Introduction Ovarian carcinosarcomas (OCS) account for approximately 5% ovarian cancer cases, but have been specifically excluded from large treatment-defining clinical trials. They present at late stage and have a poor prognosis. Originally considered to be a type of sarcoma, these biphasic tumours are now considered to be epithelial, with biphasic differentiation occurring during tumour development. Due to their exclusion from clinical trials, little is known of the biology of this disease. We sought to identify the genomic abnormalities in OCS. Materials /Patients and methods Archival FFPE samples from 18 OCS cases were identified. Next-generation sequencing (custom 377-gene panel; Illumina NextSeq500; 2 x 75bp PE) was performed on micro-dissected carcinomatous and sarcomatous components as well as metastatic deposits. Copy number (CN), indel and SNV analysis was performed on each component. Immunohistochemistry for immune cell populations and RNAseq analysis on matched carcinoma/sarcoma samples are on-going. Results Median age at diagnosis was 70.9 years (range 49.4-74.6). 11 cases were stage 3 or 4 at diagnosis. Median overall survival was 32.2 months. 17/18 cases had high-grade serous carcinoma (HGSC) morphology in the carcinoma component; the remaining case arose within a G2 endometrioid tumour. Tumours had genomic profiles similar to HGSC, including near-ubiquitous TP53 mutation (17/18 cases, including 17/17 HGSC-derived cases), CCNE1 amplification (4/18), BRCA2 loss or mutation (4/18), inactivation (by SNV) or disruption (by rearrangement) of NF1 and CDKN2A (2/18 each), deletion of RB1 and PTEN (2/18 each) and MYC amplification (1/18 cases). The endometrioid-derived case lacked TP53 mutations but contained mutations in KRAS and ERBB2 in both carcinoma and sarcoma components. As with HGSC, genomes were structurally and numerically chaotic with an average of 3.3 high-level gains and 1.4 likely homozygous losses per sample, as well as 3.9 sequence-level mutations. There were no consistent SNV/indel differences between carcinoma and sarcoma components, with at least one point mutation shared in all sample pairs. On average, carcinoma-sarcoma pairs differed by a single mutation, with 100% SNV concordance in 50% of cases. By contrast, cases showed much more diversity in CN profiles between carcinoma and sarcoma, with a mean of 10.6 genes (range 0-36) having different CN states between the two. The most commonly different genes were FGF3, which showed relative gain in the carcinoma in 4/16 cases, and MDM2, which showed relative gain in the sarcoma in 3/16 cases. Metastases showed substantial genomic divergence from their corresponding primary, indicating early seeding of metastatic sites. Conclusions There are few mutational differences between carcinoma and sarcoma components in OCS, with shared point mutations indicating a clonal origin. By contrast, there is diversity in CN profiles. Metastases show substantial genomic divergence from their corresponding primary, indicating early metastatic seeding. Overall, OCS cases with a high grade serous carcinoma component have genomic profiles that are very similar to classical HGSC, including ubiquitous TP53 mutation. These tumours should be included in clinical trials that recruit patients with HGSC.
No datasets are available for this submission.
No license information is available for this submission.