Abstract 5344: Phylogenetic and signature analyses predict treatment response in high-grade serous ovarian cancer

Jaana Oikkonen,Yilin Li,Amjad Alkodsi,Olli Carpén,Kaisa Huhtinen,Sakari Hietanen,Seija Grénman,Rainer Lehtonen,Sampsa Hautaniemi

DOI: https://doi.org/10.1158/1538-7445.am2018-5344

IF: 11.2

2018-01-01

Cancer Research

Abstract:Abstract Background. High-grade serous ovarian cancer (HGSOC) is a heterogeneous cancer where majority of the patients will eventually relapse. Phylogenetic analysis of longitudinal samples can reveal mechanisms leading to chemoresistance in different subclones. Companied with signature analysis timing of major mutations can be estimated. These methods can reveal the origin of the resistance mechanisms and bring knowledge about the heterogeneity in the surviving subclones. Methods. We employed whole genome sequencing in 48 tumor tissue samples from 9 patients with stage IIB -IVB HGSOC treated at the Turku University Hospital, Finland. DNA from 3-7 tumor tissues at diagnosis (debulking surgery or laparoscopy) and interval debulking surgery (NACT treated patients) and/or relapse from each patient was sequenced with coverage of 30-100x. The mutations were detected with MuTect2 and HaplotypeCaller and copy number alterations with Ascat, excluding germline variation using a blood control. The subclonal structures were identified with PyClone and ClonEvol using maximum of 3000 mutations in each patient. Phylogenetic trees explaining largest proportion of the variations were used. Signatures were fitted with the 30 consensus signatures from COSMIC for each tumor sample as well as for major subclones. Results. Each patient was characterized with 7-15 variant clusters (subclones). Largest divergence was detected in ovaries (11-49% of the mutations were ovary spesific) and lymph nodes compared to other tissues. Even left and right ovary can be very different from each other and ovaries exhibit earlier spreading than other sampled sites. Other tissues were spread from clones detected in the ovaries. Constant mutation rate or higher mutation rate during the treatment could not be explained by the detected number of variants between time points. Early clonal BRCA (COSMIC signature 3) with high contribution predicted better response to chemotherapy. Signature 3 has been associated with homologous recombination deficiency and sensitivity to platinum-based chemotherapy. High APOBEC (signature 2 and 13) associates with higher level of diversification of related branches. Patient level heterogeneity does not seem to predict outcome contrary to within sample heterogeneity which was higher in the patients with shorter survival times. Relapses contained late or early branching clones. In one poor outcome case a clone which was enriched in relapses dominated in the patient derived cell line indicating superior viability and cause of chemoresistance. Conclusions. Cancer evolution trees improved mutation signature analyses and relieved clinically relevant differences between HGSOC patients. Early BRCA signature as well as similar signatures between subclones predicted good clinical outcomes. Citation Format: Jaana Oikkonen, Yilin Li, Amjad Alkodsi, Olli Carpen, Kaisa Huhtinen, Sakari Hietanen, Seija E. Grenman, Rainer Lehtonen, Sampsa Hautaniemi. Phylogenetic and signature analyses predict treatment response in high-grade serous ovarian cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5344.