930PA RETROSPECTIVE SINGLE INSTITUTION STUDY EVALUATING CLINICAL OUTCOME AND PROGNOSTIC MARKERS FOR ENDOMETRIAL AND OVARIAN CARCINOSARCOMAS (CS)

Abstract Aim: To assess whether: (1) clinical outcomes for ovarian vs. endometrial CS were comparable (2) the epithelial histological subtype or presence of heterologous elements in the mesenchymal component impacted overall survival (OS) (3) a personal or family history (FHx) of cancer was associated with OS. Methods: Women treated at University College London Hospital Gynae-oncology Unit between April 2002 – 2011 were identified using electronic records. Patient clinical data and full histological reports were collected retrospectively from the medical records. Results: 82 patients (72 endometrial; 10 ovarian) with a median age of 67 years were identified. 22% had a prior cancer history and 3.6% had previously received abdominal/pelvic radiotherapy (RT). Median OS was 26.7 months. Analysis for OS according to baseline characteristicsOS #events/nHR(95% CI)p-valueOriginal Tumour siteEndometrial Ovarian33/72 6/101.0 1.33 (0.56 -3.17)0.5FIGO Stage I/II III/IV10/37 22/331.0 3.22 (1.52, 6.82)0.001Epithelial type Poorly/un-differentiated Other15/39 16/311.0 1.55 (0.75,3.18)0.23Cancer History No Yes22/56 10/181.00 1.23 (0.59, 2.55)0.58Cancer FHx No Yes26/56 7/151.00 1.08 (0.47, 2.49)0.86 All ovarian CS patients had surgical resection (SR), 90% completed 6 cycles of postoperative carboplatin/paclitaxel chemotherapy (CT) with 20% complete responses (CR) and 40% partial responses (PR). 88% of endometrial CS patients were treated with SR, CT +/- RT. The response rates (RR) after CT were 70% CR and 9% PR. Neither the epithelial subtypes, nor the presence of a heterologous component showed statistically significant difference for OS. Conclusions: Clinical outcome for ovarian (12%) vs. endometrial CS were comparable in regards to OS. FIGO stage was the only statistically significant prognostic factor. There was a trend for worse OS in CS with poorly/un- differentiated adenocarcinoma and/or absence of heterologous components although statistically not significant in this cohort. This warrants further clinical evaluation in a larger study. Disclosure: All authors have declared no conflicts of interest.