confirm the reproducibility of our PTCL signatures using FFPE specimens on the NanoString platform in order to move the assay forward commercially.

[0119] Deliverables: Lockdown of an optimized PTCL signature based on training cohort and inter-laboratory study results (Objective 2a), validation cohort results confirming suitability of the lockdown signature for progression to Phase II development (Objective 2b).

[0120] Aim 2a: To develop a lockdown diagnostic and prognostic PTCL gene signature and evaluate its inter-laboratory reproducibility. We will evaluate our refined diagnostic and prognostic signatures using a set of non-overlapping PTCL cases (approximately 100) with the same subtype distribution as the training set. Consecutive scrolls from representative FFPE blocks will be blindly evaluated using a working "lockdown" standard operating procedure (SOP) developed from our training set study (Objective 1b). Assays will be performed in diagnostic laboratories at UNMC (Dr. Timothy Greiner), City of Hope (Dr. Raju Pillai), and HealthChart, Insight Molecular Labs (Drs. Steve Morris and Dave Hout - www.insightgenetics.com; see letter) along with consulting commercial CLIA labs.

[0121] Statistical considerations: The inventors of the present application have reviewed earlier studies conducted by the LLMPP. 10~12 A portion of the samples that were part of the FFPE model will be reanalyzed (but separate from the cases studied in Aim 1). With a relatively small number of specimens (approximately 100), it is not possible to conclusively confirm every single performance parameter of the FFPE model with the desired statistical criteria (due to limitations inherent in Phase I funding). Rather, the inventors will reanalyze a portion of the samples that were part of the FFPE model (but separate from the cases studied in Aim 2). The relatively small number of specimens (approximately 100) will not allow conclusive confirmation of every single performance parameter of the FFPE model with the desired statistical criteria (due to limitations inherent in Phase I funding). Rather, the inventors will reanalyze a portion of the samples that were part of the FFPE model (but separate from the cases studied in Aim 3). 20、21 We attempted to mimic this and used modest numbers for training and validation sets (51 and 68 cases, respectively) to confirm the overall fidelity of the subtype determination. This approach has proven successful in follow-up studies of the Lymph2Cx assay. Calls were concordant in 96% of 49 repeat biopsies and 100% of 83 FFPE specimens compared to the gold standard FF / array assay, and crucially, no misclassifications (ABC-DLBCL to GCB-DLBCL or vice versa) were observed. 21To estimate the predictive ability of the FFPE tissue model in PTCL, we evaluate its performance using leave-one-out cross-validation, which includes the selection of all weights and thresholds in the model as part of the cross-validation step.

[0122] Analysis Plan: FF / microarray diagnostics are considered the gold standard, and accuracy, precision, sensitivity, and specificity are based on comparison with array results. Assay results are provided to Drs. George Wright (NCI Biometric Research Program) and Brian Ring (HealthChart LLC), who will independently determine case classification. For reproducibility analysis, we compare agreement between model scores for final predicted classifications from three laboratory sites and their distance from the cut point to provide a more accurate estimate of prediction reproducibility. Based on our previous experience, we have found that models with NanoString are highly reproducible between sites, generally with correlations greater than 0.99. 20、21 Therefore, we set a high goal and estimate that 100 samples should be sufficient to achieve this criterion with 99% power for each model.

[0123] Objective 2b: Validate “locked” diagnostic / prognostic signatures for Phase II assay transition. The goal of objective 2b is to show that the locked assay can be easily applied clinically in multiple institutional laboratories using archival material and also to provide final troubleshooting of the assay, if necessary, to ensure its readiness for phase II development. A cohort of archival FFPE PTCL specimens (approximately 100-120 cases, to include 20 cases of each diagnostic and prognostic subtype) will be selected from laboratories around the world (see letter from the coordinating investigator). In these cases, Affymetrix expression array studies were not performed, so direct comparison is not feasible and more emphasis will be placed on the operational aspects of reproducibility, reliability, and problem solving (if necessary) across multiple laboratories using locked aspects and SOPs. In addition, emphasis will be placed on quality control measurements across different laboratories, which will be noted to refine SOPs for future routine testing.

[0124] Example 2 A plan for obtaining adequate specimens to develop and optimize the assay The inventors collected approximately 70 PTCL cases from their previous GEP studies. 10Pre-selection is a critical step as representative tumor sections must be cut to confirm the presence of representative and appropriate tissue and to validate the diagnosis of PTCL cases. These cases have undergone pathology review organized by Drs. Chan and Amador. These included cases with consensus initial diagnosis and molecular signature diagnosis. Of these, we have consensus diagnosis of AITL (n=18), ALCL (n=15), ENKTL (n=4), PTCL-NOS (n=30). Additional immunostains currently used for characterization or subclassification were performed on cases (Figure 8), including stains that may help distinguish between GATA3 and TBX21 subgroups. We have developed an immunohistochemistry-based algorithm that can subclassify PTCL-NOS into GATA3 and TBX21 subgroups. We will continue to refine this IHC algorithm and cross-validate the cases with gene signature diagnosis when developed.

[0125] Optimization of DNA / RNA isolation from formalin-fixed paraffin-embedded (FFPE) tissues: Tissue specimens are typically prepared by fixation with formalin (formaldehyde), which results in cross-linking and adducts of biomolecules and reduces the signal from molecular analyses involving hybridization procedures. Generally, isolation of DNA / RNA from FFPTE tissues using commercially available kits (e.g., Qiagen AllPrep DNA / RNA FFPE kit) involves heating in Tris buffer at high temperatures (>80°C) for several hours, which damages RNA and DNA. The present inventors have found that the isolation of DNA / RNA from FFPTE tissues using commercially available kits (e.g., Qiagen AllPrep DNA / RNA FFPE kit) involves heating in Tris buffer at high temperatures (>80°C) for several hours, which damages RNA and DNA. 26We routinely use these kits for DNA / RNA isolation for nCounter assays with good success. Because we expect some FFPE cases from our retrospective series to be over 10 years old, our analytical assay requires that at least 40% of the total RNA content in the reaction mixture for the nCounter assay is over 200 bp in length. Therefore, we tested a water-soluble bifunctional catalyst (anthranilate and phosphanilate) method that removes formaldehyde adducts from RNA and DNA at lower temperatures and shorter incubation times (Karmaker et.al Nat Chem. 2015;7(9):752-758), available at https: / / celldatasci.com / products / RNAstorm / . To test the utility of this protocol, we used 10 µm scrolls (2x) of representative FFPE blocks from 1988-2017 (non-essential cases) and compared them to the Qiagen FFPE isolation kit. Total RNA yields from these kits were in a similar range in these cases, although older blocks showed comparatively lower yields (Figure 9).

[0126] Analysis of the size distribution of RNA molecules using TapeStation-2200 showed that the Strom Kit produced relatively better (larger size) RNA molecules than the Qiagen Kit in older (>10 years) cases, while there was a more modest improvement in cases from FFPE blocks from 2014 (Figures 10A and 10B). Therefore, the priority is to use cases from 2010 or later for the nCounter assay in the early stages. In conclusion, the isolation procedure from the Qiagen Kit may be a viable method for recent cases (<5 years), while the Storm Kit isolation may be a valid alternative for RNA / DNA isolation when cases older than 10 years are required for optimization, especially in rare entities such as ENKTL.

[0127] Conversion of GEP data from an Affymetrix Probe set to a NanoString to nCounter code set for a multi-analyte assay: Our main goal is to use NanoString technology to convert GEP signatures from fresh frozen tissues to FFPE tissues. Because code set design is expensive and bioinformatically challenging, we refine our gene expression signatures from our previous work, adding at least 10% more genes by number in each diagnostic signature, so that we have a sufficient number of genes available for signature refinement if some of the code sets do not work optimally in FFPE RNA without changing the accuracy of the diagnostic signature (Figure 11). The code sets were designed using nDesign™ Gateway with bioinformatics personnel from NanoString under contract UNeMEd for confidentiality of the gene list. This design includes a total of 416 code sets, including housekeeping genes (n=50), AITL diagnostic (n=50), prognostic (n=49), and PTCL-NOS subgroup (n=50) code sets to distinguish NK from gamma delta subgroups (n=50), as well as code sets for classifying ALCL (including ALK+ALCL vs. ALK-ALCL, n=100). We will receive these code sets for these genes and begin assaying on FFPE cases in a few weeks. We expect that some RNAs will be inappropriate for post-fixation studies and will need to be excluded, especially if fewer signatures are used. Therefore, we will perform initial assays on 10 fresh frozen RNAs and corresponding FFPETs to compare signal intensities using the nCounter platform, and select only transcripts with a correlation of 0.8 or higher for further evaluation.

[0128] The subject matter disclosed herein is further illustrated by further descriptions of the features, benefits, and advantages of the present invention, as well as by specific, but non-limiting examples as described in the Addendum attached hereto and incorporated herein by reference.

[0129] All publications, patents, and patent applications mentioned in this specification and any supplements are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.

[0130] It will be understood that various details of the subject matter disclosed herein can be changed without departing from the scope of the subject matter disclosed herein. Further, the foregoing description is by way of example only and not by way of limitation.

[0131] References 1. Rudiger T, Weisenburger DD, Anderson JR, et al. Peripheral T-cell lymphoma (excluding anaplastic large-cell lymphoma): results from the Non-Hodgkin's Lymphoma Classification Project. Annals of Oncology: Official journal of the European Society for Medical Oncology / ESMO.2002;13(1):140-149. 2.Bellei M, Chiattone CS, Luminari S, et al.T-cell lymphomas in South America and Europe.Rev Bras Hematol Hemoter.2012;34(1):42-47. 3.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification:Pathology and Genetics of Tumors of Haematopoietic and Lymphoid Tissues.WHO;4.Lyon,France:IARC Press;2008。 4.Vose J,Armitage J,Weisenburger D,et al.International Peripheral T-cell and Natural Killer / T-cell Lymphoma Study:pathology findings and clinical outcomes.J Clin Oncol.2008;26(25):4124-4130。 5.Morris SW,Kirstein MN,Valentine MB,et al.Fusion of a kinase gene,ALK,to a nucleolar protein gene,NPM,in non-Hodgkin’s lymphoma.Science.1994;263(5151):1281-1284。 6.Parrilla Castellar ER,Jaffe ES,Said JW,et al.ALK-negative anaplastic large cell lymphoma is a genetically heterogeneous disease with widely disparate clinical outcomes.Blood.2014;124(9):1473-1480。 7.Escalon MP,Liu NS,Yang Y,et al.Prognostic factors and treatment of patients with T-cell non-Hodgkin lymphoma:the M.D.Anderson Cancer Center experience.Cancer.2005;103(10):2091-2098. 8.Crozier JA,Sher T,Yang D,et al.Persistent disparities among patients with T-cell Non-Hodgkin Lymphomas and B-cell Diffuse Large Cell Lymphomas over 40 years:A SEER database review.Clinical Lymphoma Myeloma and Leukemia.2015;15(10):578-85。 9.Xu B,Liu P.No survival improvement for patients with angioimmunoblastic T-cell lymphoma over the past two decades:a population-based study of 1207 cases.PLoS One.2014;9(3):e92585。 10.Iqbal J,Wright G,Wang C,et al.Gene expression signatures delineate biological and prognostic subgroups in peripheral T-cell lymphoma.Blood.2014;123(19):2915-2923。 11.Iqbal J,Weisenburger DD,Greiner TC,et al.Molecular signatures to improve diagnosis in peripheral T-cell lymphoma and prognostication in angioimmunoblastic T-cell lymphoma.Blood.2010;115(5):1026-1036。 12.Iqbal J,Weisenburger DD,Chowdhury A,et al.Natural killer cell lymphoma shares strikingly similar molecular features with a group of non-hepatosplenic gamma / delta T-cell lymphoma and is highly sensitive to a novel aurora kinase A inhibitor in vitro.Leukemia.2011;25(2):348-358。 13.Moskowitz AJ,Lunning MA,Horwitz SM.How I treat the peripheral T-cell lymphomas.Blood.2014;123(17):2636-2644。 14.Gambacorti-Passerini C,Messa C,Pogliani EM.Crizotinib in anaplastic large-cell lymphoma.N Engl J Med.2011;364(8):775-776。 15.Ishida T,Joh T,Uike N,et al.Defucosylated anti-CCR4 monoclonal antibody(KW-0761)for relapsed adult T-cell leukemia-lymphoma:a multicenter phase II study.J Clin Oncol.2012;30(8):837-842。 16.O’Connor OA,Bhagat G,Ganapathi K,et al.Changing the paradigms of treatment in peripheral T-cell lymphoma:from biology to clinical practice.Clin Cancer Res.2014;20(20):5240-5254。 17.Coiffier B,Federico M,Caballero D,et al.Therapeutic options in relapsed or refractory peripheral T-cell lymphoma.Cancer Treat Rev.2014;40(9):1080-1088。 18.Iqbal J,Wilcox R,Naushad H,et al.Genomic signatures in T-cell lymphoma:How can these improve precision in diagnosis and inform prognosis?Blood Rev.2016;30(2):89-100。 19.Linton K,Howarth C,Wappett M,et al.Microarray gene expression analysis of fixed archival tissue permits molecular classification and identification of potential therapeutic targets in diffuse large B-cell lymphoma.J Mol Diagn.2012;14(3):223-232。 20.Scott DW,Wright GW,Williams PM,et al.Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue.Blood.2014;123(8):1214-1217。 21.Scott DW,Mottok A,Ennishi D,et al.Prognostic significance of diffuse large B-cell lymphoma cell of origin determined by digital gene expression in formalin-fixed paraffin-embedded tissue biopsies.J Clin Oncol.2015;33(26):2848-2856。 22.Rimsza LM,Leblanc ML,Unger JM,et al.Gene expression predicts overall survival in paraffin-embedded tissues of diffuse large B-cell lymphoma treated with R-CHOP.Blood.2008;112(8):3425-3433。 23.Rimsza LM.Diffuse large B-cell lymphoma classification tied up nicely with a“String”.Clin Cancer Res.2015;21(10):2204-2206。 24.Feldman AL,Law M,Remstein ED,et al.Recurrent translocations involving the IRF4 oncogene locus in peripheral T-cell lymphomas.Leukemia.2009;23(3):574-580。 25.Feldman AL,Dogan A,Smith DI,et al.Discovery of recurrent t(6;7)(p25.3;q32.3)translocations in ALK-negative anaplastic large cell lymphomas by massively parallel genomic sequencing.Blood.2011;117(3):915-919。 26.Crescenzo R,Abate F,Lasorsa E,et al.Convergent mutations and kinase fusions lead to oncogenic STAT3 activation in anaplastic large cell lymphoma.Cancer Cell.2015;27(4):516-532。 27.Boddicker RL,Kip NS,Xing X,et al.The oncogenic transcription factor IRF4 is regulated by a novel CD30 / NF-kappaB positive feedback loop in peripheral T-cell lymphoma.Blood.2015;125(20):3118-3127。 28.Kucuk C,Jiang B,Hu X,et al.Activating mutations of STAT5B and STAT3 in lymphomas derived from gammadelta-T or NK cells.Nat Commun.2015;6:6025。 29.Cairns RA,Iqbal J,Lemonnier F,et al.IDH2 mutations are frequent in angioimmunoblastic T-cell lymphoma.Blood.2012;119(8):1901-1903。 30.Wang C,McKeithan TW,Gong Q,et al.IDH2 R172 mutations define a unique subgroup of patients with angioimmunoblastic T-cell lymphoma.Blood.2015;126(15):1741-1752。 31.Rohr J,Guo S,Huo J,et al.Recurrent activating mutations of CD28 in peripheral T-cell lymphomas.Leukemia.2016;30(5):1062-1070。 32.Roberts RA,Sabalos CM,LeBlanc ML,et al.Quantitative nuclease protection assay in paraffin-embedded tissue replicates prognostic microarray gene expression in diffuse large-B-cell lymphoma.Lab Invest.2007;87(10):979-997。 33.Geiss GK,Bumgarner RE,Birditt B,et al.Direct multiplexed measurement of gene expression with color-coded probe pairs.Nat Biotechnol.2008;26(3):317-325。 34.Ring BZ,Hout DR,Morris SW,et al.Generation of an algorithm based on minimal gene sets to clinically subtype triple negative breast cancer patients.BMC Cancer.2016 Feb 23;16:143。 35.Alizadeh AA,Eisen MB,Davis RE,et al.Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling.Nature.2000;403(6769):503-511。 36.Dave SS,Fu K,Wright GW,et al.Molecular diagnosis of Burkitt’s lymphoma.N Engl J Med.2006;354(23):2431-2442。 37.Dave SS,Wright G,Tan B,et al.Prediction of survival in follicular lymphoma based on molecular features of tumor-infiltrating immune cells.N Engl J Med.2004;351(21):2159-2169。 38.Rosenwald A,Wright G,Chan WC,et al.The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma.N Engl J Med.2002;346(25):1937-1947。 39.Rosenwald A,Wright G,Leroy K,et al.Molecular diagnosis of primary mediastinal B cell lymphoma identifies a clinically favorable subgroup of diffuse large B cell lymphoma related to Hodgkin lymphoma.J Exp Med.2003;198(6):851-862。 40.Rosenwald A,Wright G,Wiestner A,et al.The proliferation gene expression signature is a quantitative integrator of oncogenic events that predicts survival in mantle cell lymphoma.Cancer Cell.Feb 2003;3(2):185-197。 41. Tibshirani R, Hastie T, Narasimhan B, Chu G. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proc Natl Acad Sci US A.2002;99(10):6567-6572. 42. Iqbal J, Naushad H, Bi C, et al. Genomic signatures in B-cell lymphoma: How can these improve precision in diagnosis and inform prognosis?Blood Rev.2016;30(2):73-88.

[0132] Example 3 Gene expression signatures for accurate diagnosis of peripheral T-cell lymphoma entities in routine clinical practice summary the purpose Peripheral T-cell lymphomas (PTCLs) comprise heterogeneous clinicopathological entities with numerous diagnostic and therapeutic challenges. We previously defined robust transcriptomic signatures that distinguish common PTCL entities and identified two novel biological and prognostic PTCL-NOS subtypes (PTCL-TBX21 and PTCL-GATA3). We aimed to integrate gene expression-based subclassification using formalin-fixed paraffin-embedded (FFPE) tissues to improve accuracy and precision in PTCL diagnosis.

[0133] method We assembled a well-characterized PTCL training cohort (n=105) with gene expression profiling (GEP) data and derived a diagnostic signature using fresh frozen (FF) tissues on the HG-U133plus2 platform (Affymetrix, Inc.), which was then validated using matched FFPE tissues on a digital GEP platform (NanoString, Inc.). A statistical filtering approach was applied to refine the transcriptomic signature, which was then validated in another PTCL cohort (n=140) with rigorous pathology testing and ancillary assays.

[0134] result In the training cohort, the refined transcriptomic classifier in FFPE tissues showed high sensitivity (>80%), specificity (>95%), and accuracy (>94%) for PTCL subclassification compared with the FF-derived diagnostic model, with high reproducibility across three independent laboratories. In the validation cohort, the transcriptional classifier was concordant with the pathologic diagnosis given by three expert hematopathologists in 85% (n=119) of cases, showed a "borderline association" with the molecular signature in 6% (n=8), and was discordant in 8% (n=11). The classifier improved the pathologic diagnosis in two cases that were validated by clinical findings. Of the 11 discordant cases, four had molecular classifications that could provide an improvement over pathologic diagnosis based on overall transcriptomic and morphological features. The molecular subclassification provided a comprehensive molecular characterization of PTCL subtypes, including viral etiological factors and translocation partners.

[0135] conclusion The present inventors have developed a novel transcriptomic approach for PTCL subclassification that facilitates translation into clinical practice with greater accuracy and uniformity than traditional pathologic diagnosis.

[0136] Summary of Contents Primary Objective To evaluate the role of digital gene expression signatures for the classification of peripheral T-cell lymphoma (PTCL) using formalin-fixed paraffin-embedded (FFPE) tissues and to develop a highly accurate and reproducible diagnostic assay applicable to routine clinical practice.

[0137] Generated knowledge Using digital quantification of transcripts, we defined a robust transcriptomic signature capable of distinguishing common PTCL subtypes according to the WHO classification, including two novel biological and diagnostic subgroups within PTCL-NOS. We refined the classification algorithm and standardized the assay procedure for a robust diagnostic assay that was validated in an independent PTCL cohort. The assay was reproducible across centers and showed high classification accuracy.

[0138] Related Digital transcriptome assays provide robust molecular classification of PTCL using mRNA from FFPE tissues, facilitating translation to clinical trials that provide a more accurate and uniform diagnosis than traditional pathology. The molecular assays also facilitate better stratification of cases for research studies and clinical trials.

[0139] Introduction Peripheral T-cell lymphoma (PTCL) accounts for approximately 10-15% of non-Hodgkin's lymphoma (NHL). 1 However, even for expert hematopathologists, diagnosis poses many challenges. 2~4 The World Health Organization (WHO) classification defines more than 25 different subtypes of PTCL, with angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma (ALCL), adult T-cell leukemia / lymphoma (ATLL), and extranodal NK / T-cell lymphoma, nasal type (ENKTCL) as the most frequent entities with geographic variation. 5However, 30% of PTCLs cannot be classified as any of the specific entities in the WHO classification and are classified as PTCL, not otherwise specified (PTCL-NOS). 6、7 Tumor-defining abnormalities, e.g., translocations involving the ALK gene in ALK-positive ALCL (ALK+ ALCL) 8 Human T-lymphotropic virus type 1 (HTLV-1) infection in ATLL 9 , EBV positivity in ENKTCL, and IDH2 in AITL R172 Mutations are generally rare in PTCL. PTCL generally has a poor prognosis with current therapies. 2 , more intensive regimens have not been proven superior 10 However, novel targeted therapies are currently being tested, with some notable results. 11、12 .

[0140] Due to the complexity of T cell biology with many functional subsets and functional plasticity, subclassification of PTCLs is more challenging compared to B cell lymphomas. Gene expression profiling (GEP) has helped to delineate novel biological subtypes and identify oncogenic pathways in some B-NHLs. 13~18 A similar approach in PTCL has yielded robust molecular classifiers for common PTCL entities and identified two biological and diagnostic subgroups within PTCL-NOS (PTCL-GATA3 and PTCL-TBX21). 19~21 However, these studies were performed on fresh frozen (FF) samples using transcriptome-wide arrays, and therefore their application is limited to routine clinical practice. 22、23 Formalin-fixed paraffin-embedded (FFPE) tissue samples are widely used in routine diagnostics, but formalin fixation causes RNA and DNA fragmentation, cross-linking, and chemical modification. 24Thus, effectively translating our highly accurate RNA-based PTCL diagnostic signature to FFPE tissues is challenging but essential for implementing an assay with broad clinical application. 25 Diffuse large B-cell lymphoma 26~28 Using digital quantification of RNA, as in, we have integrated our PTCL diagnostic signatures from fresh frozen RNA into a single technological platform for accurate diagnosis across the major PTCL entities. 19~21 We used a "training" PTCL cohort (n=105) and validated it in an "independent" PTCL cohort (n=140) and among the largest number of well-characterized cases investigated on a single platform for PTCL subclassification. By adding several viral transcripts to improve diagnostic accuracy, we report a diagnostic algorithm that achieved high sensitivity, specificity, and accuracy in distinguishing PTCL entities, including the novel molecular biological subtypes PTCL-GATA3 and PTCL-TBX21. 20、21 .

[0141] Materials and Methods Patient Information We included 249 diagnosed PTCL cases from multiple centers, which were analyzed using a previously generated GEP with matched FF and FFPE samples after excluding 4 cases with poor RNA quality. 19~21 The patients were divided into a training cohort (105 cases) with a history of PTCL and a validation cohort (140 cases) not previously analyzed (Tables 4 and 5 and FIG. 12A). The basic clinical and pathological characteristics of the cases are shown in Table 6. The inclusion and exclusion criteria for PTCL cases are detailed in the supplementary section.

[0142] Table 4: Performance of the PTCL diagnostic algorithm in the training cohort. TIFF2025506567000005.tif56128

[0143] (Table 5) Performance of the PTCL diagnostic algorithm in the validation cohort. TIFF2025506567000006.tif56128

[0144] Table 6. Characteristics of training and validation cohort cases and excluded cases. TIFF2025506567000007.tif242170TIFF2025506567000008.tif233170TIFF2025506567000009.tif228170TIFF2025506567000010.tif23117 0TIFF2025506567000011.tif228170TIFF2025506567000012.tif230170TIFF2025506567000013.tif232170TIFF2025506567000014.tif22617 0TIFF2025506567000015.tif221170TIFF2025506567000016.tif234170TIFF2025506567000017.tif79170AITL, angioimmunoblastic T-cell lymphoma, ALCL, anaplastic large cell lymphoma, ATLL, adult T-cell leukemia / lymphoma, ENKTCL, extranodal NK / T-cell lymphoma, PTCL-NOS, peripheral T-cell lymphoma not otherwise specified, PTCL-TFH, peripheral T-cell lymphoma with TFH phenotype, RH, reactive hyperplasia, NA: not available.

[0145] Histopathological / immunomorphological characteristics of the PTCL cohort PTCL cases were centrally examined and diagnosed according to the current WHO classification. 6 The validation cohort was thoroughly re-evaluated by three hematopathologists (CA, DDW, WCC) using a comprehensive immunostaining panel and TCRg gene rearrangement analysis as appropriate. A consensus diagnosis was reached when there was unanimous agreement on the diagnosis.

[0146] RNA extraction and digital gene expression for PTCL subclassification Details on the RNA extraction protocol, quality control measures, nCounter® assay, data processing, cross-validation, and reproducibility assessment are provided in the supplemental section and in Figure 12A. Data analysis and normalization were designed to process samples individually rather than in batches so that the protocol is suitable for processing patient samples as needed. Class prediction was based on a series of binary comparisons that were combined for a final classification call for each sample (Figure 12A and Figure 17). Detailed materials and methods are provided herein.

[0147] Survival analysis Survival data were analyzed using the survival, survminer, and coin packages in R and are detailed herein.

[0148] result Patient characteristics in the “training” and “validation” cohorts The training cohort (n=105) was recruited from previously generated GEP data in FF. 19~21 The PTCL "validation" cohort (n=140) without fresh frozen transcriptome data was based on current WHO diagnostic criteria. 6 The PTCL-NOS cases were rigorously diagnosed by three expert hematopathologists using the PTCL-NOS method (Figure 12A and Tables 4 and 5). Of note, the PTCL-NOS cases were rigorously diagnosed by three expert hematopathologists using the PTCL-NOS method (Figure 12A and Tables 4 and 5). FH Case Exclusion FH Markers were evaluated and then the recently published IHC algorithm 29 The patients were subclassified into PTCL-GATA3 and PTCL-TBX21 using the PTCL-GATA3 and PTCL-TBX21 subclassification.

[0149] The clinicopathological characteristics of the training and validation cohorts are summarized in Table 6. There were no significant differences in gender, age, and OS between the validation and training cohorts (Figure 12C). The median follow-up for survivors was 3.5 years (range 0.01-24 years) for patients with available survival data. ALK+ ALCL cases showed better outcomes than other entities, consistent with published studies (Figure 12D-E). 2、21 .

[0150] Developing transcriptomic signatures for FFPE tissues FFPE tissue blocks were selected based on the presence of sufficient tumor tissue, and RNA quality was assessed as shown in Supplementary S1. Transcriptome signatures evaluated between the two platforms (HG-U133plus2 (Affymetrix, Inc) vs. nCounter® platform) revealed high correlation (correlation coefficient r>0.4) in the majority (~60%) of signature-specific genes (Figure 19A-B). We performed a recursive filtering analysis and filtered out transcripts with correlation coefficients ≤0.4 using Pearson correlation (except for 3 transcripts) to generate 11-20 diagnostic transcripts and 16 housekeeping genes per PTCL subtype (Figure 19C-D). The use of these well-performing transcripts in nCounter® was demonstrated by comparing the RNA quality of fresh frozen RNA (HG-U133plus 2) or matched FFPE RNA in the training cohort (n=105). Neither RNA (nCounter®, Inc.) affected classification accuracy, sensitivity, or specificity. Molecular subclassification using FFPE samples was highly comparable to the FF gold standard across the various PTCL subtypes, with an error rate of less than 5%.

[0151] Accuracy and inter-laboratory reproducibility of PTCL classification using refined signatures The reduced transcript signature retained classification accuracy in HG-U133 plus2 platform data 19~21(Fig. 19C-D). The reduced diagnostic transcripts were therefore considered the "gold standard" for subsequent comparison with the nCounter® assay in the training set (Fig. 12B, left panel). The classification of the FFPE training cohort was replicated in nCounter® in 90% of cases (95 out of 105 cases) (Tables 4, 5 and Fig. 24). We observed a prognostic difference between PTCL-GATA3 and PTCL-TBX21 (Fig. 19F), but the number of samples was too small to reach statistical significance. As we moved from a larger panel of diagnostic transcripts to a smaller panel, 7 cases from different PTCL subtypes were re-evaluated on the nCounter® platform using the reduced transcript set against the original larger panel. They showed concordant results and maintained a similar diagnostic accuracy (Fig. 20). When the assay was performed at two additional CLIA sites to assess reproducibility, the inventors observed highly concordant results and the same classification as at the original site for all 24 PTCL cases studied and between observers (Figure 20).

[0152] Refined diagnostic algorithms across different PTCL entities and validation cohorts To evaluate diagnostic performance, the molecular classification obtained using the nCounter® platform was compared to the consensus pathology diagnosis in an independent validation cohort (n=140). This cohort had a similar clinical outcome and distribution of PTCL subtypes as the training cohort (Figure 12A,C). The classification obtained using the nCounter® platform was highly similar to the diagnosis made by expert pathologists, with an overall agreement of 91% (127 out of 140 cases, 95% confidence interval (CI) 0.85-0.95) in validation cases, refining the classification of difficult PTCL cases as shown below (Tables 4, 5 and Figure 25; Figure 12B, right panel; Figure 21).

[0153] AITL: The mean expression of the diagnostic signature was pan-T- FHSignificantly correlated with the gene expression signature (i.e., 6 transcripts were ranked T in the WHO FH Defined as a marker 6 ) (Figures 13A-B). Cases molecularly classified as AITL by nCounter® in the training and validation cohorts displayed immunomorphological features commonly associated with AITL. Consistent with previous studies, high expression of CD20 was associated with better OS. 20、21、30 , which was verified using immunohistochemistry (Figure 13C). AITL mutation spectrum (i.e., TET2, DNMT3A, RHOA GI7V , and IDH2 R172 ) was found in 89% of cases with available sequencing data (Figure 13D).

[0154] Using the nCounter® platform, AITL was classified with 83% concordance (20 / 24) in the training cohort and 74% (14 / 19) in the validation cohort, while the remaining cases (3 in training and 5 in validation) showed a "borderline model score" between AITL and PTCL-NOS (Tables 4 and 5; Figure 13E; Figure 21). These cases misdiagnosed the AITL molecular diagnosis based on the "threshold or cut point". Re-examination confirmed that these cases had classic AITL immunomorphological features, and mutation analysis supported the diagnosis, as shown in Figure 13D, with TET2, RHOA, and IL-1 mutations. G17V , and IDH2 R172 These cases had a classic AITL mutation spectrum, including: FH The 3 follicular T-cell lymphoma cases were all classified as AITL, as expected. The 2 PTCL-NOS cases were molecularly classified as AITL. These cases were consistent with the T FH Although the markers BCL6 and ICOS were focally positive (Fig. 13G-I), nodular PTCL-T FHThese two cases (Figure 13F) had higher AITL signature genes and T than most other PTCL-NOS cases. FH Although we clearly expressed the mRNA signature, mutations in genes commonly found in AITL (TET2, IDH2 R172 , R.H.O.A. G17V , DNMT3A). These two cases were PTCL-T, which was not classified by the immunostaining performed. FH can be represented as follows.

[0155] ALCL subtypes: Initial analysis of ALCL vs. other PTCL showed 83% (30 / 36, 95% CI: 0.67-0.94) concordance in the validation cohort, which was comparable to the training cohort (26 / 29, 90%, 95% CI: 0.73-0.98) using the refined signature (Figure 14A). The validation cohort (Tables 4 and 5, Figure 14B) showed 90% (18 / 20) concordance in ALK+ ALCL and 75% (12 / 16) concordance in ALK- ALCL. Interestingly, two ALK- ALCLs were classified as ALK+ ALCL because of the high ALK+ ALCL signature, but ALK mRNA expression was low (Figure 14C), and IHC did not detect ALK protein expression. As expected, cases classified as ALCL by ALK status showed that ALK-positive cases had better OS (Figure 14D). Notably, none of the PTCL-NOS cases, including those with strong CD30 positivity by IHC, were misclassified as ALCL. However, four ALCLs (two ALK+ ALCLs and two ALK- ALCLs) showed relatively low ALCL signature expression, while two ALK+ ALCLs had high ALK signature (Figure 14C). Two ALK- ALCL cases with low ALCL signature did not express CD30 mRNA, and the failure to classify these cases may be due to insufficient RNA quality or low tumor content (Figure 14E). These findings suggest that occasional cases with low diagnostic signatures should be diagnosed with caution.

[0156] ATLL: The ATLL molecular signature detected 100% of ATLL cases in the training cohort (7 / 7, 95% CI: 0.59-1) and 83% of the validation cohort cases (10 / 12, 95% CI: 0.52-0.98) (Figure 15A). The two discordant cases had borderline expression of the ATLL diagnostic signature, but both were confirmed positive for HTLV1 mRNA expression (i.e., HBZ by qPCR), whereas the expression of HBZ measured by nCounter® was lower than in the other ATLL cases (Figure 15B-C). Of the two HTLV1 transcripts (HZB and TAX1), HZB was consistently expressed at higher levels in ATLL cases compared to TAX, showing a positive correlation with the ATLL signature (Figure 15D-E) and a significant correlation with HBZ expression measured by qRT-PCR (Figure 15F). Interestingly, two PTCL-NOS cases from the validation cohort were molecularly classified as ATLL (Figure 15A). Reevaluation of these cases using HBZ-specific qPCR confirmed expression of HBZ (Figure 15E), and subsequent examination of clinical charts showed serological positivity for HTLV1 and clinical findings compatible with ATLL, which was unknown at the time of the original diagnosis. Morphologically, these cases consisted of CD4-positive monomorphic T-cell lymphoma and would have been erroneously diagnosed as PTCL-NOS in the absence of appropriate medical history, blood work, and serology. 7 (Figure 15B). Therefore, these two cases were reclassified as ATLL.

[0157] ENKTCL:The ENKTCL molecular classifier was able to identify 90% (9 / 10) of ENKTCL cases in the training cohort and 95% (21 / 22) in the validation cohort (Tables 4 and 5, Figure 15F-G). A subset of cases had elevated expression of CD3γ and CD3δ compared to other cases (Figure 15H), and may be derived from the T lineage. Two cases were diagnosed as "primary EBV-positive nodal T / NK cell lymphoma" according to the current WHO classification. These cases resembled ENKTCL but had a predominantly nodal finding, and both were classified as ENKTCL by molecular assays (Figure 22). One molecularly classified ENTKCL case uniquely displayed both ALCL and ENKTCL signatures, but showed strong expression of EBV transcripts (Figure 14I).

[0158] Elaboration and validation of two novel PTCL-NOS subtypes (PTCL-GATA3 and PTCL-TBX21). Among the PTCL-NOS cases in the training cohort, we isolated two novel molecular subgroups (i.e., PTCL-GATA3 or PTCL-TBX21) with 87% concordance using a reduced number of transcripts (n=19). Of the remaining PTCL-NOS groups in the validation, 40% (21 / 52) were classified as PTCL-GATA3 subtype and 52% (27 / 52) as PTCL-TBX21 subtype using the nCounter® platform (Tables 4 and 5, FIG. 16A). This classification showed good concordance with our IHC algorithm classification (overall concordance: 80%). To further validate our transcriptome signature classification, we compared the sequencing data in these two groups. We observed that genetic alterations such as TET2 mutations were frequent in the PTCL-TBX21 subtype, whereas TP53 mutations were enriched in the PTCL-GATA3 subgroup, which is consistent with our previous findings. 31 This was consistent with (Figure 16B).

[0159] Our previous observations 21、29Consistent with our previous observations, cases expressing cytotoxic transcripts were significantly enriched in the PTCL-TBX21 subtype (Figure 16C), validated by a more frequent cytotoxic immunophenotype in the PTCL-TBX21 subtype than in the PTCL-GATA3 subtype (52% vs. 16%, p=0.013, Figure 16C). An inverse correlation was observed between the mean cytotoxic CD8+ T cell signature and pan-B cell CD20 transcripts (Figure 16D). 29 Consistent with this, PTCL-TBX21 cases frequently had an enriched inflammatory background, regardless of the cytotoxic phenotype (Figure S16E). Because clinical outcomes for PTCL-NOS were only available in a limited number of cases, the combined cohort tended to have inferior OS associated with cases classified as PTCL-GATA3 vs. PTCL-TBX21 (median OS: 0.57 vs. 1.4 years, Figure S16F).

[0160] PTCL-T excluded from validation cohort FH Case evaluation PTCL-T using the nCounter® platform FH (n=12) cases were analyzed, only 2 cases (17%) showed significant association with AITL molecular signature, 4 cases had borderline scores between AITL and PTCL-NOS, and 6 cases showed clear association with PTCL-NOS, similar to PTCL-TBX21 cases (n=3) or PTCL-GATA3 (n=3). FH When examining the signatures specifically, PTCL-T has a high AITL signature. FH The cases also had a higher T FH We also included 10 cases of reactive hyperplasia, none of which showed expression of the diagnostic signatures for the subtypes of PTCL included in the assay.

[0161] Consideration The diagnosis of PTCL is one of the most challenging of the lymphomas, with more often than not inconclusive, inconsistent, or inaccurate diagnoses. 19~21、32、33 Recently, a novel therapeutic approach has been developed for CD30+ PTCL, especially ALCL, using brentuximab-vedotin (BV). 34 Crizotinib in ALK+ ALCL 35、36 , Mogamulizumab in ATLL 37 , AITL or T FH -HDACi and demethylating agents in PTCL 38 , and possibly enasidenib in IDH2 mutant AITL 39 The study showed a significant benefit in subgroups of PTCL, including pts, pts with glaucoma ... and pts with glaucoma. Accurate diagnosis may therefore be important in patient care and in clinical trials of new drugs. 40~42 The present inventors have conducted extensive GEP studies on PTCL, developed RNA-based molecular diagnostic signatures and predictors of survival, and elucidated key carcinogenic mechanisms in detail. 19~21 Some of these findings are included in the revised 2016 WHO classification. 6 To translate this molecular information into a platform suitable for clinical use, 25 We performed a systematic analysis to identify RNA from FFPE tissues in PTCL using the nCounter® platform, which correlates well with GEP data from fresh frozen tissues because this digital quantification technique is more resistant to degraded RNA typical of FFPE material. In addition, we developed a diagnostic transcriptomic signature with a minimal number of transcripts that performed comparably to previous GEP-derived diagnostics in the training set. 13~15 Preanalytical evaluation has shown that the RNA yield and quality (i.e., DV 200>50%) was improved using the RNAstorm™ kit. However, with recently acquired FFPE tissues, other isolation methods may also provide good quality RNA. In addition, the assay can be performed with limited amounts of RNA (minimum required 200 ng), which is usually obtained from a few unstained slides, depending on the size of the tissue. However, we were able to have sufficient classification using RNA extracted from core needle biopsies, which represented about 10% of the study samples. Interlaboratory comparisons of variability and reproducibility across three CLIA-certified laboratories also correlated very well.

[0162] Since we did not have GEP data on the cases for validation, we included only cases with definite pathological diagnosis after extensive IHC studies and rigorous testing. Findings were very similar to those of the training set (Figure 12), maintaining high sensitivity, specificity, and accuracy. In some cases, the signature scores were just outside the cutoff points, and these were considered "borderline" cases. The few cases in which the molecular assays made an accurate diagnosis in retrospective analysis are encouraging. However, clear discrepancies were also observed. This may be due in part to technical reasons such as tumor content and tumor heterogeneity, but some may be related to as yet unknown biology, such as the strong ALK signature in some ALK-ALCL cases that may represent the recently reported ALK-positive-like ALCL. 43 Similarly, the two PTCL-NOS are AITL and T FH showed a significant association with the transcriptomic signature, which, when examined, FH Markers (BCL6, ICOS) were expressed focally. FH The expression signatures suggest that these cases are PTCL-T FH Similar to, but with at least two T FHIt may indicate that the current criteria of strong expression of markers were not met. In these cases, the molecular assays revealed the complex and overlapping biology between PTCL-NOS and AITL, and that their boundaries are less clear. We also found that the addition of EBV and HTLV1 transcripts enhanced the diagnostic performance of the molecular assay in ENTKCL and ATLL, respectively. One important contribution of this approach is the robust definition of PTCL-GATA3 and PTCL-TBX21 cases, which have distinct biology and prognosis, as supported by recent genetic findings. 31 Although it is possible to simulate GEP classification using IHC panels, it can be difficult to optimize and interpret the staining, which can lead to substantial variability between laboratories. The assay reported here is highly reproducible between laboratories and therefore represents a major advantage.

[0163] The first GEP study to define a diagnostic signature was FH The provisional entity definitions for PTCL-T were not formally included in this study. However, 12 PTCL-Ts detected by the nCounter® assay were FH Cases had higher mean AITL and T FH The individual signatures are shown in the study by Dobay et al. 44 These exploratory studies have demonstrated that the molecular signatures of PTCL-T FH This finding indicates that it is unlikely to be a single entity as currently defined and requires further evaluation to determine how it should best be characterized. Similarly, our previous GEP studies 20、21demonstrated a cytotoxic PTCL variant within the PTCL-TBX21 subgroup. Consistent with that observation, we found that 25% of cases in PTCL-TBX21 had a cytotoxic signature and verified the association with the CD8+ phenotype by IHC. Due to the small number of cases studied, further investigation is needed to develop a robust signature for this group of cases.

[0164] In summary, we have described an approach for translating PTCL diagnostic signatures into a clinically applicable assay, which we envision as a useful tool for general and academic pathologists in the diagnostically challenging field of PTCL. In addition, we believe it can facilitate better definition of cases for research studies and ensure more homogenous cohorts for clinical trials. PTCL classification is an evolving area, and as our understanding of the underlying biology and available technologies improve, modifications will be made to make the classification clinically relevant.

[0165] References cited in this example: 1.Bellei M, Chiattone CS, Luminari S, et al.T-cell lymphomas in South america and europe.Rev Bras Hematol Hemoter.2012;34(1):42-47. 2.Vose J, Armitage J, Weisenburger D, International TCLP. International peripheral T-cell and natural killer / T-cell lymphoma study: pathology findings and clinical outcomes. J Clin Oncol. 2008;26(25):4124-4130. 3.Briski R,Feldman AL,Bailey NG,et al.The role of front-line anthracycline-containing chemotherapy regimens in peripheral T-cell lymphomas.Blood Cancer J.2014;4:e214。 4.Rudiger T,Weisenburger DD,Anderson JR,et al.Peripheral T-cell lymphoma(excluding anaplastic large-cell lymphoma):results from the Non-Hodgkin’s Lymphoma Classification Project.Ann Oncol.2002;13(1):140-149。 5.Fox CP,Civallero M,Ko YH,et al.Survival outcomes of patients with extranodal natural-killer T-cell lymphoma:a prospective cohort study from the international T-cell Project.Lancet Haematol.2020;7(4):e284-e294。 6.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues,Fourth Edition.Revised.2017;2。 7.Weisenburger DD,Savage KJ,Harris NL,et al.Peripheral T-cell lymphoma,not otherwise specified:a report of 340 cases from the International Peripheral T-cell Lymphoma Project.Blood.2011;117(12):3402-3408。 8.Morris SW,Kirstein MN,Valentine MB,et al.Fusion of a kinase gene,ALK,to a nucleolar protein gene,NPM,in non-Hodgkin’s lymphoma.Science.1994;263(5151):1281-1284。 9.Tsukasaki K,Tsushima H,Yamamura M,et al.Integration patterns of HTLV-I provirus in relation to the clinical course of ATL:frequent clonal change at crisis from indolent disease.Blood.1997;89(3):948-956。 10.Escalon MP,Liu NS,Yang Y,et al.Prognostic factors and treatment of patients with T-cell non-Hodgkin lymphoma:the M.D.Anderson Cancer Center experience.Cancer.2005;103(10):2091-2098。 11.Horwitz S,O’Connor OA,Pro B,et al.Brentuximab vedotin with chemotherapy for CD30-positive peripheral T-cell lymphoma(ECHELON-2):a global,double-blind,randomised,phase 3 trial.Lancet.2019;393(10168):229-240。 12.Pro B,Advani R,Brice P,et al.Five-year results of brentuximab vedotin in patients with relapsed or refractory systemic anaplastic large cell lymphoma.Blood.2017;130(25):2709-2717。 13.Alizadeh AA,Eisen MB,Davis RE,et al.Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling.Nature.2000;403(6769):503-511。 14.Rosenwald A,Wright G,Chan WC,et al.The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma.N Engl J Med.2002;346(25):1937-1947。 15.Dave SS,Fu K,Wright GW,et al.Molecular diagnosis of Burkitt’s lymphoma.N Engl J Med.2006;354(23):2431-2442。 16.Monti S,Savage KJ,Kutok JL,et al.Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response.Blood.2005;105(5):1851-1861。 17.Rosenwald A,Wright G,Wiestner A,et al.The proliferation gene expression signature is a quantitative integrator of oncogenic events that predicts survival in mantle cell lymphoma.Cancer Cell.2003b;3(2):185-197。 18.Shipp MA,Ross KN,Tamayo P,et al.Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning.Nat Med.2002;8(1):68-74。 19.Iqbal J,Weisenburger DD,Chowdhury A,et al.Natural killer cell lymphoma shares strikingly similar molecular features with a group of non-hepatosplenic gammadelta T-cell lymphoma and is highly sensitive to a novel aurora kinase A inhibitor in vitro.Leukemia.2011;25(2):348-358。 20.Iqbal J,Weisenburger DD,Greiner TC,et al.Molecular signatures to improve diagnosis in peripheral T-cell lymphoma and prognostication in angioimmunoblastic T-cell lymphoma.Blood.2010;115(5):1026-1036。 21.Iqbal J,Wright G,Wang C,et al.Gene expression signatures delineate biological and prognostic subgroups in peripheral T-cell lymphoma.Blood.2014;123(19):2915-2923。 22.Lone W,Alkhiniji A,Manikkam Umakanthan J,Iqbal J.Molecular Insights Into Pathogenesis of Peripheral T Cell Lymphoma:a Review.Curr Hematol Malig Rep.2018;13(4):318-328。 23.Iqbal J,Amador C,McKeithan TW,Chan WC.Molecular and Genomic Landscape of Peripheral T-Cell Lymphoma.Cancer Treat Res.2019;176:31-68。 24.Srinivasan M,Sedmak D,Jewell S.Effect of fixatives and tissue processing on the content and integrity of nucleic acids.Am J Pathol.2002;161(6):1961-1971。 25.Crescenzo R,Abate F,Lasorsa E,et al.Convergent mutations and kinase fusions lead to oncogenic STAT3 activation in anaplastic large cell lymphoma.Cancer cell.2015;27(4):516-532。 26.Scott DW,Wright GW,Williams PM,et al.Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue.Blood.2014;123(8):1214-1217。 27.Phang KC,Akhter A,Tizen NMS,et al.Comparison of protein-based cell-of-origin classification to the Lymph2Cx RNA assay in a cohort of diffuse large B-cell lymphomas in Malaysia.J Clin Pathol.2018;71(3):215-220。 28.Rimsza LM,Wright G,Schwartz M,et al.Accurate classification of diffuse large B-cell lymphoma into germinal center and activated B-cell subtypes using a nuclease protection assay on formalin-fixed,paraffin-embedded tissues.Clin Cancer Res.2011;17(11):3727-3732。 29.Amador C,Greiner TC,Heavican TB,et al.Reproducing the molecular subclassification of peripheral T-cell lymphoma-NOS by immunohistochemistry.Blood.2019;134(24):2159-2170。 30.Advani RH,Ansell SM,Lechowicz MJ,et al.A phase II study of cyclophosphamide,etoposide,vincristine and prednisone(CEOP)Alternating with Pralatrexate(P)as front line therapy for patients with peripheral T-cell lymphoma(PTCL):final results from the T-cell consortium trial.Br J Haematol.2016;172(4):535-544。 31.Heavican TB,Bouska A,Yu J,et al.Genetic drivers of oncogenic pathways in molecular subgroups of peripheral T-cell lymphoma.Blood.2019;133(15):1664-1676。 32.Maura F,Agnelli L,Leongamornlert D,et al.Integration of transcriptional and mutational data simplifies the stratification of peripheral T-cell lymphoma.Am J Hematol.2019。 33.Piccaluga PP,Fuligni F,De Leo A,et al.Molecular profiling improves classification and prognostication of nodal peripheral T-cell lymphomas:results of a phase III diagnostic accuracy study.J Clin Oncol.2013;31(24):3019-3025。 34.Van Der Weyden C,Dickinson M,Whisstock J,Prince HM.Brentuximab vedotin in T-cell lymphoma.Expert Rev Hematol.2019;12(1):5-19。 35.Andraos E,Dignac J,Meggetto F.NPM-ALK:A Driver of Lymphoma Pathogenesis and a Therapeutic Target.Cancers(Basel).2021;13(1)。 36.Gambacorti-Passerini C,Messa C,Pogliani EM.Crizotinib in anaplastic large-cell lymphoma.N Engl J Med.2011;364(8):775-776。 37.Moskowitz AJ,Lunning MA,Horwitz SM.How I treat the peripheral T-cell lymphomas.Blood.2014;123(17):2636-2644。 38.Lopez AT,Bates S,Geskin L.Current Status of HDAC Inhibitors in Cutaneous T-cell Lymphoma.Am J Clin Dermatol.2018;19(6):805-819。 39.https: / / clinicaltrials.gov / ct2 / show / NCT02273739。 40.Ishida T,Joh T,Uike N,et al.Defucosylated anti-CCR4 monoclonal antibody(KW-0761)for relapsed adult T-cell leukemia-lymphoma:a multicenter phase II study.J Clin Oncol.2012;30(8):837-842。 41.O’Connor OA,Bhagat G,Ganapathi K,et al.Changing the paradigms of treatment in peripheral T-cell lymphoma:from biology to clinical practice.Clin Cancer Res.2014;20(20):5240-5254。 42. Iqbal J, Wilcox R, Naushad H, et al. Genomic signatures in T-cell lymphoma: How can these improve precision in diagnosis and inform prognosis?Blood Rev.2016;30(2):89-100. 43.Yeung KHRA,Russell V,Choi W,et al.ALK-Negative Anaplastic Large Cell Lymphomas Encompass Distinct Subgroups Including an ALK-Positive-like Subgroup with Favorable Prognosis.Blood.2021;138 2403. 44. Dobay MP, Lemonnier F, Missiaglia E, et al. Integrative clinicopathological and molecular analyzes of angioimmunoblastic T-cell lymphoma and other nodal lymphomas of follicular helper T-cell origin. Haematologica.2017;102(4):e148-e151.

[0166] Materials and Methods Patient Information We included 249 diagnosed PTCL cases from multiple centers, which were then divided into a training cohort of 105 cases and a validation cohort of 140 cases after excluding 4 cases with poor RNA quality (Tables 4 and 5 / Figure 12A). The baseline clinical and pathological characteristics of the two cohorts are shown in Table 6. The training cohort was based on a previous GEP study using RNA from FF tissues on HG-U133 Plus 2.0 Array13-15 (Affymetrix, Santa Clara, CA; accession number GSE19069). 1~3The study included a "bridging set" of 91 cases previously evaluated in and had corresponding FFPE samples available. In addition, 1~3 Nine de novo PTCLs with available HG-U133 Plus 2.0 data not included in the study, and five well-characterized classical ENKTCL cases with confirmed diagnoses were also included in the training set. The training cohort consisted of 24 AITL, 14 ALK+ ALCL, 15 ALK- ALCL, 7 ATLL, 10 ENKTCL, and 35 PTCL-NOS (15 PTCL-GATA3 and 20 PTCL-TBX21). These were used to identify which RNA transcripts previously identified as diagnostic in FF tissues performed well in FFPE on the nCounter® platform (NanoString Inc, Seattle).

[0167] A separate validation cohort consisted of an independent series of 140 PTCL cases not previously analyzed by GEP but for which adequate FFPE tissue was available. This set included 19 AITL, 20 ALK+ ALCL, 16 ALK- ALCL, 12 ATLL, 22 ENKTCL, and 51 PTCL-NOS. We initially collected 152 PTCL cases for validation, which we determined were T follicular helper (T FH Twelve cases were excluded from this cohort because they were pathologically reclassified as PTCL with a PTCL-T phenotype. FH The case and 10 additional cases of reactive lymphadenopathy were analyzed with the NanoString-based assay for comparison. The study protocol was approved by the UNMC and COH-MC Institutional Review Boards.

[0168] Morphologic and immunophenotypic characteristics of the PTCL cohort. All cases were classified according to the current WHO classification 4Cases in the validation cohort consisted of a new cohort of cases without FF GEP data that underwent rigorous pathological review in a consensus conference by three hematopathologists (CA, DW, WCC) with expertise in T-cell lymphoma diagnosis. A consensus diagnosis was reached when there was unanimous agreement on the diagnosis. Pathological evaluation included a partial or full panel of B- and T-cell immunostains, including CD3, CD20, CD30, TFH markers (PD1, ICOS, CD10, CXCL13, and BCL6), CD4, CD8, cytotoxicity markers (TIA1, granzyme B, and perforin), TCRαβ, TCRγδ, and EBER in situ hybridization, as appropriate. PTCL-NOS cases with strong expression of two or more TFH markers were reclassified as nodal PTCL with TFH phenotype. PTCL-NOS cases were subclassified by one hematopathologist (CA) into PTCL-GATA3 and PTCL-TBX21 using a recently published IHC algorithm using antibodies against GATA3, TBX21, CCR4, and CXCR3. 5 Of the 43 AITL cases, 40 had typical AITL morphology, whereas 3 were T-FL that were grouped together with AITL for transcriptomic analysis. The ATLL cases in the validation series were compared with those in the previously established test for the presence of HTLV-1 virus in ATLL. 4、6、7According to , the HZB viral transcripts must have a positive signal (CT value) using qRT-PCR (Taq man probe) based on FFPE-extracted RNA. ENKTCL cases in the validation cohort were confirmed using either EBER in situ hybridization or PCR for the presence of LMP1 / EBNA1 in genomic DNA. Of the 32 ENKTCL cases, 30 had nasal involvement and were typical ENKTCL, and 2 had predominantly nodal involvement but otherwise had similar phenotypic features as nasal ENKTCL cases. Detailed information on the nasal involvement in these two cases was not available. The two cases could represent primary nodal EBV-positive T / NK cell lymphoma in the current WHO classification, but were grouped together with ENK / TCL for the transcriptome analysis.

[0169] RNA Extraction from FFPE Tissues and Digital Gene Expression Using the nCounter® System FFPE scrolls of 10–20 μm thickness were cut from the blocks with a surface area of ​​1 cm2 and stored at −20 °C until extraction. For the few cases where FFPE scrolls were not available, tissue was scraped from unstained slides. Deparaffinization and extraction of total RNA was performed using two commercially available kits: (a) Qiagen AllPrep DNA / RNA FFPE Kit (Qiagen GmbH, Germany) or (b) RNAStormTM RNA Isolation Kit (Cell Data Sciences, CD501). 8 RNA quality was assessed using an Agilent Tape-Station and RNA Integrity Analysis (RIN) values ​​were calculated using the Agilent ELISA Kit according to the manufacturer's instructions. 9 Also, RNA fragment sizes >200 bases DV 200 10RNA was quantified using Qubit and 200ng of total RNA (DV200:>35%) was used for hybridization, with additional RNA used for more degraded samples as recommended by NanoString, Inc. (https: / / www.nanostring.com / wp-content / uploads / 2021 / 07 / MAN-10050-05-Preparing-48 RNA-from-FFPE-Samples.pdf). Total RNA was hybridized to a custom code set (see below) overnight (16-18 hours) at 65°C in an nCounter® Prep Station and GEP data was acquired on an nCounter® Digital Analyzer at "High 50 Resolution" setting. Standard QC (as provided by nSolver™ analysis software, NanoString Technologies) was used to flag any samples where the sum of the positive spike-in controls was outside the 0.3-3x geometric mean of the total positive spike-ins for that cartridge. Signal count values ​​were normalized by dividing the counts for each gene for a given sample by the geometric mean of the counts for the housekeeping genes in that sample. The normalized counts were then log2 transformed to obtain the gene signal value utilized in all subsequent analyses. Samples with a geometric mean of housekeeping genes below 10 were excluded from the study.

[0170] Code Set Design for the nCounter® PTCL Subtyping Assay A training cohort of PTCL cases (n=105) with GEP derived from fresh frozen RNA was used to select classifier genes for the nCounter® platform using corresponding FFPE tissues. We reanalyzed the initial FF GEP data (HG-U133 Plus 2.0 Array, Affymetrix, Inc) to generate diagnostic transcript signatures including 287 classifier transcripts specific to PTCL subtypes, 50 housekeeping genes with consistent expression in PTCL, and an extra set of genes related to T cell differentiation. Genes were selected to be (1) differentially expressed (either positively or negatively) in one of the PTCL subtypes (i.e., AITL, ALCL, ENKTCL, ATLL), (2) differentially expressed between ALK+ ALCL and ALK- ALCL, or (3) characteristically differentially expressed to distinguish the PTCL-TBX21 and PTCL-GATA3 subgroups of PTCL-NOS. 50 housekeeping genes were selected from those with the lowest variance across PTCLs in the FF GEP data. These classifier transcripts were therefore designed to be the most informative and robust elements for classification, and nCounter® code sets were designed by the NanoString bioinformatics team using standard nCounter® chemistry for code set design. The barcodes corresponding to each captured transcript were imaged and counted after hybridization, thus providing digital quantification of each captured RNA (details on custom code sets for genes can be found at https: / / www.nanostring.com / products / custom-solutions / custom-CodeSets).

[0171] To minimize costs associated with repeated codeset counting and manufacturing (nCounter®, NanoString Inc), we used samples from the bridging cohort to curate an initial set of diagnostic markers and defined a set of 93 genes and 5 subtype-specific viral transcripts from EBV and HTLV-1 that showed the best performance based on the NanoString platform. The following set of criteria was used to perform this evaluation: (a) correlation between NanoString log2 signal values ​​and their corresponding HG-U133 Plus 2.0 signal intensity (SI) values, (b) T-statistic between normalized log-transformed NanoString SI values ​​for those samples in the PTCL subtype versus other PTCL subtypes, (c) T-statistic between HG-U133 Plus 2.0 signal values ​​for samples in the PTCL subtype and other PTCL subtypes for all samples. 1 , and (d) Iqbal et al. 2014 1 Correlation between NanoString signal values ​​from the NIH-10001 and model scores for that subtype. Care was taken to ensure that this curated list included both positively and negatively associated genes for each distinction. In addition, housekeeping genes were reduced to 16, with the lowest variance in the bridging cohort. We also added 40 other genes identified from the literature as important to T-cell lymphoma biology. This resulted in a final 153-gene code set, which was used in all further analyses.

[0172] Predicting subclassification in a training cohort Multiple prediction methods, including linear discriminant analysis, Lasso regression, support vector machines, and random forests, were applied to the training set in a cross-validated fashion utilizing several different predictive architectures (data not shown). The method with the best cross-validated predictive accuracy was based on a set of binary linear predictors structured as described below and shown in Figure 12A and Figure 17.

[0173] For a given pair of class and selected gene i, Calculate TIFF2025506567000018.tif13128, where μ 1i and μ 2i represents the average expression value of the gene in class 1 and class 2, and σ i 2 is the pooled within-group variance estimate. Then, for each sample j, we obtain the weighted score of discrimination between set 1 and set 2 as follows: TIFF2025506567000019.tif13128, where xij is the normalized signal value for gene i on sample j and the sum is over all genes selected for a particular discrimination.

[0174] We then plotted an ROC curve based on the relationship between this score and membership in class 1 or class 2 to identify the point where the average sensitivity and specificity were maximized. Samples with scores above this point were identified as class 1, and samples with scores below this point were identified as class 2. An exception to this rule was made for AITL subtype classification, where there was an overlapping region between the AITL model score and the PTCL-NOS model score. For samples with scores in this region, it was uncertain whether the sample was AITL. To define this region, we used a range of sensitivities + 2 * The lower cut point that maximized specificity, and specificity +2 * The upper cut point that maximized the sensitivity was considered: samples above the upper cut point were called AITL, samples below the lower cut point were called non-AITL, and samples between the two cut points were called borderline AITL.

[0175] To predict the subtype of individual samples, we first implemented six predictors, four of which were trained to distinguish PTCL-NOS from cases in one of the following four categories: AITL, ALCL, ENKTCL, or ATLL. Because we found that the ENKTCL predictor tended to identify PTCL-NOS with cytotoxic marker expression as ENKTCL, we further generated a classifier between PTCL-NOS and ENKTCL based on the viral transcript EBER. For samples predicted as ENKTCL, we required both the multigene ENKTCL and the single EBER classifier to be in agreement with the molecular diagnosis. A final predictor was also generated to distinguish ALK- ALCL from ALK+ ALCL. First, it divided ALCL samples according to ALK positivity, and second, it was used to distinguish ENKTCL samples that share some of the features of ALCL but are uniformly ALK(-). If the initial set of predictors all concluded that the sample was PTCL-NOS, a final predictor was applied that distinguished PTCL-GATA3 from PTCL-TBX21. If more than one of the four predictors resulted in a non-PTCL-NOS diagnosis (e.g., both ALCL and AITL), an additional head-to-head predictor between the types was applied to break ties (e.g., ALCL vs. AITL) using all genes characteristic of either type selected for inclusion in the model.

[0176] Pre-analytical evaluation of FFPE RNA To identify transcripts that perform well on the nCounter® platform, we correlated the signal intensity for each transcript in 10 fresh frozen RNA samples profiled by both the nCounter® platform and the Affymetrix HG-U133 plus2 array. More than 86.5% and more than 65% of the transcripts showed a correlation of more than 0.4 or more than 0.6 between the two platforms, respectively, and were selected for further analysis (Figure 18A-B). In addition, RNA isolated from FFPE samples using the RNAstorm™ method produced longer transcripts (>200 bases) compared to another commercially available kit (Figure 18C-D), resulting in a higher correlation coefficient with the fresh frozen RNA transcript profile (Figure 18E). Therefore, the RNAstorm™ isolation kit was used in the standard operating procedure for this assay on subsequent samples. We were able to isolate at least 200ng-1ug of total RNA per FFPE specimen, with more than 50% of the RNA showing a DV200>50%. There was a high correlation between FF RNA run on nCounter® (NanoString, Inc.) and the corresponding FFPE RNA (Figure 18F).

[0177] Reproducibility across locations and alignment between the original gene set (n=442) and the refined gene set (n=153) To check the reproducibility of the assay at different centers, we selected 24 samples from the training set and reanalyzed them on the 153-gene array at two different centers (City of Hope and Insight Genetics). Seven of these samples, along with one additional training sample, were also retried at the original location (UNMC) on the reduced gene set array, giving a total of 55 replicate samples. We observed that for some of the predictor scores, these appeared to be systematic biases between the models run on the 442-gene array and the 153-gene array (see FIG. 20). To correct for this, we calculated the mean and standard deviation of the scores for replicate samples on the 153-gene array (denoted by μ153 and σσ153) and the 442-gene array (denoted by μ442 and σσ442). Using these values, we adjusted the scores generated by the 153-gene array (particularly those in the validation set) to match the scores of the 442-gene array according to the following formula: TIFF2025506567000020.tif11128 Note that no validation samples were used to generate the predictive model or to define this adjustment.

[0178] statistical analysis Baseline patient characteristics were compared between groups using chi-squared and t-tests. The Kaplan-Meier method was used to estimate the distribution of overall survival. Overall survival (OS) time was calculated as the time from diagnosis to the date of death or last contact. Patients who were alive at the time of last contact were treated as censored for the overall survival analysis. "Approximated" distribution 11 Differences in outcomes between groups were assessed using a permutation-based log-rank test as implemented in the coin R package. A two-sided p < .05 was considered significant.

[0179] References cited in this method 1.Iqbal J,Wright G,Wang C,et al.Gene expression signatures delineate biological and prognostic subgroups in 181 peripheral T-cell lymphoma.Blood.2014;123(19):2915-2923.182。 2.Iqbal J,Weisenburger DD,Greiner TC,et al.Molecular signatures to improve diagnosis in peripheral T-cell 183 lymphoma and prognostication in angioimmunoblastic T-cell lymphoma.Blood.2010;115(5):1026-1036.184。 3.Iqbal J,Weisenburger DD,Chowdhury A,et al.Natural killer cell lymphoma shares strikingly similar molecular 185 features with a group of non-hepatosplenic gammadelta T-cell lymphoma and is highly sensitive to a novel 186 aurora kinase A inhibitor in vitro.Leukemia.2011;25(2):348-358.187。 4.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification of Tumours of Haematopoietic and Lymphoid 188 Tissues,Fourth Edition.Revised.2017;2.189。 5.Amador C,Greiner TC,Heavican TB,et al.Reproducing the molecular subclassification of peripheral T-cell 190 lymphoma-NOS by immunohistochemistry.Blood.2019;134(24):2159-2170.191。 6.Akbarin MM,Shirdel A,Bari A,et al.Evaluation of the role of TAX,HBZ,and HTLV-1 proviral load on the survival 192 of ATLL patients.Blood Res.2017;52(2):106-111.193。 7.Li M,Green PL.Detection and quantitation of HTLV-1 and HTLV-2 mRNA species by real-time RT-PCR.J Virol 194 Methods.2007;142(1-2):159-168.195。 8.Karmakar S,Harcourt EM,Hewings DS,et al.Organocatalytic removal of formaldehyde adducts from RNA and 196 DNA bases.Nat Chem.2015;7(9):752-758.197。 9.Schroeder A,Mueller O,Stocker S,et al.The RIN:an RNA integrity number for assigning integrity values to RNA 198 measurements.BMC Mol Biol.2006;7:3.199。 10.Wimmer I,Troscher AR,Brunner F,et al.Systematic evaluation of RNA quality,microarray data reliability and 200 pathway analysis in fresh,fresh frozen and formalin-fixed paraffin-embedded tissue samples.Sci Rep.201 2018;8(1):6351.202。 11.Hothorn T,Hornik K,van de Wiel MA,Zeileis A.Implementing a Class of Permutation Tests:The coin Package.203 Journal of Statistical Software.2008;28(8):1-23。

Claims

1. A method of distinguishing between subtypes of peripheral T-cell lymphoma (PTCL), comprising: subjecting a sample from the subject to nucleic acid isolation; obtaining a gene expression profile from said sample; and Identifying subtypes of PTCL based on the presence of specific genes within said gene expression profile.

2. The method of claim 1, wherein the PTCL subtype is angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, and the indeterminate indicates that the sample contains characteristics of at least two of the PTCL subtypes. (a) the PTCL subtype is identified as AITL if the gene expression profile includes any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2; or (b) the PTCL subtype is identified as ALK-ALCL if the gene expression profile includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101; or (c) the PTCL subtype is identified as ENKTL if the gene expression profile includes any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25; or (d) the PTCL subtype is identified as ATLL if the gene expression profile includes any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, HBZ, and ZCCHC12; or (e) the PTCL subtype is identified as ALK+ ALCL if the gene expression profile includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2; or (f) the PTCL subtype is identified as PTCL-NOS if the gene expression profile includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, WARS, TBX21, CXCR3, GATA3, and CCR4; The method of claim 2.

4. the sample is a biopsy specimen from the subject; Optionally, (a) the sample comprises formalin-fixed, paraffin-embedded tissue; or (b) the sample comprises fresh frozen tissue; The method according to any one of claims 1 to 3.

5. 1. A diagnostic kit for identifying a subtype of peripheral T-cell lymphoma (PTCL) in a sample from a subject, comprising: At least one of means for detecting the presence of one or a combination of genes representing a gene signature indicative of a particular PTCL subtype, and instructions for use. The diagnostic kit comprising:

6. The kit of claim 5, wherein the PTCL subtype is angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, and the indeterminate indicates that the sample contains characteristics of at least two of the PTCL subtypes. (a) if the gene signature includes any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2, the PTCL subtype is indicative of AITL; or (b) if the gene signature includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101, the PTCL subtype is indicative of ALK-ALCL; or (c) if the gene signature includes any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25, the PTCL subtype is indicative of ENKTL; or (d) if the gene signature includes any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, TAX, and ZCCHC12, the PTCL subtype is indicative of ATLL; or (e) if the gene signature includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2, the PTCL subtype is indicative of ALK+ ALCL; or (f) if the gene signature includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, and WARS, the PTCL subtype is indicative of PTCL-NOS; The kit of claim 6.

8. 1. A method for identifying a particular subtype of peripheral T-cell lymphoma (PTCL) in a sample, comprising: obtaining a gene expression profile from the sample; comparing the gene expression profile to a gene signature associated with a particular PTCL subtype, wherein each PTCL subtype contains a unique gene signature; and Identifying the subtype of PTCL in the sample as either angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, wherein the indeterminate indicates that the gene expression profile of the sample includes genes from the unique gene signatures of at least two of the PTCL subtypes.

9. the gene signature of AITL comprises any one or more of ACHE, ADRA2A, ARRDC4, COL4A4, EFNB2, FCAMR, GJA4, GNA14, IER3IP1, ISG15, MSR1, OLFM1, PAPLN, S1PR3, SOX8, TMEM206, TUBB2B, and VANGL2; the gene signature of ALK-ALCL comprises any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT1, and ZNF101; the gene signature of ENKTL comprises any one or more of ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF1, SH2D1B, EBER, LMP1, and TNFRSF25; the gene signature of ATLL comprises any one or more of ARSG, CCNE1, DOK5, FGF18, MYCN, NFATC1, NSMCE1, NUCB2, SAT1, SLC7A10, SPPL2A, STOM, TIAM2, UST, TAX, and ZCCHC12; the gene signature of ALK+ ALCL includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RABGAP1L, TIAM2, TMEM158, and ZADH2; the gene signature of PTCL-NOS includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC31A2, TCN2, TMEM176B, and WARS; Or a combination of these; The method of claim 8.

10. 10. The method of claim 8 or 9, wherein identifying the PTCL subtype comprises: a set of at least four binary predictors, a first binary predictor determining whether the gene expression profile of the sample is more consistent with the gene signature of AITL or PTCL-NOS; a second binary predictor determining whether the gene expression profile of the sample is more consistent with the gene signature of ALCL or PTCL-NOS; a third binary predictor determining whether the gene expression profile of the sample is more consistent with the gene signature of ATLL or PTCL-NOS; The fourth binary predictor includes determining whether the gene expression profile of the sample more closely matches the gene signature of ENKTCL or PTCL-NOS. The series of at least four binary predictors; and Assigning the PTCL subtype of the specimen based on the results from the binary predictors. **Claim 11** (a) In the first binary predictor, if the gene expression profile of the sample more closely matches the gene signature of AITL, and In the second, third, and fourth binary predictors, if the gene expression profile of the sample more closely matches the gene signature of PTCL-NOS is the PTCL subtype identified as AITL; or (b) In the second binary predictor, if the gene expression profile of the sample more closely matches the gene signature of ALCL, and In the first, third, and fourth binary predictors, if the gene expression profile of the sample more closely matches the gene signature of PTCL-NOS is the PTCL subtype identified as ALCL; The method according to claim 10. **Claim 12** The method according to claim 11(b), further comprising: A fifth binary predictor including determining whether the gene expression profile of the sample more closely matches the gene signature of ALK+ ALCL or ALK− ALCL; and Identifying the PTCL subtype as ALK+ ALCL if the gene expression profile of the sample more closely matches the gene signature of ALK+ ALCL, or Identifying the PTCL subtype as ALK− ALCL if the gene expression profile of the sample more closely matches the gene signature of ALK− ALCL. **Claim 13** (a) In the third binary predictor, if the gene expression profile of the sample more closely matches the gene signature of ATLL, and In the first, second, and fourth binary predictors, if the gene expression profile of the sample more closely matches the gene signature of PTCL-NOS is the PTCL subtype identified as ATLL; or (b) in the fourth binary predictor, if the gene expression profile of the sample is more consistent with the gene signature of ENKTL; and and wherein the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS in the first, second, and third binary predictors. wherein the PTCL subtype is identified as ENKTL, and optionally, samples identified as ENKTL are subdivided into NK or gamma / delta T cell lineages; or (c) identifying the PTCL subtype as PTCL-NOS if the gene expression profile of the sample is more consistent with PTCL-NOS in the first, second, third, and fourth binary predictors, and optionally further separating samples identified as PTCL-NOS cases into a GATA3-high subgroup and a TBX21-high subgroup; or (d) if at least two of the binary predictors are less consistent with PTCL-NOS, the PTCL subtype is identified as indeterminate, and optionally, samples identified as indeterminate are subjected to an additional binary predictor, wherein the additional binary predictor is: a first potential subtype comprising one of the PTCL subtypes less consistent with PTCL-NOS; and a second potential subtype comprising at least one other PTCL subtype less consistent with PTCL-NOS. Including; The method comprises: determining whether the gene expression profile of the sample is more consistent with a gene signature of the first potential subtype or the second potential subtype; and identifying the PTCL subtype as the first potential subtype if the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype; or identifying the PTCL subtype as the second potential subtype if the gene expression profile of the sample is more consistent with the gene signature of the second potential subtype. further comprising: Further optionally, if the PTCL subtype is identified as ALCL, the method further comprises: a fifth binary predictor comprising determining whether the gene expression profile of the sample is more consistent with the gene signature of ALK+ ALCL or ALK− ALCL. determining the ALCL subtype as either ALK+ ALCL or ALK- ALCL using further comprising: The method of claim 10.

14. A therapeutic agent for use in treating peripheral T-cell lymphoma in a subject, wherein the subject is obtaining a gene expression profile from a sample of the subject; comparing the gene expression profile to gene signatures associated with specific PTCL subtypes, wherein each PTCL subtype contains a unique gene signature; identifying the subtype of PTCL in the sample as either angioimmunoblastic T-cell lymphoma (AITL), anaplastic lymphoma kinase (ALK)-negative anaplastic large cell lymphoma (ALK- ALCL), adult T-cell leukemia / lymphoma (ATLL), extranodal natural killer / T-cell lymphoma (ENKTL), ALK-positive ALCL (ALK+ ALCL), PTCL not otherwise specified (PTCL-NOS), or indeterminate, wherein the indeterminate indicates that the gene expression profile of the sample includes genes from the unique gene signatures of at least two of the PTCL subtypes. have been identified as being in need of treatment for peripheral T-cell lymphoma based on the therapeutic agent is configured to treat the identified PTCL subtype; Optionally, (a) The therapeutic agent is a mammalian target of the present invention, comprising a histone deacetylase (HDAC) inhibitor, an antifolate, an acylating agent, a proteosome inhibitor, an antibody-drug conjugate, a phosphoinositide 3-kinase (PI3K) inhibitor, a Janus kinase (JAK) inhibitor, a signal transducer and activator of transcription (STAT) 3 inhibitor, a STAT5 inhibitor, an anaplastic lymphoma kinase (ALK) inhibitor, a hepatocyte growth factor (HGF) inhibitor, a cMET inhibitor, a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor, a platelet-derived growth factor receptor beta (PDGFRβ) inhibitor, or rapamycin. targeted (mTOR) pathway inhibitors, immune checkpoint inhibitors, hypomethylating agents, anti-cluster of differentiation 52 (CD52) antibodies, immunomodulatory agents, anti-inducible T-cell costimulator (ICOS) antibodies, CC-chemokine receptor 4 (CCR4) inhibitors, isocitrate dehydrogenase (IDH) inhibitors, B-cell lymphoma 2 inhibitors, anti-interleukin 2 receptor alpha chain (CD25) antibodies, calcineurin inhibitors, Notch signaling inhibitors, spleen tyrosine kinase (SYK) inhibitors, bispecific antibodies, chimeric antigen receptor T (CAR-T) cells, or combinations thereof; Optionally, the HDAC inhibitor comprises romidepsin, belinstat, panobinostat, or a combination thereof; the antifolate comprises pralatrexate; the akylating agent comprises bendamustine; the proteosome inhibitor comprises bortezomib; the antibody-drug conjugate comprises brentuximab vedotin; the PI3K inhibitor comprises duvelisib, tenalisib, or a combination thereof; the JAK inhibitor comprises ruxolitinib; the ALK inhibitor comprises crizotinib; the mTOR pathway inhibitor comprises everolimus; the hypomethylating agent comprises 5-azacytidine (5-Aza); the anti-CD52 antibody comprises alemtuzumab; the ImiD comprises lenalidomide; the CCR4 inhibitor comprises mogamulizumab; the IDH inhibitor comprises enasidenib; the BCL2 inhibitor comprises venetoclax; the anti-CD25 antibody comprises camidanlumabtesirin; the calcineurin inhibitor comprises cyclosporin A; the SYK inhibitor comprises celdulatinib; the bispecific antibody comprises AFM13; the chimeric antigen receptor T (CAR-T) cells comprise CD30, CD7, or both; or a combination thereof; or (b) if the PTCL subtype is identified as AITL, the therapeutic agent comprises romidepsin, 5-Aza, an isocitrate dehydrogenase (IDH) inhibitor, a calcineurin inhibitor, or a combination thereof; or (c) if the PTCL subtype is identified as ALK-ALCL, the therapeutic agent comprises brentuximab vedotin; or (d) if the PTCL subtype is identified as ALK+ ALCL, the therapeutic agent comprises an ALK inhibitor, a platelet-derived growth factor receptor beta (PDGFRβ) inhibitor, or a combination thereof; The therapeutic agent.

Images