Biomarker Based Ovarian Carcinoma Typing: A Retrospective Analysis of the OTTA Consortium
Sandra L Lee, Steve E Kalloger, Susan J Ramus, Ellen Goode, Maire Duggan, David Huntsman, Blake C Gilks, Kobel Martin. University of Calgary, Calgary, AB, Canada; University of British Columbia, Vancouver, BC, Canada; University of Southern California, Los Angeles, CA; Mayo Clinic, Rochester, MN
Background: The five major histological types of ovarian carcinoma are high-grade serous (HGSC), clear cell (CCC), endometrioid (EC), mucinous (MC), and low-grade serous (LGSC). Histological type is associated with outcome and is predictive of therapeutic response. The aim of this study was to evaluate the agreement of histological type with a 10 marker immunohistochemical (IHC) calculator of ovarian subtype probability (COSP version 2, COSPv2 based on Kalloger et al. Mod Pathol 2011:24:512) with the original diagnosis and to determine if reclassification impacts outcome.
Design: The Ovarian Tumor Tissue Analysis (OTTA) consortium is dedicated to tissue based ovarian cancer research. The final test set (N=524) was obtained from three OTTA studies: Mayo clinic, UKOPS, and HOPE studies. The following three type assessments were compared: original diagnosis, COSPv2, and a WT1 assisted review of the tissue microarray (TMA) cores (core review). Consensus type was defined as agreement between all three assessments. Statistical analysis was done with JMP version 10.0 (SAS).
Results: Table 1 depicts agreement between the three methods. For the 26% of cases in which COSPv2 disagreed with the original diagnosis, the core review and COSPv2 agreed in 13%, the core review with the original diagnosis in 7%, and in 6% of cases there was no agreement between the three methods. A consensus type was established in 66% of cases. Assuming consensus type represents the most accurate method of typing, nearly 50% of EC were overcalled in the original diagnosis. Reclassification of EC by consensus type resulted in a higher 5-year survival (57% to 93%), reduction in grade 3 EC (50% to 22%), and decreased WT1 expression (40% to 3%). Only consensus type was an independent predictor of outcome when adjusted for age and stage.
|Original Diganosis||COSPv2||Core Review|