Targeted Mutational Analysis of Circulating Tumor DNA to Decipher Temporal Heterogeneity of High-Grade Serous Ovarian Cancer

SIMPLE SUMMARY: The issue of spatial and temporal heterogeneity of high-grade serous ovarian cancer (HGS-EOC) has hampered the possibility to shape the molecular portrait of relapsed disease, which ultimately impacts our ability to develop a more rational second-line treatment. Liquid biopsy offers the unique opportunity to track tumor evolution over time and infer the dynamic changes of tumor clonal architecture. Differently from other tumors, no actionable driving lesions characterize HGS-EOC, thus genome-scale analysis like whole-exome sequencing is not compatible with the clinical turnaround time. In the present work, we provided a novel framework based on the analysis of both qualitative and quantitative features of circulating tumor DNA (ctDNA) in order to identify, at the time of molecular relapse, the early genetic vulnerabilities that will characterize the clinical recurrence and thus be amenable of a more rational second-line treatment. ABSTRACT: We have previously demonstrated that longitudinal untargeted analysis of plasma samples withdrawn from patients with high-grade serous ovarian cancer (HGS-EOC) can intercept the presence of molecular recurrence (TRm) earlier than the diagnosis of clinical recurrence (TRc). This finding opens a clinical important temporal window to acquire through plasma sample analysis a real-time picture of those emerging molecular lesions that will drive and sustain the growth of relapsed disease and ultimately will confer resistance. In this proof of principle study, the same genomic libraries obtained at the diagnosis (T0), TRm and TRc were further analyzed by targeted resequencing approach to sequence the coding region of a panel of 65 genes to provide longitudinal analysis of clonal evolution as a novel strategy to support clinical decisions for the second-line treatment. Experiments were performed on plasma and tumor tissues withdrawn on a selection of previously analyzed cohorts of cases (i.e., 33 matched primary and synchronous lesions and 43 plasma samples from 18 patients). At T0, the median concordance of mutations shared by each tumor tissue biopsy and its matched plasma sample was 2.27%. This finding confirms the limit of a single tumor biopsy to be representative of the entire disease, while plasma analysis can recapitulate most of the main molecular lesions of the disease. A comparable scenario was observed during longitudinal analysis, where, with the exception of the TP53 gene and germline mutations in BRCA1/2 genes, no other gene shared the same locus specific gene mutation across T0, TRm and TRc time points. This high level of temporal heterogeneity has important implications for planning second-line treatment. For example, in three out of 13 cases, plasma ctDNA analysis at TRm or TRc reported acquired novel variants in the TP53BP1 gene not present at T0. In particular, patient 21564, potentially eligible for PARP-inhibitor (PARPi) treatment at the time of diagnosis (BRCA1 c.5182delA mutation), would unlikely respond to these drugs in second-line therapy due to the presence of eight distinct TP53BP1 variants in plasma samples collected TRc. This study demonstrates that liquid biopsy provides a real-time molecular picture to intercept those actionable genetic vulnerabilities or drug resistance mechanisms that could be used to plan a more rational second-line treatment.