VOLTA: an enVironment-aware cOntrastive ceLl represenTation leArning for histopathology

In clinical oncology, many diagnostic tasks rely on the identification of cells in histopathology images. While supervised machine learning techniques necessitate the need for labels, providing manual cell annotations is time-consuming. In this paper, we propose a self-supervised framework (enVironment-aware cOntrastive cell represenTation learning: VOLTA) for cell representation learning in histopathology images using a technique that accounts for the cell's mutual relationship with its environment. We subject our model to extensive experiments on data collected from multiple institutions comprising over 800,000 cells and six cancer types. To showcase the potential of our proposed framework, we apply VOLTA to ovarian and endometrial cancers and demonstrate that our cell representations can be utilized to identify the known histotypes of ovarian cancer and provide insights that link histopathology and molecular subtypes of endometrial cancer. Unlike supervised models, we provide a framework that can empower discoveries without any annotation data, even in situations where sample sizes are limited.

Artificial Intelligence and Pathomics: Prostate Cancer

Artificial intelligence (AI) has the potential to transform pathologic diagnosis and cancer patient management as a predictive and prognostic biomarker. AI-based systems can be used to examine digitally scanned histopathology slides and differentiate benign from malignant cells and low from high grade. Deep learning models can analyze patient data from individual or multimodal combinations and identify patterns to be used to predict the response to different therapeutic options, the risk of recurrence or progression, and the prognosis of the newly diagnosed patient. AI-based models will improve treatment planning for patients with prostate cancer and improve the efficiency and cost-effectiveness of the pathology laboratory.

Harmonized molecular classification; assessment of a single-test ProMisE NGS tool

OBJECTIVES

Despite recommendations for integrating molecular classification of endometrial cancers (EC) into pathology reporting and clinical management, uptake is inconsistent. To assign ProMisE subtype, all molecular components must be available (POLE mutation status, mismatch repair (MMR) and p53 immunohistochemistry (IHC)) and often these are assessed at different stages of care and/or at different centres resulting in delays in treatment. We assessed a single-test DNA-based targeted next generation sequencing (NGS) molecular classifier (ProMisE NGS), comparing concordance and prognostic value to the original ProMisE classifier.

METHODS

DNA was extracted from formalin-fixed paraffin embedded (FFPE) ECs that had previously undergone ProMisE molecular classification (POLE sequencing, IHC for p53 and MMR). DNA was sequenced using the clinically validated Imagia Canexia Health Find It™ amplicon-based NGS gene panel assay to assess for pathogenic POLE mutations (unchanged from original ProMisE), TP53 mutations (in lieu of p53 IHC), and microsatellite instability (MSI) (in lieu of MMR IHC),with the same order of segregation as original ProMisE used for subtype assignment. Molecular subtype assignment of both classifiers was compared by concordance metrics and Kaplan-Meier survival statistics.

RESULTS

The new DNA-based NGS molecular classifier (ProMisE NGS) was used to determine the molecular subtype in 164 ECs previously classified with ProMisE. 159/164 cases were concordant with a kappa statistic of 0.96 and an overall accuracy of 0.97. Prognostic differences in progression-free, disease-specific and overall survival between the four molecular subtypes were observed for the new NGS classifier, recapitulating the survival curves of the original ProMisE classifier. ProMisE NGS was 100% concordant between matched biopsy and hysterectomy samples.

CONCLUSION

ProMisE NGS is feasible on standard FFPE material, demonstrates high concordance with the original ProMisE classifier and maintains prognostic value in EC. This test has the potential to facilitate implementation of molecular classification of EC at the time of first diagnosis.

Identification of a novel subtype of endometrial cancer with unfavorable outcome using artificial intelligence-based histopathology image analysis

5594

Background: Molecular subtyping of endometrial cancer (EC), unlike histopathological evaluation, offers an objective and reproducible classification system that has strong prognostic value and therapeutic implications. The Proactive Molecular risk classifier for Endometrial cancer (ProMisE) was developed by our team as a pragmatic, cost-effective, and clinically applicable molecular classifier for EC patients. ProMisE has four subtypes: (i) POLE mutant ( POLEmut), (ii) mismatch repair deficient (MMRd), (iii) p53 abnormal (p53abn) by immunohistochemistry, and (iv) NSMP (No Specific Molecular Profile), lacking any of the defining features of the other three subtypes. While ProMisE subtypes are associated with clinical outcomes, within each subtype, there are clinical/prognostic outliers. This is particularly true within the largest ProMisE subtype; NSMP (representing ̃50% of ECs), where a subset of patients experience a very aggressive disease course, comparable to what is observed in patients with p53abn ECs. Methods: We hypothesized that objective assessment of the digitized hematoxylin and eosin (H&E)-stained histopathology slides of the largest and most diverse EC subset, NSMP, could potentially identify clinical outcome outliers. As such, we developed an artificial intelligence (AI)-based image analysis model to identify the NSMP cases that had similar histopathological features to the p53abn subtype, as assessed by H&E stain. We used a discovery cohort of 182 and an external validation cohort of 195 NSMP ECs. Results: Our AI-based image analysis model, based on deep convolutional neural networks, identified 21 (11.5%) out of the 182 NSMP cases with similar histopathological features as p53abn cases. We refer to these cases as ‘p53abn-like’ NSMPs. Compared to the rest of the NSMP cases, these cases had markedly inferior disease-specific survival (DSS) (10-year DSS 58.9% vs. 93.1% ( p<3.44e-8)) and progression-free survival (PFS) (10-year PFS 55.1% vs. 91.4% ( p<3.76e-6)). These findings were confirmed in our validation cohort, with 10.7% of the 195 patients categorized as ‘p53abn-like’ tumors with 10-year DSS of 82% vs. 51.3% ( p<5.28e-5) and PFS of 89.3% vs. 56.6% (p<2.15e-4). Conclusions: Utilizing an AI-based approach for histopathology image analysis, we have discovered ‘p53abn-like’ NSMPs, a novel subtype of NSMP ECs with morphological features similar to p53abn cases. ‘p53abn-like’ NSMPs exhibit similar clinical behavior as p53abn, having noticeably inferior outcome compared to the rest of the NSMP cases in two independent cohorts. These findings warrant further molecular investigation of this novel subtype of EC to identify the biological underpinning and future therapeutic strategies.

Tumor associated antigen PRAME exhibits dualistic functions that are targetable in diffuse large B-cell lymphoma

PRAME is a prominent member of the cancer testis antigen family of proteins, which triggers autologous T cell–mediated immune responses. Integrative genomic analysis in diffuse large B cell lymphoma (DLBCL) uncovered recurrent and highly focal deletions of 22q11.22, including the PRAME gene, which were associated with poor outcome. PRAME-deleted tumors showed cytotoxic T cell immune escape and were associated with cold tumor microenvironments. In addition, PRAME downmodulation was strongly associated with somatic EZH2 Y641 mutations in DLBCL. In turn, PRC2-regulated genes were repressed in isogenic PRAME-KO lymphoma cell lines, and PRAME was found to directly interact with EZH2 as a negative regulator. EZH2 inhibition with EPZ-6438 abrogated these extrinsic and intrinsic effects, leading to PRAME expression and microenvironment restoration in vivo. Our data highlight multiple functions of PRAME during lymphomagenesis and provide a preclinical rationale for synergistic therapies combining epigenetic reprogramming with PRAME-targeted therapies.

Gastrointestinal symptoms and the severity of COVID-19: Disorders of gut-brain interaction are an outcome

BACKGROUND

Many of the studies on COVID-19 severity and its associated symptoms focus on hospitalized patients. The aim of this study was to investigate the relationship between acute GI symptoms and COVID-19 severity in a clustering-based approach and to determine the risks and epidemiological features of post-COVID-19 Disorders of Gut-Brain Interaction (DGBI) by including both hospitalized and ambulatory patients.

METHODS

The study utilized a two-phase Internet-based survey on: (1) COVID-19 patients' demographics, comorbidities, symptoms, complications, and hospitalizations and (2) post-COVID-19 DGBI diagnosed according to Rome IV criteria in association with anxiety (GAD-7) and depression (PHQ-9). Statistical analyses included univariate and multivariate tests.

RESULTS

Five distinct clusters of symptomatic subjects were identified based on the presence of GI symptoms, loss of smell, and chest pain, among 1114 participants who tested positive for SARS-CoV-2. GI symptoms were found to be independent risk factors for severe COVID-19; however, they did not always coincide with other severity-related factors such as age >65 years, diabetes mellitus, and Vitamin D deficiency. Of the 164 subjects with a positive test who participated in Phase-2, 108 (66%) fulfilled the criteria for at least one DGBI. The majority (n = 81; 75%) were new-onset DGBI post-COVID-19. Overall, 86% of subjects with one or more post-COVID-19 DGBI had at least one GI symptom during the acute phase of COVID-19, while 14% did not. Depression (65%), but not anxiety (48%), was significantly more common in those with post-COVID-19 DGBI.

CONCLUSION

GI symptoms are associated with a severe COVID-19 among survivors. Long-haulers may develop post-COVID-19 DGBI. Psychiatric disorders are common in post-COVID-19 DGBI.

The utility of color normalization for AI-based diagnosis of hematoxylin and eosin-stained pathology images

The color variation of hematoxylin and eosin (H&E)-stained tissues has presented a challenge for applications of artificial intelligence (AI) in digital pathology. Many color normalization algorithms have been developed in recent years in order to reduce the color variation between H&E images. However, previous efforts in benchmarking these algorithms have produced conflicting results and none have sufficiently assessed the efficacy of the various color normalization methods for improving diagnostic performance of AI systems. In this study, we systematically investigated eight color normalization algorithms for AI-based classification of H&E-stained histopathology slides, in the context of using images both from one center and from multiple centers. Our results show that color normalization does not consistently improve classification performance when both training and testing data are from a single center. However, using four multi-center datasets of two cancer types (ovarian and pleural) and objective functions, we show that color normalization can significantly improve the classification accuracy of images from external datasets (ovarian cancer: 0.25 AUC increase, p = 1.6 e-05; pleural cancer: 0.21 AUC increase, p = 1.4 e-10). Furthermore, we introduce a novel augmentation strategy by mixing color-normalized images using three easily accessible algorithms that consistently improves the diagnosis of test images from external centers, even when the individual normalization methods had varied results. We anticipate our study to be a starting point for reliable use of color normalization to improve AI-based, digital pathology-empowered diagnosis of cancers sourced from multiple centers. © 2021 The Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Adult-type granulosa cell tumor of the ovary: a FOXL2-centric disease

Adult-type granulosa cell tumors (aGCTs) account for 90% of malignant ovarian sex cord-stromal tumors and 2-5% of all ovarian cancers. These tumors are usually diagnosed at an early stage and are treated with surgery. However, one-third of patients relapse between 4 and 8 years after initial diagnosis, and there are currently no effective treatments other than surgery for these relapsed patients. As the majority of aGCTs (>95%) harbor a somatic mutation in FOXL2 (c.C402G; p.C134W), the aim of this study was to identify genetic mutations besides FOXL2 C402G in aGCTs that could explain the clinical diversity of this disease. Whole-genome sequencing of 10 aGCTs and their matched normal blood was performed to identify somatic mutations. From this analysis, a custom amplicon-based panel was designed to sequence 39 genes of interest in a validation cohort of 83 aGCTs collected internationally. KMT2D inactivating mutations were present in 10 of 93 aGCTs (10.8%), and the frequency of these mutations was similar between primary and recurrent aGCTs. Inactivating mutations, including a splice site mutation in candidate tumor suppressor WNK2 and nonsense mutations in PIK3R1 and NLRC5, were identified at a low frequency in our cohort. Missense mutations were identified in cell cycle-related genes TP53, CDKN2D, and CDK1. From these data, we conclude that aGCTs are comparatively a homogeneous group of tumors that arise from a limited set of genetic events and are characterized by the FOXL2 C402G mutation. Secondary mutations occur in a subset of patients but do not explain the diverse clinical behavior of this disease. As the FOXL2 C402G mutation remains the main driver of this disease, progress in the development of therapeutics for aGCT would likely come from understanding the functional consequences of the FOXL2 C402G mutation.

Synthesis of diagnostic quality cancer pathology images by generative adversarial networks

Deep learning-based computer vision methods have recently made remarkable breakthroughs in the analysis and classification of cancer pathology images. However, there has been relatively little investigation of the utility of deep neural networks to synthesize medical images. In this study, we evaluated the efficacy of generative adversarial networks to synthesize high-resolution pathology images of 10 histological types of cancer, including five cancer types from The Cancer Genome Atlas and the five major histological subtypes of ovarian carcinoma. The quality of these images was assessed using a comprehensive survey of board-certified pathologists (n = 9) and pathology trainees (n = 6). Our results show that the real and synthetic images are classified by histotype with comparable accuracies and the synthetic images are visually indistinguishable from real images. Furthermore, we trained deep convolutional neural networks to diagnose the different cancer types and determined that the synthetic images perform as well as additional real images when used to supplement a small training set. These findings have important applications in proficiency testing of medical practitioners and quality assurance in clinical laboratories. Furthermore, training of computer-aided diagnostic systems can benefit from synthetic images where labeled datasets are limited (e.g. rare cancers). We have created a publicly available website where clinicians and researchers can attempt questions from the image survey (http://gan.aimlab.ca/). © 2020 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

The coming 15 years in gynaecological pathology: digitisation, artificial intelligence, and new technologies

Surgical pathology forms the cornerstone of modern oncological medicine, owing to the wealth of clinically relevant information that can be obtained from tissue morphology. Although several ancillary testing modalities have been added to surgical pathology, the way in which we view and interpret tissue morphology has remained largely unchanged since the inception of our profession. In this review, we discuss new technological advances that promise to transform the way in which we access tissue morphology and how we use it to guide patient care.

Endometrial Cancer Molecular Risk Stratification is Equally Prognostic for Endometrioid Ovarian Carcinoma

PURPOSE

Endometrioid ovarian carcinoma (ENOC) is generally associated with a more favorable prognosis compared with other ovarian carcinomas. Nonetheless, current patient treatment continues to follow a "one-size-fits-all" approach. Even though tumor staging offers stratification, personalized treatments remain elusive. As ENOC shares many clinical and molecular features with its endometrial counterpart, we sought to investigate The Cancer Genome Atlas-inspired endometrial carcinoma (EC) molecular subtyping in a cohort of ENOC.

EXPERIMENTAL DESIGN

IHC and mutation biomarkers were used to segregate 511 ENOC tumors into four EC-inspired molecular subtypes: low-risk POLE mutant (POLEmut), moderate-risk mismatch repair deficient (MMRd), high-risk p53 abnormal (p53abn), and moderate-risk with no specific molecular profile (NSMP). Survival analysis with established clinicopathologic and subtype-specific features was performed.

RESULTS

A total of 3.5% of cases were POLEmut, 13.7% MMRd, 9.6% p53abn, and 73.2% NSMP, each showing distinct outcomes (P < 0.001) and survival similar to observations in EC. Median OS was 18.1 years in NSMP, 12.3 years in MMRd, 4.7 years in p53abn, and not reached for POLEmut cases. Subtypes were independent of stage, grade, and residual disease in multivariate analysis.

CONCLUSIONS

EC-inspired molecular classification provides independent prognostic information in ENOC. Our findings support investigating molecular subtype-specific management recommendations for patients with ENOC; for example, subtypes may provide guidance when fertility-sparing treatment is desired. Similarities between ENOC and EC suggest that patients with ENOC may benefit from management strategies applied to EC and the opportunity to study those in umbrella trials.

Abstract GMM-047: GENOMIC CHARACTERIZATION OF ADULT-TYPE GRANULOSA CELL TUMORS: IMPLICATIONS FOR PATHOGENESIS AND TREATMENT OF RECURRENT DISEASE

BACKGROUND: Adult granulosa cell tumors (AGCT) represent 3-5% of all ovarian cancers. These tumors are characterized by their slow growth and usually occur in postmenopausal women with a median age of diagnosis of 50 to 54 years. The majority of patients are diagnosed as stage I and are treated with surgery to remove their ovaries and uterus. Although these treatments are effective at first, one third of patients relapse leading to mortality in 50-80% of relapsed patients. The best course of treatment for patients with recurrent advanced stage disease is optimal debulking surgery. Currently, there are no effective treatments available for patients where surgery is not an option. Our research team previously discovered a somatic missense mutation (c.402C&gt;G; pC134W) in the transcription factor Forkhead box L2 (FOXL2) in 97% of AGCTs. We also discovered frequent activating telomerase reverse transcriptase (TERT) promoter mutations in AGCT. As the FOXL2 C134W mutation is present in essentially all AGCTs and telomerase reactivation is required for tumorigenesis, it is likely that additional mutations are responsible for the variability in clinical behaviour. This study aims to describe the mutational landscape of AGCT to further refine our understanding of the frequent recurrence of this disease.

METHODS: Using whole genome sequencing (WGS), we characterized the genomes of ten AGCTs and their matched normal blood. We observed that AGCTs have a low mutation burden and the majority of mutations are single nucleotides variants. We have collected 516 formalin-fixed paraffin-embedded AGCT specimens including primary, recurrent and metastatic tumors from seven international centres for validation of our WGS results. Allelic discrimination assays were used for hotspots mutations, in addition to targeted sequencing of 39 genes of interest using a custom amplicon-based panel in our extension cohort.

RESULTS: Of the 39 genes analyzed, the third most commonly mutated gene (FOXL2 and TERT being the first and second most common) in our preliminary analysis of 88 cases was lysine (K)-specific methyltransferase 2D (KMT2D or MLL2). We identified various missense and nonsense mutations in this gene in 16 of 88 AGCTs (18%) analyzed thus far. KMT2D is a histone methyltransferase that targets histone H3 lysine 4 (H3K4), a methylation activation mark, and has an essential role in transcriptional regulation. Using an allelic discrimination assay, we identified a known hotspot mutation (c.49G&gt;A;p.E17K) in v-akt murine thymoma viral oncogene 1 (AKT1) in 2 of 67 (3%) AGCT patients, one of which the mutation was present in all three recurrent specimens from the same patient. AKT1 E17K mutation has been reported in multiple cancers such as breast, colorectal and high grade serous ovarian cancer at a low prevalence. A recent clinical trial of AKT inhibition in solid tumors with AKT1 mutations included one recurrent AGCT patient and showed significant tumor regression in one metastatic site.

CONCLUSION: AKT1 E17K mutations are present in AGCT at a low prevalence and could represent a therapeutic target for patients with recurrent advanced stage disease harbouring this mutation.

Citation Format: Jessica A. Pilsworth, Dawn R. Cochrane, Samantha Neilson, Anniina E.M. Färkkilä, Hugo M. Horlings, Satoshi Yanagida, Janine Senz, Yi Kan Wang, Bahar Moussavi, Daniel Lai, Ali Bashashati, Jacqueline Keul, Adele Wong, Hannah van Meurs, Sara Y. Brucker, Florin-Andrei Taran, Bernhard Krämer, Annette Staebler, Esther Oliva, Sohrab P. Shah, Stefan Kommoss, Friedrich Kommoss, C. Blake Gilks and David G. Huntsman. GENOMIC CHARACTERIZATION OF ADULT-TYPE GRANULOSA CELL TUMORS: IMPLICATIONS FOR PATHOGENESIS AND TREATMENT OF RECURRENT DISEASE [abstract]. In: Proceedings of the 12th Biennial Ovarian Cancer Research Symposium; Sep 13-15, 2018; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2019;25(22 Suppl):Abstract nr GMM-047.

Regulation of pH by Carbonic Anhydrase 9 Mediates Survival of Pancreatic Cancer Cells With Activated KRAS in Response to Hypoxia

BACKGROUND & AIMS

Most pancreatic ductal adenocarcinomas (PDACs) express an activated form of KRAS, become hypoxic and dysplastic, and are refractory to chemo and radiation therapies. To survive in the hypoxic environment, PDAC cells upregulate enzymes and transporters involved in pH regulation, including the extracellular facing carbonic anhydrase 9 (CA9). We evaluated the effect of blocking CA9, in combination with administration of gemcitabine, in mouse models of pancreatic cancer.

METHODS

We knocked down expression of KRAS in human (PK-8 and PK-1) PDAC cells with small hairpin RNAs. Human and mouse (Kras^G12D/Pdx1-Cre/Tp53/Rosa^YFP) PDAC cells were incubated with inhibitors of MEK (trametinib) or extracellular signal-regulated kinase (ERK), and some cells were cultured under hypoxic conditions. We measured levels and stability of the hypoxia-inducible factor 1 subunit alpha (HIF1A), endothelial PAS domain 1 protein (EPAS1, also called HIF2A), CA9, solute carrier family 16 member 4 (SLC16A4, also called MCT4), and SLC2A1 (also called GLUT1) by immunoblot analyses. We analyzed intracellular pH (pHi) and extracellular metabolic flux. We knocked down expression of CA9 in PDAC cells, or inhibited CA9 with SLC-0111, incubated them with gemcitabine, and assessed pHi, metabolic flux, and cytotoxicity under normoxic and hypoxic conditions. Cells were also injected into either immune-compromised or immune-competent mice and growth of xenograft tumors was assessed. Tumor fragments derived from patients with PDAC were surgically ligated to the pancreas of mice and the growth of tumors was assessed. We performed tissue microarray analyses of 205 human PDAC samples to measure levels of CA9 and associated expression of genes that regulate hypoxia with outcomes of patients using the Cancer Genome Atlas database.

RESULTS

Under hypoxic conditions, PDAC cells had increased levels of HIF1A and HIF2A, upregulated expression of CA9, and activated glycolysis. Knockdown of KRAS in PDAC cells, or incubation with trametinib, reduced the posttranscriptional stabilization of HIF1A and HIF2A, upregulation of CA9, pHi, and glycolysis in response to hypoxia. CA9 was expressed by 66% of PDAC samples analyzed; high expression of genes associated with metabolic adaptation to hypoxia, including CA9, correlated with significantly reduced survival times of patients. Knockdown or pharmacologic inhibition of CA9 in PDAC cells significantly reduced pHi in cells under hypoxic conditions, decreased gemcitabine-induced glycolysis, and increased their sensitivity to gemcitabine. PDAC cells with knockdown of CA9 formed smaller xenograft tumors in mice, and injection of gemcitabine inhibited tumor growth and significantly increased survival times of mice. In mice with xenograft tumors grown from human PDAC cells, oral administration of SLC-0111 and injection of gemcitabine increased intratumor acidosis and increased cell death. These tumors, and tumors grown from PDAC patient-derived tumor fragments, grew more slowly than xenograft tumors in mice given control agents, resulting in longer survival times. In Kras^G12D/Pdx1-Cre/Tp53/Rosa^YFP genetically modified mice, oral administration of SLC-0111 and injection of gemcitabine reduced numbers of B cells in tumors.

CONCLUSIONS

In response to hypoxia, PDAC cells that express activated KRAS increase expression of CA9, via stabilization of HIF1A and HIF2A, to regulate pH and glycolysis. Disruption of this pathway slows growth of PDAC xenograft tumors in mice and might be developed for treatment of pancreatic cancer.

Molecular profiling and molecular classification of endometrioid ovarian carcinomas

OBJECTIVE

Endometrioid ovarian carcinomas (EOCs) comprise 5-10% of all ovarian cancers and commonly co-occur with synchronous endometrioid endometrial cancer (EEC). We sought to examine the molecular characteristics of pure EOCs in patients without concomitant EEC.

METHODS

EOCs and matched normal samples were subjected to massively parallel sequencing targeting 341-468 cancer-related genes (n = 8) or whole-genome sequencing (n = 28). Mutational frequencies of EOCs were compared to those of high-grade serous ovarian cancers (HGSOCs; n = 224) and EECs (n = 186) from The Cancer Genome Atlas, and synchronous EOCs (n = 23).

RESULTS

EOCs were heterogeneous, frequently harboring KRAS, PIK3CA, PTEN, CTNNB1, ARID1A and TP53 mutations. EOCs were distinct from HGSOCs at the mutational level, less frequently harboring TP53 but more frequently displaying KRAS, PIK3CA, PIK3R1, PTEN and CTNNB1 mutations. Compared to synchronous EOCs and pure EECs, pure EOCs less frequently harbored PTEN, PIK3R1 and ARID1A mutations. Akin to EECs, EOCs could be stratified into the four molecular subtypes: 3% POLE (ultramutated), 19% MSI (hypermutated), 17% copy-number high (serous-like) and 61% copy-number low (endometrioid). In addition to microsatellite instability, a subset of EOCs harbored potentially targetable mutations, including AKT1 and ERBB2 hotspot mutations. EOCs of MSI (hypermutated) subtype uniformly displayed a good outcome.

CONCLUSIONS

EOCs are heterogeneous at the genomic level and harbor targetable genetic alterations. Despite the similarities in the repertoire of somatic mutations between pure EOCs, synchronous EOCs and EECs, the frequencies of mutations affecting known driver genes differ. Further studies are required to define the impact of the molecular subtypes on the outcome and treatment of EOC patients.

Abstract 3480: TMEM30A loss-of-function mutations drive lymphomagenesis and confer therapeutically exploitable vulnerability in B-cell lymphoma

Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma subtype worldwide, accounting for 40% of all non-Hodgkin lymphomas. DLBCL presents as an aggressive disease requiring immediate treatment. Although significant improvement in outcome has been achieved, ~40% of patients still experience treatment failure. Here, we characterized the recurrent genetic alterations and transcriptomic signatures in diagnostic biopsies from a population registry-based cohort of 347 patients with de novo DLBCL uniformly treated with R-CHOP. This analysis revealed bi-allelic loss of function mutations of TMEM30A that were associated with favorable treatment outcome. TMEM30A is a chaperone protein, involved in maintaining the asymmetric distribution of phosphatidylethanolamine and phosphatidylserine, an integral component of the plasma membrane and “eat-me” signal recognized by macrophages. Using TMEM30A knockout systems by CRISPR genome editing techniques, we have functionally characterized this loss-of-function mutation in representative human and mouse DLBCL cell line models. We have discovered that TMEM30A loss is associated with increased B-cell signaling following antigen stimulation, including a two-fold increase in the diffusion rate of B-cell receptor (BCR) clustering, using high resolution Single Particle Tracking (SPT) technology. In addition, we have measured three-fold increase in chemotherapeutic drug accumulation in both knockout cell lines and randomly selected patient biopsies with TMEM30A biallelic loss. This observation was validated in a xenograft mouse model, which presented improved survival and limited tumor growth following vincristine treatment in mice injected with TMEM30A null DLBCL cell lines compared with native cell lines. This phenotype explains the improved prognosis observed in DLBCL patients following R-CHOP treatment. Furthermore, we have observed over two fold higher numbers of tumor-associated macrophages in B-cell lymphoma syngeneic mouse models with Tmem30a loss-of-function, prior to any form of treatment, suggesting the existence of “hot” and primed tumors. Our data highlight a multi-faceted role for TMEM30A and plasma membrane physiology in B-cell lymphomagenesis, and characterize intrinsic and extrinsic vulnerabilities of cancer cells that can be therapeutically exploited. Characterization of these mechanisms will address a missing link in the cancer field as related insights in lymphoma will outline therapeutic approaches that can be extended to cancer therapy in general.

Citation Format: Shannon Healy, Daisuke Ennishi, Ali Bashashati, Saeed Saberi, Christoffer Hother, Anja Mottok, Fong Chun Chan, Lauren Chong, Robert Kridel, Merrill Boyle, Barbara Meissner, Tomohiro Aoki, Katsuyoshi Takata, Bruce W. Woolcock, Elena Vigano, Libin Abraham, Michael Gold, Adele Telenius, Pedro Farinha, Graham Slack, Susana Ben-Neriah, Daniel Lai, Allen W. Zhang, Sohrab Salehi, Hennady P. Shulha, Derek S. Chiu, Sara Mostafavi, Alina S. Gerrie, Diego Villa, Laurie H. Sehn, Kerry J. J. Savage, Andrew J. J. Mungall, Andrew P. Weng, Marcel Bally, Ryan D. Morin, Gabriela V. Cohen Freue, Joseph M. Connors, Marco A. Marra, Sohrab P. Shah, Randy D. Gascoyne1, David W. Scott, Christian Steidl, Ulrich Steidl. TMEM30A loss-of-function mutations drive lymphomagenesis and confer therapeutically exploitable vulnerability in B-cell lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3480.

Targeting Hypoxia-Induced Carbonic Anhydrase IX Enhances Immune-Checkpoint Blockade Locally and Systemically

Treatment strategies involving immune-checkpoint blockade (ICB) have significantly improved survival for a subset of patients across a broad spectrum of advanced solid cancers. Despite this, considerable room for improving response rates remains. The tumor microenvironment (TME) is a hurdle to immune function, as the altered metabolism-related acidic microenvironment of solid tumors decreases immune activity. Here, we determined that expression of the hypoxia-induced, cell-surface pH regulatory enzyme carbonic anhydrase IX (CAIX) is associated with worse overall survival in a cohort of 449 patients with melanoma. We found that targeting CAIX with the small-molecule SLC-0111 reduced glycolytic metabolism of tumor cells and extracellular acidification, resulting in increased immune cell killing. SLC-0111 treatment in combination with immune-checkpoint inhibitors led to the sensitization of tumors to ICB, which led to an enhanced Th1 response, decreased tumor growth, and reduced metastasis. We identified that increased expression of CA9 is associated with a reduced Th1 response in metastatic melanoma and basal-like breast cancer TCGA cohorts. These data suggest that targeting CAIX in the TME in combination with ICB is a potential therapeutic strategy for enhancing response and survival in patients with hypoxic solid malignancies.

Pharmacological systems analysis defines EIF4A3 functions in cell-cycle and RNA stress granule formation

The RNA helicase EIF4A3 regulates the exon junction complex and nonsense-mediated mRNA decay functions in RNA transcript processing. However, a transcriptome-wide network definition of these functions has been lacking, in part due to the lack of suitable pharmacological inhibitors. Here we employ short-duration graded EIF4A3 inhibition using small molecule allosteric inhibitors to define the transcriptome-wide dependencies of EIF4A3. We thus define conserved cellular functions, such as cell cycle control, that are EIF4A3 dependent. We show that EIF4A3-dependent splicing reactions have a distinct genome-wide pattern of associated RNA-binding protein motifs. We also uncover an unanticipated role of EIF4A3 in the biology of RNA stress granules, which sequester and silence the translation of most mRNAs under stress conditions and are implicated in cell survival and tumour progression. We show that stress granule induction and maintenance is suppressed on the inhibition of EIF4A3, in part through EIF4A3-associated regulation of G3BP1 and TIA1 scaffold protein expression.

Integrated structural variation and point mutation signatures in cancer genomes using correlated topic models

Mutation signatures in cancer genomes reflect endogenous and exogenous mutational processes, offering insights into tumour etiology, features for prognostic and biologic stratification and vulnerabilities to be exploited therapeutically. We present a novel machine learning formalism for improved signature inference, based on multi-modal correlated topic models (MMCTM) which can at once infer signatures from both single nucleotide and structural variation counts derived from cancer genome sequencing data. We exemplify the utility of our approach on two hormone driven, DNA repair deficient cancers: breast and ovary (n = 755 samples total). We show how introducing correlated structure both within and between modes of mutation can increase accuracy of signature discovery, particularly in the context of sparse data. Our study emphasizes the importance of integrating multiple mutation modes for signature discovery and patient stratification, and provides a statistical modeling framework to incorporate additional features of interest for future studies.

Over time DNA accumulates mutations from a variety of sources. Some mutations result from external mutagens, such as UV radiation, while others result from processes occurring within the cell itself. Each of these sources can impart characteristic patterns of mutations on the genome, known as mutation signatures, which can be detected using computational techniques. Loss of DNA repair mechanisms can leave specific mutation signatures in the genomes of cancer cells. To identify cancers with broken DNA-repair processes, accurate methods are needed for detecting mutation signatures and, in particular, their activities or probabilities within individual cancers. In this paper, we introduce a class of statistical modeling methods used for natural language processing, known as “topic models”, that outperform standard methods for signature analysis. We show that topic models that incorporate signature probability correlations across cancers perform best, while jointly analyzing multiple mutation types improves robustness to low mutation counts.

Molecular and Genetic Characterization of MHC Deficiency Identifies EZH2 as Therapeutic Target for Enhancing Immune Recognition

We performed a genomic, transcriptomic, and immunophenotypic study of 347 patients with diffuse large B-cell lymphoma (DLBCL) to uncover the molecular basis underlying acquired deficiency of MHC expression. Low MHC-II expression defines tumors originating from the centroblast-rich dark zone of the germinal center (GC) that was associated with inferior prognosis. MHC-II-deficient tumors were characterized by somatically acquired gene mutations reducing MHC-II expression and a lower amount of tumor-infiltrating lymphocytes. In particular, we demonstrated a strong enrichment of EZH2 mutations in both MHC-I- and MHC-II-negative primary lymphomas, and observed reduced MHC expression and T-cell infiltrates in murine lymphoma models expressing mutant Ezh2 ^Y641. Of clinical relevance, EZH2 inhibitors significantly restored MHC expression in EZH2-mutated human DLBCL cell lines. Hence, our findings suggest a tumor progression model of acquired immune escape in GC-derived lymphomas and pave the way for development of complementary therapeutic approaches combining immunotherapy with epigenetic reprogramming. SIGNIFICANCE: We demonstrate how MHC-deficient lymphoid tumors evolve in a cell-of-origin-specific context. Specifically, EZH2 mutations were identified as a genetic mechanism underlying acquired MHC deficiency. The paradigmatic restoration of MHC expression by EZH2 inhibitors provides the rationale for synergistic therapies combining immunotherapies with epigenetic reprogramming to enhance tumor recognition and elimination.See related commentary by Velcheti et al., p. 472.This article is highlighted in the In This Issue feature, p. 453.

Double-Hit Gene Expression Signature Defines a Distinct Subgroup of Germinal Center B-Cell-Like Diffuse Large B-Cell Lymphoma

PURPOSE

High-grade B-cell lymphoma with MYC and BCL2 and/or BCL6 rearrangements (HGBL-DH/TH) has a poor outcome after standard chemoimmunotherapy. We sought to understand the biologic underpinnings of HGBL-DH/TH with BCL2 rearrangements (HGBL-DH/TH- BCL2) and diffuse large B-cell lymphoma (DLBCL) morphology through examination of gene expression.

PATIENTS AND METHODS

We analyzed RNA sequencing data from 157 de novo germinal center B-cell-like (GCB)-DLBCLs, including 25 with HGBL-DH/TH- BCL2, to define a gene expression signature that distinguishes HGBL-DH/TH- BCL2 from other GCB-DLBCLs. To assess the genetic, molecular, and phenotypic features associated with this signature, we analyzed targeted resequencing, whole-exome sequencing, RNA sequencing, and immunohistochemistry data.

RESULTS

We developed a 104-gene double-hit signature (DHITsig) that assigned 27% of GCB-DLBCLs to the DHITsig-positive group, with only one half harboring MYC and BCL2 rearrangements (HGBL-DH/TH- BCL2). DHITsig-positive patients had inferior outcomes after rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone immunochemotherapy compared with DHITsig-negative patients (5-year time to progression rate, 57% and 81%, respectively; P < .001), irrespective of HGBL-DH/TH- BCL2 status. The prognostic value of DHITsig was confirmed in an independent validation cohort. DHITsig-positive tumors are biologically characterized by a putative non-light zone germinal center cell of origin and a distinct mutational landscape that comprises genes associated with chromatin modification. A new NanoString assay (DLBCL90) recapitulated the prognostic significance and RNA sequencing assignments. Validating the association with HGBL-DH/TH- BCL2, 11 of 25 DHITsig-positive-transformed follicular lymphomas were classified as HGBL-DH/TH- BCL2 compared with zero of 50 in the DHITsig-negative group. Furthermore, the DHITsig was shared with the majority of B-cell lymphomas with high-grade morphology tested.

CONCLUSION

We have defined a clinically and biologically distinct subgroup of tumors within GCB-DLBCL characterized by a gene expression signature of HGBL-DH/TH- BCL2. This knowledge has been translated into an assay applicable to routinely available biopsy samples, which enables exploration of its utility to guide patient management.

Abstract PR13: TERT is frequently mutated in adult-type granulosa cell tumors of the ovary compared to other malignant sex cord-stromal tumors

The telomerase reverse transcriptase (TERT) gene is highly expressed in stem cells and silenced upon differentiation. Cancer cells can attain immortality by activating TERT to maintain telomere length and telomerase activity, a crucial step of tumorigenesis. Two somatic mutations in the TERT promoter (C228T; C250T) have been identified in multiple cancers, such as melanoma and glioblastoma, as gain-of-function mutations that promote transcriptional activation of TERT. A recent study investigating TERT promoter mutations in ovarian carcinomas found mutations in 15% of clear cell carcinomas. However, it is unknown whether these mutations are prevalent in adult-type granulosa cell tumors (AGCTs) of the ovary, Sertoli-Leydig cell tumors (SLCTs), and other malignant sex cord-stromal tumors.

We performed whole-genome sequencing on ten AGCT cases with matched normal and identified the TERT C228T promoter mutation in 50% of cases. We found that AGCT cases with mutated TERT promoter have increased expression of TERT mRNA compared to those with wild-type TERT promoter. All five TERT promoter mutated cases had high levels of TERT mRNA expression, whereas three of the five wild-type TERT cases had no measurable TERT mRNA expression. There was a tendency towards longer telomere lengths in AGCT cases with the TERT promoter mutation relative to those without, although it was not statistically significant. These results suggest that telomerase may be activated by a different method in the cases with no TERT promoter mutations but have TERT mRNA expression. The remaining cases with neither TERT promoter mutations nor TERT mRNA expression likely maintain their telomeres using a telomerase-independent method, such as the alternative lengthening of telomeres pathway. TERT C228T allelic discrimination analysis of 331 AGCTs, 5 SLCTs, and 18 other malignant sex cord-stromal tumors detected the mutation in 56/247 (23%) of primary AGCTs, 22/84 (26%) of recurrent AGCTs, 1/5 (20%) of SLCTs and (0/18) 0% of other malignant sex cord-stromal tumors. The single SLCT case with the TERT promoter mutation was poorly differentiated and harbored the pathognomonic FOXL2 mutation of AGCT, suggesting this SLCT case may actually be an AGCT. In 204 AGCT cases with available survival data, there was a trend towards worse disease-specific survival in patients with the TERT promoter mutation compared to those without; however, statistical significance was not reached (p = 0.128, log-ranked test). In 5 AGCT cases with primary and recurrent tissues, we found that the TERT promoter mutation was absent in the primary tumors but present in the recurrent tumors, suggesting that TERT C228T mutation may play an active role in progression of AGCTs.

Overall, we found that TERT C228T promoter mutation was most common in AGCTs among the different malignant sex cord-stromal tumors. Our data confirm the activation of telomerase in AGCTs via TERT C228T promoter mutation, although alternative telomerase activation methods in AGCTs may exist. Our results suggest that TERT activation may play a role in AGCT recurrence. As such, telomere biology may be important for the progression of AGCTs.

This abstract is also being presented as Poster B54.

Citation Format: Jessica A. Pilsworth, Dawn R. Cochrane, Zhouchunyang Xia, Geraldine Aubert, Anniina E. M. Färkkilä, Hugo M. Horlings, Satoshi Yanagida, Winnie Yang, Jamie L. P. Lim, Yikan Wang, Ali Bashashati, Jacqueline Keul, Adele Wong, Esther Oliva, Sohrab P. Shah, Stefan Kommoss, Friedrich Kommoss, Peter M. Lansdorp, Duncan M. Baird, David G. Huntsman. TERT is frequently mutated in adult-type granulosa cell tumors of the ovary compared to other malignant sex cord-stromal tumors. [abstract]. In: Proceedings of the AACR Conference: Addressing Critical Questions in Ovarian Cancer Research and Treatment; Oct 1-4, 2017; Pittsburgh, PA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(15_Suppl):Abstract nr PR13.

Abstract B22: Capturing L1 retrotransposon-mediated DNA transductions in endometriosis associated ovarian cancers as a way to track tumor development

Endometrioid ovarian cancer (ENOC) and clear cell ovarian cancer (CCOC) share a common precursor lesion, endometriosis (ectopic growth of uterine lining), hence the designation endometriosis-associated ovarian cancer (EAOC). Women with endometriosis have up to three-fold increased risk of developing ENOC and CCOC. Efforts have been made to look for biomarkers that can help identifying women at risk of developing cancer; however, there are currently no biomarkers that stratify risk of cancer development. We performed whole-genome sequencing (WGS) on 29 ENOC and 35 CCOC cases and observed a frequent transduction event originating from an active LINE-1 (L1) retrotransposable element in the TTC28 gene. Such event occurred in 34% (10/29) of ENOC, and 31% (11/35) of CCOC cases. L1 retrotransposons are repetitive, mobile genetic elements capable of taking downstream DNA fragments and inserting them into random genomic locations via a process called 3’ transduction. Approximately 70-100 different potentially active L1s are epigenetically silenced in normal tissues, but tend to be reactivated in cancers. We subsequently used PCR to validate these TTC28-L1 transductions, and compared their presence to single nucleotide variations (SNVs) and frameshift mutations in formalin-fixed, paraffin-embedded (FFPE) tumor tissues from different tumor sites for 4 ENOC and 3 CCOC cases. We found that these transduction events along with classical driver mutations were almost ubiquitous across the tumor sites, suggesting these L1 events likely occurred early in the malignant transformation of EAOCs.

We developed a low-input, probe-based capture assay to test the presence of TTC28-L1 transductions as an alternative method to performing WGS. Oligonucleotide probes tiling 1 kb downstream of active L1s are used to capture DNA fragments containing the transduced DNA, and the fragments are sequenced on the MiSeq next-generation sequencing platform. Analyses are performed using the Geneious software and the published bioinformatics tool Socrates, specific for detecting DNA fragments with split reads (fragments with ends aligning to different parts of the genome). We successfully validated the assay on 9 cases with WGS data: 7 EAOC cases with TTC28-L1 transductions and 2 EAOC cases without TTC28-L1 transductions. DNA extracted from frozen tumor and buffy coat (normal control) were used for each case, and FFPE tissues were used for selected cases. All reads containing the transduction events aligned to genomic coordinates corresponding to the WGS data.

While L1-mediated DNA transductions are often passenger events during tumorigenesis, our data suggest that they likely occur early in ovarian cancer tumorigenesis. Our data show that this probe-based capture assay provides an alternative method to WGS, and may be useful in detecting active 3’ transductions in novel cases to track the development of ovarian tumors.

Citation Format: Zhouchunyang Xia, Dawn Cochrane, Michael Anglesio, Winnie Yang, Miguel Alcaide, Tayyebeh Nazeran, Janine Senz, Amy Lum, Ali Bashashati, Yikan Wang, Ryan Morin, Sohrab Shah, David Huntsman. Capturing L1 retrotransposon-mediated DNA transductions in endometriosis associated ovarian cancers as a way to track tumor development. [abstract]. In: Proceedings of the AACR Conference: Addressing Critical Questions in Ovarian Cancer Research and Treatment; Oct 1-4, 2017; Pittsburgh, PA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(15_Suppl):Abstract nr B22.

TERT promoter mutation in adult granulosa cell tumor of the ovary

The telomerase reverse transcriptase (TERT) gene is highly expressed in stem cells and silenced upon differentiation. Cancer cells can attain immortality by activating TERT to maintain telomere length and telomerase activity, which is a crucial step of tumorigenesis. Two somatic mutations in the TERT promoter (C228T; C250T) have been identified as gain-of-function mutations that promote transcriptional activation of TERT in multiple cancers, such as melanoma and glioblastoma. A recent study investigating TERT promoter mutations in ovarian carcinomas found C228T and C250T mutations in 15.9% of clear cell carcinomas. However, it is unknown whether these mutations are frequent in other ovarian cancer subtypes, in particular, sex cord-stromal tumors including adult granulosa cell tumors. We performed whole-genome sequencing on ten adult granulosa cell tumors with matched normal blood and identified a TERT C228T promoter mutation in 50% of tumors. We found that adult granulosa cell tumors with mutated TERT promoter have increased expression of TERT mRNA and exhibited significantly longer telomeres compared to those with wild-type TERT promoter. Extension cohort analysis using allelic discrimination revealed the TERT C228T mutation in 51 of 229 primary adult granulosa cell tumors (22%), 24 of 58 recurrent adult granulosa cell tumors (41%), and 1 of 22 other sex cord-stromal tumors (5%). There was a significant difference in overall survival between patients with TERT C228T promoter mutation in the primary tumors and those without it (p = 0.00253, log-rank test). In seven adult granulosa cell tumors, we found the TERT C228T mutation present in recurrent tumors and absent in the corresponding primary tumor. Our data suggest that TERT C228T promoter mutations may have an important role in progression of adult granulosa cell tumors.

Interfaces of Malignant and Immunologic Clonal Dynamics in Ovarian Cancer

High-grade serous ovarian cancer (HGSC) exhibits extensive malignant clonal diversity with widespread but non-random patterns of disease dissemination. We investigated whether local immune microenvironment factors shape tumor progression properties at the interface of tumor-infiltrating lymphocytes (TILs) and cancer cells. Through multi-region study of 212 samples from 38 patients with whole-genome sequencing, immunohistochemistry, histologic image analysis, gene expression profiling, and T and B cell receptor sequencing, we identified three immunologic subtypes across samples and extensive within-patient diversity. Epithelial CD8+ TILs negatively associated with malignant diversity, reflecting immunological pruning of tumor clones inferred by neoantigen depletion, HLA I loss of heterozygosity, and spatial tracking between T cell and tumor clones. In addition, combinatorial prognostic effects of mutational processes and immune properties were observed, illuminating how specific genomic aberration types associate with immune response and impact survival. We conclude that within-patient spatial immune microenvironment variation shapes intraperitoneal malignant spread, provoking new evolutionary perspectives on HGSC clonal dispersion.

Kronos: a workflow assembler for genome analytics and informatics

Background:

The field of next-generation sequencing informatics has matured to a point where algorithmic advances in sequence alignment and individual feature detection methods have stabilized. Practical and robust implementation of complex analytical workflows (where such tools are structured into “best practices” for automated analysis of next-generation sequencing datasets) still requires significant programming investment and expertise.

Results:

We present Kronos, a software platform for facilitating the development and execution of modular, auditable, and distributable bioinformatics workflows. Kronos obviates the need for explicit coding of workflows by compiling a text configuration file into executable Python applications. Making analysis modules would still require programming. The framework of each workflow includes a run manager to execute the encoded workflows locally (or on a cluster or cloud), parallelize tasks, and log all runtime events. The resulting workflows are highly modular and configurable by construction, facilitating flexible and extensible meta-applications that can be modified easily through configuration file editing. The workflows are fully encoded for ease of distribution and can be instantiated on external systems, a step toward reproducible research and comparative analyses. We introduce a framework for building Kronos components that function as shareable, modular nodes in Kronos workflows.

Conclusions:

The Kronos platform provides a standard framework for developers to implement custom tools, reuse existing tools, and contribute to the community at large. Kronos is shipped with both Docker and Amazon Web Services Machine Images. It is free, open source, and available through the Python Package Index and at https://github.com/jtaghiyar/kronos.

CDK12 regulates alternative last exon mRNA splicing and promotes breast cancer cell invasion

CDK12 (cyclin-dependent kinase 12) is a regulatory kinase with evolutionarily conserved roles in modulating transcription elongation. Recent tumor genome studies of breast and ovarian cancers highlighted recurrent CDK12 mutations, which have been shown to disrupt DNA repair in cell-based assays. In breast cancers, CDK12 is also frequently co-amplified with the HER2 (ERBB2) oncogene. The mechanisms underlying functions of CDK12 in general and in cancer remain poorly defined. Based on global analysis of mRNA transcripts in normal and breast cancer cell lines with and without CDK12 amplification, we demonstrate that CDK12 primarily regulates alternative last exon (ALE) splicing, a specialized subtype of alternative mRNA splicing, that is both gene- and cell type-specific. These are unusual properties for spliceosome regulatory factors, which typically regulate multiple forms of alternative splicing in a global manner. In breast cancer cells, regulation by CDK12 modulates ALE splicing of the DNA damage response activator ATM and a DNAJB6 isoform that influences cell invasion and tumorigenesis in xenografts. We found that there is a direct correlation between CDK12 levels, DNAJB6 isoform levels and the migration capacity and invasiveness of breast tumor cells. This suggests that CDK12 gene amplification can contribute to the pathogenesis of the cancer.

The driver landscape of sporadic chordoma

Chordoma is a malignant, often incurable bone tumour showing notochordal differentiation. Here, we defined the somatic driver landscape of 104 cases of sporadic chordoma. We reveal somatic duplications of the notochordal transcription factor brachyury (T) in up to 27% of cases. These variants recapitulate the rearrangement architecture of the pathogenic germline duplications of T that underlie familial chordoma. In addition, we find potentially clinically actionable PI3K signalling mutations in 16% of cases. Intriguingly, one of the most frequently altered genes, mutated exclusively by inactivating mutation, was LYST (10%), which may represent a novel cancer gene in chordoma.Chordoma is a rare often incurable malignant bone tumour. Here, the authors investigate driver mutations of sporadic chordoma in 104 cases, revealing duplications in notochordal transcription factor brachyury (T), PI3K signalling mutations, and mutations in LYST, a potential novel cancer gene in chordoma.

Abstract 3381: TERT promoter mutation in granulosa cell tumours of the ovary: Prevalence and prognostic significance

Granulosa cell tumours (GCTs) of the ovary account for 90% of sex cord-stromal tumours and have a high recurrence rate up to 50%. A missense mutation in the FOXL2 gene (c.402C&gt;G; pC134W) is a defining feature of GCT and is used as a robust marker for diagnosis. However, other than the FOXL2 mutation the pathogenesis and the driving pathways remain unknown. Determining secondary genetic events in GCTs is essential to understanding and improving prognosis.

In a pilot study, we completed an analysis of whole genome sequencing of ten GCTs and matched normal cases to generate a comprehensive catalogue of coding and non-coding events. We identified a TERT promoter mutation (c.228C&gt;T) in 50% of these cases. TERT is normally inactivated in somatic tissues; however, this promoter mutation has been shown to re-activate transcription of TERT. We validated this TERT mutation in an international cohort of 300 GCTs and found it was present in approximately 25% of cases overall. These TERT promoter mutations have been used to revise the molecular classification of other cancer types such as gliomas. In GCT, we found that this TERT mutation was correlated with a significantly worse survival outcome in patients with primary GCT (p&lt;0.005). Further, we found that this TERT mutation was present in a larger proportion of recurrent cases. Thus, this mutation may denote a novel subtype of GCT with a worse prognosis.

Previous research has shown that TERT activation is evident in over 90% of cancers and is a fundamental step in tumourigenesis that enables unlimited proliferation. This TERT promoter mutation in GCT provides an explanation of how granulosa cells escape atresia and attain immortality. Thus, we hypothesize a mechanism in which the FOXL2 mutation prevents apoptosis and the TERT mutation allows limitless proliferation for oncogenes to transform granulosa cells. However, the current cell models of GCT lack relevant functional pathways and do not recapitulate the biology of these tumours. Therefore, we are developing more suitable cell models to test our hypothesis. We believe that understanding the interaction between these TERT and FOXL2 mutations may lead to novel cancer cell-specific targeted therapies.

Citation Format: Jessica A. Pilsworth, Dawn R. Cochrane, Zhouchunyang Xia, Hugo M. Horlings, Winnie Yang, Melissa K. McConechy, Satoshi Yanagida, Anniina E. Färkkilä, Adele P. Wong, Genny Trigo-Gonzalez, S.W. Grace Cheng, Yikan Wang, Ali Bashashati, Gregg B. Morin, Esther Oliva, Sohrab P. Shah, David G. Huntsman. TERT promoter mutation in granulosa cell tumours of the ovary: Prevalence and prognostic significance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3381. doi:10.1158/1538-7445.AM2017-3381

Abstract DPOC-014: BEYOND CODING MUTATIONS: USING RETROTRANSPOSONS TO PREDICT OVARIAN CANCER DEVELOPMENT

PURPOSE: Women with endometriosis, a painful condition caused by displaced endometrial tissue, have a 3-fold increased risk of developing endometrioid ovarian cancer (ENOC) and clear cell ovarian cancer (CCOC). How two distinct cancers arise from the same precursor lesion is unknown. Sensitive biomarkers are needed to identify women with endometriosis at risk of developing cancer. We performed whole genome sequencing on 29 ENOC and 36 CCOC cases and observed a highly frequent insertion event originating from an active LINE-1 (L1) retrotransposon in the TTC28 gene. L1 retrotransposons are mobile genetic elements that can take downstream DNA pieces and insert them into random genomic locations in a process called 3' transduction. L1s are epigenetically silenced in normal tissues, but are known to become activated in a variety of cancers. A recent study showed a stepwise loss of methylation across various L1 loci between normal endometrium, contiguous endometriosis (endometriosis adjacent to tumor), and ENOC/CCOC tissues. We hypothesize that TTC28 L1 retrotransposon is an early event in the transformation of endometriosis into ENOC and CCOC and such events could be used as biomarkers for endometriosis with high cancer risk.

METHODS: We compared the presence of TTC28 L1 3' transductions to six SNVs and frame shifts mutations in normal, endometriosis, and tumor tissues from different anatomical sites in four ENOC and four CCOC cases. PCR followed by Sanger sequencing was used to detect TTC28 L1 insertions, and micro-fluidic PCR assay followed by MiSeq sequencing used to detect SNV/frameshift mutations. To broaden the analysis we will use a target capture sequencing method to track novel TTC28 and other L1 transductions. In these experiments probes tiling 1kbp downstream of L1s will be used to capture DNA fragments containing L1 transductions; the captured fragments will be sequenced on the MiSeq. We will assess the difference in TTC28 L1 methylation status between normal, endometriosis, and tumor tissues via the sequencing of bi-sulfite treated DNA.

RESULTS: TTC28 L1 retrotransposition insertion is present at all 5 tumor sites in 75% (6/8) of cases, and is present in 3/5 or 4/5 tumor sites in the remaining cases. Analysis shows that TTC28 L1 insertion preceded some SNV and/or frameshift mutations. Preliminary results show that TTC28 L1 promoters are unmethylated in tumors with L1 insertions. Future experiments involving additional cases with endometriosis tissues will be performed. We expect to see L1 promoter hypomethylation and L1 transductions in endometriosis tissues.

CONCLUSION: TTC28 L1 promoter hypomethylation and TTC28 L1 transductions may be early events in the transformation of endometriosis to cancer that can be explored as a method to predict tumor development. The development of a target capture assay to detect novel L1 transductions will be crucial for investigating cases without whole genome sequencing data. Ultimately, we hope to detect L1 insertions in plasma samples, and use L1 insertions as a biomarker to identify high-risk endometriosis cases.

Citation Format: Zhouchunyang Xia, Dawn Cochrane, Michael S Anglesio, Tayyebeh M Nazeran, Janine Senz, Amy Lum, Ali Bashashati, Yi Kan Wang, Sohrab P Shah, David Huntsman. BEYOND CODING MUTATIONS: USING RETROTRANSPOSONS TO PREDICT OVARIAN CANCER DEVELOPMENT [abstract]. In: Proceedings of the 11th Biennial Ovarian Cancer Research Symposium; Sep 12-13, 2016; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2017;23(11 Suppl):Abstract nr DPOC-014.

Divergent modes of clonal spread and intraperitoneal mixing in high-grade serous ovarian cancer

We performed phylogenetic analysis of high-grade serous ovarian cancers (68 samples from seven patients), identifying constituent clones and quantifying their relative abundances at multiple intraperitoneal sites. Through whole-genome and single-nucleus sequencing, we identified evolutionary features including mutation loss, convergence of the structural genome and temporal activation of mutational processes that patterned clonal progression. We then determined the precise clonal mixtures comprising each tumor sample. The majority of sites were clonally pure or composed of clones from a single phylogenetic clade. However, each patient contained at least one site composed of polyphyletic clones. Five patients exhibited monoclonal and unidirectional seeding from the ovary to intraperitoneal sites, and two patients demonstrated polyclonal spread and reseeding. Our findings indicate that at least two distinct modes of intraperitoneal spread operate in clonal dissemination and highlight the distribution of migratory potential over clonal populations comprising high-grade serous ovarian cancers.

Synchronous Endometrial and Ovarian Carcinomas: Evidence of Clonality

Many women with ovarian endometrioid carcinoma present with concurrent endometrial carcinoma. Organ-confined and low-grade synchronous endometrial and ovarian tumors (SEOs) clinically behave as independent primary tumors rather than a single advanced-stage carcinoma. We used 18 SEOs to investigate the ancestral relationship between the endometrial and ovarian components. Based on both targeted and exome sequencing, 17 of 18 patient cases of simultaneous cancer of the endometrium and ovary from our series showed evidence of a clonal relationship, ie, primary tumor and metastasis. Eleven patient cases fulfilled clinicopathological criteria that would lead to classification as independent endometrial and ovarian primary carcinomas, including being of FIGO stage T1a/1A, with organ-restricted growth and without surface involvement; 10 of 11 of these cases showed evidence of clonality. Our observations suggest that the disseminating cells amongst SEOs are restricted to physically accessible and microenvironment-compatible sites yet remain indolent, without the capacity for further dissemination.

The genomic landscape of epithelioid sarcoma cell lines and tumours

We carried out whole genome and transcriptome sequencing on four tumour/normal pairs of epithelioid sarcoma. These index cases were supplemented with whole transcriptome sequencing of three additional tumours and three cell lines. Unlike rhabdoid tumour (the other major group of SMARCB1-negative cancers), epithelioid sarcoma shows a complex genome with a higher mutational rate, comparable to that of ovarian carcinoma. Despite this mutational burden, SMARCB1 mutations remain the most frequently recurring event and are probably critical drivers of tumour formation. Several cases show heterozygous SMARCB1 mutations without inactivation of the second allele, and we explore this further in vitro. Finding CDKN2A deletions in our discovery cohort, we evaluated CDKN2A protein expression in a tissue microarray. Six out of 16 cases had lost CDKN2A in greater than or equal to 90% of cells, while the remaining cases had retained the protein. Expression analysis of epithelioid sarcoma cell lines by transcriptome sequencing shows a unique profile that does not cluster with any particular tissue type or with other SWI/SNF-aberrant lines. Evaluation of the levels of members of the SWI/SNF complex other than SMARCB1 revealed that these proteins are expressed as part of a residual complex, similarly to previously studied rhabdoid tumour lines. This residual SWI/SNF is susceptible to synthetic lethality and may therefore indicate a therapeutic opportunity.

Abstract AS18: The somatic mutational landscape of endometriosis associated ovarian cancers and precursor lesions

Introduction: Clear Cell ovarian carcinomas (CCC) represent 10% of ovarian carcinomas, with outcomes for high-stage significantly worse than high-grade serous form. The complete mutational landscape, the molecular basis of the transformation of endometriosis, the putative precursor, and patterns of clonal evolution in CCC are not well understood.

Methods: Whole genome sequencing and gene expression profiling was done to uncover somatic alterations and measure effects on transcriptional networks. Targeted deep sequencing of primary tumors, metastases and endometriosis was also performed and statistical modeling approaches were used to validate mutations, quantify clonal diversity, and trace patterns of selection.

Results: Mutations in ARID1A (11/19) and PIK3CA (8/19) were by far the most frequent aberrations seen. Three other SWI/SNF components also showed somatic alteration: two in non-ARID1A mutant cases and a truncating mutation of ARID1B in an ARID1A-null case. Amongst 24 significant “candidate drivers” impacting expression, five “cancer genes” have been previously described: PIK3CA, ARID1A, CTNNB1, TP53, and PPP2R1A. We observed no association between PIK3CA or ARID1A status with disease stage, genomic instability, or mutation load. Deep sequencing data suggested multiple clones in every case. In cases with matching precursor lesions, we observed multiple mutations in at least one such lesion. In precursor lesions where tumor-matched somatic mutations were observed, ARID1A and PIK3CA mutations were also always present, if observed in the matched tumor.

Conclusions: ARID1A and PIK3CA mutations appear as early and histo-type defining events in CCC. Pattern of endometriosis transformation can be associated with somatic mutations in all cases, including histologically “Atypical” and non-atypical endometriosis. Finally, patterns of mutational conservation across the series of precursor lesions may present an opportunity for early screening of endometriosis tissues as an indicator of transformation potential.

Citation Format: Michael S Anglesio, Ali Bashashati, Yikan Wang, Gavin Ha, Janine Senz, Winnie Yang, Steve E Kalloger, Leah M Prentice, Satoshi Yanagida, Clara Salamanca, Galina Soukhatcheva, Anthony Kazernis, Hector Chang, Anne-Marie Mes-Mason, Aikou Okamoto, Marco A Marra, Blake Gilks, Sohrab P Shah, David G Huntsman. The somatic mutational landscape of endometriosis associated ovarian cancers and precursor lesions [abstract]. In: Proceedings of the 10th Biennial Ovarian Cancer Research Symposium; Sep 8-9, 2014; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(16 Suppl):Abstract nr AS18.

Comparing Different Classifiers in Sensory Motor Brain Computer Interfaces

A problem that impedes the progress in Brain-Computer Interface (BCI) research is the difficulty in reproducing the results of different papers. Comparing different algorithms at present is very difficult. Some improvements have been made by the use of standard datasets to evaluate different algorithms. However, the lack of a comparison framework still exists. In this paper, we construct a new general comparison framework to compare different algorithms on several standard datasets. All these datasets correspond to sensory motor BCIs, and are obtained from 21 subjects during their operation of synchronous BCIs and 8 subjects using self-paced BCIs. Other researchers can use our framework to compare their own algorithms on their own datasets. We have compared the performance of different popular classification algorithms over these 29 subjects and performed statistical tests to validate our results. Our findings suggest that, for a given subject, the choice of the classifier for a BCI system depends on the feature extraction method used in that BCI system. This is in contrary to most of publications in the field that have used Linear Discriminant Analysis (LDA) as the classifier of choice for BCI systems.

Multifocal endometriotic lesions associated with cancer are clonal and carry a high mutation burden

Endometriosis is a significant risk factor for clear cell and endometrioid ovarian cancers and is often found contiguous with these cancers. Using whole‐genome shotgun sequencing of seven clear cell ovarian carcinomas (CCC) and targeted sequencing in synchronous endometriosis, we have investigated how this carcinoma may evolve from endometriosis. In every case we observed multiple tumour‐associated somatic mutations in at least one concurrent endometriotic lesion. ARID1A and PIK3CA mutations appeared consistently in concurrent endometriosis when present in the primary CCC. In several cases, one or more endometriotic lesions carried the near‐complete complement of somatic mutations present in the index CCC tumour. Ancestral mutations were detected in both tumour‐adjacent and ‐distant endometriotic lesions, regardless of any cytological atypia. These findings provide objective evidence that multifocal benign endometriotic lesions are clonally related and that CCCs arising in these patients progress from endometriotic lesions that may already carry sufficient cancer‐associated mutations to be considered neoplasms themselves, albeit with low malignant potential. We speculate that genomically distinct classes of endometriosis exist and that ovarian endometriosis with high mutational burden represents one class at high risk for malignant transformation. © 2015 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.

Dynamics of genomic clones in breast cancer patient xenografts at single-cell resolution

Human cancers, including breast cancers, comprise clones differing in mutation content. Clones evolve dynamically in space and time following principles of Darwinian evolution, underpinning important emergent features such as drug resistance and metastasis. Human breast cancer xenoengraftment is used as a means of capturing and studying tumour biology, and breast tumour xenografts are generally assumed to be reasonable models of the originating tumours. However, the consequences and reproducibility of engraftment and propagation on the genomic clonal architecture of tumours have not been systematically examined at single-cell resolution. Here we show, using deep-genome and single-cell sequencing methods, the clonal dynamics of initial engraftment and subsequent serial propagation of primary and metastatic human breast cancers in immunodeficient mice. In all 15 cases examined, clonal selection on engraftment was observed in both primary and metastatic breast tumours, varying in degree from extreme selective engraftment of minor (<5% of starting population) clones to moderate, polyclonal engraftment. Furthermore, ongoing clonal dynamics during serial passaging is a feature of tumours experiencing modest initial selection. Through single-cell sequencing, we show that major mutation clusters estimated from tumour population sequencing relate predictably to the most abundant clonal genotypes, even in clonally complex and rapidly evolving cases. Finally, we show that similar clonal expansion patterns can emerge in independent grafts of the same starting tumour population, indicating that genomic aberrations can be reproducible determinants of evolutionary trajectories. Our results show that measurement of genomically defined clonal population dynamics will be highly informative for functional studies using patient-derived breast cancer xenoengraftment.

Abstract LB-312: The somatic mutational landscape of ovarian clear cell carcinoma and its precursor lesions

Clear Cell ovarian carcinomas (CCC) represent ∼10% of ovarian carcinomas, with outcomes for high-stage cases significantly worse than the more common high-grade serous form. Response to standard chemotherapies are poor and efforts to improve treatment strategies are confounded by studies grouping ovarian histologies together, as well as a general lack of molecular background data on CCC. CCC frequently occurs in a background of endometriosis. The complete mutational landscape, the molecular basis of the transformation of endometriosis and patterns of clonal evolution in CCC are not understood.

We performed whole genome sequencing and gene expression profiling on 19 CCC to uncover candidate somatic alterations (mutations and, copy number aberrations) and measure their effect on transcriptional networks as candidates for driver mutations. We then performed targeted deep sequencing on the primary tumor samples, metastases and, from a subset of cases, adjacent or distant typical and atypical endometriosis. We used statistical modeling approaches to validate mutations, quantify the degree of clonal diversity and trace patterns of selection through oncogenic transformation.

Mutations in ARID1A and PIK3CA were by far the most frequent aberrations seen in the cohort (ARID1A: 10/19 cases; PIK3CA: 8/19 cases). The majority of ARID1A mutant cases exhibited bi-allelic loss of function. Two non-ARID1A mutant cases showed alterations in other SWI/SNF complex components. Amongst the 24 most significant candidate drivers impacting expression, five genes (PIK3CA, CTNNB1, TP53, PPP2R1A and KRAS) were known drivers. No association between PIK3CA or ARID1A status with disease stage, genomic instability, or mutation load was observed. Analysis of deep sequencing data suggested the presence of multiple clones in every case. For each case with matching precursor lesions, we observed multiple mutations in at least one such lesion. Cases with ARID1A and PIK3CA mutations always showed evidence of these mutations in their precursor lesions. The proportion of mutations from the primary tumor that were also present in precursor lesions varied widely across the cohort from approximately 10% to nearly 100%.

Our data support both ARID1A and PIK3CA mutations as early events in CCC. The pattern of endometriosis transformation could be associated with somatic mutations in all cases. This suggests that candidate tumor-initiating mutations and global- or individually- targetable features should be a focus to improve management of this disease. Finally, we suggest that patterns of mutational conservation across the series of precursor lesions presents an opportunity for early screening of endometriosis tissues as an indicator of transformation potential.

Citation Format: Ali Bashashati, Michael Anglesio, Yikan Wang, Gavin Ha, Janine Senz, Winnie Yang, Steve Kalloger, Leah Prentice, Satoshi Yanagida, Clara Salamanca, Galina Soukhatcheva, Anthony Karnezis, Hector Chang, Martin Hirst, Anne-Marie Mes-Mason, Aikou Okamoto, Marco Marra, Blake Gilks, Sohrab Shah, David Huntsman. The somatic mutational landscape of ovarian clear cell carcinoma and its precursor lesions. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-312. doi:10.1158/1538-7445.AM2014-LB-312

TITAN: inference of copy number architectures in clonal cell populations from tumor whole-genome sequence data

The evolution of cancer genomes within a single tumor creates mixed cell populations with divergent somatic mutational landscapes. Inference of tumor subpopulations has been disproportionately focused on the assessment of somatic point mutations, whereas computational methods targeting evolutionary dynamics of copy number alterations (CNA) and loss of heterozygosity (LOH) in whole-genome sequencing data remain underdeveloped. We present a novel probabilistic model, TITAN, to infer CNA and LOH events while accounting for mixtures of cell populations, thereby estimating the proportion of cells harboring each event. We evaluate TITAN on idealized mixtures, simulating clonal populations from whole-genome sequences taken from genomically heterogeneous ovarian tumor sites collected from the same patient. In addition, we show in 23 whole genomes of breast tumors that the inference of CNA and LOH using TITAN critically informs population structure and the nature of the evolving cancer genome. Finally, we experimentally validated subclonal predictions using fluorescence in situ hybridization (FISH) and single-cell sequencing from an ovarian cancer patient sample, thereby recapitulating the key modeling assumptions of TITAN.

Distinct evolutionary trajectories of primary high-grade serous ovarian cancers revealed through spatial mutational profiling

High-grade serous ovarian cancer (HGSC) is characterized by poor outcome, often attributed to the emergence of treatment-resistant subclones. We sought to measure the degree of genomic diversity within primary, untreated HGSCs to examine the natural state of tumour evolution prior to therapy. We performed exome sequencing, copy number analysis, targeted amplicon deep sequencing and gene expression profiling on 31 spatially and temporally separated HGSC tumour specimens (six patients), including ovarian masses, distant metastases and fallopian tube lesions. We found widespread intratumoural variation in mutation, copy number and gene expression profiles, with key driver alterations in genes present in only a subset of samples (eg PIK3CA, CTNNB1, NF1). On average, only 51.5% of mutations were present in every sample of a given case (range 10.2-91.4%), with TP53 as the only somatic mutation consistently present in all samples. Complex segmental aneuploidies, such as whole-genome doubling, were present in a subset of samples from the same individual, with divergent copy number changes segregating independently of point mutation acquisition. Reconstruction of evolutionary histories showed one patient with mixed HGSC and endometrioid histology, with common aetiologic origin in the fallopian tube and subsequent selection of different driver mutations in the histologically distinct samples. In this patient, we observed mixed cell populations in the early fallopian tube lesion, indicating that diversity arises at early stages of tumourigenesis. Our results revealed that HGSCs exhibit highly individual evolutionary trajectories and diverse genomic tapestries prior to therapy, exposing an essential biological characteristic to inform future design of personalized therapeutic solutions and investigation of drug-resistance mechanisms.

DriverNet: uncovering the impact of somatic driver mutations on transcriptional networks in cancer

Simultaneous interrogation of tumor genomes and transcriptomes is underway in unprecedented global efforts. Yet, despite the essential need to separate driver mutations modulating gene expression networks from transcriptionally inert passenger mutations, robust computational methods to ascertain the impact of individual mutations on transcriptional networks are underdeveloped. We introduce a novel computational framework, DriverNet, to identify likely driver mutations by virtue of their effect on mRNA expression networks. Application to four cancer datasets reveals the prevalence of rare candidate driver mutations associated with disrupted transcriptional networks and a simultaneous modulation of oncogenic and metabolic networks, induced by copy number co-modification of adjacent oncogenic and metabolic drivers. DriverNet is available on Bioconductor or at http://compbio.bccrc.ca/software/drivernet/.

B cells with high side scatter parameter by flow cytometry correlate with inferior survival in diffuse large B-cell lymphoma

Despite advances in the understanding of diffuse large B-cell lymphoma (DLBCL) biology, only the clinically based International Prognostic Index (IPI) is used routinely for risk stratification at diagnosis. To find novel prognostic markers, we analyzed flow cytometric data from 229 diagnostic DLBCL samples using an automated multiparameter data analysis approach developed in our laboratory. By using the developed automated data analysis pipeline, we identified 71 of 229 cases as having more than 35% B cells with a high side scatter (SSC) profile, a parameter reflecting internal cellular complexity. This high SSC B-cell feature was associated with inferior overall and progression-free survival (P = .001 and P = .01, respectively) and remained a significant predictor of overall survival in multivariate Cox regression analysis (IPI, P = .001; high SSC, P = .004; rituximab, P = .53). This study suggests that high SSC among B cells may serve as a useful biomarker to identify patients with DLBCL at high risk for relapse. This is of particular interest because this biomarker is readily available in most clinical laboratories without significant alteration to existing routine diagnostic strategies or incurring additional costs.

The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups

The elucidation of breast cancer subgroups and their molecular drivers requires integrated views of the genome and transcriptome from representative numbers of patients. We present an integrated analysis of copy number and gene expression in a discovery and validation set of 997 and 995 primary breast tumours, respectively, with long-term clinical follow-up. Inherited variants (copy number variants and single nucleotide polymorphisms) and acquired somatic copy number aberrations (CNAs) were associated with expression in ~40% of genes, with the landscape dominated by cis- and trans-acting CNAs. By delineating expression outlier genes driven in cis by CNAs, we identified putative cancer genes, including deletions in PPP2R2A, MTAP and MAP2K4. Unsupervised analysis of paired DNA–RNA profiles revealed novel subgroups with distinct clinical outcomes, which reproduced in the validation cohort. These include a high-risk, oestrogen-receptor-positive 11q13/14 cis-acting subgroup and a favourable prognosis subgroup devoid of CNAs. Trans-acting aberration hotspots were found to modulate subgroup-specific gene networks, including a TCR deletion-mediated adaptive immune response in the ‘CNA-devoid’ subgroup and a basal-specific chromosome 5 deletion-associated mitotic network. Our results provide a novel molecular stratification of the breast cancer population, derived from the impact of somatic CNAs on the transcriptome.

JointSNVMix: a probabilistic model for accurate detection of somatic mutations in normal/tumour paired next-generation sequencing data

MOTIVATION

Identification of somatic single nucleotide variants (SNVs) in tumour genomes is a necessary step in defining the mutational landscapes of cancers. Experimental designs for genome-wide ascertainment of somatic mutations now routinely include next-generation sequencing (NGS) of tumour DNA and matched constitutional DNA from the same individual. This allows investigators to control for germline polymorphisms and distinguish somatic mutations that are unique to the tumour, thus reducing the burden of labour-intensive and expensive downstream experiments needed to verify initial predictions. In order to make full use of such paired datasets, computational tools for simultaneous analysis of tumour-normal paired sequence data are required, but are currently under-developed and under-represented in the bioinformatics literature.

RESULTS

In this contribution, we introduce two novel probabilistic graphical models called JointSNVMix1 and JointSNVMix2 for jointly analysing paired tumour-normal digital allelic count data from NGS experiments. In contrast to independent analysis of the tumour and normal data, our method allows statistical strength to be borrowed across the samples and therefore amplifies the statistical power to identify and distinguish both germline and somatic events in a unified probabilistic framework.

AVAILABILITY

The JointSNVMix models and four other models discussed in the article are part of the JointSNVMix software package available for download at http://compbio.bccrc.ca

CONTACT

sshah@bccrc.ca

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Automated analysis of multidimensional flow cytometry data improves diagnostic accuracy between mantle cell lymphoma and small lymphocytic lymphoma

Mantle cell lymphoma (MCL) and small lymphocytic lymphoma (SLL) exhibit similar but distinct immunophenotypic profiles. Many cases can be diagnosed readily by flow cytometry (FCM) alone; however, ambiguous cases are frequently encountered and necessitate additional studies, including immunohistochemical staining for cyclin D1 and fluorescence in situ hybridization for IgH-CCND1 rearrangement. To determine if greater diagnostic accuracy could be achieved from FCM data alone, we developed an unbiased, machine-based algorithm to identify features that best distinguish between the 2 diseases. By applying conventional diagnostic criteria to the flow cytometry data, we were able to assign 28 of 44 (64%) MCL and 48 of 70 (69%) SLL cases correctly. In contrast, we were able to assign all 44 (100%) MCL and 68 of 70 (97%) SLL cases correctly using a novel set of criteria, as identified by our automated approach. The most discriminating feature was the CD20/CD23 mean fluorescence intensity ratio, and we found unexpectedly that inclusion of FMC7 expression in the diagnostic algorithm actually reduced its accuracy. This study demonstrates that computational methods can be used on existing clinical FCM data to improve diagnostic accuracy and suggests similar computational approaches could be used to identify novel prognostic markers and perhaps subdivide existing or define new diagnostic entities.

Feature-based classifiers for somatic mutation detection in tumour-normal paired sequencing data

Motivation: The study of cancer genomes now routinely involves using next-generation sequencing technology (NGS) to profile tumours for single nucleotide variant (SNV) somatic mutations. However, surprisingly few published bioinformatics methods exist for the specific purpose of identifying somatic mutations from NGS data and existing tools are often inaccurate, yielding intolerably high false prediction rates. As such, the computational problem of accurately inferring somatic mutations from paired tumour/normal NGS data remains an unsolved challenge.

Results: We present the comparison of four standard supervised machine learning algorithms for the purpose of somatic SNV prediction in tumour/normal NGS experiments. To evaluate these approaches (random forest, Bayesian additive regression tree, support vector machine and logistic regression), we constructed 106 features representing 3369 candidate somatic SNVs from 48 breast cancer genomes, originally predicted with naive methods and subsequently revalidated to establish ground truth labels. We trained the classifiers on this data (consisting of 1015 true somatic mutations and 2354 non-somatic mutation positions) and conducted a rigorous evaluation of these methods using a cross-validation framework and hold-out test NGS data from both exome capture and whole genome shotgun platforms. All learning algorithms employing predictive discriminative approaches with feature selection improved the predictive accuracy over standard approaches by statistically significant margins. In addition, using unsupervised clustering of the ground truth ‘false positive’ predictions, we noted several distinct classes and present evidence suggesting non-overlapping sources of technical artefacts illuminating important directions for future study.

Availability: Software called MutationSeq and datasets are available from http://compbio.bccrc.ca.

Contact:saparicio@bccrc.ca

Supplementary information:Supplementary data are available at Bioinformatics online.

Per-channel basis normalization methods for flow cytometry data

Between-sample variation in high-throughput flow cytometry data poses a significant challenge for analysis of large-scale data sets, such as those derived from multicenter clinical trials. It is often hard to match biologically relevant cell populations across samples because of technical variation in sample acquisition and instrumentation differences. Thus, normalization of data is a critical step before analysis, particularly in large-scale data sets from clinical trials, where group-specific differences may be subtle and patient-to-patient variation common. We have developed two normalization methods that remove technical between-sample variation by aligning prominent features (landmarks) in the raw data on a per-channel basis. These algorithms were tested on two independent flow cytometry data sets by comparing manually gated data, either individually for each sample or using static gating templates, before and after normalization. Our results show a marked improvement in the overlap between manual and static gating when the data are normalized, thereby facilitating the use of automated analyses on large flow cytometry data sets. Such automated analyses are essential for high-throughput flow cytometry.