Biomarkers for treatment of peripheral t-cell lymphoma and methods of use thereof
Identifying specific markers through gene expression profiling analysis and accurately diagnosing the subtype of PTCL, solving the problem of diagnosis in the prior art and improving the personalized selection of treatment and patient prognosis.
Patent Information
- Application Number
- CN202380029738.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-02-08
- Filing Date
- 2023-02-08
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art is difficult to accurately diagnose and classify subtypes of peripheral T-cell lymphoma (PTCL), resulting in difficult treatment choices and poor prognosis.
Through gene expression profiling analysis, specific gene markers are identified to distinguish the subtypes of PTCL, including the presence of ACHE, ADRA2A, ARRDC4 and other genes to identify the presence of AITL, BATF3, C9orf89 and other genes to identify ALK-ALCL, etc., and then provide subtypes Specific treatment.
Accurate diagnosis and classification of PTCL subtypes are achieved, personalized treatment choices are improved, and patient diagnosis and prognosis are improved.
Smart Images

Figure CN118974283A_ABST
Abstract
Description
[0001] This application is an international application that claims priority to U.S. Provisional Patent Application No. 63 / 307,905, filed on February 8, 2022, which is incorporated herein by reference in its entirety.
[0002] All patents, patent applications, and publications cited herein are hereby incorporated by reference in their entirety. The disclosures of these publications are hereby incorporated by reference in their entirety into this application in order to more fully describe the state of the art as 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 files and records, but otherwise reserves any and all copyrights whatsoever.
[0004] Government interests
[0005] This invention was made with government support under Grants U01 CA157581 and UH3 CA206127 from the National Institutes of Health (NIH). The government has certain rights in this invention. Technical Field
[0006] The subject matter of the present disclosure relates to methods for the diagnosis, prognosis and / or treatment of peripheral T-cell lymphoma (PTCL). Additionally, the subject matter of the present disclosure relates to kits and reagents for the diagnosis, prognosis and / or treatment of PTCL. In addition, the subject matter of the present disclosure relates to methods for identifying a group of biomarkers for the diagnosis, prognosis and / or treatment of PTCL using software and analysis systems.
[0007] introduce
[0008] Malignancies derived from T cells and natural killer (NK) cell lineages account for 10% of all non-Hodgkin lymphomas (NHL) in Western countries, and are more common in Asia. Peripheral T-cell lymphoma (PTCL) is a heterogeneous cancer, accounting for up to 20% of all non-Hodgkin lymphomas (NHL). The World Health Organization (WHO) classification identifies many unique subtypes of peripheral T-cell lymphoma (PTCL), including angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma (ALCL), adult T-cell leukemia / lymphoma (ATLL), and entities mainly derived from NK cells, including extranodal NK / nasal T-cell lymphoma (ENKTL). There are other rare PTCLs, most of which are extranodal tumors. The morphology of most subtypes is diverse and lacks clear diagnostic markers, making the diagnosis and classification of these diseases challenging and hindering their research. With current immunophenotypes and molecular markers, approximately 30% to 50% of PTCL cases are unclassifiable and are classified as PTCL not otherwise specified (PTCL-NOS). Although some recurrent genetic abnormalities have been reported in PTCL, no specific genetic abnormality is diagnostic of a specific PTCL subtype, with the exception of t(2;5)(p23;q35) in ALK(1)ALCL. Patients with PTCL generally have a poor prognosis under current standard of care therapies, and outcomes for patients with PTCL have not improved over the past two decades. Therefore, accurate diagnosis of PTCL and its subtypes is needed to ensure that the most appropriate and effective treatment is selected for each individual patient. In particular, the creation of a commercial assay for routine clinical use would improve the diagnosis, treatment, and outcomes of patients with PTCL. Summary of the invention
[0009] In the first aspect, the present invention includes a method for distinguishing the subtypes of peripheral T-cell lymphoma (PTCL). In an embodiment, the method includes isolating nucleic acids from a sample from a subject; 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 undetermined, wherein the undetermined type indicates that the sample contains the characteristics of at least two of these PTCL subtypes.
[0010] 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 can 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 can be identified as ALK-ALCL. In an embodiment, 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 is identified as ENKTL. 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, the PTCL subtype can be identified as ATLL.
[0011] The method of claim 2, wherein 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, the PTCL subtype is identified as ALK+ALCL. In certain embodiments, 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 PTCL subtype is identified as PTCL-NOS.
[0012] 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.
[0013] The method may also include providing an effective amount of a subtype-specific therapeutic to the subject. In an embodiment, the subtype-specific therapeutic includes a histone deacetylase (HDAC) inhibitor, an antifolate, an alkylating agent, a proteasome 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 p (PDGFRβ) inhibitor , mammalian target of rapamycin (mTOR) pathway inhibitors, immune checkpoint inhibitors, hypomethylating agents, anti-cluster of differentiation 52 (CD52) antibodies, immunomodulatory drugs, anti-inducible T-cell co-stimulator (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 a combination thereof. In an embodiment, the HDAC inhibitor comprises romidepsin, belinostat, panobinostat, or a combination thereof; the antifolate comprises pralatrexate; the alkylating agent comprises bendamustine; the proteasome inhibitor comprises bortezomib; the antibody-drug conjugate comprises brentuximab; the PI3K inhibitor comprises duvelixib, tenalixib, 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- zacitidine; the anti-CD52 antibody includes alemtuzumab; the ImiD includes lenalidomide; the CCR4 inhibitor includes moglizumab; the IDH inhibitor includes ensidipine; the BCL2 inhibitor includes venetoclax; the anti-CD25 antibody includes tecamimumab; the calcineurin inhibitor includes cyclosporine A; the SYK inhibitor includes selegiline; the bispecific antibody includes AFM13; the chimeric antigen receptor T (CAR-T) cell comprises CD30, CD7 or both; or a combination thereof.
[0014] Subtype-specific treatments for PTCL-NOS may include romidepsin, belinostat, brentuximab, duvelixibu, or a combination thereof. Subtype-specific treatments for AITL may include romidepsin, 5-Aza, isocitrate dehydrogenase (IDH) inhibitors, calcineurin inhibitors, or a combination thereof. Subtype-specific treatments for ALK-ALCL may include brentuximab. Subtype-specific treatments for ALK+ALCL may include ALK inhibitors, platelet-derived growth factor receptor β (PDGFRβ) inhibitors, or a combination thereof. Subtype-specific treatments for ATLL may include NOTCH inhibitors, hepatocyte growth factor (HGF) inhibitors, cMET inhibitors, moglizumab, or a combination thereof. Subtype-specific treatments for ENKTL may include platelet-derived growth factor receptor α (PDGFRαα) inhibitors.
[0015] 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 a tool for detecting the presence of a gene or a combination of genes and instructions for use, the genes representing a gene signature indicative of a particular PTCL subtype. In embodiments, 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 specified (PTCL-NOS), or undetermined, wherein the undetermined type indicates that the sample contains features of at least two of these 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 indicate 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 indicated as ALK-ALCL. In embodiments, 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 indicates ENKTL. 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 may indicate ATLL. 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 indicate 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 indicate PTCL-NOS.
[0016] Another aspect includes a method for identifying a specific subtype of peripheral T-cell lymphoma (PTCL) in a sample, the method comprising: obtaining a gene expression profile from the sample; comparing the gene expression profile to a gene signature associated with a specific PTCL subtype, wherein each subtype of PTCL comprises a unique gene signature; and identifying the subtype of the PTCL within the sample as 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 undetermined, wherein the undetermined type indicates that the gene expression profile of the sample includes genes from the unique gene signatures of at least two of the PTCL subtypes.
[0017] In the embodiment, 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 ALK-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 ATP8B4, CD3D, CD5, CTSW, FAM102A, FASLG, GZMB, KLRC2, KLRC3, KLRC4, PRF 1, SH2D 1B, EBER, LMP 1 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; the gene signature of ALK+ALCL includes ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GA any one or more of S1, 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 thereof.
[0018] In an embodiment, identifying a subtype of PTCL comprises: a series of at least four binary predictors, wherein: a first binary predictor comprises 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 comprises 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 comprises determining whether the gene expression profile of the sample is more consistent with the gene signature of ATLL or PTCL-NOS; a fourth binary predictor comprises determining whether the gene expression profile of the sample is more consistent with the gene signature of ENKTL or PTCL-NOS; and assigning the PTCL subtype of the sample based on the results from the binary predictors. A PTCL subtype may be identified as AITL when: in the first binary predictor, the gene expression profile of the sample is more consistent with the gene signature of AITL; and in the second, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS. The PTCL subtype can be identified as ALCL when: in the second binary predictor, the gene expression profile of the sample is more consistent with the gene signature of ALCL; and in the first, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS. In one embodiment, the method also includes a fifth binary predictor, wherein the fifth binary predictor includes determining whether the gene expression profile of the sample is more consistent with the gene signature of ALK+ALCL or ALK-ALCL; and when the gene expression profile of the sample is more consistent with the gene signature of ALK+ALCL, identifying the PTCL subtype as ALK+ALCL; or when the gene expression profile of the sample is more consistent with the gene signature of ALK-ALCL, identifying the PTCL subtype as ALK-ALCL.
[0019] In one embodiment, the PTCL subtype is identified as ATLL when the following conditions occur: in the third binary predictor, the gene expression profile of the sample is more consistent with the gene characteristics of ATLL; and in the first, second, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene characteristics of PTCL-NOS. When the following conditions occur, the PTCL subtype can be identified as ENKTL: in the fourth binary predictor, the gene expression profile of the sample is more consistent with the gene characteristics of ENKTL; and in the first, second, and third binary predictors, the gene expression profile of the sample is more consistent with the gene characteristics of PTCL-NOS. When the following conditions occur, the PTCL subtype can be identified as PTCL-NOS: in the first, second, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with PTCL-NOS.
[0020] When at least two of the binary predictors are not more consistent with PTCL-NOS, the PTCL subtype can be identified as undetermined. In one embodiment, the sample identified as undetermined is subjected to an additional binary predictor, the additional binary predictor comprising: a first potential subtype comprising one of the PTCL subtypes that is not more consistent with PTCL-NOS and a second potential subtype comprising at least one other PTCL subtype that is not more consistent with PTCL-NOS. In such embodiments, the method may also include: determining whether the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype or the second potential subtype; and when the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype, identifying the PTCL subtype as the first potential subtype; or when the gene expression profile of the sample is more consistent with the gene signature of the second potential subtype, identifying the PTCL subtype as the second potential subtype. If the PTCL subtype is identified as ALCL, the method may also include determining the ALCL subtype as ALK+ALCL or ALK-ALCL using a fifth binary predictor as disclosed herein.
[0021] In an embodiment, samples identified as PTCL-NOS cases are further separated into GATA3- and TBX21-high subgroups. Samples identified as ENKTL can be subdivided into NK or gamma / delta T cell lineages.
[0022] Another aspect includes a method for treating peripheral T-cell lymphoma. In an embodiment, the method includes: obtaining a gene expression profile from a sample of a subject; comparing the gene expression profile to a gene signature associated with a specific PTCL subtype, wherein each subtype of PTCL comprises a unique gene signature; identifying the subtype of PTCL within the sample as 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 undetermined, wherein the undetermined type indicates that the gene expression profile of the sample includes genes from the unique gene signature of at least two of the PTCL subtypes; and administering an effective amount of a therapeutic agent to the subject, wherein the therapeutic agent is configured to treat the identified PTCL subtype. In an embodiment, the therapeutic agent includes a histone deacetylase (HDAC) inhibitor, an antifolate, an alkylating agent, a proteasome inhibitor, an antibody-drug conjugate, a phosphoinositide 3-kinase (P13K) 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 lactation Animal mTOR pathway inhibitors, immune checkpoint inhibitors, hypomethylating agents, anti-cluster of differentiation 52 (CD52) antibodies, immunomodulatory drugs, anti-inducible T cell co-stimulator (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 a combination thereof.
[0023] In one embodiment, the HDAC inhibitor includes romidepsin, belinostat, panobinostat, or a combination thereof; the antifolate includes pralatrexate; the alkylating agent includes bendamustine; the proteasome inhibitor includes bortezomib; the antibody-drug conjugate includes brentuximab; the PI3K inhibitor includes duvelixib, tenalixib, or a combination thereof; the JAK inhibitor includes ruxolitinib; the ALK inhibitor includes crizotinib; the mTOR pathway inhibitor includes everolimus; the hypomethylating agent includes 5-azacitidine (5-Aza); the anti-CD52 antibody includes alemtuzumab; the ImiD includes lenalidomide; the CCR4 inhibitor includes moglizumab; the IDH inhibitor includes ensidipine; the BCL2 inhibitor includes venetoclax; the anti-CD25 antibody includes tecamimub; the calcineurin inhibitor includes cyclosporine A; the SYK inhibitor includes selegiline; the bispecific antibody includes AFM13; the chimeric antigen receptor T (CAR-T) cell comprises CD30, CD7 or both; or a combination thereof.
[0024] In the embodiment, when the PTCL subtype is identified as PTCL-NOS, the therapeutic agent includes romidepsin, belinostat, brentuximab, duvelixibu 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, and moglizumab. In the embodiment, when the PTCL subtype is identified as ALK-ALCL, the therapeutic agent includes brentuximab. When the PTCL subtype is identified as ALK+ALCL, the therapeutic agent may include ALK inhibitors, platelet-derived growth factor receptor β (PDGFRβ) inhibitors or a combination thereof. In one embodiment, when the PTCL subtype is identified as ENKTL, the therapeutic agent includes a platelet-derived growth factor receptor α (PDGFRα) inhibitor. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] This patent or application document contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawings will be provided by the 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 herein by reference in its entirety.
[0026] Certain features of the present invention are particularly set forth in the appended claims. A better understanding of the features and advantages of the present invention will be obtained by referring to the following detailed description of illustrative embodiments that illustrate embodiments utilizing the principles of the present invention and the accompanying drawings, in which:
[0027] Figure 1 Molecular diagnostic features of PTCL subgroups are shown. Unique gene expression signatures of major PTCL entities were identified using a mixed covariate prediction model (see references 10, 11 for details). Each column represents a PTCL patient, and each row represents a unique gene of the classifier.
[0028] Figure 2 Two major molecular subgroups within PTCL-NOS are shown. (A) Bayesian predictors for GATA3 and TBX21 subtypes were derived. Classification accuracy was cross-validated using leave-one-out cross-validation. Approximately 18% of cases were not clearly defined (UC, unclassifiable). (B) Overall survival (OS) analysis of GATA3 and TBX21 subtypes showed significant differences in clinical outcomes (P = .01) (Ref. 10).
[0029] Figure 3 The relation of error rate and gene quantity in the model is shown.The Y axis shows the average cross validation error rate of 500 simulated data sets that change with the X axis, wherein noise is added to the influence of approximate platform change, and the X axis indicates the available candidate gene quantity.Each colored line represents different diagnosis distinctions (referring to the legend in the figure).The curve drops sharply from left to right, and wherein each curve flattens at 15 to 20 gene places.These results show, the gene of this quantity is enough to carry out every kind of diagnosis distinction, and additional gene can not further reduce error rate.
[0030] Figure 4The performance of the Lymph2Cx assay in an independent validation cohort is shown. (A) The Lymph2Cx assay for 67 patients is shown as a gene expression heat map. The 20 genes contributing to the model are shown on the left, including 5 housekeeping genes. The cell-of-origin assignment of the assay is shown and compared to a gold standard method that used a previously published algorithm for gene expression on fresh frozen tissue and 3 immunohistochemistry-based algorithms. 20,23 (B) Comparison of Lymph2Cx scores in validation cohorts from two independent laboratories (MoCha, CLC). Horizontal and vertical light gray lines indicate the thresholds between GCB subtype, unclassified subtype, and ABC subtype. 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, Lymphoma Center, BC Cancer Agency (Vancouver, BC).
[0031] Figure 5 Included are histograms showing the distribution of correlations between Nanostring and Affymetrix data for 667 diagnostic and prognostic genes in DLBCL (A). (B) Genes present in the "academic" PTCL signature. Most of the 40 genes were well correlated, with a median correlation of 0.76.
[0032] Figure 6 Assay development using U133 array data from FF samples independent of the validation set to develop a linear predictor for subtypes including cut-off points is shown. In the training set, replicate samples will be used to adjust gene weights and additive shifts to account for platform differences, and to mimic the FF predictor score as closely as possible with FFPE samples / NanoString analysis, thereby establishing a locked version for validation. A Phase II study that will follow the current proposal is also indicated.
[0033] 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) institution. The overlap of these lines indicates excellent reproducibility between institutions, with some variability at minimum signal intensity.
[0034] Figure 8 Exemplary immunohistochemical staining of subclassification of ALCL and PTCL-NOS is shown. ALCL can be subdivided into CD30+ or ALK+, and GATA3 and TBX21 represent two subgroups of PTCL-NOS.
[0035] Fig. 9A histogram comparison of RNA yields from FFPE tissues collected at different times using different DNA / RNA isolation protocols is provided. RNA isolation was performed using the STORM protocol, the modified STORM protocol, or the Quiagen protocol. RNA recovery from older blocks was generally lower, but there was no difference using the newer kit.
[0036] Fig.10 Representative gel images of RNA isolated from FFPE and fresh frozen samples using different protocols are included in A. The tissues were collected in the indicated years. Fig.10 B provides a histogram comparing the effectiveness of different protocols in isolating larger RNA molecules (greater than 200 base pairs in length) from samples of different ages.
[0037] Fig.11 Display and Figure 1 In comparison, an improved gene signature (modified from ref. 10) was added with additional genes in the diagnostic signature list.
[0038] Fig.12 A PTCL molecular classifier is provided. Fig.12 A provides a study design and schematic diagram for molecular diagnosis of PTCL. HG-U133plus2.0 array data from PTCL with fresh frozen (FF) tissue (n=109), designated as a training cohort, was used to derive a molecular classifier for PTCL subclassification. The molecular classifier had 442 different genes, including housekeeping genes and other genes involved in T cell biology. Using the corresponding matched FFPE samples, transcriptomic features were considered for Analysis (NanoString, Inc). Transcripts showing high correlation between FF and FFPE data were selected. The algorithm was further refined to have a minimum number of transcripts for subclassification, and the FF predictor score was simulated with the NanoString platform (see M&M for details). The final diagnostic model generated 153 transcripts (99 diagnostic transcripts, 5 viral transcripts and 16 housekeeping transcripts and 33 T cell biology related transcripts) and was validated in an independent cohort of PTCL cases rigorously characterized by pathology and other ancillary methods. The classification algorithm is based on a series of several binary predictors to distinguish one entity from another, such as Fig.17 Described in detail. Fig.12 B provides the final Heatmap of classifiers. EBV and HTLV-1 viral transcripts commonly expressed in specific PTCL subtypes are shown, and housekeeping genes used for normalization and that do not vary between PTCL subtypes are shown for comparison. Fig.12C shows Kaplan-Meier curves for overall survival for 89 of the 105 training cohort cases and 66 of the 140 validation cohort cases with available outcome data. Fig.12 D shows Kaplan-Meier curves of OS for PTCL subtypes included in the training cohort (molecular classification by HG-U133plus2.0 array). Fig.12 E shows the Kaplan-Meier curves for OS of PTCL entities included in the validation cohort (pathological classification).
[0039] Fig.13 AITL classification is shown. Fig.13 A provides a scatter plot of the AITL diagnostic score versus the mean expression of the five TFH-related genes in the training and validation AITL. Fig.13 B provides box plots of EBER and EBNA1 transcript expression in AITL (training / validation cohort). Fig.13 C provides the Kaplan-Meier curve of OS of AITL in the combined group according to CD20mRNA expression. The solid line is the case classified as AITL according to molecular classification (p=0.02); the dotted line is AITL classified by pathology (p=0.06). CD20 expression in representative low-expression and high-expression AITL cases (400x). Fig.13 D shows the mutation status of cases with available sequencing data. Cases that were AITL-PTCL-NOS intermediate or not classified as AITL on the NanoString classifier are marked with asterisks. Fig.13 E provides Figure 2 Violin plot of AITL classification diagnostic scores in AITL and PTCL-NOS cases analyzed above. Discordant cases between AITL and PTCL-NOS are colored in red, and intermediate AITL cases are colored in gray. Fig.13 F provides a heat map of AITL, which shows the Disagreement in classification. The average characteristics of the concordant cases are shown. For cases labeled as intermediate, those diagnosed as AITL by consensus pathology are on the left, and intermediate cases diagnosed as PTCL-NOS are on the right. Figure GI provides an example of focal expression of BCL6 and ICOS (400x; G, H & I; B, BCL-6 and C, ICOS) observed in a PTCL-NOS case, which was classified as AITL according to The platform classification is AITL.
[0040] Fig.14 ALCL classification is shown. Fig.14 A provides a violin-shaped scatter plot of ALCL classification scores and PTCL-NOS. Fig.14 B shows ALK-positive ALCL and ALK-negative ALCL on the platform. Cases with discordance between ALCL and PTCL-NOS or between ALK-negative and ALK-positive are colored in red. Fig.14 C provides a heat map of ALK-negative and ALK-positive ALCL cases showing the differences in the validation cohort according to Disagreement in classification. H&E staining and CD30 and ALK immunostaining of a representative ALK-negative ALCL case classified as ALK-positive ALCL according to the NanoString assay are shown (400x). Fig.14 D provides Kaplan-Meier curves for OS of ALCL in the training and validation cohorts according to NanoString classification. Fig.14 E provides a heat map of CD30 and cytotoxic transcript expression in ALCL and PTCL-NOS cases. Discrepant cases are marked with red lines.
[0041] Fig.15 ATLL and ENKTCL classifications are shown. Fig.15 A provides Violin plots of ATLL classification scores in ATLL and PTCL-NOS cases analyzed above. Fig.15 B provides the verification group according to Heat map of inconsistently classified ATLL cases. H&E staining and CD4 staining of a case classified as ATLL (lower panel). Fig.15 C provides a scatter plot of the expression of the HTLV-1 specific transcript HBZ in training and validation ATLL cases versus ATLL score. The fitted solid line represents the training data, and the dashed line represents the validation data. Fig.15 D provides a heat map of HBZ and TAX expression in the ATLL and PTCL-NOS validation cohorts. Discrepant cases are marked with red asterisks. Fig.15 E provides a method for Scatter plot of measured HBZ expression. Fig.15 FG provides Violin plots of ENKTCL classification scores (F) or EBER scores (G) in ENKTCL and PTCL-NOS cases analyzed in the training and validation cohorts on Figure 2. Discordant cases are in red. Fig.15H provides heat maps of the expression of CD3γ and δ and EBV transcripts in ENKTCL and PTCL-NOS cases in the training and validation cohorts. Fig.15 I provides a heat map of the expression of relevant features in ENKTCL discordant cases compared with the mean features in the validation cohort.
[0042] Fig.16 PTCL-NOS subclassifications are shown. Fig.16 A provides Violin plots of PTCL-GATA3 classification scores in PTCL-NO S cases analyzed in the training and validation cohorts on Figure 5. Fig.16 B shows the mutation status of the PTCL-NOS cohort with available sequencing data. Fig.16 C provides heat maps of the expression of CD4, CD8, CD20, and cytotoxic genes in training (upper) and validation (lower) cases. Fig.16 D provides a scatter plot of the mean expression of cytotoxic genes versus CD20 in cases classified as PTCL-TBX21 according to NanoString. Fig.16 E provides H&E staining and IHC staining of a representative PTCL-GATA3 case showing GATA3 (left) and a PTCL-TBX21 case showing TBX21 and CD8 expression (right). Fig.16 F shows 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.
[0043] Fig.17 A schematic diagram of the algorithm design for PTCL subclassification is provided: six paired submodels are combined to generate the final predictor for classification. Stage 1: First, four categories (AITL, ALCL, ENKTCL or ATLL) are distinguished from PTCL-NOS. Subsequently, EBV viral transcripts (EBER) are added to the ENKTCL classifier (relative to PTCL-NOS), and ALCL is finally subdivided into ALK-ALCL and ALK+ALCL. Stage 2: In the second stage, samples considered as PTCL-NOS are distinguished as PTCL-GATA3 or PTCL-TBX21 subtypes. If more than one predictor in the 4 predictors produces a non-PTCL-NOS diagnosis (e.g., ALCL and AITL), additional head-to-head predictors are applied between those types to break the balance (e.g., ALCL and AITL), wherein all gene features of any type are selected to be included in the model.
[0044] Fig.18 A provides a scatter plot of log2(counts) by nCounter442 Gene Assay (NanoString, Inc) versus log2(probe intensity) values of corresponding transcripts of the same RNA extracted from fresh frozen tissue measured by HG-U133 plus2 (Affymetrix, Inc). Fig.18 B provides a histogram of the correlation values of each mRNA transcript of 10 cases analyzed by two platforms, nCounter (NanoString, Inc) and HG-U133 plus 2 (Affymetrix, Inc). Fig.18 C provides a gel image comparing RNA extracted with Qiagen (Q) or Storm kit (S). Fig.18 D provides a bar graph comparing the % RNA >200 bases in RNA extracted using the Qiagen (Q) or Storm kit (S). Fig.18 E provides scatter plots and histograms comparing log2(counts) generated by nCounter analysis of RNA extracted from fresh frozen (FF) or FFPE from corresponding biopsies using Qiagen or Storm kits. Numbers in boxes indicate correlation coefficients. Fig.18 F provides a scatter plot of log2(counts) generated by nCounter analysis of RNA extracted from FF or FFPE tissues of the corresponding biopsies.
[0045] Fig.19 A provides a histogram of the correlation values of each mRNA transcript for FFPE RNA analyzed by nCounter (NanoString, Inc) and FF RNA analyzed by HG-U133plus 2 (Affymetrix, Inc) for 100 cases analyzed by both platforms. Fig.19 B provides a histogram of the correlation values for each mRNA transcript in the classifier for FFPE RNA analyzed by nCounterand and FF RNA analyzed by HG-U133plus 2 for 100 cases analyzed by both platforms. Fig.19 CD provides heat maps of Affymetrix U133plus2 data from FF RNA of cases included in the training cohort, which depicts the original but expanded classifier gene list (C) and the simplified classifier gene list (D) ultimately used in the nCounter platform classification model. Housekeeping genes that do not vary between PTCL subtypes are shown for comparison. Fig.19E provides box plots and dot plots of the variance of housekeeping genes and model genes analyzed by nCounter in FFPE tissues in training and validation case runs. Fig.19 F provides Kaplan-Meier curves for overall survival of PTCL-NOS cases classified as GATA3 or TBX21 using FF RNA analyzed by HG-U 133plus2 Affymetrix array (top) or FFPE RNA analyzed by nCounter (bottom).
[0046] Fig. 20 The model score (FFPE RNA) produced at the University of Nebraska Medical Center (UNMC) (initial design) is shown with the comparison of the model score on the final diagnostic model run at 3 different sites (UNMC, Insight.Inc and COH). The threshold for classifying cases as PTCL entities is marked as green. The light green line describes the threshold of the intermediate AITL / PTCL-NOS score. The red line indicates the result of orthogonal regression, which can be compared with the gray line indicating the diagonal line of equal model score. *nCounterpanel uses a simplified gene design comprising 153 code sets. **nCounterpanel uses an initial design comprising 442 code sets. All cases are run on the initial 442 code set design at UNMC, and run on the final 153 code set design at other sites, while 7 samples are also repeated using 153 final design code set panels at UNMC. All scores are generated using the final model.
[0047] Fig.21 Violin plots of the classification scores in all PTCL cases analyzed in the training and validation cohorts on nCounter are provided. The x-axis is the known diagnosis. Cases where the nCounter classification disagrees with the expected diagnosis are colored according to how they were classified, while consistent cases are gray.
[0048] Fig. 22 Shown are H&E and IHC staining of one of the nodular ENKTL cases with strong expression of EBER and cytotoxic (TIA-1 and perforin) markers.
[0049] Fig.23 A provides heat maps of the expression of TFH, AITL, GATA3, and TBX21 signatures in AITL, PTCL-TFH and PTCL-TBX21, PTCL-GATA3, and reactive hyperplasia. Fig.23 B provides a box plot of the mean expression of TFH-specific genes in the entities.
[0050] Fig.24 The agreement between the gold standard diagnosis and the improved diagnostic signature in the training cohort is shown. AITL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ATLL, adult T-cell lymphoma / leukemia; 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. # Combination of pathological and molecular diagnosis based on fresh frozen GEP.
[0051] Fig.25 The consistency between the gold standard diagnosis and the improved diagnostic features in the validation group is shown. In 85% (119 / 140) of the case, the transcription classifier matches the pathological diagnosis provided by 3 hematopathology experts, and in 6% (8 / 140) of the case, " critical correlation " with molecular features is demonstrated. In general, between consensus pathological examination and molecular diagnosis, it is believed that 127 cases (green) in 140 cases are consistent. The classifier improves the pathological diagnosis of 2 cases (that is, PTCL-NOS cases (aqua) classified as ATLL in molecules), which is consistent with subsequent clinical pathological information when reviewed. In the remaining 11 cases with " inconsistent ", 4 cases have molecular classification (yellow), which can provide an improvement that is superior to pathological diagnosis based on overall transcriptomic features and morphological features. AITL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ATLL, adult T-cell lymphoma / leukemia; ENKTCL, extranodal NK / T-cell lymphoma; PTCL-NOS, peripheral T-cell lymphoma not otherwise specified. *Borderline cases were not considered discordant. +PTCL-NOS cases were molecularly classified as ATLL on review and were concordant with subsequent clinicopathological information. ++ PTCL-NOS cases were molecularly classified as AITL, which showed TFH marker expression on reexamination. +++ ALK-ALCL cases may represent ALK-positive-like ALCL based on the overall transcriptomic signature. DETAILED DESCRIPTION
[0052] The details of one or more embodiments of the subject matter of the present disclosure are provided herein. However, it should be understood that the subject matter of the present disclosure can be embodied in various forms. After studying the information provided in this document, the modification of the embodiment described in this document and other embodiments will be apparent to those of ordinary skill in the art. The information provided in this document, particularly the specific details of the exemplary embodiments described, are provided primarily for clear understanding and should not be construed as limiting. In the event of a conflict, the specification (including definitions) of this document shall prevail.
[0053] Although it is believed that the terms used herein are well understood by those of ordinary skill in the art, certain definitions are set forth in order to explain the subject matter of the present disclosure. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those of ordinary skill in the art to which the subject matter of the present disclosure belongs.
[0054] Although any methods, devices, and materials similar or equivalent to those described herein can be used in the practice or testing of the presently disclosed subject matter, representative methods, devices, and materials are now described.
[0055] In some cases, the nucleotides and polypeptides disclosed herein are included in publicly available databases, such as and SWISSPROT. The information including sequences and other information related to such nucleotides and polypeptides contained in such publicly available databases is expressly incorporated herein by reference. Unless otherwise indicated or apparent, references to such publicly available databases are references to the most recent version of the database as of the filing date of this application.
[0056] Unless otherwise defined, all technical terms and scientific terms used herein have the same meaning as the meaning commonly understood by those skilled in the art to which the present invention belongs. Unless otherwise stated, all patents, patent applications, published applications and publications, GenBank sequences, databases, websites and other published materials mentioned in the entire disclosure of this article are incorporated by reference in their entirety. In the case where the term of this article has multiple definitions, the definition in this chapter and section shall prevail. In the case of quoting URL or other such identifiers or addresses, it should be understood that although such identifiers may change and the specific information on the Internet may vary, equivalent information can be found by searching the Internet. The availability and public dissemination of such information have been demonstrated by quoting such information.
[0057] Although any methods, devices, and materials similar or equivalent to those described herein can be used in the practice or testing of the presently disclosed subject matter, representative methods, devices, and materials are now described.
[0058] Following long-standing patent law convention, when used in this application, including the claims, the terms "a," "an," and "the" mean "one or more." Thus, for example, reference to "a cell" includes a plurality of such cells, and so forth.
[0059] The term "substantially" permits deviations from the descriptor that do not adversely affect the intended purpose. Even if the word "substantially" is not explicitly recited, the descriptive term should be understood to be modified by the term "substantially." Thus, for example, the phrase "wherein the lever extends vertically" means "wherein the lever extends substantially vertically," as long as a precise vertical arrangement is not necessary for the lever to perform its function.
[0060] The terms "comprising" and "including" and "having" and "involving" (and similarly "comprises," "includes," "has," and "involves") and the like are used interchangeably and have the same meaning. Specifically, each of these terms is defined to be consistent with the common U.S. Patent law definition of "comprising," and thus should be interpreted as open terms meaning "at least the following," and should also be interpreted as not excluding additional features, limitations, aspects, etc. Thus, for example, "a process involving steps a, b, and c" means a process that includes at least steps a, b, and c. Whenever the term "a" or "an" is used, it should be understood to mean "one or more" unless such interpretation does not make sense in the context.
[0061] Unless otherwise indicated, all numbers expressing quantities of ingredients, properties such as 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 indicated to the contrary, 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 of the present disclosure.
[0062] 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 specified amount in some embodiments ±20%, from the specified amount in some embodiments ±10%, from the specified amount in some embodiments ±5%, from the specified amount in some embodiments ±1%, from the specified amount in some embodiments ±0.5%, and from the specified amount in some embodiments ±0.1%, as such variations are suitable for performing the disclosed methods.
[0063] As used herein, ranges may be expressed as from "about" one particular value and / or to "about" another particular value. It is also understood that a number of values are disclosed herein, and that 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 every 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.
[0064] As used herein, "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 in the form of, for example, a tissue biopsy, such as a puncture biopsy, a brush biopsy, a surface biopsy, a needle biopsy, a punch biopsy, an excisional biopsy, an open biopsy, an incisional biopsy, and an endoscopic biopsy. "Biological sample" can refer to a sample of tissue or fluid separated from a subject, including, but not limited to, for example, blood, plasma, serum, tumor biopsy, urine, feces, sputum, cerebrospinal fluid, pleural effusion, nipple aspirate, lymph fluid, skin, respiratory tract, intestine, and genitourinary tract external sections, tears, saliva, milk, cells, tumors, organs, and samples of in vitro cell culture components. Non-limiting examples of biological samples include formalin-fixed paraffin-embedded (FFPE) tissues, fresh frozen (FF) samples, or other prepared sample types.
[0065] Technicians will recognize that biological samples from subjects can be obtained by a variety of conventional techniques. The methods described herein generally involve obtaining biological samples from subjects (such as subjects suspected of having cancer or subjects suffering from cancer). As used herein, the phrase "obtaining a biological sample" refers to any process for obtaining a biological sample directly or indirectly from a subject. For example, a biological sample can be obtained by obtaining a tissue or fluid sample (e.g., a biopsy, a blood draw, a bone marrow sample, a spinal tap) from a subject. Alternatively, a biological sample can be obtained by receiving a biological sample (e.g., in a laboratory facility) from one or more people who directly obtain a sample from a subject. The biological sample can be, for example, a tissue (e.g., a biopsy, blood), a cell (e.g., a hematopoietic cell, such as a hematopoietic stem cell, a leukocyte or a reticular cell, a stem cell or a plasma cell, a cancer cell), a vesicle, a biomolecule aggregate, or a platelet from a subject.
[0066] In one embodiment, the sample is from a resection, biopsy or core needle biopsy of a primary or metastatic tumor. In addition, a fine needle is used to aspirate the sample. The sample can be paraffin-embedded or frozen tissue.
[0067] In embodiments, the biological sample can be a biopsy, wherein the biopsy contains at least one cancer cell or cell suspected of being a cancer cell. The term "biopsy" can refer to a tissue sample removed from a subject for the purpose of determining whether the sample contains cancerous tissue or examining cancerous tissue. In some embodiments, the biopsy is obtained because the subject is suspected of having cancer. The biopsy is then examined for the presence of cancer. In other embodiments, the biopsy is obtained because the subject is known to have cancer, and the biopsy is then examined for markers indicating the stage of cancer and / or treatment options.
[0068] As used herein, "diagnosis" refers to detecting and identifying a disease in a subject. The term may also encompass assessing or evaluating the disease status (progression, regression, stability, response to treatment, etc.) of a patient known to have a disease (eg, PTCL).
[0069] As used herein, the term "prognosis" refers to information that provides information about the impact of the presence of cancer (e.g., as determined by the diagnostic methods of the present invention) on the future health of the subject (e.g., expected morbidity or mortality, the likelihood of suffering from cancer, and the risk of metastasis). In other words, the term "prognosis" refers to a prediction that provides the possible course and outcome of cancer or the likelihood of recovery from cancer. The term "prognosis" is well-known in the art and encompasses predictions about the possible course of a disease or disease progression, particularly predictions about the likelihood of disease remission, disease exacerbation, tumor recurrence, metastasis, and death. A "good prognosis" may refer to the likelihood that a patient with cancer will remain cancer-free after treatment. A "poor prognosis" may refer to the likelihood that a potential cancer will deteriorate or recur after treatment, the likelihood of developing metastasis, and / or the likelihood of death. In a particular embodiment, the time frame for assessing prognosis is, for example, less than one year, one year, two years, three years, four years, five years, six years, seven years, eight years, nine years, ten years, fifteen years, twenty years, or more years.
[0070] As used herein, the terms "treat," "treatment," or "treating" refer to any type of treatment that brings benefit to a patient suffering from a disease, including improving the patient's condition (eg, one or more symptoms), slowing the progression of the condition, and the like.
[0071] Aspects of the invention relate to a method of treating cancer in a patient in need thereof. As used herein, treatment of cancer may refer to partially or completely inhibiting, delaying, or preventing the progression of cancer, including cancer metastasis; inhibiting, delaying, or preventing cancer recurrence; or preventing the onset or development of cancer in a mammal (e.g., a human) (chemoprevention). Treating cancer may be indicated by preventing or reducing tumor development, tumor growth, proliferation, angiogenesis, and / or metastasis.
[0072] Aspects of the present invention also include providing diagnostic information of cancer in a subject. For example, aspects of the present invention may include providing diagnostic information to a subject suspected of having cancer or a subject at risk of having cancer. The term "subject suspected of having cancer" may refer to a subject who exhibits one or more symptoms indicating cancer (e.g., an obvious lump or lumps) or is undergoing cancer screening (e.g., during a routine physical examination). Subjects suspected of having cancer may also have one or more risk factors. Subjects suspected of having cancer have not generally been tested for cancer. However, "subject suspected of having cancer" encompasses individuals who have received a preliminary diagnosis but whose cancer stage is unclear. The term also includes people who have once had cancer (e.g., individuals in remission). As used herein, the term "subject at risk of cancer" refers to a subject with one or more risk factors for developing a specific cancer. Risk factors include, but are not limited to, sex, age, genetic susceptibility, environmental exposure, previous cancer incidence, preexisting non-cancer diseases, and lifestyle.
[0073] The phrase "effective amount" refers to the amount of a therapeutic agent that results in an improvement in a patient's condition. In certain embodiments, an effective amount increases survival by at least one year. An effective amount can be measured by the number of years in which the disease has not substantially progressed after treatment has begun. In embodiments, an effective amount can be an amount sufficient to cause at least one year of progression-free survival (PFS). An effective amount can be sufficient to cause more than one year of PFS. In embodiments, an effective amount is a dose sufficient to cause 1 year PFS, 2 year PFS, 3 year PFS, 4 year PFS, or 5 year PFS. In embodiments, the therapeutic agent is an anticancer agent.
[0074] As used herein, the term "subject" refers to the target of administration. The subject of the methods disclosed herein can be a vertebrate, such as a mammal, fish, bird, reptile, or amphibian. Thus, the subject of the methods disclosed herein can be human or non-human. Therefore, veterinary therapeutic uses are provided according to the subject matter of the present disclosure.
[0075] "Subject" may refer to mammals, such as humans and non-human primates, as well as mammals of importance due to endangerment, such as Siberian tigers; mammals 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; pigs, including piglets, swine, and wild boars; ruminants and / or ungulates, such as cattle, bulls, sheep, giraffes, deer, goats, bison, and camels; rabbits, guinea pigs, and rodents. Treatments of birds are also provided, including those species of birds that are endangered and / or kept in zoos, as well as poultry, more specifically domesticated poultry, i.e., poultry, such as turkeys, chickens, ducks, geese, guinea fowl, etc., because they are also of economic importance to humans. Thus, treatments of livestock are also provided, including, but not limited to, domestic pigs, ruminants, ungulates, horses (including racehorses), poultry, etc.
[0076] In embodiments, the subject has previously been diagnosed with cancer and may have received cancer treatment. In alternative embodiments, the subject has not previously been diagnosed with cancer. The present invention can be used for all patients at risk of cancer. Although each type of cancer has its own risk factors, the risk of developing cancer increases with age, sex, race, and personal and family history. Other risk factors are primarily related to lifestyle choices, while certain infections, occupational exposures, and some environmental factors may also be relevant to developing cancer.
[0077] As used herein, the terms "administering" and "administration" refer to any method of providing a pharmaceutical formulation 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, inhalation administration, nasal administration, topical administration, intravaginal administration, ocular administration, intraoral administration, intracerebral administration, rectal administration, and parenteral administration, including injections such as intravenous administration, intraarterial administration, intramuscular administration, and subcutaneous administration. Administration may be continuous or intermittent. The formulation may be administered therapeutically; that is, administered to treat an existing condition of interest. The formulation may be administered prophylactically; that is, administered to prevent a condition of interest.
[0078] The disclosed subject matter relates to methods of 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 embodiments, the cancer is non-Hodgkin's lymphoma. The cancer may be PTCL.
[0079] In one embodiment, the method includes: obtaining a biological sample from a subject, wherein the biological sample contains at least one cell that is a cancer cell or suspected of being a cancer cell; measuring or determining the expression level of a gene from the biological sample; comparing the expression level in the biological sample with a control sample; and providing diagnostic, prognostic or predictive information based on the measuring step.
[0080] In embodiments, the measurement result (such as the level of nucleotides or proteins of a gene) can be compared to a control, and if the level is different from the control, the subject may have an increased likelihood of suffering from and / or developing cancer, a reduced likelihood of suffering from or developing cancer, an increased likelihood of responding to a given treatment, or a reduced likelihood of responding to a given treatment. A sample or subject that is "different from the level of the control" should be understood as having a level of an analyte to be detected or a diagnostic or therapeutic indicator (e.g., a marker) that is statistically different from the level of a sample from a normal, untreated, or abnormal state control sample. Those skilled in the art have the ability to determine statistical significance, such as the number of standard deviations constituting the mean of a positive or negative result. For example, a control sample may include one or more non-cancerous cells, or a biological sample from a patient who has not been diagnosed with cancer.
[0081] In embodiments, the measurement results (such as the level of mRNA and / or protein) can be compared to a threshold value to determine whether the subject has cancer, diagnose the type or subtype of cancer, provide a prognosis for the patient, determine a specific treatment regimen, or a combination thereof. The term "threshold value" can refer to a value of a biomarker derived from a plurality of biological samples, above which the threshold value is associated with an increased likelihood of having and / or developing cancer, or an increased likelihood of responding to a given treatment.
[0082] According to certain embodiments of the inventive method, the presence or amount of a gene product (e.g., a polypeptide or nucleic acid) encoded by a gene is detected in a sample derived from a subject (e.g., a sample of tissue or cells obtained from a tumor or a blood sample obtained from a subject). Various processing steps may be performed on the sample before or during detection.
[0083] For the purposes of the present invention, the term "gene" has the meaning understood in the art. However, it will be appreciated by those of ordinary skill in the art that the term "gene" has a variety of 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 also be understood that the definition of "gene" includes nucleic acids that do not encode proteins but encode functional RNA molecules (such as tRNA). For the sake of clarity, we note that, as used in this application, the term "gene" generally refers to a portion of a nucleic acid that encodes a protein; the term may optionally encompass 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 in this document refers to protein-coding nucleic acids.
[0084] A gene product or expression product is typically the RNA transcribed from the gene or a polypeptide encoded by the RNA transcribed from the gene.
[0085] The expression of a gene can be measured by a variety of techniques known in the art. Certain techniques can utilize polynucleotides corresponding to part or all of a gene, rather than antibodies that bind to the polypeptide encoded by the gene. Suitable techniques include, but are not limited to, in situ hybridization, Northern blotting, and various nucleic acid amplification techniques such as PCR, quantitative PCR, and ligase chain reaction.
[0086] In some embodiments, the present application provides methods 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-NO S, 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 their target genes.
[0087] In certain embodiments, the biomarkers include gene signatures that are uniquely present in different subtypes of PTCL. The gene signature of each subtype of PTCL may include one or more genes listed in Table 1 below.
[0088] Table 1: Exemplary PTCL subtype gene signatures
[0089]
[0090]
[0091] In some embodiments, the one or more biomarkers include GATA, TBX21 or their related target genes. The one or more biomarkers may include NK or γ / δT cell lineages. In certain embodiments, the NK lineage includes EBV viral transcription (EBER and LMP1). In embodiments, the gene signature of ATLL subtypes also includes HTLV-1 viral transcript expression (TAX, HBZ or its combination).
[0092] 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 set of genes.
[0093] In some embodiments, the present application provides a method for predicting the survival of a subject diagnosed with a specific subgroup of PTCL. In some embodiments, the method involves detecting the expression profile of a specific gene group.
[0094] In some embodiments of the presently disclosed subject matter, methods for treating a specific subgroup of PTCL in a subject are also provided. In embodiments, the treatment method comprises administering an effective amount of an anticancer agent to a subject and administering the anticancer agent to the subject. The anticancer agent may vary depending on the PTCL subtype of the subject. In embodiments, treatment is optimized to specifically target the PTCL subtype of the subject. Non-limiting examples of such agents include chemotherapy, immunotherapy, toxin therapy, radiotherapy, or a combination thereof. Anticancer agents may include pralatrexate, romidepsin, belinstat, brentuximab, duvelixib, decitabine and related compounds, EZH1 / 2 inhibitors, other PTCL subtype-specific treatments currently known or later discovered, or a combination thereof,
[0095] Additionally, the subject matter of the present disclosure relates to kits and reagents for diagnosis, prognosis and / or treatment of PTCL. In some embodiments, the kit includes a device for detecting the expression profile of a specific gene group. In some embodiments, the kit includes one or more probes.
[0096] In addition, the presently disclosed subject matter relates to methods of identifying a set of biomarkers for diagnosis, prognosis, and / or treatment of PTCL using software and analysis systems.In some embodiments, the presently disclosed subject matter provides methods for subtyping PTCL by comparing the expression levels of multiple genes.
[0097] Embodiment adopts method to analyze a large amount of nucleic acid in a single reaction.In embodiments, method or kit adopts high-throughput nucleic acid sequencing technology, such as next generation sequencing (NGS).In embodiments, targeted RNA sequence group can be provided on NGS platform.Non-limiting examples of such NGS technology include instruments and schemes from Illumina, Inc, San Diego, CA, USA, Thermo Fischer Scientific, Waltham, MA, USA, and Qiagen, Venlo, Netherlands.In other embodiments, some platforms based on multiple PCR are adopted.Embodiment uses a platform based on qPCR configured to accommodate a large amount of analytes, such as Qiagen Modaplex (Venlo, Netherlands) or similar platforms.Embodiment can adopt conventional real-time PCR platform. Alternative embodiments utilize any of various platforms based on expression microarrays, including but not limited to Affymetrix from Santa Clara, CA, USA, Aknonni Biosystems from Frederick, Maryland, USA, MD, USA, Biofire Diagnostics from Salt Lake City, Utah, USA, UT, USA, or similar platforms. Some embodiments use non-array platforms and schemes. In certain non-array embodiments, quantitative nuclease protection assays are utilized, and other non-array embodiments use platforms and schemes from NanoString Technologies, Inc., Seattle, Washington, USA, WA, USA. The examples provided herein are merely exemplary and should not be considered to be limited in any way. Embodiments use any of various molecular diagnostic platforms.
[0098] Certain embodiments employ a DNA chip, which is a device that facilitates simultaneous comparison of the expression levels of multiple genes. DNA chip-based expression profiling can be performed by methods disclosed in, for example, "Microarray Biochip Technology" (Mark Schena, Eaton Publishing, 2000).
[0099] DNA chips contain fixed high-density probes to detect multiple genes. Therefore, the expression levels of many genes can be estimated simultaneously through a single round of analysis. That is, DNA chips can be used to determine the expression profile of a sample.
[0100] In some embodiments, the subject matter of the present disclosure relates to methods for identifying biomarkers that are specific to different cancer subtypes (such as PTCL). Embodiments include steps of performing a diagnostic algorithm based on GEP analysis of a biological sample. Certain embodiments use a stepwise binary model algorithm for subgroup classification. Embodiments provide methods for substantially reducing the number of genes in a cancer subgroup classifier without significantly compromising diagnostic accuracy. In embodiments, 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 between 15-20, including end values. 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, about 30. The method also includes identifying different oncogenic pathways, which include potential therapeutic targets in these different molecular PTCL entities.
[0101] In some embodiments, the presently disclosed subject matter provides methods of distinguishing PTCL subtypes in a subject using refined diagnostic and prognostic biomarkers.
[0102] The methods described herein can also be used to select a treatment for cancer. For example, embodiments may include providing a cancer sample or obtaining (e.g., isolating) a cancer sample from a subject; and obtaining a cancer sample from a subject; measuring at least one of the expression, activity, or product of a gene in a PTCL feature; and selecting a treatment based on the measurement step. Embodiments may also include a step of comparing the measured level to a level or threshold of a control sample.
[0103] Aspects of the invention also encompass kits, such as kits for treating and / or diagnosing cancer.
[0104] In one embodiment, a kit may include one or more compositions described herein, such as an inhibitor.
[0105] For example, the kit may include one or more reagents for detecting gene levels in a PTCL signature or other protein in a sample. The one or more reagents may be fixed to a solid support. Non-limiting examples of the composition of the solid support structure include plastic, cardboard, glass, plexiglass, tin, paper, or a combination thereof. The solid support may also include a dipstick, a spoon, a scoopula, a filter paper, or a swab.
[0106] 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 device for determining the amount of gene expression in a sample; and a device for comparing the amount of gene expression in a sample with a standard (such as a control sample or a threshold value). In some embodiments, the standard is a non-cancerous cell or a cell extract thereof. The compound or reagent may be packaged in a suitable container. The kit may also include instructions for using the kit to detect cancer in a sample.
[0107] In an embodiment, the kit may further 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 nucleic acids that hybridize with the primers.
[0108] The kits may also include sample collection devices, such as those used to collect biological fluids or tumor biopsies.
[0109] In one embodiment, the kit may include a container containing one or more reagents and optionally (b) informational material. The informational material may be descriptive, instructional, marketing or other material related to the methods described herein and / or the use of the reagents for diagnostic purposes. In one embodiment, the kit also includes one or more anti-cancer therapeutic agents.
[0110] The informational material of the kit is not limited in its form. In one embodiment, the informational material may include information about the production of the components of the kit, such as molecular weight, concentration, expiration date, batch or production 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 forms, including printed text, computer readable material, video recording or audio recording, or information providing links or addresses to substantive materials.
[0111] The kit may include other ingredients, such as solvents or buffers, stabilizers or preservatives. Optionally, the kit may include therapeutic agents, which may be provided in any form, such as liquid, dry or lyophilized form, preferably substantially pure and / or sterile. When the medicament is provided in a liquid solution, the liquid solution is preferably an aqueous solution. When the medicament is provided in a dry form, it is usually reconstituted by adding a suitable solvent. Solvents, such as sterile water or buffer, may optionally be provided in the kit.
[0112] Other Implementations
[0113] While the invention has been described in conjunction with specific embodiments thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages and modifications are within the scope of the appended claims.
[0114] The present invention will be further described in the following examples, which do not limit the scope of the invention described in the claims.
[0115] Example
[0116] Examples are provided below to facilitate a more complete understanding of the present invention. The following examples illustrate exemplary modes for preparing and practicing the present invention. However, the scope of the present invention is not limited to the specific embodiments disclosed in these examples for illustrative purposes only, because similar results can be obtained using alternative methods.
[0117] Example 1
[0118] The primary goal of this study is to transition an "academic" PTCL diagnostic and prognostic gene expression (GE) signature originally developed using fresh frozen samples analyzed with expression microarrays to an optimized, locked signature suitable for formalin-fixed paraffin-embedded (FFPE) samples analyzed on the Nanostring nCounter platform. The two objectives of this proposal are: Objective 1. Identify a simplified set of transcripts and create a new model capable of replicating the existing "academic" diagnostic and prognostic PTCL signature; Objective 2. Confirm the applicability of the refined diagnostic and prognostic PTCL gene signature set on the NanoString nCounter platform for analyzing FFPE samples - The feasibility of the optimized FFPE / nCounter assay as a candidate for further commercial test development will be determined, which will be conducted in a Phase II study immediately following the current proposal and will include a comprehensive evaluation and validation of the analytical and clinical performance and clinical utility of the assay.
[0119] Specific goals
[0120] Phase I: "Academic" Gene Expression Profiling (GEP) for the Diagnosis and Prognosis of Peripheral T-cell Lymphoma (PTCL) Optimization of the optimized features and preliminary assessment of their feasibility for commercial assay development .
[0121] Objective. The primary objective of this study is to transition an “academic” PTCL diagnostic and prognostic GEP signature initially developed using fresh frozen (FF) samples analyzed with expression microarrays to an optimized, locked signature suitable for use with formalin-fixed paraffin-embedded (FFPE) samples analyzed on the Nanostring nCounter platform. This work is divided into two objectives, and specific activities for each objective are detailed below. These Phase I studies will be used to determine the feasibility of the optimized FFPE sample / nCounter assay as a candidate for further commercial test development (including its provision in CLIA-certified laboratories and further development as an in vitro diagnostic (IVD) kit and companion diagnostic (CDx) assay for new or existing PTCL therapies), and if the Phase I objectives are achieved, this will be followed by a Phase II follow-up study, including a comprehensive evaluation and validation of the analytical and clinical performance and clinical utility of the assay.
[0122] Objective 1: To identify a reduced set of transcripts and create a library of transcripts that can replicate existing “academic” diagnostic and prognostic features of PTCL. New Model of Signs .
[0123] 1a. Identification of a reduced set of transcripts required to optimally distinguish each PTCL entity Deliverable: Identify gene transcripts from the original PTCL subtyping signature that are transitionable from the fresh frozen (FF) sample / expression array platform to the FFPE sample / nCounter platform. These studies will determine the performance of the codeset on the nCounter platform in RNA from approximately 60 FFPE samples compared to data from FF samples from the same cases previously characterized using arrays.
[0124] 1b. Refinement of diagnostic and prognostic features that can be applied to FFPE tissue to differentiate PTCL entities Deliverable: Identification of a single, workable minimal gene set that replicates the original “academic” PTCL signature with as close to uniform consistency as possible. Transcripts that transition well from the FF / microarray to the FFPE / nCounter platform as shown in Objective 1a will be used in a training set of approximately 120 PTCL cases to develop classification models and algorithms consisting of the smallest possible number of genes to distinguish clinically relevant diagnostic and prognostic PTCL subtypes.
[0125] Objective 2: To confirm a refined diagnostic and prognostic PTCL gene signature set for analysis of FFPE samples Applicability on the NanoString nCounter Platform .
[0126] 2a. To create a locked diagnostic and prognostic PTCL gene signature and evaluate interlaboratory reproducibility. Deliverables: Target an optimized PTCL signature based on initial validation cohort and interlaboratory study results. Approximately 100 PTCL FFPE cases will be subtyped independently using refined diagnostic and prognostic gene signatures at the molecular diagnostic laboratories of UNMC, City of Hope, and HealthChart in conjunction with a consulting commercial CLIA laboratory (InsightMolecular Labs), and results will be compared to FF / microarray gold standard subtyping previously obtained from matched samples. Interlaboratory assay reproducibility will be evaluated, assay algorithms finalized, and final assay SOPs derived.
[0127] 2b. Validation of a "lock-in" diagnostic / prognostic signature for phase II assay transition Deliverable: Validation cohort results confirm suitability of locking signature for progression to Phase II development. A PTCL validation cohort of approximately 30 FFPE cases will be analyzed on the nCounter platform in multiple independent laboratories worldwide to confirm the suitability of the optimized subtyping signature, algorithm, and SOP for progression to Phase II assay development.
[0128] A. Scientific and Clinical Significance
[0129] Peripheral T-cell lymphomas (PTCLs) are a heterogeneous entity that presents therapeutic challenges: PTCLs constitute up to 20% of non-Hodgkin lymphomas. 1,2 The World Health Organization (WHO) recognizes many 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 PTCL is challenging to diagnose even for hematopathologists. 4 , and 30%-50% of PTCL cannot be classified by current immunophenotypes and molecular markers and are therefore classified as PTCL-NOS. 3 In addition to ALK(+) ALCL (characterized by co-lead Steve Morris, MD) 5 Patients with PTCL, except ALK(-)ALCL with DUSP22 rearrangement, have a poor prognosis with CHOP-like chemotherapy or more intensive regimens. 4,6,7 In fact, 5-year progression-free survival (PFS) has remained at only 20%-25% over the past two decades. 8-12 However, new treatments have begun to be tested in specific patient groups, some with striking results (see Table 2). 13-16Therefore, accurate PTCL classification will be crucial both for therapeutic trials and for selecting the best treatment for patients in the clinic. 17 (See, for example, letters of support from Celgene and Seattle Genetics).
[0130] Table 2: Phase II studies of new therapeutic agents by subtype response (adapted from reference 13) .
[0131]
[0132] Improved PTCL diagnosis requires new approaches: We have performed extensive gene expression profiling (GEP) studies and defined robust molecular signatures that classify the major subtypes of PTCL 10-12 We were also able to reclassify many PTCL-NOS into distinct entities (15% AITL, 10% ALCL, and 9% γδ-PTCL). Importantly, we also identified two new major subgroups within PTCL-NOS that showed enrichment of distinct signaling pathways and different prognoses. Furthermore, we identified an expression signature that predicts survival in AITL patients. 10,11,18 .
[0133] Among the several platforms tested by LLMPP for FFPE samples, Affymetrix 19 、NanoString 20,21 and quantitative nuclease protection assay (qNPA; HTG Molecular Diagnostics, Inc.) 22 The FDA-approved NanoStringnCounter system is superior in sample handling, does not require enzymatic reactions, and directly quantifies mRNA expression, while also providing high accuracy and high sensitivity. 23 Of note, LLMPP established a Lymph2Cx expression-based assay for DLBCL classification on the nCounter platform using a similar approach as described herein. 20,21 The assay was subsequently exclusively licensed to Nano String and is currently being developed in collaboration with Celgene (with up to $45 million in upfront payments, development and regulatory milestones, and commercial payments to Nano String) 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).
[0134] Innovation
[0135] This approach will bring practice changes to the diagnosis, prognosis, and management of PTCL by creating a high-content, quantitative, reproducible molecular assay that will provide greater diagnostic precision and accuracy for currently defined PTCL subtypes as well as new, previously unrecognized entities.
[0136] We propose to translate 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 subsequently as IVD kits for broad clinical application.
[0137] Because new agents in clinical trials show differences in response rates in different PTCL subtypes (e.g., see Table 2), it is critical to evaluate molecular features in experimental trials using standardized and validated assays. The molecular assays developed here will facilitate the practice of precision medicine in PTCL, just as the Lymph2Cx assay has helped guide individualized treatment options for DLBCL. 20,23 .
[0138] In summary, we will develop and validate a molecular diagnostic / prognostic assay that reliably identifies clinically relevant PTCL subtypes in routinely available samples, thereby enabling optimal personalized treatment and facilitating trials of new agents through accurate patient stratification.
[0139] B. Preliminary Research
[0140] B.1. Sample cohort for assay development: We have access to more than 1,000 FFPE cases 4 and more than 300 archived fresh frozen (FF) PTCL tissues with pathological and clinical data 10-12 FFPE blocks corresponding to the FF samples used in our original GEP study will be used for the initial assay refinement and optimization performed herein. We will obtain additional retrospectively collected PTCL samples from multiple European and Asian collaborators as needed (see individual profiles / letters of support) for the initial validation studies performed as part of this Phase I proposal, as well as for subsequent Phase II studies. These tissue and data resources are unparalleled and, combined with cutting-edge technology and sophisticated bioinformatics analyses, will allow us to further optimize characterization and validate the PTCL assay for clinical application.
[0141] B.2. Robust diagnostic and prognostic signatures for PTCL: During the development of the academic version of the PTCL assay, Dr. Chan and Dr. Iqbal and extended LLMPP researchers used the HG U133 plus 2 (U133) microarray (Affymetrix) 11,12Gene expression profiling (GEP) was performed on 160 fresh-frozen PTCL tissue biopsies. Subsequently, the researchers used a larger series of cases (n = 372) 10 To refine these findings. Figure 1 Distinct gene expression profiles of the major classes of PTCL were shown. This work clearly demonstrates: (a) the successful performance of a diagnostic algorithm based on GEP analysis of fresh frozen tissue from several institutions using standardized operating procedures, (b) a stepwise binary model algorithm can be used for subgroup classification in complex entities, and (c) the number of genes in the subgroup classifier can be significantly reduced (e.g., from >200 to 22 genes for AITL) without significantly compromising diagnostic accuracy 10 , and (d) different oncogenic pathways including potential therapeutic targets can be identified in these various molecular PTCL entities.
[0142] Some additional notable findings include: (i) ALK(-) ALCL is a distinct entity (a finding that was later independently supported by genomic and mutational data) 24-27 , and 11% of PTCL-NOS can be reclassified into this entity, which is highly responsive to the anti-CD30 therapy brentuximab (Table 2). (ii) ENKTL can be subdivided into NK and γδ-T cell subtypes 10,12 , and approximately 9% of PTCL-NOS can be reclassified as γδ-PTCL 10 (subtype containing activating STAT5B mutations and sensitive to JAK inhibitors 28 ). Equally important, Aurora A or NOTCH1 inhibition was shown to induce significant cytotoxicity in NK lymphoma cell lines 12 (iii) AITL prognosis is highly dependent on the tumor microenvironment (i.e., B cell, dendritic cell, and monocyte content), and the quartile of patients with the most favorable characteristics had a 5-year overall survival (OS) rate of 55%, while the OS in the least favorable quartile was only 15% 10,11 .
[0143] B.3. Identification of different subgroups in PTCL-NOS: PTCL-NOS, the largest part of PTCL diagnosed by current histopathological methods 4 , can now be divided into two major molecular subtypes, characterized by high expression of the transcription factors GATA3 (about 33%) or TBX21 (about 49%) and many of their target genes 10 ( Figure 2 A). Cases with high expression of GATA3 showed poor clinical outcomes ( Figure 2 B) 10 .
[0144] B.4. Mutational analysis to cross-validate molecular classification. Our mutational analysis of more than 90 PTCL cases demonstrated the presence of only IDH2 in AITL classified or reclassified (from PTCL-NOS) by GEP. R172 mutation 29,30 , and CD28 and RHOA G17V Mutations also occur preferentially 31 In contrast, JAK1 and STAT3 mutations were more prevalent in ALK(-) ALCL. 26 , and STAT5B mutations are common in γδ-PTCL 28 These observations support the validity of molecular classification, and these mutations provide independent markers to evaluate the performance of molecular classifiers.
[0145] B.5. GEP signatures in FFPE tissues: There is concern that FFPE tissues may not yield RNA of sufficient quality for GEP assays. We have demonstrated that GEP signatures derived from FFPE tissues using expression arrays can robustly subclassify DLBCL 19 In addition, we evaluated 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 signatures in archived DLBCL FFPE samples. In addition, Dr. Morris, industry co-lead, and Dr. Brian Z. Ring, lead industry bioinformatician for the current proposal, used RNA-seq to perform GEP on triple-negative breast cancer FFPE samples to develop the commercially available Insight TNBCtype TM 101-gene assay that identifies six molecular TNBC subtypes, each with clinical relevance 34 .
[0146] B.6. Gene signatures can be refined into smaller groups: Dr. Chan, Dr. Iqbal and colleagues screened 345 genes from their previous array study as the most powerful predictors of PTCL diagnosis and prognosis 10-12 We are now in the process of further refining this gene list to build classifiers with a smaller number of genes. Figure 3 Computational modeling was presented, indicating that the error rate of distinguishing the two PTCL subtypes using a smaller panel of 15-20 genes did not change significantly.
[0147] B.7. NanoString Technologies Platform for Profiling FFPE Tissues: We have demonstrated that using the NanoString nCounter system, signatures derived from FF samples can be reduced to a 15-gene set on FFPE tissue for DLBCL stratification: the "Lymph2Cx" signature 20,23 ( Figure 4 A), and the assay can be performed in multiple laboratories with high reproducibility ( Figure 4 B). The Lymph2Cx assay was recently selected as the CDx for FDA evaluation in the current Phase III ROBUST trial (Celgene; ClinicalTrials.gov ID: NCT02285062), which is examining the effect of R-CHOP ± lenalidomide in ABC-DLBCL. The Lymph2Cx developed by Dr. Chan, Dr. Iqbal, and their LLMPP colleagues will be replicated for the assay presented 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 corporate partner HealthChart LLC and our consulting molecular diagnostic company Insight Genetics, Inc., which will participate in the validation study conducted in Aim 2a, as described below.
[0148] C. Methods / Specific Objectives
[0149] Note on technical terminology / statistical considerations: Transcripts are referred to as "probe sets" on the Affymetrix U133 array and "code sets" on the NanoString platform. The "training set" consists of cases for which corresponding GEP data are available for FF samples analyzed on the U133 array; the "validation set" consists of non-overlapping cases against which the locking algorithm will be 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 WuConsulting, Inc., a biostatistics company with extensive practice in the pharmaceutical and biotechnology fields since 1993 (www.wuconsulting.com; see letter).
[0150] Objective 1: To identify a reduced set of transcripts and create a library of transcripts that can replicate existing “academic” diagnostic and prognostic features of PTCL. New Model of Signs .
[0151] Overview. The primary goal of this Phase I STTR proposal is to transition an academic PTCL diagnostic and prognostic GEP signature developed using fresh frozen samples analyzed with expression arrays to an optimized, locked signature suitable for use with FFPE samples analyzed with the Nanostring nCounter platform. Current Signatures 10larger than necessary and would be impractical and uneconomical for clinical assay development. Therefore, our initial study (Aim 1) will reduce the diagnostic set to as few genes as possible using the same linear model approach that was successfully used to develop the Lymph2Cx assay for DLBCL. 20 In addition, however, we will explore complementary approaches if needed (e.g., the centroid-based models we used previously). 34 , to help ensure successful creation of a robust model using a minimum number of genes.
[0152] Deliverable: Create a reduced gene set for the derivation of an optimized, commercially applicable PTCL classification model.
[0153] Objective 1a: Identify a reduced set of transcripts required to optimally distinguish each PTCL entity .
[0154] Pre-analytical considerations for FFPE samples : We have previously compared the performance of 667 diagnostic and prognostic genes for DLBCL identified using the U133 array with data obtained from analysis of FFPE tissues from the same cases (n=88) using the NanoString code set. Figure 5 A shows the correlation between the codeset and probesets in these transcripts. Most genes correlated well, with a median correlation of 0.75, although some performed poorly. We specifically examined the correlation of 40 genes present in both this DLBCL gene set and our academic PTCL signature and observed a correlation of 0.76 ( Figure 5 B). We expected that similar correlations would apply to the remaining genes in our PTCL signature, and considering that only about 15-20 genes were sufficient to distinguish each subtype in the binary model ( Figure 3 ), we should start from our microarray studies 10-12 Sufficient numbers of transcripts were obtained from the 345 genes identified in the study to model a significant reduction in gene number.
[0155] Calculation of test force: The number of matched FF and FFPE PTCL cases required for Aim 1a was relatively small because these studies sought only to determine the performance of the codeset in RNA from FFPE samples on the nCounter platform compared to data from corresponding FF samples previously characterized using arrays. We anticipated that 60 samples would be sufficient to constrain the correlation of well-performing transcripts with a correlation of 0.80 between sample and platform type to between 0.68 and 0.88 with 95% probability, and also to constrain the correlation of poorly performing transcripts with a correlation of 0.4 to between 0.155 and 0.60, thereby allowing easy discrimination between well- and poorly performing transcripts. This sample size would also enable appropriate scaling and additive adjustment such that for a typical sample, the error due to inappropriate adjustment would be less than 1 / 5 of the error due to noise between modeling of FF and FFPE samples.
[0156] Aim 1b: Refine diagnostic and prognostic features that can be applied to FFPE tissue to differentiate PTCL entities .
[0157] Methods to Reproduce Gene Signature Predictors into FFPE Tissue / NanoString Assays Our original FFRNA PTCL signature consists of a model score consisting of a linear combination of expression values with individual weights based on the discriminatory power of each gene and a set of cut-off points that separate samples into PTCL molecular subtypes. 10-12 . We will generate a "training set" using FF PTCL samples analyzed on the U133 array and matched FFPE tissues analyzed on the Nano String platform. This training set will accurately estimate the correlation between any given Affymetrix probe set used in a FF tissue-based model and the corresponding NanoString code set on the FFPE sample. This analysis will indicate the extent to which changes in the platform introduce noise into expression estimates, and therefore indicate the extent to which the weights of specific genes should be reduced to indicate a reduction or complete elimination of predictive power in the model. Second, once the weighted scores are generated, the training set will be used to estimate changes in the center and dynamic range of the final model scores to determine appropriate cutoff points. This should be easily achieved by estimating the mean and SD of the model on the FFPE sample / NanoString assay and comparing the results to the mean and SD of the FF sample / array model of the matched sample. To achieve these goals, the required training set is estimated to be 120 cases ( Figure 6 ).
[0158] Development of algorithms for PTCL subtype classification and prognosis prediction : Our team has previously targeted B-cell lymphoma 20,23,35-40 and PTCL 10-12The detailed methods for subtype classification are described. In brief, for the current study, a series of 4 binary predictors will be used to distinguish four separate subtypes from PTCL-NOS, namely AITL, ALCL, ATLL, and ENKTL. Based on the results of these predictors, we classified the samples into one of 6 possible categories (AITL, ALCL, ATLL, ENKTL, PTCL-NOS, or undetermined) (Table 3). "Undetermined" samples have characteristics of two or more clearly defined PTCL subtypes; based on our previous experience 10-12 , undetermined cases should be a very small fraction of the sample (1%-2%). PTCL-NOS cases were then divided into high GATA3 subgroups and high TBX21 subgroups, and ENKTL was also subdivided into NK and γ / δ T cell lineages. AITL prognostic subgroups as previously defined 10,11 Transformation will be performed as described for diagnostic features. We will identify well-performing code sets and adjust the weights of each code set to best reproduce the model scores, thereby stratifying patients into subgroups that show the greatest survival differences. To ensure that prognostic models are robust, we will perform locking and validation only if they can demonstrate differentiation in survival with a one-sided p value of <0.05.
[0159] Table 3: Stepwise binary classification scheme for subtype analysis .
[0160]
[0161]
[0162] - Sample size considerations: At this stage of the project, we propose to collect additional new samples to bring the total number of paired FF and FFPE samples to 120 cases (20 AITL, 10 ALK(+)ALCL, 10 ALK(-)ALCL, 20 ATLL, 20 ENKTL, 20 GATA3PTCL, and 20 TBX21PTCL). With this number of patients, we should be able to estimate the noise introduced by variation in the platform for well-measured genes (r>0.8) to within 12% of its true value with 95% confidence. Estimates of the scaling and mean shift required to normalize our FFPE model scores to their frozen counterparts with 95% confidence indicate that the difference between the estimated model thresholds and their optimized values is less than 20% of the expected noise in a single sample due to platform variation. Furthermore, 20 samples per subtype for each binary distinction would also allow us to normalize our model scores to their frozen counterparts, such that the difference between the estimated model cut-points and their optimized values would be only 12% of the individual sample variability at 95% confidence. Simulations suggest that, at 95% confidence, this degree of inaccuracy would result in at most 4% of samples being predicted differently than would be predicted using a similar NanoString model based on an unlimited number of samples.
[0163] Potential problems and alternatives: Reclassification Discordant cases will be re-examined to determine if morphological differences between FF and FFPE tissues could explain the discordance. 10,20,23 The molecular classification will be cross-validated by analyzing somatic mutations and / or DNA copy number changes that are unique or enriched in specific subtypes. 10-12,24-31 . Clinical outcomes will be used as validation of the prognostic signature. Classification Model There are many other classification models that could be considered (e.g., centroid-based models, which we have previously used successfully to develop similar expression-based commercial diagnostic assays). 34,41 ); however, our PTCL signature has been tested and validated in FF tissue using a linear model, and we expect the same model to be robust in FFPE tissue, as demonstrated in other studies 18,42 . Number of genes in the model : The current "academic" signature contains 345 genes. Our preliminary bioinformatics studies indicate that it is easy to reduce the transcripts to a total of 50-60 or less for the entire signature. It is worth noting that the NanoString platform can measure up to 800 transcripts in a single test.
[0164] Specific Aim 2: Confirm the refined diagnostic and prognostic PTCL gene signature set for analysis of FFPE samples Applicability on the NanoString nCounter Platform .
[0165] Overview: The primary goal of Aim 2 will be 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 exchange of frozen material run on the Affymetrix platform ( Figure 7 ), demonstrating the ability to develop SOPs and obtain reproducible results across testing sites; here we will conduct similar studies to confirm the reproducibility of our PTCL characterization using FFPE samples on the NanoString platform to advance the commercialization of this assay.
[0166] Deliverables: Target an optimized PTCL signature based on training cohort and interlaboratory study results (Aim 2a); Validation cohort results confirm suitability of the targeted signature for progression to Phase II development (Aim 2b).
[0167] Objective 2a: Create a targeted diagnostic and prognostic PTCL gene signature and evaluate interlaboratory reproducibility .
[0168] We will evaluate our refined diagnostic and prognostic features using a series of non-overlapping PTCL cases (approximately 100 cases) with the same subtype distribution as the training set. Continuous rolls from representative FFPE blocks will be blindly evaluated using a working "locked" standard operating procedure (SOP) created based on our training set study (Goal 1b). The assay will be performed in a commercial CLIA laboratory (Insight Molecular Labs) (Dr. Steve Morris and Dave Hout - www.insightgenetics.com; see letter) in conjunction with the diagnostic laboratories of UNMC (Dr. Timothy Greiner), City of Hope (Dr. Raju Pillai), and HealthChart.
[0169] Statistical considerations : We will reanalyze a portion of those samples that were part of the initial study conducted by LLMPP 10-12 (but separate from those cases studied in Aim 1). The relatively small number of samples (approximately 100) will not allow us to unequivocally confirm every single performance parameter of the FFPE model at the desired statistical metric (due to inherent limitations of Phase I funding). Instead, we will attempt to mimic previous LLMPP studies that involved optimizing and transitioning ABC vs GCB DLBCL subtyping characterization from FF / array assays to FFPE / nCounter platforms. 20,21, where modest numbers were used for training and validation sets (51 and 68 cases, respectively) to confirm the overall fidelity of subtyping. This approach has been shown to be successful in subsequent studies of the Lymph2Cx assay: 96% concordance in 49 repeat biopsies, 100% concordance in 83 FFPE samples, and importantly, no misclassification (ABC-DLBCL to GCB-DLBCL or vice versa) was observed when compared to the gold standard FF / array assay. 21 To assess the predictive power of the FFPE tissue model in PTCL, we will evaluate its performance using leave-one-out cross-validation, which involves selecting all weights and thresholds in the model as part of the cross-validation step.
[0170] Analysis plan :FF / microarray diagnostics are considered the gold standard, and accuracy, precision, sensitivity, and specificity will be based on comparisons with array results. The results will be provided to Dr. George Wright (NCI BiometricResearch Program) and Dr. Brian Ring (HealthChart LLC), who will independently determine the classification of these cases. For reproducibility analysis, we will compare the final predicted classification from three laboratory sites and the consistency between the model score and its distance from the cutoff point, thereby providing a more accurate estimate of the reproducibility of the predictions. Based on our previous experience, we found that the model on NanoString is highly reproducible between sites, typically with >0.99 20,21 Therefore, we set a high bar and estimated that 100 samples should be sufficient to achieve this metric with 99% power for each model.
[0171] Objective 2b: Validation of “lock-in” diagnostic / prognostic signatures for phase II assay transition .
[0172] The goal of Aim 2b is to demonstrate that the locked assay can be readily adapted for clinical use across multiple institutional laboratories using archival material and, if needed, to provide final troubleshooting of the assay to ensure its readiness for Phase II development. A cohort of archived FFPE PTCL samples (approximately 100 to 120 cases, including 20 cases of each diagnostic and prognostic subtype) will be selected from laboratories around the world (see letter from cooperating investigators). Because these cases will not undergo Affymetrix expression array studies, direct comparison will not be feasible and the focus will be more on operational aspects of reproducibility, reliability, and problem solving (if needed) between multiple laboratories using the locked version and SOP. Additionally, emphasis will be placed on quality control measures between different laboratories, which will be documented in order to refine the SOP for future routine testing.
[0173] Example 2
[0174] A plan for obtaining suitable samples for assay development and optimization :
[0175] We have learned from our previous GEP research 10 Approximately 70 PTCL cases were collected in the preselection. Preselection was a critical step because representative tumor sections needed to be cut, the presence of representative and sufficient tissue needed to be confirmed, and the diagnosis of the PTCL cases needed to be validated. These cases had undergone pathological review organized by Dr. Chan and Dr. Amador. Included were cases with consistent initial diagnosis and molecular characterization. Among them, we had consensus diagnosis for the following: AITL (n=18); ALCL (n=15); ENKTL (n=4); PTCL-NOS (n=30). Cases were subjected to additional immunostaining currently used for characterization or subclassification, including staining for GATA3 and TBX21, which helps distinguish subgroups ( Figure 8 ). We have developed an immunohistochemistry-based algorithm that subdivides PTCL-NOS into GATA3 and TBX21 subgroups. We will continue to refine this IHC algorithm and cross-validate cases with genetically characterized diagnoses as we develop it.
[0176] Optimization of DNA / RNA Isolation from Formalin-Fixed Paraffin-Embedded (FFPE) Tissue :
[0177] Since tissue samples are usually prepared by fixation with formalin (formaldehyde), this produces biomolecular cross-links and adducts and reduces the signal from molecular analyses involving hybridization procedures. Typically, DNA / RNA isolation from FFPTE tissue using commercial kits (e.g., Qiagen AllPrep DNA / RNA FFPE Kit) involves heating at high temperatures (>80°C) for several hours in Tris buffer, which damages both RNA and DNA. In previous studies, 26We have been using these kits routinely to isolate DNA / RNA for nCounter assays with good success in our retrospective series. Because we expect many of the FFPE cases from our retrospective series to be older than 10 years, and our analytical assay requires that at least 40% of the total RNA content in the reaction mixture for the nCounter assay be greater than 200 bp in length, we therefore tested a water-soluble bifunctional catalyst (anthranilate and phosphonobenzoate) 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 effectiveness of this protocol, we used 10 μm coils (2x) of representative FFPE blocks from 1988 to 2017 (optional cases) and compared with the Qiagen FFPE isolation kit. In these cases, total RNA yields from these kits were in a similar range, but older blocks showed relatively lower yields ( Fig. 9 ).
[0178] Analysis of the size distribution of RNA molecules using the TapeStation-2200 showed that the Strom kit produced relatively better (larger size) RNA in older cases (>10 years) compared with the Qiagen kit, with more modest improvements in cases from 2014 FFPE blocks. Fig.10 A and Fig.10 B). Therefore, it is imperative to perform nCounter testing initially using cases after 2010. In conclusion, for recent cases (<5 years), the isolation procedure from the Qiagen kit may be a feasible approach, whereas if optimization is needed for cases >10 years old, especially in rare entities such as ENKTL, the Storm kit isolation may be an effective alternative for isolating RNA / DNA.
[0179] Convert GEP data from Affymetrix probe sets to nCounter code groups from NanoString, For multi-analyte determination :
[0180] Our primary goal was to translate GEP signatures from fresh frozen tissue to FFPE tissue using NanoString technology. Since code sets are expensive and bioinformatically challenging to design, we refined our gene expression signatures from our previous work and added more genes (at least 10% more in number) to each diagnostic signature so that we had a sufficient number of genes available for signature refinement in cases where they did not work optimally in FFPE RNA without altering the accuracy of the diagnostic signature ( Fig.11 ). These code sets are created using nDesign TM The gene lists were designed by Gateway and kept confidential by NanoString’s bioinformatician staff under UNeMEd protocols. The design includes a total of 416 codebases, including housekeeping genes (n=50), AITL diagnostic (n=50), prognostic (n=49), and PTCL-NOS subgroup (n=50) code sets, as well as code sets for classifying ALCL (including ALK+ALCL vs ALK-ALCL, n=100) to distinguish between NK and γδ subgroups (n=50). We have received these code sets for these genes and will begin assaying FFPE cases within a few weeks. We anticipate that some RNAs will not be suitable for study after fixation and will need to be excluded, especially with smaller signals. Therefore, we will perform an initial analysis of 10 fresh frozen RNAs and corresponding FFPETs using the nCounter platform to compare signal intensity and select only transcripts with correlations ≥0.8 for further evaluation.
[0181] The subject matter of the present disclosure is further illustrated by further explanation of the features, benefits and advantages of the present invention and by specific but non-limiting examples set forth in the appendix, which is attached hereto and incorporated herein by reference.
[0182] All publications, patents, and patent applications mentioned herein and in the appendices are incorporated herein 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.
[0183] It should be understood that various details of the presently disclosed subject matter can be changed without departing from the scope of the subject matter disclosed herein.In addition, the foregoing description is for the purpose of illustration only, and not for the purpose of limitation.
[0184] References
[0185] 1.Rudiger T,Weisenburger DD,Anderson JR,et al.Peripheral T-celllymphoma(excluding anaplastic large-cell 1ymphoma):results from the Non-Hodgkin′s Lymphoma Classification Project.Annals of Oncology:Official journalof the European Society for Medical Oncology / ESMO.2002;13(1):140-149.
[0186] 2.Bellei M,Chiattone CS,Luminari S,et al.T-cell lymphomas in SouthAmerica and Europe.Rev Bras Hematol Hemoter.2012;34(1):42-47.
[0187] 3.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification:Pathologyand Genetics of Tumors of Haematopoietic and Lymphoid Tissues.WHO;4.Lyon,France:IARC Press;2008.
[0188] 4.Vose J,Armitage J,Weisenburger D,et al.International Peripheral T-cell and Natural Killer / T-cell Lymphoma Study:pathology findings and clinicaloutcomes.J Clin Oncol.2008;26(25):4124-4130.
[0189] 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.
[0190] 6.Parrilla Castellar ER,Jaffe ES,Said JW,et al.ALK-negativeanaplastic large cell lymphoma is a genetically heterogeneous disease withwidely disparate clinical outcomes.Blood.2014;124(9):1473-1480.
[0191] 7.Escalon MP,Liu NS,Yang Y,et al.Prognostic factors and treatment ofpatients with T-cell non-Hodgkin lymphoma:the M.D.Anderson Cancer Centerexperience.Cancer.2005;103(10):2091-2098.
[0192] 8.Crozier JA,Sher T,Yang D,et al.Persistent disparities amongpatients with T-cell Non-Hodgkin Lymphomas and B-cell Diffuse Large CellLymphomas over 40 years:A SEER database review.Clinical Lymphoma Myeloma andLeukemia.2015;15(10):578-85.
[0193] 9.Xu B,Liu P.No survival improvement for patients withangioimmunoblastic T-cell lymphoma over the past two decades:a population-based study of 1207 cases.PLoS One.2014;9(3):e92585.
[0194] 10.Iqbal J,Wright G,Wang C,et al.Gene expression signatures delineatebiological and prognostic subgroups in peripheral T-cell lymphoma.Blood.2014;123(19):2915-2923.
[0195] 11.Iqbal J,Weisenburger DD,Greiner TC,et al.Molecular signatures toimprove diagnosis in peripheral T-cell lymphoma and prognostication inangioimmunoblastic T-cell lymphoma.Blood.2010;115(5):1026-1036.
[0196] 12.Iqbal J,Weisenburger DD,Chowdhury A,et al.Natural killer celllymphoma shares strikingly similar molecular features with a group of non-hepatosplenic gamma / delta T-cell lymphoma and is highly sensitive to a novelaurora kinase A inhibitor in vitro.Leukemia.2011;25(2):348-358.
[0197] 13.Moskowitz AJ,Lunning MA,Horwitz SM.How I treat the peripheral T-cell lymphomas.Blood.2014;123(17):2636-2644.
[0198] 14.Gambacorti-Passerini C,Messa C,Pogliani EM.Crizotinib inanaplastic large-cell lymphoma.N Engl J Med.2011;364(8):775-776.
[0199] 15.Ishida T,Joh T,Uike N,et al.Defucosylated anti-CCR4 monoclonalantibody(KW-0761)for relapsed adult T-cell leukemia-lymphoma:a multicenterphase II study.J Clin Oncol.2012;30(8):837-842.
[0200] 16.O′Connor OA,Bhagat G,Ganapathi K,et al.Changing the paradigms oftreatment in peripheral T-cell lymphoma:from biology to clinicalpractice.Clin Cancer Res.2014;20(20):5240-5254.
[0201] 17.Coiffier B,Federico M,Caballero D,et al.Therapeutic options inrelapsed or refractory peripheral T-cell lymphoma.Cancer Treat Rev.2014;40(9):1080-1088.
[0202] 18.Iqbal J,Wilcox R,Naushad H,et al.Genomic signatures in T-celllymphoma:How can these improve precision in diagnosis and inform prognosis?Blood Rev.2016;30(2):89-100.
[0203] 19.Linton K,Howarth C,Wappett M,et al.Microarray gene expressionanalysis of fixed archival tissue permits molecular classification andidentification of potential therapeutic targets in diffuse large B-celllymphoma.J Mol Diagn.2012;14(3):223-232.
[0204] 20.Scott DW,Wright GW,Williams PM,et al.Determining cell-of-originsubtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue.Blood.2014;123(8):1214-1217.
[0205] 21.Scott DW,Mottok A,Ennishi D,et al.Prognostic significance ofdiffuse large B-cell lymphoma cell of origin determined by digital geneexpression in formalin-fixed paraffin-embedded tissue biopsies.J ClinOncol.2015;33(26):2848-2856.
[0206] 22.Rimsza LM,Leblanc ML,Unger JM,et al.Gene expression predictsoverall survival in paraffin-embedded tissues of diffuse large B-celllymphoma treated with R-CHOP.Blood.2008;112(8):3425-3433.
[0207] 23.Rimsza LM.Diffuse large B-cell lymphoma classification tied upnicely with a″String″.Clin Cancer Res.2015;21(10):2204-2206.
[0208] 24.Feldman AL,Law M,Remstein ED,et al.Recurrent translocationsinvolving the IRF4 oncogene locus in peripheral T-celllymphomas.Leukemia.2009;23(3):574-580.
[0209] 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 lymphomasby massively parallel genomic sequencing.Blood.2011;117(3):915-919.
[0210] 26.Crescenzo R,Abate F,Lasorsa E,et al.Convergent mutations andkinase fusions lead to oncogenic STAT3 activation in anaplastic large celllymphoma.Cancer Cell.2015;27(4):516-532.
[0211] 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.
[0212] 28.Kucuk C,Jiang B,Hu X,et al.Activating mutations of STAT5B andSTAT3 in lymphomas derived from gammadelta-T or NK cells.Nat Commun.2015:6:6025.
[0213] 29.Cairns RA,Iqbal J,Lemonnier F,et al.IDH2 mutations are frequent inangioimmunoblastic T-cell lymphoma.Blood.2012:119(8):1901-1903.
[0214] 30.Wang C,McKeithan TW,Gong Q,et al.IDH2 R172 mutations define a uniquesubgroup of patients with angioimmunoblastic T-cell lymphoma.Blood.2015:126(15):1741-1752.
[0215] 31.Rohr J,Guo S,Huo J,et al.Recurrent activating mutations of CD28 peripheral T-cell lymphomas.Leukemia.2016:30(5):1062-1070.
[0216] 32.Roberts RA,Sabalos CM,LeBlanc ML,et al.Quantitative nucleaseprotection assay in paraffin-embedded tissue replicates prognostic microarraygene expression in diffuse large-B-cell lymphoma.Lab Invest.2007;87(10):979-997.
[0217] 33.Geiss GK,Bumgarner RE,Birditt B,et al.Direct multiplexedmeasurement of gene expression with color-coded probe pairs.NatBiotechnol.2008;26(3):317-325.
[0218] 34.Ring BZ,Hout DR,Morris SW,et al.Generation of an algorithm basedon minimal gene sets to clinically subtype triple negative breast cancerpatients.BMC Cancer.2016 Feb 23;16:143.
[0219] 35.Alizadeh AA,Eisen MB,Davis RE,et al.Distinct types of diffuselarge B-cell lymphoma identified by gene expression profiling.Nature.2000;403(6769):503-511.
[0220] 36.Dave SS,Fu K,Wright GW,et al.Molecular diagnosis of Burkitt′slymphoma.N Engl J Med.2006;354(23):2431-2442.
[0221] 37.Dave SS,Wright G,Tan B,et al.Prediction of survival in follicularlymphoma based on molecular features of tumor-infiltrating immune cells.NEngl J Med.2004;351(21):2159-2169.
[0222] 38.Rosenwald A,Wright G,Chan WC,et al.The use of molecular profilingto predict survival after chemotherapy for diffuse large-B-cell lymphoma.NEngl J Med.2002;346(25):1937-1947.
[0223] 39.Rosenwald A,Wright G,Leroy K,et al.Molecular diagnosis of primarymediastinal B cell lymphoma identifies a clinically favorable subgroup ofdiffuse large B cell lymphoma related to Hodgkin lymphoma.J Exp Med.2003;198(6):851-862.
[0224] 40.Rosenwald A,Wright G,Wiestner A,et al.The proliferation geneexpression signature is a quantitative integrator of oncogenic events thatpredicts survival in mantle cell lymphoma.Cancer Cell.Feb2003;3(2):185-197.
[0225] 41. Tibshirani R, Hastie T, Narasimhan B, Chu G. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proc Natl Acad Sci U SA. 2002; 99(10): 6567-6572.
[0226] 42. Iqbal J, Naushad H, Bi C, et al. Genomic signatures in B-celllymphoma: How can these improve precision in diagnosis and inform prognosis? Blood Rev.2016;30(2):73-88.
[0227] Example 3
[0228] Gene expression signatures for accurate diagnosis of peripheral T-cell lymphoma entities in routine clinical practice
[0229] summary
[0230] Purpose
[0231] Peripheral T-cell lymphomas (PTCLs) comprise heterogeneous clinicopathological entities with many 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 improve the accuracy and precision of PTCL diagnosis by integrating gene expression-based subclassification using formalin-fixed paraffin-embedded (FFPE) tissues.
[0232] method
[0233] We assembled a well-characterized PTCL training cohort (n=105) with gene expression profiling (GEP) data to derive diagnostic signatures using fresh frozen (FF) tissue on the HG-U133plus2 platform (Affymetrix, Inc), followed by validation using matched FFPE tissue on the digital GEP platform (NanoString, Inc). Statistical filtering methods were applied to refine the transcriptomic signatures, which were then validated in another PTCL cohort (n=140) by rigorous pathological review and ancillary assays.
[0234] result
[0235] In the training cohort, the improved transcriptome classifier in FFPE tissue showed high sensitivity (>80%), specificity (>95%) and accuracy (>94%) for PTCL subclassification compared with the FF-derived diagnostic model, and showed high reproducibility between 3 independent laboratories. In the validation cohort, in 85% (n=119) of the cases, the transcriptional classifier matched the pathological diagnosis provided by 3 hematopathologists, showed "critical correlation" with molecular features in 6% (n=8) of the cases, and showed inconsistency in 8% (n=11) of the cases. The classifier improved the pathological diagnosis of 2 cases and was verified by clinical results. Among the 11 cases with inconsistency, 4 cases had molecular classification, which could provide an improvement over pathological diagnosis based on overall transcriptomic features and morphological features. Molecular subclassification provides a comprehensive molecular characterization of PTCL subtypes, including viral etiology factors and translocation partners.
[0236] in conclusion
[0237] We developed a novel transcriptomic approach for PTCL subclassification that facilitates translation into clinical practice with greater precision and uniformity compared with conventional pathological diagnosis.
[0238] Context Summary
[0239] Key Purpose
[0240] 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 suitable for routine clinical practice.
[0241] The knowledge formed
[0242] Using digital quantification of transcripts, we have defined a robust transcriptomic signature that distinguishes common PTCL subtypes according to the WHO classification, including two new biologic and prognostic subgroups within PTCL-NOS. We improved the classification algorithm and normalized the assay to achieve a robust diagnostic test validated in an independent PTCL cohort. The assay was reproducible across institutions and showed high classification accuracy.
[0243] Relevance
[0244] Using mRNA from FFPE tissue, digital transcriptomic analysis has generated a robust molecular classification of PTCL, which facilitates its translation into clinical testing, providing a more precise and uniform diagnosis than conventional pathology. This molecular assay will also help better stratify cases for research and clinical trials.
[0245] introduce
[0246] Peripheral T-cell lymphoma (PTCL) accounts for approximately 10% to 15% of non-Hodgkin lymphoma (NHL) 1 Even for hematopathologists, diagnosis presents many challenges 2-4 The World Health Organization (WHO) classification identifies more than 25 different subtypes of PTCL, of which angioimmunoblastic T-cell lymphoma (AITL), anaplastic large cell lymphoma (ALCL), adult T-cell leukemia / lymphoma (ATLL), and extranodal NK / T-cell lymphoma of nasal type (ENKTCL) are the most common entities with geographical variation. 5 However, 30% of PTCLs cannot be classified into any specific entity in the WHO classification and are classified as PTCL-NOS. 6,7 Tumor-defining abnormalities, such as translocations involving the ALK gene in ALK-positive ALCL (ALK+ALCL) 8 , Human T-lymphotropic virus (HTLV-1) infection in ATLL 9 , EBV-positive in ENKTCL and IDH2 in AITL R172 Mutations are usually uncommon in PTCL. PTCL generally has a poor prognosis with current therapies 2 , and more intensive treatment options have not been shown to be superior 10 However, new targeted therapies are currently being tested with some remarkable results. 11,12 .
[0247] Compared with B-cell lymphomas, subclassification of PTCL is more challenging due to the complexity of T-cell biology, with numerous functional subsets and functional plasticity. Gene expression profiling (GEP) has helped delineate new biological subtypes and identify oncogenic pathways in several B-NHLs. 13-18 A similar approach in PTCL generated a robust molecular classifier for common PTCL entities and identified two biological and prognostic subgroups within PTCL-NOS (PTCL-GATA3 and PTCL-TBX21) 19-21 However, these studies were performed on fresh-frozen (FF) samples with extensive transcriptome arrays, thus limiting their application to routine clinical practice.22,23 Formalin-fixed paraffin-embedded (FFPE) tissue samples are widely used for routine diagnosis, but formalin fixation can lead to fragmentation, cross-linking, and chemical modification of RNA and DNA. 24 Therefore, effective translation of our highly accurate RNA-based PTCL diagnostic signature to FFPE tissue is challenging but essential for implementing an assay with broad clinical application. 25 Using digital quantification of RNA, such as in diffuse large B-cell lymphoma 26-28 We integrated PTCL diagnostic signatures from fresh frozen RNA into a single technology platform for accurate diagnosis among major PTCL entities 19-21 We used a "trained" PTCL cohort (n=105) and validated it in an "independent" PTCL cohort (n=140) and in the largest number of well-characterized cases studied on a single platform for PTCL subclassification. Several viral transcripts were added to improve diagnostic accuracy, and we report a diagnostic algorithm with high sensitivity, specificity, and accuracy in distinguishing PTCL entities, including novel molecular biological subtypes of PTCL-GATA3 and PTCL-TBX21. 20,21 .
[0248] Materials and methods
[0249] Patient Information
[0250] We included 249 diagnosed PTCL cases from multiple institutions and, after excluding 4 cases with poor RNA quality, divided the remaining cases into previously generated GEPs with matched FF and FFPE samples. 19-21 The training cohort (105 cases) and the previously unanalyzed validation cohort (140 cases) (Tables 4 and 5 and Fig.12 A). The basic clinical and pathological characteristics of these cases are shown in Table 6. The inclusion and exclusion criteria of PTCL cases are described in detail in the Supplementary section.
[0251] Table 4: Performance of the PTCL diagnostic algorithm in the training cohort .
[0252]
[0253]
[0254] Table 5: Performance of the PTCL diagnostic algorithm in the validation cohort .
[0255]
[0256] Table 6: Characteristics of cases in the training and validation cohorts and excluded cases.
[0257]
[0258]
[0259]
[0260]
[0261]
[0262]
[0263]
[0264]
[0265]
[0266]
[0267]
[0268]
[0269]
[0270]
[0271]
[0272]
[0273]
[0274]
[0275] AITL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ATLL, adult T-cell lymphoma / leukemia; 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.
[0276] Histopathological / immunomorphological features of the PTCL cohort
[0277] According to the current WHO classification 6PTCL cases were centrally reviewed and diagnosed. The validation cohort was thoroughly re-evaluated by three hematopathologists (CA, DDW, WCC), and a comprehensive immunostaining panel and TCRg gene rearrangement analysis were used when necessary. When there was agreement on the diagnosis, a consensus diagnosis was reached.
[0278] RNA extraction and digital gene expression for PTCL subclassification
[0279] About RNA extraction protocol, quality control measures, Details of the assay, data processing, cross-validation, and reproducibility assessment are provided in the Supplementary Sections and Fig.12 A. Data analysis and normalization are designed to process samples individually rather than in batches so that the protocol will be adaptable to process patient samples as needed. Class prediction is based on a series of binary comparisons that are combined for the final classification call for each sample ( Fig.12 A and Fig.17 ). Detailed materials and methods are provided in this article.
[0280] Survival analysis
[0281] Survival data were analyzed using the survival package, survminer package, and coin package in R and are explained in detail in this article.
[0282] result
[0283] Patient characteristics in the "training" and "validation" cohorts
[0284] Use the previous 19-21 A training cohort (n = 105) of GEP data was generated from matched FFPE tissues to select Classifier genes of the platform. A "validation" cohort of PTCLs without fresh frozen transcriptomic data (n=140) was evaluated by 3 hematopathologists using the current WHO diagnostic criteria. 6 A strict diagnosis was performed ( Fig.12 A and Tables 4 and 5). It is worth noting that the FH Marker evaluation of PTCL-NOS cases to exclude nodular PTCL-T FH cases, followed by the use of a recently published IHC algorithm 29 Cases were subdivided into PTCL-GATA3 and PTCL-TBX21.
[0285] 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 ( Fig.12C). For patients with survival data, the median follow-up for survivors was 3.5 years (range, 0.01 to 24 years). ALK+ALCL cases showed superior outcomes compared with other entities, which is consistent with published studies ( Fig.12 DE) 2,21 .
[0286] Development of transcriptomic signatures for FFPE tissues
[0287] As shown in Supplementary S 1, FFPE tissue blocks were selected based on the presence of sufficient tumor tissue and the assessed RNA quality. Transcriptomic signatures evaluated between platforms revealed a high degree of correlation (correlation coefficient r>0.4) for most (~60%) signature-specific genes ( Fig.19 AB). We performed recursive filtering analysis to exclude transcripts with correlation coefficients <0.4 using Pearson correlation (except for 3 transcripts), resulting in 11 to 20 diagnostic transcripts / PTCL subtype and 16 housekeeping genes ( Fig.19 CD). The use of these well-represented transcripts did not affect the fresh frozen RNA (HG-U133plus 2) or matched FFPE RNA ( Inc) in classification. Molecular subclassification using FFPE samples was highly comparable to the FF gold standard, with an error rate of <5% between various PTCL subtypes.
[0288] Accuracy and interlaboratory reproducibility of PTCL classification using improved signatures
[0289] Data on the HG-U133plus2 platform 19-21 The reduced transcript signatures preserved classification accuracy ( Fig.19 CD). Therefore, the reduced diagnostic transcripts were considered to be subsequently compared with those in the training set. The "gold standard" for comparison Fig.12 B, left). On the same platform, classification of the FFPE training cohort was reproduced in 90% (95 / 105) of cases (Tables 4, 5, and Fig.24 We observed a prognostic difference between PTCL-GATA3 and PTCL-TBX21 ( Fig.19 F), but the sample numbers were too small to reach statistical significance. As we transitioned from a larger to a smaller diagnostic transcript set, Seven cases from different PTCL subtypes were re-evaluated using the reduced transcript set versus the original larger transcript panel on the platform. They demonstrated consistent results, maintaining similar diagnostic accuracy ( Fig. 20 When the assay was performed at two additional CLIA sites to assess reproducibility, we observed high interobserver agreement for all 24 PTCL cases studied, with classifications identical to those at the original sites ( Fig. 20 ).
[0290] Improved diagnostic algorithms among different PTCL entities and validation cohorts
[0291] To evaluate the diagnostic performance, The molecular classification obtained by the platform was compared with the consensus pathological diagnosis in an independent validation cohort (n = 140). This cohort had similar clinical outcomes and PTCL subtype distribution as the training cohort ( Fig.12 A, C). Use The classification obtained by the platform was highly comparable to the diagnosis proposed by the pathologists, with an overall agreement of 91% in the validation cohort cases (127 / 140, 95% confidence interval (CI) 0.85-0.95), and improved the classification of challenging PTCL cases as shown below (Tables 4, 5, and Fig.25 ; Fig.12 B, right; Fig.21 ).
[0292] AITL : Mean expression of diagnostic features and pan-T- FH Gene expression signature (i.e., WHO defined T FH Six transcripts of markers 6 ) is significantly correlated ( Fig.13 AB). In the training group and the validation group, according to Cases molecularly classified as AITL showed immunomorphologic features commonly associated with AITL. Consistent with previous studies, high expression of CD20 was associated with better OS. 20,21,30 , which can be verified using immunohistochemistry ( Fig.13 C). AITL mutation profiles (ie, TET2, DNMT3A, RHOA) were observed in 89% of cases with available sequencing data. GI7V and IDH2 R172 )( Fig.13 D).
[0293] use The platform classified AITL as 83% consistent in the training cohort (20 / 24) and 74% consistent in the validation cohort (14 / 19), while the remaining cases (3 in the training cohort and 5 in the validation cohort) showed a “borderline model score” between AITL and PTCL-NOS (Tables 4 and 5 Fig.13 E; Fig.21 ). These cases missed the molecular diagnosis of AITL based on the “threshold or cut-off point”. On review, these cases were confirmed to have the classic immunomorphological features of AITL, and mutation analysis supported the diagnosis, such as Fig.13 As shown in D, the classic AITL mutation spectrum includes TET2, RHOA G17V and IDH2 R172 These cases showed that the expression level of AITL diagnostic features was relatively higher than that of other PTCLs and slightly lower than that of AITL, and T FH The expression of the feature is high. Fig.13 A, F). As expected, all three cases of follicular T-cell lymphoma were classified as AITL. Two cases of PTCL-NOS were molecularly classified as AITL. These cases showed T FH Focal positivity for the markers BCL6 and ICOS ( Fig.13 GI), but not meeting the nodal PTCL-T FH Although these two cases ( Fig.13 F) Significantly expressed AITL characteristic genes and T FH The mRNA signature was higher than in most other PTCL-NOS cases, but common gene mutations in AITL (TET2, IDH2 R172 , RHOA G17V , DNMT3A) are not present. These two cases may represent PTCL-T that was not classified according to the immunostaining performed. FH .
[0294] ALCL subtypes Initial analysis of ALCL and other PTCLs 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 improved signature. Fig.14 A). Validation group (Tables 4 and 5; Fig.14 B) shows 90% (18 / 20) concordance in ALK+ALCL and 75% (12 / 16) concordance in ALK-ALCL. Interestingly, two ALK-ALCLs were classified as ALK+ALCLs due to high ALK+ALCL signature, but with low ALK mRNA expression ( Fig.14C), and ALK protein expression was not detected by IHC. As expected, ALCL cases classified by ALK status showed that ALK-positive cases had better OS ( Fig.14 D). Of note, no 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 expression of the ALCL signature, although two ALK+ ALCLs had high ALK signatures ( Fig.14 C). The two ALK-ALCL cases with low ALCL features did not express CD30 mRNA, and the inability to classify these cases may be due to insufficient RNA quality or low tumor content ( Fig.14 E) These findings suggest that occasional cases with subdiagnostic features should be diagnosed with caution.
[0295] ATLL : ATLL molecular signature detected 100% of ATLL cases in the training cohort (7 / 7, 95% CI: 0.59-1) and 83% of ATLL cases in the validation cohort (10 / 12, 95% CI: 0.52-0.98) ( Fig.15 A). The 2 cases that showed discordance had minimal expression of features diagnostic of adult leukemia (ATLL), but both were confirmed to be positive for HTLV1 mRNA expression (i.e., HBZ measured by qPCR), although in one case, The measured HBZ expression was lower than that of other ATLL ( Fig.15 BC). Of the two HTLV1 transcripts (HZB and TAX1), HZB was consistently expressed at higher levels in ATLL cases compared with TAX and showed a positive correlation with ATLL features ( Fig.15 DE) and had a significant correlation with HBZ expression measured by qRT-PCR ( Fig.15 F). Interestingly, two PTCL-NOS cases from the validation cohort were molecularly classified as ATLL ( Fig.15 A). Re-evaluation of these cases using HZB-specific qPCR confirmed the expression of HBZ ( Fig.15 E), subsequent clinical chart review revealed serology positivity for HTLV1 and clinical presentation consistent with ATLL, which was unknown at the time of initial diagnosis. Morphologically, these cases consisted of CD4-positive monomorphic T-cell lymphomas that were misdiagnosed as PTCL-NOS in the absence of appropriate clinical history, blood work, and serologic testing. 7 ( Fig.15 B) Therefore, both cases were reclassified as ATLL.
[0296] ENKTCL The ENKTCL molecular classifier was able to identify 90% (9 / 10) of ENKTCL cases in the training cohort and 95% (21 / 22) of ENKTCL cases in the validation cohort (Tables 4 and 5, Fig.15 FG). A subset of these cases had elevated CD3γ and CD3δ expression compared with other cases ( Fig.15 H), and may be derived from the T lineage. Two cases were diagnosed as "primary EBV-positive nodular T / NK cell lymphoma" according to the current WHO classification. These cases are similar to ENKTCL, but the manifestations are mainly nodular, and both are classified as ENKTCL by molecular assays ( Fig. 22 One case molecularly classified as ENTKCL uniquely showed features of both ALCL and ENKTCL but with strong expression of EBV transcripts ( Fig.14 I).
[0297] Refinement and validation of two new PTCL-NOS subtypes (PTCL-GATA3 and PTCL-TBX21)
[0298] In the PTCL-NOS cases in the training cohort, we isolated two new molecular subgroups (i.e., PTCL-GATA3 or PTCL-TBX21) with 87% concordance and a reduced number of transcripts (n=19). In the remaining PTCL-NOS group in the validation cohort, we used The platform classified 40% (21 / 52) of cases as PTCL-GATA3 subtype and 52% (27 / 52) of cases as PTCL-TBX21 subtype ( Tables 4 and 5 ; Fig.16 A). This classification showed good agreement with our IHC algorithm classification (overall agreement: 80%). To further validate our transcriptomic signature classification, we compared sequencing data in these two groups. We observed that genetic alterations such as TET2 mutations were common in the PTCL-TBX21 subtype, while TP53 mutations were enriched in the PTCL-GATA3 subgroup, which is consistent with our previous findings. 31 is consistent ( Fig.16 B).
[0299] Together with our previous observations 21,29 Consistently, cases with expression of cytotoxic transcripts were significantly enriched in the PTCL-TBX21 subtype ( Fig.16 C), and was validated by a cytotoxic immunophenotype that was more common in the PTCL-TBX21 subtype than in the PTCL-GATA3 subtype (52% vs. 16%, p = 0.013, Fig.16C). There is an inverse correlation between the average cytotoxic CD8+ T cell signature and the pan-B cell CD20 transcript ( Fig.16 D) Compared with our previous observations 29 Consistently, regardless of the cytotoxic phenotype, PTCL-TBX21 cases often had a rich inflammatory background ( Fig.16 E). Since clinical outcomes for PTCL-NOS were available only in a limited number of cases, the combined cohort associated with cases classified as PTCL-GATA3 vs. PTCL-TBX21 trended toward a worse OS (median OS: 0.57 vs. 1.4 years; Fig.16 F).
[0300] PTCL-T excluded from the validation cohort FH Case evaluation
[0301] use Platform Analysis PTCL-T FH Of the 12 cases (n=12), only 2 cases (17%) showed significant association with AITL molecular features, while 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 cases (n=3). FH PTCL-T with high AITL features FH The cases also showed a higher T than the rest of the cases FH feature( Fig.23 AB). We also included 10 cases of reactive hyperplasia, none of which showed expression of diagnostic features including PTCL subtypes in this assay.
[0302] discuss
[0303] The diagnosis of PTCL is one of the most challenging among lymphomas and more often results in uncertain, inconsistent, or incorrect diagnoses. 19-21,32,33 Recently, novel therapeutic approaches have shown significant benefits in PTCL subgroups, including brentuximab-vedotin (BV) for CD30+ PTCL, particularly ALCL. 34 , crizotinib for ALK+ALCL 35,36 ; For ATLL 37 Moglizumab, for AITL or T FH -HDACi and demethylating agents for PTCL 38 , and possibly ensidipine for IDH2 mutant AITL 39 Therefore, accurate diagnosis may be important for patient treatment and clinical trials of new drugs.40-42 We have performed extensive GEP studies on PTCL and constructed RNA-based molecular diagnostic signatures and survival predictors, and have delineated key oncogenic mechanisms 19-21 Some of these findings have been included in the revised 2016 WHO classification 6 In order to transform this molecular information into a platform suitable for clinical application 25 , we used , which correlates well with GEP data from fresh frozen tissues The platform was systematically analyzed to identify RNA from FFPE tissue in PTCL, as this digital quantification technique is more tolerant to the 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 In pre-analytical evaluation, RNAstorm TM The kit improves RNA yield and quality from older samples (i.e., DV 200 >50%). However, in recently obtained FFPE tissues, other isolation methods can also provide good quality RNA. Additionally, the assay can be performed using limited RNA amounts (a minimum of 200 ng is required), usually obtained from several unstained slides depending on the size of the tissue. However, we were able to have adequate classification even with RNA extracted from core needle biopsies, which accounted for approximately 10% of the samples studied. Interlaboratory comparisons of variability and reproducibility between the three CLIA-certified laboratories were also very relevant.
[0304] Since we did not have GEP data for the cases used for validation, we included only cases with a confirmed pathological diagnosis after extensive IHC studies and rigorous review. These findings were very similar to those of the training set ( Fig.12 ), all of which maintained high sensitivity, specificity, and accuracy. In some cases, the signature score fell just outside the cutoff point, and these cases were considered "borderline" cases. Fortunately, in the retrospective analysis, molecular testing made the correct diagnosis in a few cases. -. However, significant differences were also observed. This may be partly due to technical reasons, such as tumor content and tumor heterogeneity, but some reasons may be related to unknown biology, such as the strong ALK signature in some ALK-ALCL cases may represent the recently reported ALK-positive ALCL 43 Similarly, two PTCL-NOS showed similarities to AITL and T FH The transcriptomic signatures were significantly correlated, which showed T FH Focal expression of markers (BCL6, ICOS). FHExpression characteristics may indicate that these cases are different from PTCL-T FH More similar, but does not satisfy at least two T FH In these cases, molecular assays revealed a complex and overlapping biology between PTCL-NOS and AITL as well as their poorly defined borderline. We also found that the addition of EBV and HTLV1 transcripts enhanced the diagnostic performance of molecular assays in ENTKCL and ATLL, respectively. An important contribution of this approach is the robust definition of PTCL-GATA3 and PTCL-TBX21 cases, which have distinct biology and prognosis, supported by recent genetic findings. 31 Although GEP classification can be simulated using IHC panels, staining can be challenging to optimize and interpret, which can lead to large variability between institutions. The assay reported here is highly reproducible across laboratories and therefore has the major advantage.
[0305] The initial GEP study that defined diagnostic features was in PTCL-T FH The definition of temporary entities was not formally incorporated into this study prior to this study. The assay explored 12 PTCL-T FH Cases showed higher mean AITL and T FH features, and the individual features overlapped with the molecular features of AITL or PTCL subtypes, which is consistent with the study of Dobay et al. 44 These exploratory studies have shown that PTCL-T FH It is unlikely to be a single entity as currently defined and further evaluation is needed to determine how it should be best characterized. 20,21 Indicating the presence of cytotoxic PTCL variants within the PTCL-TBX21 subgroup. Consistent with this observation, we found that 25% of PTCL-TBX21 cases had cytotoxic features and validated the correlation with the CD8+ phenotype by IHC. Due to the small number of cases studied, further studies are needed to develop a robust characterization of this group of cases.
[0306] In summary, we describe a method to translate PTCL diagnostic features into a clinically applicable assay that we envision will become a useful tool for both general and academic pathologists in the diagnostically challenging area of PTCL. Additionally, we believe it may facilitate better case definitions for studies and ensure more uniform clinical trial cohorts. PTCL classification is an evolving field, and as our understanding of the underlying biology and available technology improves, modifications will begin to be made to make this classification clinically relevant.
[0307] References cited in this example :
[0308] 1.Bellei M,Chiattone CS,Luminari S,et al.T-cell lymphomas in Southamerica and europe.Rev Bras Hematol Hemoter.2012;34(1):42-47.
[0309] 2.Vose J,Armitage J,Weisenburger D,International TCLP.Internationalperipheral T-cell and natural killer / T-cell lymphoma study:pathology findingsand clinical outcomes.J Clin Oncol.2008;26(25):4124-4130.
[0310] 3.Briski R,Feldman AL,Bailey NG,et al.The role of front-lineanthracycline-containing chemotherapy regimens in peripheral T-celllymphomas.Blood Cancer J.2014;4:e214.
[0311] 4.Rudiger T,Weisenburger DD,Anderson JR,et al.Peripheral T-celllymphoma(excluding anaplastic large-cell lymphoma):results from the Non-Hodgkin′s Lymphoma Classification Project.Ann Oncol.2002;13(1):140-149.
[0312] 5.Fox CP,Civallero M,Ko YH,et a1.Survival outcomes of patients withextranodal natural-killer T-cell lymphoma:a prospective cohort study from theinternational T-cell Project.Lancet Haematol.2020;7(4):e284-e294.
[0313] 6.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification of Tumoursof Haematopoietic and Lymphoid Tissues,Fourth Edition.Revised.2017;2.
[0314] 7.Weisenburger DD,Savage KJ,Harris NL,et al.Peripheral T-celllymphoma,not otherwise specified:a report of 340 cases from the InternationalPeripheral T-cell Lymphoma Project.Blood.2011;117(12):3402-3408.
[0315] 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.
[0316] 9.Tsukasaki K,Tsushima H,Yamamura M,et al.Integration patterns ofHTLV-I provirus in relation to the clinical course of ATL:frequent clonalchange at crisis from indolent disease.Blood.1997;89(3):948-956.
[0317] 10.Escalon MP,Liu NS,Yang Y,et al.Prognostic factors and treatment ofpatients with T-cell non-Hodgkin lymphoma:the M.D.Anderson Cancer Centerexperience.Cancer.2005;103(10):2091-2098.
[0318] 11.Horwitz S,O′Connor OA,Pro B,et al.Brentuximab vedotin withchemotherapy for CD30-positive peripheral T-cell lymphoma(ECHELON-2):aglobal,double-blind,randomised,phase 3 trial.Lancet.2019;393(10168):229-240.
[0319] 12.Pro B,Advani R,Brice P,et al.Five-year results of brentuximabvedotin in patients with relapsed or refractory systemic anaplastic largecell lymphoma.Blood.2017;130(25):2709-2717.
[0320] 13.Alizadeh AA,Eisen MB,Davis RE,et al.Distinct types of diffuselarge B-cell lymphoma identified by gene expression profiling.Nature.2000;403(6769):503-511.
[0321] 14.Rosenwald A,Wright G,Chan WC,et al.The use of molecular profilingto predict survival after chemotherapy for diffuse large-B-cell lymphoma.NEngl J Med.2002;346(25):1937-1947.
[0322] 15.Dave SS,Fu K,Wright GW,et al.Molecular diagnosis of Burkitt′slymphoma.N Engl J Med.2006;354(23):2431-2442.
[0323] 16.Monti S,Savage KJ,Kutok JL,et al.Molecular profiling of diffuselarge B-cell lymphoma identifies robust subtypes including one characterizedby host inflammatory response.Blood.2005;105(5):1851-1861.
[0324] 17.Rosenwald A,Wright G,Wiestner A,et al.The proliferation geneexpression signature is a quantitative integrator of oncogenic events thatpredicts survival in mantle cell lymphoma.Cancer Cell.2003b;3(2):185-197.
[0325] 18.Shipp MA,Ross KN,Tamayo P,et al.Diffuse large B-cell lymphomaoutcome prediction by gene-expression profiling and supervised machinelearning.Nat Med.2002;8(1):68-74.
[0326] 19.Iqbal J,Weisenburger DD,Chowdhury A,et al.Natural killer celllymphoma shares strikingly similar molecular features with a group of non-hepatosplenic gammadelta T-cell lymphoma and is highly sensitive to a novelaurora kinase A inhibitor in vitro.Leukemia.2011;25(2):348-358.
[0327] 20.Iqbal J,Weisenburger DD,Greiner TC,et al.Molecular signatures toimprove diagnosis in peripheral T-cell lymphoma and prognostication inangioimmunoblastic T-cell lymphoma.Blood.2010;115(5):1026-1036.
[0328] 21.Iqbal J,Wright G,Wang C,et al.Gene expression signatures delineatebiological and prognostic subgroups in peripheral T-cell lymphoma.Blood.2014;123(19):2915-2923.
[0329] 22.Lone W,Alkhiniji A,Manikkam Umakanthan J,Iqbal J.MolecularInsights Into Pathogenesis of Peripheral T Cell Lymphoma:a Review.CurrHematol Malig Rep.2018;13(4):318-328.
[0330] 23.Iqbal J,Amador C,McKeithan TW,Chan WC.Molecular and GenomicLandscape of Peripheral T-Cell Lymphoma.Cancer Treat Res.2019;176:31-68.
[0331] 24.Srinivasan M,Sedmak D,Jewell S.Effect of fixatives and tissueprocessing on the content and integrity of nucleic acids.Am J Pathol.2002;161(6):1961-1971.
[0332] 25.Crescenzo R,Abate F,Lasorsa E,et al.Convergent mutations andkinase fusions lead to oncogenic STAT3 activation in anaplastic large celllymphoma.Cancer cell.2015;27(4):516-532.
[0333] 26.Scott DW,Wright GW,Williams PM,et al.Determining cell-of-originsubtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue.Blood.2014;123(8):1214-1217.
[0334] 27.Phang KC,Akhter A,Tizen NMS,et al.Comparison of protein-basedcell-of-origin classification to the Lymph2Cx RNA assay in a cohort ofdiffuse large B-cell lymphomas in Malaysia.J Clin Pathol.2018;71(3):215-220.
[0335] 28.Rimsza LM,Wright G,Schwartz M,et al.Accurate classification ofdiffuse large B-cell lymphoma into germinal center and activated B-cellsubtypes using a nuclease protection assay on formalin-fixed,paraffin-embedded tissues.Clin Cancer Res.2011;17(11):3727-3732.
[0336] 29.Amador C,Greiner TC,Heavican TB,et al.Reproducing the molecularsubclassification of peripheral T-cell lymphoma-NOS byimmunohistochemistry.Blood.2019;134(24):2159-2170.
[0337] 30.Advani RH,Ansell SM,Lechowicz MJ,et al.A phase II study ofcyclophosphamide,etoposide,vincristine and prednisone(CEOP)Alternating withPralatrexate(P)as front line therapy for patients with peripheral T-celllymphoma(PTCL):final results from the T-cell consortium trial.Br JHaematol.2016;172(4):535-544.
[0338] 31.Heavican TB,Bouska A,Yu J,et al.Genetic drivers of oncogenicpathways in molecular subgroups of peripheral T-cell lymphoma.Blood.2019;133(15):1664-1676.
[0339] 32.Maura F,Agnelli L,Leongamornlert D,et al.Integration oftranscriptional and mutational data simplifies the stratification ofperipheral T-cell lymphoma.Am J Hematol.2019.
[0340] 33.Piccaluga PP,Fuligni F,De Leo A,et al.Molecular profiling improvesclassification and prognostication of nodal peripheral T-cell lymphomas:results of a phase III diagnostic accuracy study.J Clin Oncol.2013;31(24):3019-3025.
[0341] 34.Van Der Weyden C,Dickinson M,Whisstock J,Prince HM.Brentuximabvedotin in T-cell lymphoma.Expert Rev Hematol.2019;12(1):5-19.
[0342] 35.Andraos E,Dignac J,Meggetto F.NPM-ALK:A Driver of LymphomaPathogenesis and a Therapeutic Target.Cancers(Basel).2021;13(1).
[0343] 36.Gambacorti-Passerini C,Messa C,Pogliani EM.Crizotinib inanaplastic large-cell lymphoma.N Engl J Med.2011;364(8):775-776.
[0344] 37.Moskowitz AJ,Lunning MA,Horwitz SM.How I treat the peripheral T-cell lymphomas.Blood.2014;123(17):2636-2644.
[0345] 38.Lopez AT,Bates S,Geskin L.Current Status of HDAC Inhibitors inCutaneous T-cell Lymphoma.Am J Clin Dermatol.2018;19(6):805-819.39.https: / / clinicaltrials.gov / ct2 / show / NCT02273739.
[0346] 40.Ishida T,Joh T,Uike N,et al.Defucosylated anti-CCR4 monoclonalantibody(KW-0761)for relapsed adult T-cell leukemia-lymphoma:a multicenterphase II study.J Clin Oncol.2012;30(8):837-842.
[0347] 41.O′Connor OA,Bhagat G,Ganapathi K,et al.Changing the paradigms oftreatment in peripheral T-cell lymphoma:from biology to clinicalpractice.Clin Cancer Res.2014;20(20):5240-5254.
[0348] 42.Iqbal J,Wilcox R,Naushad H,et al.Genomic signatures in T-celllymphoma:How can these improve precision in diagnosis and inform prognosis?Blood Rev.2016;30(2):89-100.
[0349] 43.Yeung KHRA,Russell V,Choi W,et al.ALK-Negative Anaplastic LargeCell Lymphomas Encompass Distinct Subgroups Including an ALK-Positive-likeSubgroup with Favorable Prognosis.Blood.2021;1382403.
[0350] 44.Dobay MP,Lemonnier F,Missiaglia E,et al.Integrativeclinicopathological and molecular analyses of angioimmunoblastic T-celllymphoma and other nodal lymphomas of follicular helper T-cellorigin.Haematologica.2017;102(4):e148-e151.
[0351] Materials and methods
[0352] Patient Information
[0353] We included 249 diagnosed PTCL cases from multiple institutions, after excluding 4 cases with poor RNA quality, and then divided the remaining cases into a training cohort of 105 cases and a validation cohort of 140 cases (Tables 4 and 5 / Fig.12 A). The basic clinical and pathological characteristics of the two cohorts are shown in Table 6. The training cohort consisted of a "bridging group" of 91 cases that had been previously included in the previous GEP study. 1-3 13–15 (Affymetrix, Santa Clara, CA; accession number GSE19069) were used to evaluate RNA from FF tissues for which corresponding FFPE samples were available. In addition, 9 new PTCLs had available HG-U133Plus 2.0 data that were not included in the earlier GEP study. 1-3 , and 5 well-characterized classic ENKTCL cases with a definitive diagnosis 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 cases were used to identify which RNA transcripts previously identified as diagnostic on FF tissue were It also performs well in FFPE on the NanoString platform (NanoString Inc, Seattle).
[0354] A separate validation cohort consisted of an independent series of 140 PTCL cases that had not been previously analyzed with GEP but had sufficient FFPE tissue available. This group included 19 AITL, 20 ALK+ALCL, 16 ALK-ALCL, 12 ATLL, 22 ENKTCL, and 51 PTCL-NOS. Although we initially assembled 152 PTCL cases for validation, 12 cases were excluded from this cohort because they were pathologically reclassified as having T-follicular helper type (T FH These 12 PTCL-T phenotypes were analyzed using NanoString-based assays. FH cases and 10 additional cases with reactive lymphadenopathy for comparison. The study protocol was approved by the UNMC and COH-MC Institutional Review Boards.
[0355] Morphological and immunophenotypic characteristics of the PTCL cohort
[0356] According to the current WHO classification 4All cases were centrally reviewed and diagnosed. The cases 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 with expertise in the diagnosis of T-cell lymphomas (CA, DW, WCC). A consensus diagnosis was reached when there was agreement on the diagnosis. Pathological review included partial or complete panels of B-cell and T-cell immunostaining, as needed, 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. PTCL-NOS cases with strong expression of two or more TFH markers were reclassified as nodal PTCL with a TFH phenotype. One hematopathologist (CA) 5 Using a recently published IHC algorithm, PTCL-NOS cases were subdivided into PTCL-GATA3 and PTCL-TBX21 using antibodies against GATA3, TBX21, CCR4, and CXCR3. Of the 43 AITL cases, 40 had typical AITL morphology, but three were T-FLs that were grouped together with AITL for transcriptomic analysis. 4,6,7 , ATLL cases in the validation series must produce a positive signal (CT value) with HZB virus transcripts using qRT-PCR (Taq man probe) based on FFPE-extracted RNA. For the presence of LMP1 / EBNA1 in genomic DNA, EBER in situ hybridization or PCR was used to confirm ENKTCL cases in the validation cohort. Of the 32 ENKTCL cases, 30 had nasal metastases and were typical ENKTCL, and two cases had mainly lymph node metastases but otherwise had phenotypic features similar to nasal ENKTCL cases. Detailed information on nasal metastases for these 2 cases is not available. In the current WHO classification, these two cases may represent primary nodal EBV-positive T / NK cell lymphomas, but are grouped together with ENK / TCL for transcriptomic analysis.
[0357] RNA was extracted from FFPE tissue and used Digital Gene Expression
[0358] FFPE rolls with a thickness of 10 μm to 20 μm were cut from the blocks to a surface area of 1 cm2 and stored at -20°C until extraction. For some cases where FFPE rolls were not available, tissues were scraped from unstained slides. Two commercial kits (a) Qiagen AllPrep DNA / RNA FFPE Kit (Qiagen GmbH, Germany) or (b) RNAStormTM RNA Isolation Kit (Cell Data Sciences, CD501) were used according to the manufacturer's instructions. 8 Total RNA was deparaffinized and extracted. RNA quality was assessed using an Agilent Tape-Station and RNA Integrity Number (RIN) values were used. 9 and by RNA fragment size > 200 bases (DV 200 ) to estimate 10 . RNA was quantified using Qubit and 200ng of total RNA (DV200: >35%) was used for hybridization, and for more degraded samples, additional RNA was used following the recommendations of NanoString, Inc (https: / / www.nanostring.com / wp-content / uploads / 2021 / 07 / MAN-10050-05-Preparing-48RNA-from-FFPE-Samples.pdf). The preparation station hybridized total RNA with a custom coding set (see below) at 65°C overnight (16 to 18 hours) and GEP data were acquired on a digital analyzer at "high 50 resolution" setting. Standard QC (nSolver TM Any sample was marked where the total number of positive spike controls exceeded the geometric mean of the total positive spike controls for the box by 0.3 to 3 times (provided by the analysis software). The signal count values were normalized by dividing the counts of each gene of a given sample by the geometric mean of the counts of the housekeeping genes in that sample. The normalized counts were then log2 transformed to obtain gene signal values for all subsequent analyses. Samples with a geometric mean of 10 or less for housekeeping genes were excluded from the study.
[0359] Used for Code set design for PTCL subtyping assay
[0360] A training cohort of PTCL cases (n=105) with GEP derived from fresh frozen RNA was used to select Classifier genes for the platform. We reanalyzed earlier FF GEP data (HG-U133Plus 2.0 arrays, Affymetrix, Inc) to generate a diagnostic transcript signature that includes 287 classifier transcripts specific to PTCL subtypes, 50 housekeeping genes with consistent expression in PTCL, and an additional set of genes associated with T cell differentiation. These genes were selected as characteristically differentially expressed (positive or negative): (1) one of the PTCL subtypes (i.e., AITL, ALCL, ENKTCL, ATLL), (2) differentially expressed between ALK+ and ALK-ALCL, or (3) distinguishing the PTCL-TBX21 and PTCL-GATA3 subgroups of PTCL-NOS. The 50 housekeeping genes were selected from the genes with the smallest variance in PTCL in the FF GEP data. Therefore, these classifier transcripts were designed to be the most informative and robust classification elements, and Codesets are designed using the standards used by the NanoString Bioinformatics team for codesets. The barcode corresponding to each captured transcript is imaged and counted after hybridization, 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-Code Sets).
[0361] In order to minimize the time and effort associated with repetitive code set design and manufacturing ( NanoString Inc.), we used samples from the bridging cohort to curate an initial set of diagnostic markers to define a panel of 93 genes and five subtype-specific viral transcripts from EBV and HTLV-1 that showed the best performance on the NanoString platform. This evaluation was performed using the following sets of criteria: (a) correlation between NanoString log2 signal values and their corresponding HG-U133Plus 2.0 signal intensity (SI) values; (b) T-statistic between normalized log-transformed NanoString SI values for those samples in the PTCL subtype and those in the other PTCL subtypes; (c) T-statistic between HG-U133Plus 2.0 signal values for all samples in the PTCL subtype and those in the other PTCL subtypes 1 , and (d) from Iqbal et al. 2014 1Correlation between NanoString signal values and model scores for subtypes of . . Care was taken to make this curated list include both positively and negatively correlated genes for each distinction. Additionally, housekeeping genes were reduced to 16 with minimal variance in the bridging cohort. We also added 40 additional genes identified in the literature as being critical to T-cell lymphoma biology. This resulted in a final 153-gene code set for all subsequent analyses.
[0362] Prediction of subclasses in the training cohort
[0363] A variety of prediction methods, including linear discriminant analysis, lasso regression, support vector machines, and random forests, were applied to the training set in a cross-validated manner using several different prediction architectures (data not shown). The method with the best cross-validated prediction accuracy was based on a series of binary linear predictors, whose structure is described below and shown in Fig.12 A and Fig.17 middle.
[0364] For a given class pair and a selected gene i, we compute
[0365]
[0366] where μ 1i and μ 2i represents the average expression value of the gene in category 1 and category 2, and σ i 2 is the pooled within-group variance estimate. Then, for each sample j, we compute the weighted score for the distinction between set 1 and set 2 as
[0367]
[0368] where xij is the normalized signal value of gene i on sample j and the sum is the sum of all genes selected for a particular distinction.
[0369] We then plotted the ROC curve based on the relationship between the score and the members in category 1 or category 2, and identified the point where the average values of sensitivity and specificity were maximized. Samples with scores above this point were identified as category 1, and samples with scores below this point were identified as category 2. AITL subtype classification is an exception to this rule, in which there is an overlapping area between AITL and PTCL-NOS model scores. For samples with scores in this area, it is uncertain whether the sample is AITL. In order to define this area, we considered the lower limit cutoff point that maximizes sensitivity + 2 * specificity and the upper limit cutoff point that maximizes specificity + 2 * sensitivity. Samples above the upper limit cutoff point are called AITL, and samples below the lower limit cutoff point are called non-AITL, and samples between these two cutoff points are called critical AITL.
[0370] To predict the subtype of a single sample, we first ran 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. We found that the ENKTCL predictor tended to identify PTCL-NOS with cytotoxic marker expression as ENKTCL, so we additionally generated a classifier between PTCL-NOS and ENKTCL based on the viral transcript EBER. For samples predicted to be ENKTCL, we required that both the multigene ENKTCL and the single EBER classifiers were consistent with the molecular diagnosis. The final predictor for distinguishing ALK-ALCL from ALK+ALCL was also generated. First, ALCL samples were divided according to ALK positivity, and secondly, ENKTCL samples with some features of ALCL but all ALK (-) were used to distinguish. If the initial set of predictors all suggested that the sample was PTCL-NOS, the final predictor for distinguishing PTCL-GATA3 from PTCL-TBX21 was applied. If more than one of the four predictors yielded a non-PTCL-NOS diagnosis (e.g., both ALCL and AITL), an additional head-to-head predictor was applied between those types to tip the balance (e.g., ALCL vs. AITL), where all gene signatures of either type were selected for inclusion in the model.
[0371] Pre-analytical evaluation of FFPE RNA
[0372] In order to identify We have selected transcripts that performed well on the platform. The signal intensity of each transcript in 10 fresh frozen RNA samples analyzed by the platform and HG-U133plus2 array was correlated. More than 86.5% and 65% of the transcripts showed a correlation of >0.4 or >0.6 between the two platforms, respectively, and these transcripts were selected for further analysis ( Fig.18 AB). Additionally, RNA isolated from FFPE samples using the RNAstormTM method produced longer transcripts (>200 bases) compared to another commercial kit ( Fig.18 CD), and resulted in a higher correlation coefficient with fresh frozen RNA transcript profiles ( Fig.18 E). Therefore, the RNAstormTM Isolation Kit is used in the standard operating procedure for this assay for subsequent samples. We are able to isolate at least 200ng to 1μg of total RNA / FFPE sample, where >50% of the RNA indicates a DV200 >50%. There was a high correlation between FF RNA and the corresponding FFPE RNA run on NanoString, Inc. Fig.18 F).
[0373] Reproducibility between loci and alignment between original gene set (n=442) and improved gene set (n=153)
[0374] To examine the reproducibility of the assay across institutions, we selected 24 samples from the training set and re-ran these samples on the 153-gene array at two different institutions (City of Hope and Insight Genetics). Seven of these samples, along with one additional training sample, were also re-run at the original site (UNMC) on the reduced gene set array, yielding a total of 55 replicates. We observed that for some predictor scores, there appeared to be systematic biases between the models run on the 442-gene array and the models run on 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 153) and the 442 gene array (denoted by μ442 and 442). Using these values, we adjusted the scores generated by the 153 gene array (particularly those in the validation set) to match those generated by the 442 gene array according to the following formula:
[0375]
[0376] It should be noted that no validation samples were used to generate the prediction model or define this adjustment.
[0377] Statistical analysis
[0378] Baseline patient characteristics were compared between groups using χ2 and t-tests. Overall survival distribution was estimated using the Kaplan-Meier method. Overall survival (OS) time was calculated as the time from diagnosis to date of death or date of last contact. Overall survival analysis was performed on patients who were alive at the time of last contact. Differences in outcomes between groups were assessed using the permutation-based log-rank test implemented in the coin R-package, using the "approximate" distribution. 11 Two-sided p < .05 was considered significant.
[0379] References cited in this method :
[0380] 1.Iqbal J, Wright G, Wang C, et al.Gene expression signatures delineatebiological and prognostic subgroups in 181peripheral T-celllymphoma.Blood2014;123(19):2915-2923.182
[0381] 2.Iqbal J, Weisenburger DD, Greiner TC, et a1.Molecular signaturestoimprove diagnosis in peripheral T-cell 183lymphoma and prognostication inangioimmunoblastic T-cell lymphoma.Blood2010;115(5):1026-1036.184
[0382] 3.Iqbal J, Weisenburger DD, Chowdhury A, et al.Natural killer celllymphoma 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):348-358.187
[0383] 4.Swerdlow SH,Campo E,Harris NL,et al.WHO Classification of Tumoursof Haematopoietic and Lymphoid 188 Tissues,Fourth Edition.Revised.2017;2.189
[0384] 5.Amador C,Greiner TC,Heavican TB,et al.Reproducing the molecularsubclassification of peripheral T-cell 190 lymphoma-NOS byimmunohistochemistry.Blood.2019;134(24):2159-2170.191
[0385] 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.BloodRes.2017;52(2):106-111.193
[0386] 7.Li M,Green PL.Detection and quantitation of HTLV-1 and HTLV-2mRNAspecies by real-time RT-PCR.J Virol 194 Methods.2007;142(1-2):159-168.195
[0387] 8.Karmakar S,Harcourt EM,Hewings DS,et al.Organocatalytic removal offormaldehyde adducts from RNA and 196 DNA bases.Nat Chem.2015;7(9):752-758.197
[0388] 9.Schroeder A,Mueller O,Stocker S,et al.The RIN:an RNA integritynumber for assigning integrity values to RNA 198 measurements.BMC MolBiol.2006;7:3.199
[0389] 10.Wimmer I,Troscher AR,Brunner F,et al.Systematic evaluation of RNAquality,microarray data reliability and 200 pathway analysis in fresh,freshfrozen and formalin-fixed paraffin-embedded tissue samples.Sci Rep.201 2018;8(1):6351.202
[0390] 11.Hothorn T,Hornik K,van de Wiel MA,Zeileis A.Implementing a Classof Permutation Tests:The coin Package.203 Journal of StatisticalSoftware.2008;28(8):1-23.
Claims
1. A method for distinguishing subtypes of peripheral T-cell lymphoma (PTCL), the method comprising: isolating nucleic acid from a sample from a subject; obtaining a gene expression profile from the sample; as well as The subtype of PTCL is identified based on the presence of specific genes within the 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 undetermined, wherein the undetermined type indicates that the sample contains characteristics of at least two of the PTCL subtypes.
3. The method of claim 2, wherein 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 is identified as AITL.
4. The method of claim 2, wherein 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 is identified as ALK-ALCL.
5. The method of claim 2, wherein 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 is identified as ENKTL.
6. The method of claim 2, wherein 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, the PTCL subtype is identified as ATLL.
7. The method of claim 2, wherein if the gene expression profile includes any one or more of ALK, CACNA2D2, CCDC64, CCNA1, CCR4, DMBX1, GALNT2, GAS1, GATA3, HTRA3, IL1RAP, PCOLCE2, PFKFB3, RBGAP1L, TIAM2, TMEM158, and ZADH2, the PTCL subtype is identified as ALK+ALCL.
8. The method of claim 2, wherein if the gene expression profile includes any one or more of CASP2, EZH2, FLJ37453, IFI30, IFNG, LAGE3, LAP3, NR1H3, PLA2G7, SLAMF7, SLC3 1A2, TCN2, TMEM1 76B, WARS, TBX21, CXCR3, GATA3, and CCR4, the PTCL subtype is identified as PTCL-NOS.
9. The method according to any one of claims 1 to 8, wherein the sample is a biopsy specimen from the subject.
10. The method of claim 9, wherein the sample comprises formalin-fixed paraffin-embedded tissue.
11. The method of claim 9, wherein the sample comprises fresh frozen tissue.
12. The method of any one of claims 1 to 11, further comprising providing to the subject an effective amount of a subtype-specific therapeutic.
13. The method of claim 12, wherein the subtype-specific therapeutic comprises a histone deacetylase (HDAC) inhibitor, an antifolate, an alkylating agent, a proteasome 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 (PDGF Rβ) inhibitors, mammalian target of rapamycin (mTOR) pathway inhibitors, immune checkpoint inhibitors, hypomethylating agents, anti-differentiation cluster 52 (CD52) antibodies, immunomodulatory drugs, anti-inducible T cell co-stimulator (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 a combination thereof.
14. The method according to claim 13, wherein: The HDAC inhibitor includes romidepsin, belinostat, panobinostat or a combination thereof; The antifolates include pralatrexate; The alkylating agents include bendamustine; The proteasome inhibitors include bortezomib; The antibody-drug conjugate comprises brentuximab; The PI3K inhibitor includes duvelixibu, tenalixibu or a combination thereof; the JAK inhibitor includes ruxolitinib; The ALK inhibitor includes crizotinib; The mTOR pathway inhibitor includes everolimus; The hypomethylating agent includes 5-azacitidine; The anti-CD52 antibody includes alemtuzumab; The ImiD includes lenalidomide; The CCR4 inhibitors include moglizumab; The IDH inhibitors include ensidipine; The BCL2 inhibitors include venetoclax; The anti-CD25 antibody includes tecamimumab; The calcineurin inhibitors include cyclosporine A; The SYK inhibitor includes selenitinib; The bispecific antibody comprises AFM1 3; The chimeric antigen receptor T (CAR-T) cells comprise CD30, CD7, or both; or a combination thereof.
15. The method of claim 12, wherein the subtype-specific treatment for PTCL-NOS comprises romidepsin, belinostat, brentuximab, duveliximab, or a combination thereof.
16. The method of claim 12, wherein the subtype-specific treatment for AITL comprises romidepsin, 5-Aza, isocitrate dehydrogenase (IDH) inhibitors, calcineurin inhibitors, or a combination thereof.
17. The method of claim 12, wherein the subtype-specific treatment for ALK-ALCL comprises brentuximab.
18. The method of claim 12, wherein the subtype-specific treatment for ALK+ALCL comprises an ALK inhibitor, a platelet-derived growth factor receptor β (PDGFRβ) inhibitor, or a combination thereof.
19. The method of claim 12, wherein the subtype-specific treatment for ATLL comprises a NOTCH inhibitor, a hepatocyte growth factor (HGF) inhibitor, a cMET inhibitor, moglizumab, or a combination thereof.
20. The method of claim 12, wherein the subtype-specific treatment for ENKTL comprises a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor.
21. 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 a gene or combination of genes representing a genetic signature indicative of a particular PTCL subtype and instructions for use.
22. The kit of claim 21, 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 undetermined, wherein the undetermined type indicates that the sample contains characteristics of at least two of the PTCL subtypes.
23. The kit of claim 22, wherein 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 indicates AITL.
24. The kit of claim 22, wherein if the gene signature includes any one or more of BATF3, C9orf89, CD28, DUSP2, ICMT, LGALS1, PIK3IP1, PRKCQ, TNFRSF8, TRAT 1, and ZNF 101, the PTCL subtype is indicated as ALK-ALCL.
25. The kit of claim 22, wherein 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 indicates ENKTL.
26. The kit of claim 22, wherein 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 indicates ATLL.
27. The kit of claim 22, wherein 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, TMEM1 58, and ZADH2, the PTCL subtype indicates ALK+ALCL.
28. The kit of claim 22, wherein 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 indicates PTCL-NOS.
29. A method for identifying a specific subtype of peripheral T-cell lymphoma (PTCL) in a sample, the method comprising: obtaining a gene expression profile from a sample; The gene expression profile is compared to a gene signature associated with a specific PTCL subtype, wherein each subtype of PTCL comprises a unique gene signature; and the subtype of PTCL within the sample is identified as 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 undetermined, wherein the undetermined type indicates that the gene expression profile of the sample includes genes from the unique gene signature of at least two of the PTCL subtypes.
30. The method of claim 29, wherein 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 ALK-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, SH2D 1B, EBER, LMP1, and TNFRSF25; The gene signature of ATLL includes any one or more of ARSG, CCNE1, DOK5, FGF1 8, 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 thereof.
31. The method of claim 29 or claim 30, wherein identifying the subtype of PTCL comprises: A series of at least four binary predictors where: A first binary predictor comprises 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 comprises 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 comprises determining whether the gene expression profile of the sample is more consistent with the gene signature of ATLL or PTCL-NOS; A fourth binary predictor comprises determining whether the gene expression profile of the sample is more consistent with the gene signature of ENKTL or PTCL-NOS; and The PTCL subtype of the sample is assigned based on the results from the binary predictor.
32. The method of claim 31, wherein the PTCL subtype is identified as AITL when: Among the first binary predictors, the gene expression profile of the sample is more consistent with the gene signature of AITL; and Of the second, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS.
33. The method of claim 31, wherein the PTCL subtype is identified as ALCL when: in said second binary predictor, said gene expression profile of said sample is more consistent with said gene signature of ALCL; and Of the first, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS.
34. The method of claim 33, further comprising a fifth binary predictor, wherein the fifth binary predictor comprises determining whether the gene expression profile of the sample is more consistent with the gene signature of ALK+ALCL or ALK-ALCL; and When the gene expression profile of the sample is more consistent with the gene signature of ALK+ALCL, identifying the PTCL subtype as ALK+ALCL; or When the gene expression profile of the sample is more consistent with the gene signature of ALK-ALCL, the PTCL subtype is identified as ALK-ALCL.
35. The method of claim 31, wherein the PTCL subtype is identified as ATLL when: In the third binary predictor, the gene expression profile of the sample is more consistent with the gene signature of ATLL; and Of the first, second, and fourth binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS.
36. The method of claim 31, wherein the PTCL subtype is identified as ENKTL when: In the fourth binary predictor, the gene expression profile of the sample is more consistent with the gene signature of ENKTL; and Of the first, second and third binary predictors, the gene expression profile of the sample is more consistent with the gene signature of PTCL-NOS.
37. The method of claim 31, wherein the PTCL subtype is identified as PTCL-NO S when: Of the first, second, third, and fourth binary predictors, the gene expression profile of the sample is more consistent with PTCL-NOS.
38. The method of claim 31, wherein the PTCL subtype is identified as undetermined when at least two of the binary predictors are not more consistent with PTCL-NOS.
39. The method of claim 38, wherein samples identified as undetermined are subjected to additional binary predictors comprising: a first potential subtype comprising one of said PTCL subtypes that is not more consistent with PTCL-NOS and a second potential subtype comprising at least one other PTCL subtype that is not more consistent with PTCL-NOS; The method further comprises: determining whether the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype or the second potential subtype; and When the gene expression profile of the sample is more consistent with the gene signature of the first potential subtype, the PTCL subtype is identified as the first potential subtype; or when the gene expression profile of the sample is more consistent with the gene signature of the second potential subtype, the PTCL subtype is identified as the second potential subtype.
40. The method of claim 39, wherein: If the PTCL subtype is identified as ALCL, the method further comprises determining the ALCL subtype as ALK+ALCL or ALK-ALCL using the fifth binary predictor according to claim 34.
41. The method of claim 37, wherein samples identified as PTCL-NOS cases are further separated into a GATA3-high subgroup and a TBX21-high subgroup.
42. The method of claim 36, wherein samples identified as ENKTL are subdivided into NK or gamma / delta T cell lineages.
43. A method for treating peripheral T-cell lymphoma, the method comprising: obtaining a gene expression profile from a sample from the subject; comparing the gene expression profile to a gene signature associated with a specific subtype of PTCL, wherein each subtype of PTCL comprises a unique gene signature; identifying the subtype of PTCL within the sample as 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 undetermined, wherein undetermined indicates that the gene expression profile of the sample includes genes from the unique gene signature of at least two of the PTCL subtypes; as well as An effective amount of a therapeutic agent is administered to the subject, wherein the therapeutic agent is configured to treat the identified PTCL subtype.
44. The method of claim 43, wherein the therapeutic agent comprises a histone deacetylase (HDAC) inhibitor, an antifolate, an alkylating agent, a proteasome 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, ) inhibitors, mammalian target of rapamycin (mTOR) pathway inhibitors, immune checkpoint inhibitors, hypomethylating agents, anti-cluster of differentiation 52 (CD52) antibodies, immunomodulatory drugs, anti-inducible T-cell co-stimulator (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 a combination thereof.
45. The method of claim 44, wherein: The HDAC inhibitor includes romidepsin, belinostat, panobinostat or a combination thereof; The antifolates include pralatrexate; The alkylating agents include bendamustine; The proteasome inhibitors include bortezomib; The antibody-drug conjugate comprises brentuximab; The PI3K inhibitor includes duvelixibu, tenalixibu or a combination thereof; the JAK inhibitor includes ruxolitinib; The ALK inhibitor includes crizotinib; The mTOR pathway inhibitor includes everolimus; The hypomethylating agent includes 5-azacitidine (5-Aza); The anti-CD52 antibody includes alemtuzumab; The ImiD includes lenalidomide; The CCR4 inhibitors include moglizumab; The IDH inhibitors include ensidipine; The BCL2 inhibitors include venetoclax; The anti-CD25 antibody includes tecamimumab; The calcineurin inhibitors include cyclosporine A; The SYK inhibitor includes selenitinib; The bispecific antibody comprises AFM1 3; The chimeric antigen receptor T (CAR-T) cells comprise CD30, CD7, or both; or a combination thereof.
46. The method of claim 43, wherein when the PTCL subtype is identified as PTCL-NOS, the therapeutic agent comprises romidepsin, belinostat, brentuximab, duvelixib, or a combination thereof.
47. The method of claim 43, wherein when 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.
48. The method of claim 43, wherein when the PTCL subtype is identified as ATLL, the therapeutic agent comprises a NOTCH inhibitor, a hepatocyte growth factor (HGF) inhibitor, a cMET inhibitor, or muglizumab.
49. The method of claim 43, wherein when the PTCL subtype is identified as ALK-ALCL, the therapeutic agent comprises brentuximab.
50. The method of claim 43, wherein when the PTCL subtype is identified as ALK+ALCL, the therapeutic agent comprises an ALK inhibitor, a platelet-derived growth factor receptor β (PDGFRβ) inhibitor, or a combination thereof.
51. The method of claim 43, wherein when the PTCL subtype is identified as ENKTL, the therapeutic agent comprises a platelet-derived growth factor receptor alpha (PDGFRα) inhibitor.
