Millstein, Joshua

Abstract Background: High-grade serous ovarian cancer (HGSOC) is the most common and aggressive epithelial ovarian cancer. HGSOC has poor survival, with five-year survival of less than 40%, as it is often diagnosed at a late stage. The median overall survival (OS) is approximately four years. Prognostic gene expression signatures for HGSOC have been inconsistent and have not been translated into clinical practice. The aim of this study was to develop a robust prognostic signature for HGSOC, in order to identify high-risk patients requiring alternative treatments. Methods: A meta-analysis of 1,455 tumors was performed and 200 prognostic genes were selected for expression profiling. An additional 315 candidate genes were selected from the literature, drug-targetable pathways, and other hypotheses. NanoString expression analysis of the 513 genes was performed on formalin-fixed, paraffin-embedded (FFPE) tumor tissue from 3,769 women with HGSOC from the Ovarian Tumor Tissue Analysis (OTTA) Consortium. A prognostic model for 5-year survival was developed using regression-based and machine learning methods. The model was trained on 2,702 tumors and evaluated on an independent set of 1,067 tumors. A NanoString expression analysis was performed on 2,692 additional tumors, for a subset of 218 of the genes, including the prognostic signature. Results: Expression of 276 genes was associated with OS (false discovery rate (FDR) < 0.05) in single-gene analyses adjusted for age, stage, and race. The top five genes were TAP1, ZFHX4, CXCL9, FBN1, and PTGER3 (p << 0.001). A 101-gene prognostic signature was established, where a gain of one standard deviation in the gene expression score conferred a greater than two-fold increase in risk of death (HR (hazard ratio) = 2.35 [2.02, 2.71]; p << 0.001). Median survival by quintile group was 9.5, 5.4, 3.8, 3.2, and 2.3 years. Sensitivity analyses, applying the signature to specific patient groups, showed that the prognostic power of the signature is not explained by residual disease, treatment, BRCA status, or CD8 score. The prognostic signature was enriched in pathways with treatment implications, such as PI3K, GPCR, and immune. This signature is suitable for use in FFPE tumor material and future cases can be assigned an HR from the gene expression score. The signature was developed in cases that did not receive neoadjuvant treatment (NACT); however, application to NACT cases was assessed in the second NanoString expression analysis. Conclusion: The OTTA-SPOT (Ovarian Tumor Tissue Analysis consortium – Stratified Prognosis of Ovarian Tumors) gene expression signature may improve risk stratification in clinical trials by identifying patients who are least likely to achieve 5-year survival. Citation Format: Joshua Millstein, Timothy Budden, Ellen L. Goode, Michael S. Anglesio, Aline Talhouk, OTTA consortium, David G. Huntsman, David D. Bowtell, James D. Brenton, Jennifer A. Doherty, Paul P.D. Pharoah, Susan J. Ramus. The Ovarian Tumour Tissue Analysis Consortium: Stratified Prognosis of Ovarian Tumors (OTTA-SPOT) signature for high-grade serous ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr IA12.

Abstract Our understanding of ovarian cancer is complicated by disease heterogeneity. There are five major histological types, each with alterations in different pathways and distinct clinical outcomes. Even within these histological types there is clinical and molecular heterogeneity. The Ovarian Tumor Tissue Analysis (OTTA) consortium was established to validate prognostic markers for ovarian cancer in a large population of patients to allow stratification by histological type and molecular subtype. OTTA is currently a collaboration among 38 studies and has tissue microarrays (TMAs) from over 8,000 tumors available for immunohistochemical (IHC) analysis. In addition formalin fixed paraffin embedded (FFPE) tissue on slides or cores are available from the majority of OTTA studies for DNA and RNA analysis. Approximatley 60% of the cases are part of the Ovarian Cancer Association Consortium (OCAC) and have genotype data on over 200,000 single nucleotide polymorphisms and extensive epidemiological data. We recently performed IHC analysis of 12 markers using centralised staining and scoring to reduce heterogeneity and improve power to validate biomarkers. We showed that estrogen receptor and progesterone receptor were prognostic markers for endometrioid and high-grade serous carcinomas by performing the first robust subtype-specific analyses of these biomarkers in nearly 3000 patients. In contrast, FOLR1 expression was not associated with survival in 1507 high-grade serous carcinomas. We have shown that FFPE material from OTTA is suitable for RNA expression analysis on high through-put platforms such as the NanoString GX system. Expression analysis of 50 candidate genes in 244 high grade serous ovarian cancer cases has validated prognostic markers identified by The Cancer Genome Atlas (TCGA) project and investigated a series of novel candidate genes. IHC and RNA expression profiling from formalin fixed paraffin embedded (FFPE) tumor tissue is being used to better subclassify the ovarian tumors for further analysis. We would welcome groups with collections of ovarian tumor samples and corrsponding clinical data to join OTTA and participate in ongoing collaborative projects. Citation Format: Susan J. Ramus, Martin Köbel, Weiva Sieh, Michael S. Anglesio, Joshua Millstein, David D. Bowtell, James D. Brenton, Estrid Høgdall, Paul DP Pharoah, Ellen L. Goode, David G. Huntsman. The ovarian tumor tissue analysis (OTTA) consortium. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research: From Concept to Clinic; Sep 18-21, 2013; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2013;19(19 Suppl):Abstract nr A27.