3159 Background: PARPis induce synthetic lethality in tumors that have homologous repair deficiencies (HRD). Due to relative rarity of HRD-related genomic and germline alterations across solid malignancies, it remains unclear which patients (pts) may derive most benefit from PARPis across different tumor types. Methods: We analyzed clinicopathologic and genomic characteristics of pts with advanced ovarian (OC, n=33), prostate (PC, n=28) and breast cancer (BC, n=13) treated with PARPi monotherapy at our institution. Only pts with HRD-related pathogenic alterations detected via next-generation sequencing of tumor tissue, with or without genetic susceptibility testing, were included. Results: Forty-one (55%) pts had a confirmed germline HRD-related gene alteration, of which 33 had BRCA1/2 and 8 had non-BRCA1/2 gene alterations. Average age of pts with BC was 54 years, OC 62 years, and PC 68 years. Median number of prior lines of therapy in the advanced setting was 1 (1-4) for OC and BC (0-7) and 3 (1-6) for PC. Median time to progression (TTP) in pts with OC was 8.3 months, BC 4.9 months and PC 3.6 months, likely due to i) majority of pts with OC received maintenance PARPi following favorable response to first-line chemotherapy, ii) pts with PC were older, and iii) more pts with PC had somatic rather than germline HRD-related mutations. Pts with germline vs somatic HRD mutation had longer median TTP in PC (5.0 vs 3.3 months) and BC (6.1 vs 4.0 months), but the opposite was true for OC (7.9 vs 10.8 months). Pathogenic TP53 mutations were identified in all but one pt with OC (32/33), 10/13 pts with BC, and 6/28 PCs. [table] Average TP53gene variant allele fraction (VAF) was similar in OC and BC (49.8% and 51.1%, respectively) but significantly lower in PC (26.4%), suggesting that TP53 mutations in PC were subclonal. No specific TP53 VAF or HRD-to- TP53 VAF ratio cut-off was associated with TTP or overall survival in any of the 3 cancer types. In addition, 7/28 (25%) of pts with PC had PTEN loss associated with modified response to PARPis, whereas none of the pts with BC had PTEN loss and only 3/13 had alterations of the PTEN/PI3K/AKTpathway. Conclusions: Double-strand DNA breaks caused by synergistic effects of aberrant TP53 and HRD-related genes can possibly explain the more favorable response of pts with OC and BC to PARPi, compared to those with PC. Co-alterations in other signaling pathways, epigenetic factors and/or tumor microenvironment likely play a role in the degree of response to PARPis, rather than TP53 clonality alone. In the era of expansion of genomic testing and new indications for PARPi larger genomic studies are needed to identify biomarkers of response to PARPi beyond HRD. [Table: see text]
Read full abstract