Search

Abstract Investigation of histopathology slides by pathologists is an indispensable component of the routine diagnosis of cancer. Artificial intelligence (AI) has the potential to enhance diagnostic accuracy, improve efficiency, and patient outcomes in clinical pathology. However, variations in tissue preparation, staining protocols, and histopathology slide digitization could result in over-fitting of deep learning models when trained on the data from only one center, thereby underscoring the necessity to generalize deep learning networks for multi-center use. Several techniques, including the use of grayscale images, color normalization techniques, and Adversarial Domain Adaptation (ADA) have been suggested to generalize deep learning algorithms, but there are limitations to their effectiveness and discriminability. Convolutional Neural Networks (CNNs) exhibit higher sensitivity to variations in the amplitude spectrum, whereas humans predominantly rely on phase-related components for object recognition. As such, we propose Adversarial fourIer-based Domain Adaptation (AIDA) which applies the advantages of a Fourier transform in adversarial domain adaptation. We conducted a comprehensive examination of subtype classification tasks in four cancers, incorporating cases from multiple medical centers. Specifically, the datasets included multi-center data for 1113 ovarian cancer cases, 247 pleural cancer cases, 422 bladder cancer cases, and 482 breast cancer cases. Our proposed approach significantly improved performance, achieving superior classification results in the target domain, surpassing the baseline, color augmentation and normalization techniques, and ADA. Furthermore, extensive pathologist reviews suggested that our proposed approach, AIDA, successfully identifies known histotype-specific features. This superior performance highlights AIDA’s potential in addressing generalization challenges in deep learning models for multi-center histopathology datasets.

Abstract Endometrial cancer (EC) has four molecular subtypes with strong prognostic value and therapeutic implications. The most common subtype (NSMP; No Specific Molecular Profile) is assigned after exclusion of the defining features of the other three molecular subtypes and includes patients with heterogeneous clinical outcomes. In this study, we employ artificial intelligence (AI)-powered histopathology image analysis to differentiate between p53abn and NSMP EC subtypes and consequently identify a sub-group of NSMP EC patients that has markedly inferior progression-free and disease-specific survival (termed ‘p53abn-like NSMP’), in a discovery cohort of 368 patients and two independent validation cohorts of 290 and 614 from other centers. Shallow whole genome sequencing reveals a higher burden of copy number abnormalities in the ‘p53abn-like NSMP’ group compared to NSMP, suggesting that this group is biologically distinct compared to other NSMP ECs. Our work demonstrates the power of AI to detect prognostically different and otherwise unrecognizable subsets of EC where conventional and standard molecular or pathologic criteria fall short, refining image-based tumor classification. This study’s findings are applicable exclusively to females.

Abstract We performed a deep proteogenomic analysis of bulk tumor and laser microdissection enriched tumor cell populations from high-grade serous ovarian cancer (HGSOC) tissue specimens spanning a broad spectrum of purity. We identified patients with longer progression-free survival had increased immune-related signatures and validated proteins correlating with tumor-infiltrating lymphocytes in 65 tumors from an independent cohort of HGSOC patients, as well as with overall survival in an additional 126 HGSOC patient cohort. We identified that homologous recombination deficient (HRD) tumors are enriched in pathways associated with metabolism and oxidative phosphorylation that we validated in independent patient cohorts. We further identified that polycomb complex protein BMI-1 is elevated in HR proficient (HRP) tumors, that elevated BMI-1 correlates with poor overall survival in HRP but not HRD HGSOC patients, and that HRP HGSOC cells are uniquely sensitive to BMI-1 inhibition.

Abstract Survival from ovarian cancer depends on the resection status after primary surgery. We performed genome-wide association analyses for resection status of 7705 ovarian cancer patients, including 4954 with high-grade serous carcinoma (HGSOC), to identify variants associated with residual disease. The most significant association with resection status was observed for rs72845444, upstream of MGMT, in HGSOC (p = 3.9 × 10−8). In gene-based analyses, PPP2R5C was the most strongly associated gene in HGSOC after stage adjustment. In an independent set of 378 ovarian tumours from the AGO-OVAR 11 study, variants near MGMT and PPP2R5C correlated with methylation and transcript levels, and PPP2R5C mRNA levels predicted progression-free survival in patients with residual disease. MGMT encodes a DNA repair enzyme, and PPP2R5C encodes the B56γ subunit of the PP2A tumour suppressor. Our results link heritable variation at these two loci with resection status in HGSOC.

Abstract SMARCA4 (BRG1) and SMARCA2 (BRM) are the two paralogous ATPases of the SWI/SNF chromatin remodeling complexes frequently inactivated in cancers. Cells deficient in either ATPase have been shown to depend on the remaining counterpart for survival. Contrary to this paralog synthetic lethality, concomitant loss of SMARCA4/2 occurs in a subset of cancers associated with very poor outcomes. Here, we uncover that SMARCA4/2-loss represses expression of the glucose transporter GLUT1, causing reduced glucose uptake and glycolysis accompanied with increased dependency on oxidative phosphorylation (OXPHOS); adapting to this, these SMARCA4/2-deficient cells rely on elevated SLC38A2, an amino acid transporter, to increase glutamine import for fueling OXPHOS. Consequently, SMARCA4/2-deficient cells and tumors are highly sensitive to inhibitors targeting OXPHOS or glutamine metabolism. Furthermore, supplementation of alanine, also imported by SLC38A2, restricts glutamine uptake through competition and selectively induces death in SMARCA4/2-deficient cancer cells. At a clinically relevant dose, alanine supplementation synergizes with OXPHOS inhibition or conventional chemotherapy eliciting marked antitumor activity in patient-derived xenografts. Our findings reveal multiple druggable vulnerabilities of SMARCA4/2-loss exploiting a GLUT1/SLC38A2-mediated metabolic shift. Particularly, unlike dietary deprivation approaches, alanine supplementation can be readily applied to current regimens for better treatment of these aggressive cancers.

Abstract Our objective was to test whether p53 expression status is associated with survival for women diagnosed with the most common ovarian carcinoma histotypes (high‐grade serous carcinoma [HGSC], endometrioid carcinoma [EC], and clear cell carcinoma [CCC]) using a large multi‐institutional cohort from the Ovarian Tumor Tissue Analysis (OTTA) consortium. p53 expression was assessed on 6,678 cases represented on tissue microarrays from 25 participating OTTA study sites using a previously validated immunohistochemical (IHC) assay as a surrogate for the presence and functional effect of TP53 mutations. Three abnormal expression patterns (overexpression, complete absence, and cytoplasmic) and the normal (wild type) pattern were recorded. Survival analyses were performed by histotype. The frequency of abnormal p53 expression was 93.4% (4,630/4,957) in HGSC compared to 11.9% (116/973) in EC and 11.5% (86/748) in CCC. In HGSC, there were no differences in overall survival across the abnormal p53 expression patterns. However, in EC and CCC, abnormal p53 expression was associated with an increased risk of death for women diagnosed with EC in multivariate analysis compared to normal p53 as the reference (hazard ratio [HR] = 2.18, 95% confidence interval [CI] 1.36–3.47, p = 0.0011) and with CCC (HR = 1.57, 95% CI 1.11–2.22, p = 0.012). Abnormal p53 was also associated with shorter overall survival in The International Federation of Gynecology and Obstetrics stage I/II EC and CCC. Our study provides further evidence that functional groups of TP53 mutations assessed by abnormal surrogate p53 IHC patterns are not associated with survival in HGSC. In contrast, we validate that abnormal p53 IHC is a strong independent prognostic marker for EC and demonstrate for the first time an independent prognostic association of abnormal p53 IHC with overall survival in patients with CCC.

IntroductionTargeted-immunotherapies such as antibody-drug conjugates (ADC), chimeric antigen receptor (CAR) T cells or bispecific T-cell engagers (eg, BiTE®) all aim to improve cancer treatment by directly targeting cancer cells while sparing healthy tissues. Success of these therapies requires tumor antigens that are abundantly expressed and, ideally, tumor specific. The CD34-related stem cell sialomucin, podocalyxin (PODXL), is a promising target as it is overexpressed on a variety of tumor types and its expression is consistently linked to poor prognosis. However, PODXL is also expressed in healthy tissues including kidney podocytes and endothelia. To circumvent this potential pitfall, we developed an antibody, named PODO447, that selectively targets a tumor-associated glycoform of PODXL. This tumor glycoepitope is expressed by 65% of high-grade serous ovarian carcinoma (HGSOC) tumors.MethodsIn this study we characterize these PODO447-expressing tumors as a distinct subset of HGSOC using four different patient cohorts that include pre-chemotherapy, post-neoadjuvant chemotherapy (NACT) and relapsing tumors as well as tumors from various peritoneal locations.ResultsWe find that the PODO447 epitope expression is similar across tumor locations and negligibly impacted by chemotherapy. Invariably, tumors with high levels of the PODO447 epitope lack infiltrating CD8+ T cells and CD20+ B cells/plasma cells, an immune phenotype consistently associated with poor outcome.DiscussionWe conclude that the PODO447 glycoepitope is an excellent biomarker of immune “cold” tumors and a candidate for the development of targeted-therapies for these hard-to-treat cancers.

DICER1 syndrome is associated with a predisposition to multiple tumor types. Here, the authors identify and characterize 3 molecular subgroups of mesenchymal tumors with DICER1 mutations.

Abstract In this study, we evaluate the impact of whole genome and transcriptome analysis (WGTA) on predictive molecular profiling and histologic diagnosis in a cohort of advanced malignancies. WGTA was used to generate reports including molecular alterations and site/tissue of origin prediction. Two reviewers analyzed genomic reports, clinical history, and tumor pathology. We used National Comprehensive Cancer Network (NCCN) consensus guidelines, Food and Drug Administration (FDA) approvals, and provincially reimbursed treatments to define genomic biomarkers associated with approved targeted therapeutic options (TTOs). Tumor tissue/site of origin was reassessed for most cases using genomic analysis, including a machine learning algorithm (Supervised Cancer Origin Prediction Using Expression [SCOPE]) trained on The Cancer Genome Atlas data. WGTA was performed on 652 cases, including a range of primary tumor types/tumor sites and 15 malignant tumors of uncertain histogenesis (MTUH). At the time WGTA was performed, alterations associated with an approved TTO were identified in 39 (6%) cases; 3 of these were not identified through routine pathology workup. In seven (1%) cases, the pathology workup either failed, was not performed, or gave a different result from the WGTA. Approved TTOs identified by WGTA increased to 103 (16%) when applying 2021 guidelines. The histopathologic diagnosis was reviewed in 389 cases and agreed with the diagnostic consensus after WGTA in 94% of non‐MTUH cases (n = 374). The remainder included situations where the morphologic diagnosis was changed based on WGTA and clinical data (0.5%), or where the WGTA was non‐contributory (5%). The 15 MTUH were all diagnosed as specific tumor types by WGTA. Tumor board reviews including WGTA agreed with almost all initial predictive molecular profile and histopathologic diagnoses. WGTA was a powerful tool to assign site/tissue of origin in MTUH. Current efforts focus on improving therapeutic predictive power and decreasing cost to enhance use of WGTA data as a routine clinical test.

Abstract Background To drive translational medicine, modern day biobanks need to integrate with other sources of data (clinical, genomics) to support novel data-intensive research. Currently, vast amounts of research and clinical data remain in silos, held and managed by individual researchers, operating under different standards and governance structures; a framework that impedes sharing and effective use of data. In this article, we describe the journey of British Columbia’s Gynecological Cancer Research Program (OVCARE) in moving a traditional tumour biobank, outcomes unit, and a collection of data silos, into an integrated data commons to support data standardization and resource sharing under collaborative governance, as a means of providing the gynecologic cancer research community in British Columbia access to tissue samples and associated clinical and molecular data from thousands of patients. Results Through several engagements with stakeholders from various research institutions within our research community, we identified priorities and assessed infrastructure needs required to optimize and support data collections, storage and sharing, under three main research domains: (1) biospecimen collections, (2) molecular and genomics data, and (3) clinical data. We further built a governance model and a resource portal to implement protocols and standard operating procedures for seamless collections, management and governance of interoperable data, making genomic, and clinical data available to the broader research community. Conclusions Proper infrastructures for data collection, sharing and governance is a translational research imperative. We have consolidated our data holdings into a data commons, along with standardized operating procedures to meet research and ethics requirements of the gynecologic cancer community in British Columbia. The developed infrastructure brings together, diverse data, computing frameworks, as well as tools and applications for managing, analyzing, and sharing data. Our data commons bridges data access gaps and barriers to precision medicine and approaches for diagnostics, treatment and prevention of gynecological cancers, by providing access to large datasets required for data-intensive science.

Abstract Ovarian carcinoma histotypes are distinct diseases with variable clinical outcomes and response to treatment. There is a need for new subtype‐specific treatment modalities, especially for women with widespread and chemo‐resistant disease. Stimulator of interferon genes (STING) is a part of the cGAS–STING pathway that mediates innate immune defence against infectious DNA‐containing pathogens and also detects tumour‐derived DNA and generates intrinsic antitumour immunity. The STING signalling pathway is suppressed by several mechanisms in a variety of malignant diseases and, in some cancers that may be a requirement for cellular transformation. The aim of this study was to use immunohistochemistry to evaluate STING protein expression across normal tissue, paratubal and ovarian cysts, and ovarian tumour histotypes including ovarian carcinomas. Herein, we show that the fallopian tube ciliated cells express STING protein, whereas the secretory cells are negative. STING expression differs among ovarian cancer histotypes; low‐grade serous ovarian carcinomas and serous borderline tumours have uniform high STING expression, while high‐grade serous and endometrioid carcinomas have heterogeneous expression, and clear cell and mucinous carcinomas show low expression. As low‐grade serous carcinomas are known to be genomically stable and typically lack a prominent host immune response, the consistently high STING expression is unexpected. High STING expression may reflect pathway activation or histogenesis and the mechanisms may be different in different ovarian carcinoma histotypes. Further studies are needed to determine whether the STING signalling pathway is active and whether these tumours would be candidates for therapeutic interventions that trigger innate immunity activation.

Abstract 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.

The role of non-coding somatic mutations in ovarian cancer is unclear. Here, the authors integrate genomic and epigenomic data from patient samples to show that these mutations frequently converge on the PAX8 transcriptional network.

Abstract Early relapse after platinum chemotherapy in epithelial ovarian cancer (EOC) portends poor survival. A‐priori identification of platinum resistance is therefore crucial to improve on standard first‐line carboplatin–paclitaxel treatment. The DNA repair pathway homologous recombination (HR) repairs platinum‐induced damage, and the HR recombinase RAD51 is overexpressed in cancer. We therefore designed a REMARK‐compliant study of pre‐treatment RAD51 expression in EOC, using fluorescent quantitative immunohistochemistry (qIHC) to overcome challenges in quantitation of protein expression in situ. In a discovery cohort (n = 284), RAD51‐High tumours had shorter progression‐free and overall survival compared to RAD51‐Low cases in univariate and multivariate analyses. The association of RAD51 with relapse/survival was validated in a carboplatin monotherapy SCOTROC4 clinical trial cohort (n = 264) and was predominantly noted in HR‐proficient cancers (Myriad HRDscore

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Similarities in cancers can be studied to interrogate their etiology. Here, the authors use genome-wide association study summary statistics from six cancer types based on 296,215 cases and 301,319 controls of European ancestry, showing that solid tumours arising from different tissues share a degree of common germline genetic basis.

Abstract Background Although low-grade serous ovarian cancer (LGSC) is rare, case-fatality rates are high as most patients present with advanced disease and current cytotoxic therapies are not overly effective. Recognizing that these cancers may be driven by MAPK pathway activation, MEK inhibitors (MEKi) are being tested in clinical trials. LGSC respond to MEKi only in a subgroup of patients, so predictive biomarkers and better therapies will be needed. Methods We evaluated a number of patient-derived LGSC cell lines, previously classified according to their MEKi sensitivity. Two cell lines were genomically compared against their matching tumors samples. MEKi-sensitive and MEKi-resistant lines were compared using whole exome sequencing and reverse phase protein array. Two treatment combinations targeting MEKi resistance markers were also evaluated using cell proliferation, cell viability, cell signaling, and drug synergism assays. Results Low-grade serous ovarian cancer cell lines recapitulated the genomic aberrations from their matching tumor samples. We identified three potential predictive biomarkers that distinguish MEKi sensitive and resistant lines: KRAS mutation status, and EGFR and PKC-alpha protein expression. The biomarkers were validated in three newly developed LGSC cell lines. Sub-lethal combination of MEK and EGFR inhibition showed drug synergy and caused complete cell death in two of four MEKi-resistant cell lines tested. Conclusions KRAS mutations and the protein expression of EGFR and PKC-alpha should be evaluated as predictive biomarkers in patients with LGSC treated with MEKi. Combination therapy using a MEKi with EGFR inhibition may represent a promising new therapy for patients with MEKi-resistant LGSC.

ARID1A is a core DNA-binding subunit of the BAF chromatin remodeling complex, and is lost in up to 7% of all cancers. The frequency of ARID1A loss increases in certain cancer types, such as clear cell ovarian carcinoma where ARID1A protein is lost in about 50% of cases. While the impact of ARID1A loss on the function of the BAF chromatin remodeling complexes is likely to drive oncogenic gene expression programs in specific contexts, ARID1A also binds genome stability regulators such as ATR and TOP2. Here we show that ARID1A loss leads to DNA replication stress associated with R-loops and transcription-replication conflicts in human cells. These effects correlate with altered transcription and replication dynamics in ARID1A knockout cells and to reduced TOP2A binding at R-loop sites. Together this work extends mechanisms of replication stress in ARID1A deficient cells with implications for targeting ARID1A deficient cancers.

Small cell carcinoma of the ovary, hypercalcemic type (SCCOHT) is a rare and aggressive form of ovarian cancer. SCCOHT tumors have inactivating mutations in SMARCA4 (BRG1), one of the two mutually exclusive ATPases of the SWI/SNF chromatin remodeling complex. To address the role that BRG1 loss plays in SCCOHT tumorigenesis, we performed integrative multi-omic analyses in SCCOHT cell lines +/- BRG1 reexpression. BRG1 reexpression induced a gene and protein signature similar to an epithelial cell and gained chromatin accessibility sites correlated with other epithelial originating TCGA tumors. Gained chromatin accessibility and BRG1 recruited sites were strongly enriched for transcription-factor-binding motifs of AP-1 family members. Furthermore, AP-1 motifs were enriched at the promoters of highly upregulated epithelial genes. Using a dominant-negative AP-1 cell line, we found that both AP-1 DNA-binding activity and BRG1 reexpression are necessary for the gene and protein expression of epithelial genes. Our study demonstrates that BRG1 reexpression drives an epithelial-like gene and protein signature in SCCOHT cells that depends upon by AP-1 activity.

Dedifferentiated endometrial carcinoma (DDEC) is a rare but highly aggressive type of endometrial cancer, in which an undifferentiated carcinoma arises from a low-grade endometrioid endometrial carcinoma. The low-grade component is often eclipsed, likely due to an outgrowth of the undifferentiated component, and the tumor may appear as a pure undifferentiated endometrial carcinoma (UEC). We and others have recently identified inactivating mutations of SMARCA4, SMARCB1 or ARID1B, subunits of the SWI/SNF chromatin-remodeling complex, that are unique to the undifferentiated component and are present in a large portion of DDEC and UEC. However, the understanding of whether and how these mutations drive cancer progression and histologic dedifferentiation is hindered by lack of cell line models of DDEC or UEC. Here, we established the first UEC cell line, VOA1066, which is highly tumorigenic in vivo. This cell line has a stable genome with very few somatic mutations, which do include inactivating mutations of ARID1A and ARID1B (2 mutations each), and a heterozygous hotspot DICER1 mutation in its RNase IIIb domain. Immunohistochemistry staining confirmed the loss of ARID1B, but ARID1A staining was retained due to the presence of a truncating non-functional ARID1A protein. The heterozygous DICER1 hotspot mutation has little effect on microRNA biogenesis. No additional DICER1 hotspot mutations have been identified in a cohort of 33 primary tumors. Therefore, we have established the first UEC cell line with dual inactivation of both ARID1A and ARID1B as the main genomic feature. This cell line will be useful for studying the roles of ARID1A and ARID1B mutations in the development of UEC.

Abstract Background Ovarian carcinoma is the most lethal gynecological malignancy due to early dissemination and acquired resistance to platinum-based chemotherapy. Reliable markers that are independent and complementary to clinical parameters are needed to improve the management of patients with this disease. The Canadian Ovarian Experimental Unified Resource (COEUR) provides researchers with biological material and associated clinical data to conduct biomarker validation studies. Using standards defined by the Canadian Tissue Repository Network (CTRNet), we have previously demonstrated the quality of the biological material from this resource. Here we describe the clinical characteristics of the COEUR cohort. Methods With support from 12 Canadian ovarian cancer biobanks in Canada, we created a central retrospective cohort comprised of more than 2000 patient tissue samples with associated clinical data, including 1246 high-grade serous, 102 low-grade serous, 295 endometrioid, 259 clear cell and 89 mucinous carcinoma histotypes. A two-step reclassification process was applied to assure contemporary histological classification (histotyping). For each histotypes individually, we evaluated the association between the known clinico-pathological parameters (stage, cytoreduction, chemotherapy treatment, BRCA1 and BRCA2 mutation) and patient outcome by using Kaplan-Meier and Cox proportional hazard regression analyses. Results The median follow-up time of the cohort was 45 months and the 5-year survival rate for patients with high-grade serous carcinomas was 34%, in contrast to endometrioid carcinomas with 80% at 5 years. Survival profiles differed by histotype when stratified by stage or cytoreduction. Women with mucinous or clear cell carcinomas at advanced stage or with non-optimally debulked disease had the worst outcomes. In high-grade serous carcinoma, we observed significant association with longer survival in women harboring BRCA1 or BRCA2 mutation as compared to patients without detectable mutation. Conclusions Our results show the expected survival rates, as compared with current literature, in each histotype suggesting that the cohort is an unbiased representation of the five major histotypes. COEUR, a one stop comprehensive biorepository, has collected mature outcome data and relevant clinical data in a comprehensive manner allowing stratified analysis.