Feasibility of a platform trial based on molecular analysis in rare gynecologic cancers.

Abstract

e14585 Background: Challenges in patient accrual limit successful conduct of clinical trials in rare gynecologic malignancies. A platform trial using molecular analysis to guide therapy choice would optimize enrollment but would require actionable mutations (AM) to be feasible. Therefore, we sought to determine if enough AMs exist in rare gynecologic cancers (CA) to justify the design of a platform trial. Methods: We compiled all molecular profiling data performed at Caris Life Sciences through February 2019. Non-epithelial ovarian cancer (OC), rare epithelial OC, non-endometrioid uterine, neuroendocrine gynecologic, and vulvar CA were compiled. AMs were identified. Results: Among 280 cases of non-epithelial OC, 228 malignant stromal tumors and 52 germ cell tumors were identified. AMs in stromal tumors included KMT2D (9.8%), PIK3CA (7.8%), and 16 other AMs ( < 1% each). AMs in germ cell tumors included PIK3CA (20%), KIT (10%), PTEN (10%), TMB (8.2%), CDKN2A (8%), ARID1A (6%), and 16 other AMs ( < 6% each). Among 702 rare epithelial OCs, 132 mucinous CAs, 91 low-grade serous/endometrioid, 367 clear cell, and 112 carcinosarcoma (CS) cases were identified. AMs in mucinous CAs included KRAS (61.8%), CDKN2A (15.3%), PIK3CA (12.7%), ARID1A (9.9%), GNAS (6.9%), and 20 other AMs ( < 5% each). AMs in low-grade serous/endometrioid included KRAS (28.9%), ARID1A (13.3%), PTEN (8.9%), PIK3CA (7.8%), MSI (6.2%), and 15 other AMs ( < 6% each). AMs in clear cell OCs included ARID1A (53.4%), PIK3CA (47.5%), KRAS (11.6%), PTEN (6.8%), and 36 other AMs ( < 6% each). AMs in ovarian CSs included PIK3CA (7.2%), ARID1A (6.2%), and 22 other AMs ( < 5% each). Among 4864 rare uterine CA, 4670 sarcomas, 173 serous CAs, and 21 clear cell CAs were identified. AMs in uterine sarcomas included PTEN (40.1%), PIK3CA (37.1%), ARID1A (31.6%), MSI (20.7%), KRAS (16.8%), TMB (14%), PIK3R1 (10%), and 41 other AMs ( < 10% each). Among 102 gynecologic neuroendocrine tumors, AMs included ARID1A (29.2%), PIK3CA (18%), PTEN (18%), MSI (6.5%), TMB (6%), and 4 other AMs ( < 6% each). Among 59 vulvar CAs, AMs included CDKN2A (28.7%), PIK3CA (17.8%), NOTCH1 (9.6%), TMB (6.8%), and 24 other AMs ( < 5% each). Conclusions: A variety of actionable mutations are present in every type of rare gynecologic cancer evaluated, but mutations specific to individual histologies are not reliably present. This supports the use of molecular profiling to identify potential targets and supports a platform trial strategy to study rare gynecologic cancers.