Search

Covalent inhibitors of Bruton tyrosine kinase (BTK) have transformed the therapy of chronic lymphocytic leukemia (CLL), but continuous therapy has been complicated by the development of resistance. The most common resistance mechanism in patients whose disease progresses on covalent BTK inhibitors (BTKis) is a mutation in the BTK 481 cysteine residue to which the inhibitors bind covalently. Pirtobrutinib is a highly selective, noncovalent BTKi with substantial clinical activity in patients whose disease has progressed on covalent BTKi, regardless of BTK mutation status. Using in vitro ibrutinib-resistant models and cells from patients with CLL, we show that pirtobrutinib potently inhibits BTK-mediated functions including B-cell receptor (BCR) signaling, cell viability, and CCL3/CCL4 chemokine production in both BTK wild-type and C481S mutant CLL cells. We demonstrate that primary CLL cells from responding patients on the pirtobrutinib trial show reduced BCR signaling, cell survival, and CCL3/CCL4 chemokine secretion. At time of progression, these primary CLL cells show increasing resistance to pirtobrutinib in signaling inhibition, cell viability, and cytokine production. We employed longitudinal whole-exome sequencing on 2 patients whose disease progressed on pirtobrutinib and identified selection of alternative-site BTK mutations, providing clinical evidence that secondary BTK mutations lead to resistance to noncovalent BTKis., (© 2023 by The American Society of Hematology. Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0), permitting only noncommercial, nonderivative use with attribution. All other rights reserved.)

This study investigated ibrutinib plus obinutuzumab in relapsed/refractory CLL, evaluating tolerability of 3 sequencing regimens as well as overall safety and efficacy. Fifty-two patients were initially randomized 1:1:1 to receive either obinutuzumab 1 month before ibrutinib initiation, ibrutinib 1 month prior to obinutuzumab initiation, or to start both drugs concomitantly. Higher rates of infusion-related reactions were observed with the first sequence, and only the latter 2 cohorts were expanded. Grade 4 hematologic toxicity was uncommon, and notable all-grade non-hematologic toxicities included bruising (58%), hypertension (46%), arthralgia (38%), diarrhea (37%), transaminitis (35%), atrial fibrillation (21%), and serious infection (17%). Best overall response rate was 96% (including 40% CR and 56% PR). Best rates of undetectable minimal residual disease in peripheral blood and bone marrow were 27% and 19%, respectively. With a median follow-up of 41.5 months, four-year progression-free and overall survival rates are 74% and 93%, respectively. Correlative studies demonstrated that serum CCL4 and CXCL13 levels were associated with clinical response, and BH3 profiling revealed increased BCL-2 and BCL-xL dependence in CLL cells from patients on treatment. Overall, ibrutinib plus obinutuzumab was highly active, with a manageable safety profile, supporting further investigation of this type of approach in relapsed/refractory CLL., (© 2023. The Author(s), under exclusive licence to Springer Nature Limited.)

Prostate cancer has a heterogeneous prognosis. Most previous studies have focused on the identification of prognostic biomarkers in the prostate cancer tumor. However, it is increasingly recognized that the tumor microenvironment contributes to prostate cancer aggressiveness and progression. We therefore examined whole transcriptome expression of the prostate stroma and associations with aggressive and lethal prostate cancer. We performed RNA sequencing (Illumina TruSeq Exome Capture) of 272 tumor-adjacent and 120 benign-adjacent macrodissected prostate stromal samples from 293 men with prostate cancer from the Health Professionals Follow-up Study and Physicians' Health Study. We performed differential expression analysis comparing gene expression and pathways by Gleason score and lethal outcome. We also tested a previously developed stromal gene signature of Gleason score in these datasets. Comparing high- with low-Gleason score cancers, 26 genes (P < 0.001) and 12 pathways (FDR < 0.20) were significantly differentially expressed in tumor-adjacent stroma, including pathways related to stroma composition remodeling and DNA repair, with 73 genes and 65 pathways significant in benign-adjacent stroma. Comparing lethal with nonlethal prostate cancer, 11 genes were differentially expressed in tumor-adjacent and 15 genes in benign-adjacent stroma, and pathways involved in inflammatory response were differentially enriched in both tumor and benign-adjacent stroma. In addition, our previously identified Gleason stromal gene signature was validated to be associated with Gleason score in these data. Implications: Our study uncovers stroma-specific genes and pathways that are differentially enriched with high Gleason score and lethal prostate cancer, demonstrating that the molecular investigation of the tumor microenvironment can provide additional information about prostate cancer prognosis., (©2022 American Association for Cancer Research.)

Purpose: Germline missense variants of unknown significance in cancer-related genes are increasingly being identified with the expanding use of next-generation sequencing. The ataxia telangiectasia-mutated ( ATM ) gene on chromosome 11 has more than 1,000 germline missense variants of unknown significance and is a tumor suppressor. We aimed to determine if rare germline ATM variants are more frequent in chronic lymphocytic leukemia (CLL) compared with other hematologic malignancies and if they influence the clinical characteristics of CLL., Methods: We identified 3,128 patients (including 825 patients with CLL) in our hematologic malignancy clinic who had received clinical-grade sequencing of the entire coding region of ATM . We ascertained the comparative frequencies of germline ATM variants in categories of hematologic neoplasms, and, in patients with CLL, we determined whether these variants affected CLL-associated characteristics such as somatic 11q deletion., Results: Rare germline ATM variants are present in 24% of patients with CLL, significantly greater than that in patients with other lymphoid malignancies (16% prevalence), myeloid disease (15%), or no hematologic neoplasm (14%). Patients with CLL with germline ATM variants are younger at diagnosis and twice as likely to have 11q deletion. The ATM variant p.L2307F is present in 3% of patients with CLL, is associated with a three-fold increase in rates of somatic 11q deletion, and is a hypomorph in cell-based assays., Conclusion: Germline ATM variants cluster within CLL and affect the phenotype of CLL that develops, implying that some of these variants (such as ATM p.L2307F) have functional significance and should not be ignored. Further studies are needed to determine whether these variants affect the response to therapy or account for some of the inherited risk of CLL.

PI3Kδ inhibitors are approved for the therapy of B cell malignancies, but their clinical use has been limited by unpredictable autoimmune toxicity, despite promising efficacy and evidence that toxicity is associated with improved clinical outcomes. Prior phenotypic evaluation by CyTOF has identified increases in activated CD8 T cells with activation of Th17 T cells, as well as decreases in Tregs, particularly in patients with toxicity. Here we sought to further understand the effects of idelalisib and duvelisib in vitro, and demonstrate that both idelalisib and duvelisib can inhibit T cell proliferation as well as Th1 and Treg differentiation in vitro, while promoting Th2 and Th17 differentiation. We further demonstrate directly using intracellular flow cytometry that autoimmune toxicity in patients is associated with higher absolute numbers of CD4 and CD8 T cells with Th17 differentiation in peripheral blood prior to therapy, and that gastrointestinal tissues from patients with active autoimmune complications of PI3Kδ inhibitors show infiltration with Th17 +  T cells. These same tissues show depletion of Tregs as compared to CLL patients without toxicity, suggesting that loss of Tregs may be permissive for Th17 activation to lead to autoimmune toxicity. Clinical trials to restore this balance are warranted., (© 2023. The Author(s).)

Genomic and transcriptomic data have been generated across a wide range of prostate cancer (PCa) study cohorts. These data can be used to better characterize the molecular features associated with clinical outcomes and to test hypotheses across multiple, independent patient cohorts. In addition, derived features, such as estimates of cell composition, risk scores, and androgen receptor (AR) scores, can be used to develop novel hypotheses leveraging existing multi-omic datasets. The full potential of such data is yet to be realized as independent datasets exist in different repositories, have been processed using different pipelines, and derived and clinical features are often not provided or unstandardized. Here, we present the curatedPCaData R package, a harmonized data resource representing >2900 primary tumor, >200 normal tissue, and >500 metastatic PCa samples across 19 datasets processed using standardized pipelines with updated gene annotations. We show that meta-analysis across harmonized studies has great potential for robust and clinically meaningful insights. curatedPCaData is an open and accessible community resource with code made available for reproducibility., Competing Interests: COMPETING INTERESTS The authors declare the following competing interests: J.C.C. is co-founder of PrecisionProfile and OncoRX Insights. All other authors declare no competing interests.

Older age at diagnosis is consistently associated with worse clinical outcomes in prostate cancer. We sought to characterize gene expression profiles of prostate tumor tissue by age at diagnosis. We conducted a discovery analysis in The Cancer Genome Atlas prostate cancer dataset (n = 320; 29% of men >65 years at diagnosis), using linear regressions of age at diagnosis and mRNA expression and adjusting for TMPRSS2:ERG fusion status and race. This analysis identified 13 age-related candidate genes at FDR < 0.1, six of which were also found in an analysis additionally adjusted for Gleason score. We then validated the 13 age-related genes in a transcriptome study nested in the Health Professionals Follow-up Study and Physicians' Health Study (n = 374; 53% of men >65 years). Gene expression differences by age in the 13 candidate genes were directionally consistent, and age at diagnosis was weakly associated with the 13-gene score. However, the age-related genes were not consistently associated with risk of metastases and prostate cancer-specific death. Collectively, these findings argue against tumor genomic differences as a main explanation for age-related differences in prostate cancer prognosis., Prevention Relevance: Older age at diagnosis is consistently associated with worse clinical outcomes in prostate cancer. This study with independent discovery and validation sets and long-term follow-up suggests that prevention of lethal prostate cancer should focus on implementing appropriate screening, staging, and treatment among older men without expecting fundamentally different tumor biology., (©2022 American Association for Cancer Research.)

Importance: It is unclear whether ERBB2-low breast cancer should be considered an individual biologic subtype distinct from ERBB2-0 breast cancer., Objective: To investigate whether low ERBB2 expression is associated with distinct clinicopathologic characteristics and prognosis among patients with hormone receptor (HR)-positive and triple-negative breast cancer (TNBC)., Design, Setting, and Participants: This cohort study was conducted using data from a prospectively maintained institutional database on all consecutive patients with breast cancer undergoing surgery between January 2016 and March 2021 at Dana-Farber Brigham Cancer Center. The study included 5235 patients with stage I through III, ERBB2-negative invasive breast cancer. Tumors were classified as ERBB2-low if they had an ERBB2 immunohistochemical (IHC) score of 1+ or 2+ with negative in situ hybridization assay and ERBB2-0 if they had an ERBB2 IHC score of 0. Data were analyzed from September 2021 through January 2022., Exposures: Standard treatment according to institutional guidelines., Main Outcomes and Measures: Comparison of clinicopathologic characteristics and disease outcomes (pathologic complete response rate [pCR], disease-free survival, distant disease-free survival, and overall survival) between patients with ERBB2-low and ERBB2-0 breast cancer., Results: Among 5235 patients with ERBB2-negative invasive breast cancer (5191 [99.2%] women; median [range] age at primary surgery, 59.0 [21.0-95.0] years), 2917 patients (55.7%) and 2318 patients (44.3%) had ERBB2-low and ERBB2-0 tumors, respectively. Expression of HR was significantly more common among ERBB2-low compared with ERBB2-0 tumors (2643 patients [90.6%] vs 1895 patients [81.8%]; P < .001). The rate of ERBB2-low tumors increased progressively, from 296 of 739 estrogen receptor (ER)-negative tumors (40.1%) to 31 of 67 ER-low (ie, ER 1%-9%) tumors (46.3%), 37 of 67 ER-moderate (ie, ER, 10%-49%) tumors (55.2%), 2047 of 3542 ER-high (ie, ER, 50%-95%) tumors (57.8%), and 499 of 803 ER-very high (ie, ER > 95%) tumors (62.1%) (P < .001). Among 675 patients receiving neoadjuvant chemotherapy, those with ERBB2-0 tumors experienced higher pCR rates (95 patients [26.8%] vs 53 patients [16.6%]; P = .002). However, there were no statistically significant differences in pCR rate between ERBB2-low and ERBB2-0 tumors when separately analyzing HR-positive, ER-low, HR-positive without ER-low, or TNBC tumors. In exploratory survival analysis, no differences by ERBB2-low expression in disease-free survival, distant disease-free survival, or overall survival were observed among patients with HR-positive tumors or TNBC., Conclusions and Relevance: The results of this cohort study did not support the interpretation of ERBB2-low breast cancer as a distinct biologic subtype. ERBB2-low expression was positively associated with level of ER expression, and ER-low tumors were enriched among ERBB2-0 tumors, suggesting that, given the worse prognosis of ER-low tumors, they may be associated with confounding of prognostic analyses of ERBB2-low expression.

Several PI3Kδ inhibitors are approved for the therapy of B cell malignancies, but their clinical use has been limited by unpredictable autoimmune toxicity. We have recently reported promising efficacy results in treating chronic lymphocytic leukemia (CLL) patients with combination therapy with the PI3Kδγ inhibitor duvelisib and fludarabine cyclophosphamide rituximab (FCR) chemoimmunotherapy, but approximately one-third of patients develop autoimmune toxicity. We show here that duvelisib FCR treatment in an upfront setting modulates both CD4 and CD8 T cell subsets as well as pro-inflammatory cytokines. Decreases in naive and central memory CD4 T cells and naive CD8 T cells occur with treatment, while activated CD8 T cells, granzyme positive Tregs, and Th17 CD4 and CD8 T cells all increase with treatment, particularly in patients with toxicity. Cytokines associated with Th17 activation (IL-17A and IL-21) are also relatively elevated in patients with toxicity. The only CLL feature associated with toxicity was increased priming for apoptosis at baseline, with a significant decrease during the first week of duvelisib. We conclude that an increase in activated CD8 T cells with activation of Th17 T cells, in the context of lower baseline Tregs and greater CLL resistance to duvelisib, is associated with duvelisib-related autoimmune toxicity., (© 2021. The Author(s), under exclusive licence to Springer Nature Limited.)

Inhibitors of Bruton tyrosine kinase (BTK) and phosphatidylinositol 3-kinase δ (PI3Kδ) that target the B-cell receptor (BCR) signaling pathway have revolutionized the treatment of chronic lymphocytic leukemia (CLL). Mutations associated with resistance to BTK inhibitors have been identified, but limited data are available on mechanisms of resistance to PI3Kδ inhibitors. Here we present findings from longitudinal whole-exome sequencing of cells from patients with multiply relapsed CLL (N = 28) enrolled in trials of PI3K inhibitors. The nonresponder subgroup was characterized by baseline activating mutations in MAP2K1, BRAF, and KRAS genes in 60% of patients. PI3Kδ inhibition failed to inhibit ERK phosphorylation (pERK) in nonresponder CLL cells with and without mutations, whereas treatment with a MEK inhibitor rescued ERK inhibition. Overexpression of MAP2K1 mutants in vitro led to increased basal and inducible pERK and resistance to idelalisib. These data demonstrate that MAPK/ERK activation plays a key role in resistance to PI3Kδ inhibitors in CLL and provide a rationale for therapy with a combination of PI3Kδ and ERK inhibitors., (© 2021 by The American Society of Hematology.)

Fludarabine, cyclophosphamide, and rituximab (FCR) is highly effective initial therapy for younger patients with chronic lymphocytic leukemia (CLL); however, most eventually relapse. Duvelisib is a delta/gamma PI3K inhibitor approved for relapsed/refractory CLL. We conducted an investigator-initiated, phase 1b/2 study of duvelisib + FCR (DFCR) as initial treatment for CLL patients aged ≤65. A standard 3 + 3 design included two dose levels of duvelisib (25 mg qd and 25 mg bid). Duvelisib was given for 1 week, then with standard FCR added for up to six 28-day cycles, then up to 2 years of duvelisib maintenance. Thirty-two patients were enrolled. The phase 2 dose of duvelisib was identified as 25 mg bid. Hematologic toxicity was common, and all-grade non-hematologic toxicities included transaminitis (28%), febrile neutropenia (22%), pneumonia (19%), and colitis (6%). The best overall response rate by ITT was 88% (56% CR/CRi and 32% PR). The best rate of bone marrow undetectable minimal residual disease (BM-uMRD) by ITT was 66%. The rate of CR with BM-uMRD at end of combination treatment (primary endpoint) was 25%. Three-year PFS and OS are 73 and 93%, respectively. DFCR is active as initial therapy of younger CLL patients. Immune-mediated and infectious toxicities occurred and required active management.

Purpose: The tumor microenvironment is complex, comprising heterogeneous cellular populations. As molecular profiles are frequently generated using bulk tissue sections, they represent an admixture of multiple cell types (including immune, stromal, and cancer cells) interacting with each other. Therefore, these molecular profiles are confounded by signals emanating from many cell types. Accurate assessment of residual cancer cell fraction is crucial for parameterization and interpretation of genomic analyses, as well as for accurately interpreting the clinical properties of the tumor., Materials and Methods: To benchmark cancer cell fraction estimation methods, 10 estimators were applied to a clinical cohort of 333 patients with prostate cancer. These methods include gold-standard multiobserver pathology estimates, as well as estimates inferred from genome, epigenome, and transcriptome data. In addition, two methods based on genomic and transcriptomic profiles were used to quantify tumor purity in 4,497 tumors across 12 cancer types. Bulk mRNA and microRNA profiles were subject to in silico deconvolution to estimate cancer cell-specific mRNA and microRNA profiles., Results: We present a systematic comparison of 10 tumor purity estimation methods on a cohort of 333 prostate tumors. We quantify variation among purity estimation methods and demonstrate how this influences interpretation of clinico-genomic analyses. Our data show poor concordance between pathologic and molecular purity estimates, necessitating caution when interpreting molecular results. Limited concordance between DNA- and mRNA-derived purity estimates remained a general pan-cancer phenomenon when tested in an additional 4,497 tumors spanning 12 cancer types., Conclusion: The choice of tumor purity estimation method may have a profound impact on the interpretation of genomic assays. Taken together, these data highlight the need for improved assessment of tumor purity and quantitation of its influences on the molecular hallmarks of cancers., Competing Interests: The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO’s conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center. Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments). Peter J. ParkHonoraria: Pfizer Consulting or Advisory Role: Neuroinflammation Newco Patents, Royalties, Other Intellectual Property: Patent on mutational signature-based detection of homologous recombination deficiencyPeter W. LairdConsulting or Advisory Role: Progenity, AnchorDx Patents, Royalties, Other Intellectual Property: Received royalties annually through 2018 for inventions licensed to Epigenomics AG by USC Travel, Accommodations, Expenses: AnchorDxWenyi WangStock and Other Ownership Interests: Genomic HealthFrancesca DemichelisPatents, Royalties, Other Intellectual Property: Co-inventor on a patent filed by the University of Michigan and the Brigham and Women’s Hospital covering the diagnostic and therapeutic fields for ETS fusions in prostate cancer. The diagnostic field has been licensed to Gen-Probe.Paul C. BoutrosConsulting or Advisory Role: BioSymetrics Patents, Royalties, Other Intellectual Property: Holds patents on multiple biomarkers No other potential conflicts of interest were reported., (© 2020 by American Society of Clinical Oncology.)

Benign normal (NL), premalignant (endometrial intraepithelial neoplasia, EIN) and malignant (cancer, EMCA) endometria must be precisely distinguished for optimal management. EIN was objectively defined previously as a regression model incorporating manually traced histologic variables to predict clonal growth and cancer outcomes. Results from this early computational study were used to revise subjective endometrial precancer diagnostic criteria currently in use. We here use automated feature segmentation and updated machine learning algorithms to develop a new classification algorithm. Endometrial tissue from 148 patients was randomly separated into 72-patient training and 76-patient validation cohorts encompassing all 3 diagnostic classes. We applied image analysis software to keratin stained endometrial tissues to automatically segment whole-slide digital images into epithelium, cells, and nuclei and extract corresponding variables. A total of 1413 variables were culled to 75 based on random forest classification performance in a 3-group (NL, EIN, EMCA) model. This algorithm correctly classifies cases with 3-class error rates of 0.04 (training set) and 0.058 (validation set); and 2-class (NL vs. EIN+EMCA) error rate of 0.016 (training set) and 0 (validation set). The 4 most heavily weighted variables are surrogates of those previously identified in manual-segmentation machine learning studies (stromal and epithelial area percentages, and normalized epithelial surface lengths). Lesser weighted predictors include gland and lumen axis lengths and ratios, and individual cell measures. Automated image analysis and random forest classification algorithms can classify normal, premalignant, and malignant endometrial tissues. Highest predictive variables overlap with those discovered independently in early models based on manual segmentation.

Heterogeneity in tumors is recognized as a key contributor to drug resistance and spread of advanced disease, but deep characterization of genetic variation within tumors has only recently been quantifiable with the advancement of next generation sequencing and single cell technologies. These data have been essential in developing molecular models of how tumors develop, evolve, and respond to environmental changes, such as therapeutic intervention. A deeper understanding of tumor evolution has subsequently opened up new research efforts to develop mathematical models that account for evolutionary dynamics with the goal of predicting drug response and resistance in cancer. Here, we describe recent advances and limitations of how models of tumor evolution can impact treatment strategies for cancer patients.

Background: Fludarabine, cyclophosphamide, and rituximab (FCR) can improve disease-free survival for younger (age ≤65 years) fit patients with chronic lymphocytic leukaemia with mutated IGHV. However, patients with unmutated IGHV rarely have durable responses. Ibrutinib is active for patients with chronic lymphocytic leukaemia irrespective of IGHV mutation status but requires continuous treatment. We postulated that time-limited ibrutinib plus FCR would induce durable responses in younger fit patients with chronic lymphocytic leukaemia., Methods: We did a multicentre, open-label, non-randomised, single-arm phase 2 trial at seven sites in the USA. We enrolled patients aged 65 years or younger with previously untreated chronic lymphocytic leukaemia. Our initial cohort (original cohort) was not restricted by prognostic marker status and included patients who had del(17p) or TP53 aberrations. After a protocol amendment (on March 21, 2017), we enrolled an additional cohort (expansion cohort) that included patients without del(17p). Ibrutinib was given orally (420 mg/day) for 7 days, then up to six 28-day cycles were administered intravenously of fludarabine (25 mg/m 2 , days 1-3), cyclophosphamide (250 mg/m 2 , days 1-3), and rituximab (375 mg/m 2 day 1 of cycle 1; 500 mg/m 2 day 1 of cycles 2-6) with continuous oral ibrutinib (420 mg/day). Responders continued on ibrutinib maintenance for up to 2 years, and patients with undetectable minimal residual disease in bone marrow after 2 years were able to discontinue treatment. The primary endpoint was the proportion of patients who achieved a complete response with undetectable minimal residual disease in bone marrow 2 months after the last cycle of ibrutinib plus FCR. Analyses were done per-protocol in all patients who received at least one dose of study treatment. This trial is registered with ClinicalTrials.gov (NCT02251548) and is ongoing., Findings: Between Oct 23, 2014, and April 23, 2018, 85 patients with chronic lymphocytic leukaemia were enrolled. del(17p) was detected in four (5%) of 83 patients and TP53 mutations were noted in three (4%) of 81 patients; two patients had both del(17p) and TP53 mutations. Median patients' age was 55 years (IQR 50-58). At data cutoff, median follow-up was 16·5 months (IQR 10·6-34·1). A complete response and undetectable minimal residual disease in bone marrow 2 months after the last cycle of ibrutinib plus FCR was achieved by 28 (33%, 95% CI 0·23-0·44) of 85 patients (p=0·0035 compared with a 20% historical value with FCR alone). A best response of undetectable minimal residual disease in bone marrow was achieved by 71 (84%) of 85 patients during the study. One patient had disease progression and one patient died (sudden cardiac death after 17 months of ibrutinib maintenance, assessed as possibly related to ibrutinib). The most common all-grade toxic effects were haematological, including thrombocytopenia in 63 (74%) patients, neutropenia in 53 (62%), and anaemia in 41 (49%). Grade 3 or 4 non-haematological serious adverse events included grade 3 atrial fibrillation in three (4%) patients and grade 3 Pneumocystis jirovecii pneumonia in two (2%)., Interpretation: The proportion of patients who achieved undetectable minimal residual disease in bone marrow with ibrutinib plus FCR is, to our knowledge, the highest ever published in patients with chronic lymphocytic leukaemia unrestricted by prognostic marker status. Ibrutinib plus FCR is promising as a time-limited combination regimen for frontline chronic lymphocytic leukaemia treatment in younger fit patients., Funding: Pharmacyclics and the Leukemia & Lymphoma Society., (Copyright © 2019 Elsevier Ltd. All rights reserved.)

Transcriptome deconvolution in cancer and other heterogeneous tissues remains challenging. Available methods lack the ability to estimate both component-specific proportions and expression profiles for individual samples. We present DeMixT, a new tool to deconvolve high-dimensional data from mixtures of more than two components. DeMixT implements an iterated conditional mode algorithm and a novel gene-set-based component merging approach to improve accuracy. In a series of experimental validation studies and application to TCGA data, DeMixT showed high accuracy. Improved deconvolution is an important step toward linking tumor transcriptomic data with clinical outcomes. An R package, scripts, and data are available: https://github.com/wwylab/DeMixTallmaterials., (Published by Elsevier Inc.)