A genomic and transcriptomic approach to distinguish primary and metastatic ovary tumors
e22150 Background: Distinction of primary ovarian tumors from metastatic tumors involving the ovary is in some cases challenging for final pathologic diagnosis and for treatment with efficient chemotherapy Methods: We gathered from our biobank 16 pairs of breast/ovarian tumors for some of which the diagnosis was uncertain. We investigated the possibility to improve diagnosis using genomic and transcriptomic tools. The pangenomic profiles of 16 pairs of primary breast carcinoma and ovary tumors (primary tumors or metastases from breast carcinoma) from the same patients were analyzed using the Affymetrix GeneChip Mapping 50K (XbaI) SNP arrays. The data were normalized with ITALICS algorithm. We analyzed the transcriptome of the paired samples using Affymetrix U133plus2 arrays. The data were normalized with GCRMA algorithm and a hierarchical clustering of these samples was performed, together with a dataset of primary and secondary ovary tumors. Results: Primary infiltrating lobular carcinoma (ILC) was observed in 6 patients, infiltrating ductal carcinoma (IDC) in 7 patients, 1 patient had one ILC and one IDC, 1 patient had a mixed IDC+ILC and 1 patient an undifferentiated cancer. All patients received adequate loco-regional and systemic therapies. Median time to diagnosis of ovarian tumor was 54 months. Ovarian tumors were considered as primary in 7 patients, and metastatic in 9 patients, and diagnosis was ambiguous in 4 of them. Four patients developed extra-abdominal metastases. Median survival from breast cancer diagnosis was 78 months, and from ovarian tumor diagnosis was 29 months. CGH array: In 12 pairs, the comparison of genomic profiles one with each other, confirmed the pathological characteristics of the tumors. In 4 cases, the genomic profiles established clearly the status of the ovary tumor whereas the pathological analysis did not. Transcriptome: The status established by the transcriptomic analysis was for each sample in total agreement with the status established by the genomic analysis. Conclusions: We clearly established in this training series that CGH array analysis could help to discriminate between primary and secondary ovarian tumors from breast cancer. Transcriptomic data confirmed these results. Further validation data are warranted. No significant financial relationships to disclose.