Concurrent detection of targeted copy number variants and mutations using a myeloid malignancy next generation sequencing panel allows comprehensive genetic analysis using a single testing strategy.

Currently, comprehensive genetic testing of myeloid malignancies requires multiple testing strategies with high costs. Somatic mutations can be detected by next generation sequencing (NGS) but copy number variants (CNVs) require cytogenetic methods including karyotyping, fluorescence in situ hybidization and microarray. Here, we evaluated a new method for CNV detection using read depth data derived from a targeted NGS mutation panel. In a cohort of 270 samples, we detected pathogenic mutations in 208 samples and targeted CNVs in 68 cases. The most frequent CNVs were 7q deletion including LUC7L2 and EZH2, TP53 deletion, ETV6 deletion, gain of RAD21 on 8q, and 5q deletion, including NSD1 and NPM1. We were also able to detect exon-level duplications, including so-called KMT2A (MLL) partial tandem duplication, in 9 cases. In the 63 cases that were negative for mutations, targeted CNVs were observed in 4 cases. Targeted CNV detection by NGS had very high concordance with single nucleotide polymorphism microarray, the current gold standard. We found that ETV6 deletion was strongly associated with TP53 alterations and 7q deletion was associated with mutations in TP53, KRAS and IDH1. This proof-of-concept study demonstrates the feasibility of using the same NGS data to simultaneously detect both somatic mutations and targeted CNVs.

Targeted next-generation sequencing identifies a subset of idiopathic hypereosinophilic syndrome with features similar to chronic eosinophilic leukemia, not otherwise specified.

The distinction between chronic eosinophilic leukemia, not otherwise specified and idiopathic hypereosinophilic syndrome largely relies on clonality assessment. Prior to the advent of next-generation sequencing, clonality was usually determined by cytogenetic analysis. We applied targeted next-generation sequencing panels designed for myeloid neoplasms to bone marrow specimens from a cohort of idiopathic hypereosinophilic syndrome patients (n=51), and assessed the significance of mutations in conjunction with clinicopathological features. The findings were further compared with those of 17 chronic eosinophilic leukemia, not otherwise specified patients defined by their abnormal cytogenetics and/or increased blasts. Mutations were detected in 14/51 idiopathic hypereosinophilic syndrome patients (idiopathic hypereosinophilic syndrome/next-generation sequencing-positive) (28%), involving single gene in 7 and ≥2 in 7 patients. The more frequently mutated genes included ASXL1 (43%), TET2 (36%), EZH2 (29%), SETBP1 (22%), CBL (14%), and NOTCH1 (14%). Idiopathic hypereosinophilic syndrome/next-generation sequencing-positive patients showed a number of clinical features and bone marrow findings resembling chronic eosinophilic leukemia, not otherwise specified. Chronic eosinophilic leukemia, not otherwise specified patients showed a disease-specific survival of 14.4 months, markedly inferior to idiopathic hypereosinophilic syndrome/next-generation sequencing-negative (P<0.001), but not significantly different from idiopathic hypereosinophilic syndrome/next-generation sequencing-positive (P=0.117). These data suggest that targeted next-generation sequencing helps to establish clonality in a subset of patients with hypereosinophilia that would otherwise be classified as idiopathic hypereosinophilic syndrome. In conjunction with other diagnostic features, mutation data can be used to establish a diagnosis of chronic eosinophilic leukemia, not otherwise specified in patients presenting with hypereosinophilia.

Detection of BCR-ABL1 mutations that confer tyrosine kinase inhibitor resistance using massively parallel, next generation sequencing.

Detection of BCR-ABL1 mutations that confer resistance to tyrosine kinase inhibitors is important for management of patients with t(9;22);BCR-ABL1-positive (Ph+) leukemias. Testing is often performed using Sanger sequencing (SS) which has relatively poor sensitivity. Given the widespread adoption of next generation sequencing (NGS), we sought to reevaluate the testing in the context of NGS methods. We developed an NGS-based BCR-ABL1 mutation test on the Ion Torrent Personal Genome Machine (PGM) to test for resistance mutations, primarily in the kinase domain in BCR-ABL1. We analyzed 508 clinical samples from patients with Ph+ leukemias. In a subset of these samples (n = 97), we conducted a comparison of the NGS results to a classical SS-based test. NGS facilitated detection of low-level mutations (<20 % allele frequency) that were not detectable by SS. In a subset of cases with multiple mutations, NGS was also able to determine if two mutations were on the same molecule (compound) or on separate molecules (polyclonal) but this was limited by the distance between mutated positions and by the effects of apparent distance-dependent PCR recombination. We found 22 compound mutations that centered on one or two key residues including two novel compound mutants: Q252H/Y253H and F311Y/F359I. The advantages of NGS make it a superior method for inventorying BCR-ABL1 resistance mutations. However, data analysis may be complicated by short read lengths and the effects of PCR recombination.

BCR-ABL1 compound mutations in tyrosine kinase inhibitor-resistant CML: frequency and clonal relationships.

BCR-ABL1 compound mutations can confer high-level resistance to imatinib and other ABL1 tyrosine kinase inhibitors (TKIs). The third-generation ABL1 TKI ponatinib is effective against BCR-ABL1 point mutants individually, but remains vulnerable to certain BCR-ABL1 compound mutants. To determine the frequency of compound mutations among chronic myeloid leukemia patients on ABL1 TKI therapy, in the present study, we examined a collection of patient samples (N = 47) with clear evidence of 2 BCR-ABL1 kinase domain mutations by direct sequencing. Using a cloning and sequencing method, we found that 70% (33/47) of double mutations detected by direct sequencing were compound mutations. Sequential, branching, and parallel routes to compound mutations were common. In addition, our approach revealed individual and compound mutations not detectable by direct sequencing. The frequency of clones harboring compound mutations with more than 2 missense mutations was low (10%), whereas the likelihood of silent mutations increased disproportionately with the total number of mutations per clone, suggesting a limited tolerance for BCR-ABL1 kinase domain missense mutations. We conclude that compound mutations are common in patients with sequencing evidence for 2 BCR-ABL1 mutations and frequently reflect a highly complex clonal network, the evolution of which may be limited by the negative impact of missense mutations on kinase function.

A novel approach to quantitating leukemia fusion transcripts by qRT-PCR without the need for standard curves.

Acute myeloid leukemia patients with recurrent cytogenetic abnormalities including inv(16);CBFB-MYH11 and t(15;17);PML-RARA may be assessed by monitoring the levels of the corresponding abnormal fusion transcripts by quantitative reverse transcription-PCR (qRT-PCR). Such testing is important for evaluating the response to therapy and for the detection of early relapse. Existing qRT-PCR methods are well established and in widespread use in clinical laboratories but they are laborious and require the generation of standard curves. Here, we describe a new method to quantitate fusion transcripts in acute myeloid leukemia by qRT-PCR without the need for standard curves. Our approach uses a plasmid calibrator containing both a fusion transcript sequence and a reference gene sequence, representing a perfect normalized copy number (fusion transcript copy number/reference gene transcript copy number; NCN) of 1.0. The NCN of patient specimens can be calculated relative to that of the single plasmid calibrator using experimentally derived PCR efficiency values. We compared the data obtained using the plasmid calibrator method to commercially available assays using standard curves and found that the results obtained by both methods are comparable over a broad range of values with similar sensitivities. Our method has the advantage of simplicity and is therefore lower in cost and may be less subject to errors that may be introduced during the generation of standard curves.

T-cell clonality assessment by next-generation sequencing improves detection sensitivity in mycosis fungoides.

BACKGROUND: T-cell receptor (TCR) clonality assessment is a principal diagnostic test in the management of mycosis fungoides (MF). However, current polymerase chain reaction-based methods may produce ambiguous results, often because of low abundance of clonal T lymphocytes, resulting in weak clonal peaks that cannot be size-resolved by contemporary capillary electrophoresis (CE). OBJECTIVE: We sought to determine if next-generation sequencing (NGS)-based detection has increased sensitivity for T-cell clonality over CE-based detection in MF. METHODS: Clonality was determined by an NGS-based method in which the TCR-γ variable region was polymerase chain reaction amplified and the products sequenced to establish the identity of rearranged variable and joining regions. RESULTS: Of the 35 MF cases tested, 29 (85%) showed a clonal T-cell rearrangement by NGS, compared with 15 (44%) by standard CE detection. Three patients with MF had follow-up testing that showed identical, clonal TCR sequences in subsequent skin biopsy specimens. LIMITATIONS: Clonal T-cell populations have been described in benign conditions; evidence of clonality alone, by any method, is not sufficient for diagnosis. CONCLUSION: TCR clonality assessment by NGS has superior sensitivity compared with CE-based detection. Further, NGS enables tracking of specific clones across multiple time points for more accurate identification of recurrent MF.

Targeted next generation sequencing of clinically significant gene mutations and translocations in leukemia.

Leukemias are currently subclassified based on the presence of recurrent cytogenetic abnormalities and gene mutations. These molecular findings are the basis for risk-adapted therapy; however, such data are generally obtained by disparate methods in the clinical laboratory, and often rely on low-resolution techniques such as fluorescent in situ hybridization. Using targeted next generation sequencing, we demonstrate that the full spectrum of prognostically significant gene mutations including translocations, single nucleotide variants (SNVs), and insertions/deletions (indels) can be identified simultaneously in multiplexed sequence data. As proof of concept, we performed hybrid capture using a panel of 20 genes implicated in leukemia prognosis (covering a total of 1 Mbp) from five leukemia cell lines including K562, NB4, OCI-AML3, kasumi-1, and MV4-11. Captured DNA was then sequenced in multiplex on an Illumina HiSeq. Using an analysis pipeline based on freely available software we correctly identified DNA-level translocations in three of the three cell lines where translocations were covered by our capture probes. Furthermore, we found all published gene mutations in commonly tested genes including NPM1, FLT3, and KIT. The same methodology was applied to DNA extracted from the bone marrow of a patient with acute myeloid leukemia, and identified a t(9;11) translocation with single base accuracy as well other gene mutations. These results indicate that targeted next generation sequencing can be successfully applied in the clinical laboratory to identify a full spectrum of DNA mutations ranging from SNVs and indels to translocations. Such methods have the potential to both greatly streamline and improve the accuracy of DNA-based diagnostics.

Dasatinib overcomes stroma-based resistance to the FLT3 inhibitor quizartinib using multiple mechanisms.

FLT3-ITD mutations occur in 20-30% of AML patients and are associated with aggressive disease. Patients with relapsed FLT3-mutated disease respond well to 2nd generation FLT3 TKIs but inevitably relapse within a short timeframe. In this setting, until overt relapse occurs, the bone marrow microenvironment facilitates leukemia cell survival despite continued on-target inhibition. We demonstrate that human bone marrow derived conditioned medium (CM) protects FLT3-ITD+ AML cells from the 2nd generation FLT3 TKI quizartinib and activates STAT3 and STAT5 in leukemia cells. Extrinsic activation of STAT5 by CM is the primary mediator of leukemia cell resistance to FLT3 inhibition. Combination treatment with quizartinib and dasatinib abolishes STAT5 activation and significantly reduces the IC50 of quizartinib in FLT3-ITD+ AML cells cultured in CM. We demonstrate that CM protects FLT3-ITD+ AML cells from the inhibitory effects of quizartinib on glycolysis and that this is partially reversed by treating cells with the combination of quizartinib and dasatinib. Using a doxycycline-inducible STAT5 knockdown in the FLT3-ITD+ MOLM-13 cell line, we show that dasatinib-mediated suppression of leukemia cell glycolytic activity is STAT5-independent and provide a preclinical rationale for combination treatment with quizartinib and dasatinib in FLT3-ITD+ AML.

Induction of granzyme B expression in T-cell receptor/CD28-stimulated human regulatory T cells is suppressed by inhibitors of the PI3K-mTOR pathway.

BACKGROUND: Regulatory T cells (Tregs) can employ a cell contact- and granzyme B-dependent mechanism to mediate suppression of bystander T and B cells. Murine studies indicate that granzyme B is involved in the Treg-mediated suppression of anti-tumor immunity in the tumor microenvironment and in the Treg-mediated maintenance of allograft survival. In spite of its central importance, a detailed study of granzyme B expression patterns in human Tregs has not been performed. RESULTS: Our data demonstrated that natural Tregs freshly isolated from the peripheral blood of normal adults lacked granzyme B expression. Tregs subjected to prolonged TCR and CD28 triggering, in the presence of IL-2, expressed high levels of granzyme B but CD3 stimulation alone or IL-2 treatment alone failed to induce granzyme B. Treatment of Tregs with the mammalian target of rapamycin (mTOR) inhibitor, rapamycin or the PI3 kinase (PI3K) inhibitor LY294002 markedly suppressed granzyme B expression. However, neither rapamycin, as previously reported by others, nor LY294002 inhibited Treg proliferation or induced significant cell death in TCR/CD28/IL-2 stimulated cells. The proliferation rate of Tregs was markedly higher than that of CD4+ conventional T cells in the setting of rapamycin treatment. Tregs expanded by CD3/CD28/IL-2 stimulation without rapamycin demonstrated increased in vitro cytotoxic activity compared to Tregs expanded in the presence of rapamycin in both short term (6 hours) and long term (48 hours) cytotoxicity assays. CONCLUSION: TCR/CD28 mediated activation of the PI3K-mTOR pathway is important for granyzme B expression but not proliferation in regulatory T cells. These findings may indicate that suppressive mechanisms other than granzyme B are utilized by rapamycin-expanded Tregs.

Genome-Wide Copy Number Variation Detection Using NGS: Data Analysis and Interpretation.

Copy number variants (CNVs) and copy neutral loss of heterozygosity (CN-LOH) represent important types of genomic abnormalities in cancer. Genomic DNA microarray serves as the current gold standard method for detecting genome-wide CNVs and CN-LOH. However, as next-generation sequencing (NGS) is widely used to detect gene variants in clinical testing, the ability of NGS to detect CNVs and CN-LOH has also been demonstrated. This chapter describes a protocol for detecting genome-wide large somatic CNVs and CN-LOH using a single nucleotide polymorphism (SNP) sequencing backbone. When combined with a targeted gene mutation panel, this strategy allows for simultaneous detection of somatic gene mutations and genome-wide CNVs and CN-LOH.

Genetic Testing in the Diagnosis and Biology of Myeloid Neoplasms (Excluding Acute Leukemias).

OBJECTIVES: The 2017 Workshop of the Society for Hematopathology/European Association for Haematopathology reviewed the role of genetic testing in the diagnosis of hematopoietic neoplasms, including non-acute leukemia myeloid malignancies. METHODS: The workshop panel assigned 98 submitted cases to the category of non-acute leukemia myeloid neoplasms, of which 13 were selected for oral presentation. RESULTS: Data from both conventional karyotyping and genetic sequencing had important impact on diagnosis, classification, and prognostication. However, some cases had genetic results that appeared discordant from the morphology and/or clinical features. Thus, the workshop underscored the need for careful management of genetic data by the pathologist and clinician, in the context of other findings. CONCLUSIONS: The workshop cases highlighted the significance of genetic aberrations in the diagnosis and treatment of non-acute leukemia myeloid neoplasms. Many genetic data have already been incorporated in the most recent World Health Organization classification, and undoubtedly they will factor increasingly in future classifications.

Molecular Fingerprinting of Anatomically and Temporally Distinct B-Cell Lymphoma Samples by Next-Generation Sequencing to Establish Clonal Relatedness.

CONTEXT.­: B-cell lymphomas exhibit balanced translocations that involve immunoglobulin loci and result from aberrant V(D)J recombination, class switch recombination, or somatic hypermutation. Although most of the breakpoints in the immunoglobulin loci occur in defined regions, those in the partner genes vary; therefore, it is unlikely that 2 independent clones would share identical breakpoints in both partners. Establishing whether a new lesion in a patient with history of lymphoma represents recurrence or a new process can be relevant. Polymerase chain reaction (PCR)-based clonality assays used in this setting rely only on evaluating the length of a given rearrangement. In contrast, next-generation sequencing (NGS) provides the exact translocation breakpoint at single-base resolution. OBJECTIVE.­: To determine if translocation breakpoint coordinates can serve as a molecular fingerprint unique to a distinct clonal population. DESIGN.­: Thirty-eight follicular lymphoma/diffuse large B-cell lymphoma samples collected from different anatomic sites and/or at different time points from 18 patients were analyzed by NGS. For comparison, PCR-based B-cell clonality and fluorescence in situ hybridization studies were performed on a subset of cases. RESULTS.­: IGH-BCL2 rearrangements were detected in all samples. The breakpoint coordinates on derivative chromosome(s) were identical in all samples from a given patient, but distinct between samples derived from different patients. Additionally, 5 patients carried a second rearrangement also with conserved breakpoint coordinates in the follow-up sample(s). CONCLUSIONS.­: Breakpoint coordinates in the immunoglobulin and partner genes can be used to establish clonal relatedness of anatomically/temporally distinct lesions. Additionally, an NGS-based approach has the potential to detect secondary translocations that may have prognostic and therapeutic significance.

Transcription factor Oct1 protects against hematopoietic stress and promotes acute myeloid leukemia.

A better understanding of the development and progression of acute myelogenous leukemia (AML) is necessary to improve patient outcome. Here we define roles for the transcription factor Oct1/Pou2f1 in AML and normal hematopoiesis. Inappropriate reactivation of the CDX2 gene is widely observed in leukemia patients and in leukemia mouse models. We show that Oct1 associates with the CDX2 promoter in both normal and AML primary patient samples, but recruits the histone demethylase Jmjd1a/Kdm3a to remove the repressive H3K9me2 mark only in malignant specimens. The CpG DNA immediately adjacent to the Oct1 binding site within the CDX2 promoter exhibits variable DNA methylation in healthy control blood and bone marrow samples, but complete demethylation in AML samples. In MLL-AF9-driven mouse models, partial loss of Oct1 protects from myeloid leukemia. Complete Oct1 loss completely suppresses leukemia but results in lethality from bone marrow failure. Loss of Oct1 in normal hematopoietic transplants results in superficially normal long-term reconstitution; however, animals become acutely sensitive to 5-fluorouracil, indicating that Oct1 is dispensable for normal hematopoiesis but protects blood progenitor cells against external chemotoxic stress. These findings elucidate a novel and important role for Oct1 in AML.

Nuclear-Cytoplasmic Transport Is a Therapeutic Target in Myelofibrosis.

PURPOSE: Myelofibrosis is a hematopoietic stem cell neoplasm characterized by bone marrow reticulin fibrosis, extramedullary hematopoiesis, and frequent transformation to acute myeloid leukemia. Constitutive activation of JAK/STAT signaling through mutations in JAK2, CALR, or MPL is central to myelofibrosis pathogenesis. JAK inhibitors such as ruxolitinib reduce symptoms and improve quality of life, but are not curative and do not prevent leukemic transformation, defining a need to identify better therapeutic targets in myelofibrosis. EXPERIMENTAL DESIGN: A short hairpin RNA library screening was performed on JAK2V617F-mutant HEL cells. Nuclear-cytoplasmic transport (NCT) genes including RAN and RANBP2 were among top candidates. JAK2V617F-mutant cell lines, human primary myelofibrosis CD34+ cells, and a retroviral JAK2V617F-driven myeloproliferative neoplasms mouse model were used to determine the effects of inhibiting NCT with selective inhibitors of nuclear export compounds KPT-330 (selinexor) or KPT-8602 (eltanexor). RESULTS: JAK2V617F-mutant HEL, SET-2, and HEL cells resistant to JAK inhibition are exquisitely sensitive to RAN knockdown or pharmacologic inhibition by KPT-330 or KPT-8602. Inhibition of NCT selectively decreased viable cells and colony formation by myelofibrosis compared with cord blood CD34+ cells and enhanced ruxolitinib-mediated growth inhibition and apoptosis, both in newly diagnosed and ruxolitinib-exposed myelofibrosis cells. Inhibition of NCT in myelofibrosis CD34+ cells led to nuclear accumulation of p53. KPT-330 in combination with ruxolitinib-normalized white blood cells, hematocrit, spleen size, and architecture, and selectively reduced JAK2V617F-mutant cells in vivo. CONCLUSIONS: Our data implicate NCT as a potential therapeutic target in myelofibrosis and provide a rationale for clinical evaluation in ruxolitinib-exposed patients with myelofibrosis.

Combining the Allosteric Inhibitor Asciminib with Ponatinib Suppresses Emergence of and Restores Efficacy against Highly Resistant BCR-ABL1 Mutants.

BCR-ABL1 point mutation-mediated resistance to tyrosine kinase inhibitor (TKI) therapy in Philadelphia chromosome-positive (Ph+) leukemia is effectively managed with several approved drugs, including ponatinib for BCR-ABL1T315I-mutant disease. However, therapy options are limited for patients with leukemic clones bearing multiple BCR-ABL1 mutations. Asciminib, an allosteric inhibitor targeting the myristoyl-binding pocket of BCR-ABL1, is active against most single mutants but ineffective against all tested compound mutants. We demonstrate that combining asciminib with ATP site TKIs enhances target inhibition and suppression of resistant outgrowth in Ph+ clinical isolates and cell lines. Inclusion of asciminib restores ponatinib's effectiveness against currently untreatable compound mutants at clinically achievable concentrations. Our findings support combining asciminib with ponatinib as a treatment strategy for this molecularly defined group of patients.

Similar expression profiles in CD34+ cells from chronic phase chronic myeloid leukemia patients with and without deep molecular responses to nilotinib.

The life expectancy of patients with chronic phase chronic myeloid leukemia on tyrosine kinase inhibitor therapy now approaches that of the general population. Approximately 60% of patients treated with second generation tyrosine kinase inhibitors achieve a deep molecular response, the prerequisite for a trial of treatment-free remission. Those patients unlikely to achieve deep molecular response may benefit from more intensive therapy up front. To identify biomarkers predicting deep molecular response we performed transcriptional profiling on CD34+ progenitor cells from newly diagnosed chronic phase chronic myeloid leukemia patients treated with nilotinib on a prospective clinical trial. Using unsupervised and targeted analytical strategies, we show that gene expression profiles are similar in patients with and without subsequent deep molecular response. This result is in contrast to the distinct expression signature of CD34+ chronic phase chronic myeloid leukemia patients failing to achieve a cytogenetic response on imatinib and suggests that deep molecular response to second-generation tyrosine kinase inhibitors is governed by the biology of more primitive chronic myeloid leukemia cells or extrinsic factors.

Detection of genome-wide copy number variants in myeloid malignancies using next-generation sequencing.

AIMS: Genetic abnormalities, including copy number variants (CNV), copy number neutral loss of heterozygosity (CN-LOH) and gene mutations, underlie the pathogenesis of myeloid malignancies and serve as important diagnostic, prognostic and/or therapeutic markers. Currently, multiple testing strategies are required for comprehensive genetic testing in myeloid malignancies. The aim of this proof-of-principle study was to investigate the feasibility of combining detection of genome-wide large CNVs, CN-LOH and targeted gene mutations into a single assay using next-generation sequencing (NGS). METHODS: For genome-wide CNV detection, we designed a single nucleotide polymorphism (SNP) sequencing backbone with 22 762 SNP regions evenly distributed across the entire genome. For targeted mutation detection, 62 frequently mutated genes in myeloid malignancies were targeted. We combined this SNP sequencing backbone with a targeted mutation panel, and sequenced 9 healthy individuals and 16 patients with myeloid malignancies using NGS. RESULTS: We detected 52 somatic CNVs, 11 instances of CN-LOH and 39 oncogenic mutations in the 16 patients with myeloid malignancies, and none in the 9 healthy individuals. All CNVs and CN-LOH were confirmed by SNP microarray analysis. CONCLUSIONS: We describe a genome-wide SNP sequencing backbone which allows for sensitive detection of genome-wide CNVs and CN-LOH using NGS. This proof-of-principle study has demonstrated that this strategy can provide more comprehensive genetic profiling for patients with myeloid malignancies using a single assay.

The ratio of FOXP3+ regulatory T cells to granzyme B+ cytotoxic T/NK cells predicts prognosis in classical Hodgkin lymphoma and is independent of bcl-2 and MAL expression.

We studied the prognostic importance of tumor-infiltrating regulatory T lymphocytes (Tregs) and cytotoxic T/NK lymphocytes (CTLs) in 98 diagnostic biopsy specimens from patients with classical Hodgkin lymphoma (cHL). Immunohistochemical analysis was performed for FOXP3 to identify Tregs and for granzyme B (GrB) to identify activated CTLs. Failure-free survival (FFS) and overall survival (OS) were clinical end points. Patients with fewer than 25 FOXP3+ cells per high-power field (HPF) had a mean +/- SD 5-year FFS of 64% +/- 7% vs 85% +/- 5% for patients with 25 or more FOXP3+ cells/HPF (P = .05). A FOXP3/GrB ratio of 1 or less was associated with poor FFS (46% +/- 10% vs 86% +/- 4%; P < .001) and OS (67% +/- 10% vs 93% +/- 3%; P < .001). When prior available MAL and bcl-2 expression data were included in a multivariate analysis of all clinical and biologic factors, a FOXP3/GrB ratio of 1 or less and tumor cell expression of MAL and bcl-2 all independently predicted poor FFS. This demonstrates the importance of evaluating tumor cell markers and the tumor immune infiltrate when considering biologic prognostic markers in cHL.