Genomic and transcriptomic characterization of ovarian cancer under neoadjuvant and adjuvant chemotherapy.

e17571 Background: Neoadjuvant chemotherapy (NACT) has become a routine selection for late-staged ovarian cancer patients to reduce the tumor load before debulking surgery; and chemotherapy remains as the mainstream therapy for ovarian cancer. However, the effects of NACT on the cancer genome and the molecular determinates of tumor response to chemotherapy in ovarian cancer have not been characterized systematically. Methods: We obtained 121 clinical samples from 44 ovarian cancer patients (19 underwent NACT and 25 did not) and collected the key clinical features for these samples such as age, tumor stage, platinum-free interval, and platinum response. We then performed whole-exome sequencing for all the 121 samples and RNA-sequencing for 54 samples. We conducted a systematic comparative analysis between samples collected from the patients with and without NACT to characterize the effects of NACT on the cancer genome and transcriptomes. Focusing on the genomic and transcriptomic differences of treatment-naive samples between platinum sensitive and resistant patients, we also did a comparative analysis to identify molecular features that can predict the response to platinum. Results: We observed significantly elevated mutational signatures associated with platinum chemotherapy in the samples from patients who have undergone NACT. We also found increased microsatellite instability (MSI) in the post-NACT primary samples compared to treatment-naive ones. Interestingly, the prevalence of CCNE1 amplification significantly decreased in the post-NACT metastatic samples of patients. Consistently, the mRNA expression of CCNE1 appeared to down-regulated in such samples. Furthermore, the expression signature of cancer hallmarks related to this gene decreased in post-NACT metastatic samples, such as E2F targets, G2M checkpoint, and mitotic spindle. Remarkably, we found that homologous recombination deficiency (HRD) is a strong indicator for platinum sensitivity, which was also supported by HRD-associated mutational signatures. The expression signature of immune-related hallmarks was significantly upregulated in sensitive primary samples, including interferon alpha response, interferon gamma response, and IL6 JAK STAT3 signaling. Conclusions: Based on comprehensive profiling for genomic and transcriptomic characterization of a large cohort of ovarian cancer patients, we found that (i) NACT likely increased MSI and decreased prevalence of CCNE1 amplification; and (ii) a high HRD score and elevated immune-related expression signatures can strongly predict the sensitivity to platinum-based chemotherapy. These results lays a key foundation for develop novel, effective therapies for ovarian cancer.