<b>Objectives:</b> Poly (ADP-ribose) polymerase inhibitors (PARPis) are often incorporated into epithelial ovarian cancer (EOC) treatment. PARPis are most effective for patients with a <i>BRCA</i> mutation (germline or somatic), followed by homologous recombination deficiency (HRD). We aimed to determine if genetic mutational status impacts toxicity to PARPis or next-line (NL) cytotoxic chemotherapy, as this could inform early interventions to support and maintain patients on treatment. <b>Methods:</b> A retrospective review was conducted of patients with EOC receiving PARPi from March 1, 2014, to May 31, 2021. Medical records were reviewed for clinical-pathologic, genetic, treatment and toxicity data. Toxicities were graded using Common Terminology Criteria for Adverse Events Version 5.0 and coded by organ system. Statistics included descriptive, Chi-square, and Fisher's exact. <b>Results:</b> A total of 230 patients were identified. The mean age at diagnosis was 60.9 years (27-93). Most patients had high-grade serous histology (89.1%) and advanced stage (III 63.0%; IV 23.0%). PARPi indications included frontline maintenance (FM) (18.7%), platinum-sensitive recurrent maintenance (PSRM) (43.9%) and active therapy (AT) (37.4%). Genetic variants included 27.8% germline <i>BRCA</i> (g<i>BRCA</i>+), 5.2% somatic <i>BRCA</i> (s<i>BRCA</i>+), 4.8% HRD, 55.7% wild type and 6.5% not tested. PARPi toxicity was common (overall 94.4%); gastrointestinal (GI) (67.8%), constitutional (64.8%) and hematologic (59.1%) were most prevalent. The myelodysplastic syndrome was rare (0.87%). <i>gBRCA+</i> was not associated with increased PARPi overall (p=0.13), Gr 3/4 (p=0.40), or any specific category of toxicity (Table 1). Stratifying by PARPi indication, <i>gBRCA+</i> was only predictive of higher neurologic toxicity with PARPi used as PSRM (57.1% vs 26.3%, p=0.01). Total 117 patients (50.9%) received NL chemotherapy after PARPi. Toxicity was common (overall 95.7%), with 71.8% hematologic, 63.2% GI and 57.3% constitutional. <i>gBRCA+</i> did not predict worse NL chemotherapy overall (p=0.25), Gr 3/4 (p=0.43), or any category of toxicity (Table 1). Stratifying by the first PARPi (PARPi-1) indication, <i>gBRCA+</i> did not predict worse NL chemotherapy toxicity (overall, Gr 3/4, or any specific category). Broadening the above analyses to incorporate patients with <i>sBRCA</i> mutations +/- HRD did not significantly impact the aforementioned PARPi-1 or NL chemotherapy toxicity (Table 1). Nineteen patients (8.3%) received a second PARPi (PARPi-2); six (2.6% cohort / 31.6% PARP-2) changed agents due to PARPi-1 toxicity and 13 (5.7% cohort / 68.4% PARP-2) used as a new treatment line. PARPi-2 toxicity occurred often (overall 89.5%; Gr 3/4 15.8%); GI (68.4%), constitutional (42.1%) and hematologic (42.1%) were most common. <i>gBRCA+</i> did not predict PARPi-2 overall (87.5% vs 90.9%, p=0.68), Gr 3/4 (25.0% vs 9.1%, p=0.38) or any specific category of toxicity; this held true when analyzed as <i>gsBRCA+</i> (overall 88.9% vs 90.0%, p=0.74; Gr 3/4 22.2% vs 10.0%, p=0.46) and gs<i>BRCA</i>+HRD+ (overall 81.8% vs 100%, p=0.32; Gr 3/4 18.2% vs 12.5%, p=0.62). <b>Conclusions:</b> In this cohort, genetic variant status did not predict worse toxicity to PARPi-1, PARPi-2, or NL chemotherapy. Overall, this data supports the continued use of PARPis across indications and genetic mutational status with low concern for treatment-limiting toxicities.
Read full abstract