Use of Gynecologic Cancer Intergroup (GCIG) CA125 criteria to evaluate cell cycle checkpoint kinase 1 inhibitor (CHK1i) prexasertib response and disease progression (PD) in recurrent high-grade serous ovarian cancer (HGSOC).

Abstract

e17536 Background: Data suggest discordance between serum biomarker CA125 and Response Evaluation Criteria in Solid Tumors (RECIST) response and/or PD in chemotherapies and PARP inhibitors. Here, we investigated the impact of CA125 on prognosis and its relationship with RECIST response/PD in HGSOC patients treated with a CHK1i, prexasertib. Methods: All relapsed HGSOC patients in this post-hoc analysis were treated with prexasertib at 105 mg/m2 IV every 14 days in a 28-day cycle until RECIST v1.1 PD, unacceptable toxicity, or consent withdrawal (NCT02203513). CA125 was assessed every cycle and CT scans obtained every two cycles. GCIG CA125 response was defined as >50% reduction confirmed after 28 days and PD as > 2X upper limit of normal (ULN) if baseline was normal or doubling of nadir if baseline was elevated. Patients with baseline CA125 < 2X ULN or nadir > 750 units/mL were not eligible for analysis. Clinical benefit rate (CBR; CR + PR + SD ≥ 6 months) was analyzed by Fisher’s exact test. Progression-free survival (PFS) and overall survival (OS) were estimated by Kaplan-Meier curves and univariate (UV) and multivariate (MV) hazard ratios by Cox regression models. Positive predictive value (PPV) and negative predictive value (NPV) confidence intervals (CI) were estimated by the Wilson-Brown method. Results: Eighty-one patients received at least two cycles of prexasertib between January 2015 and November 2020, out of which 72 were evaluable by both RECIST and GCIG CA125 response criteria. CBR in the CA125 response (CA125-R) group (24/32; 75%) was significantly higher than the CA125 non-responder (CA125-N) group (12/40; 30%; p<0.001). CA125 response tended to occur prior to RECIST response (median: 28 days; IQR: 28, 30). PFS was improved in CA125-R at 8.0 months vs. CA125-N at 3.5 months (UV HR: 0.34, 95% CI: 0.21-0.56; MV HR: 0.35, 95% CI: 0.20-0.59). OS was improved in CA125-R at 19.8 months vs. CA125-N at 10.3 months (HR: 0.43, 95% CI: 0.26-0.71; MV HR: 0.45, 95% CI: 0.27-0.75). No significant differences in PFS or OS were observed between subgroups by BRCA mutation status, platinum-sensitivity, FIGO stage, baseline CA125 value, or previous PARP inhibitor therapy in MV analysis. Out of 70 patients evaluable by RECIST and CA125 PD criteria, 43 patients had CA125 PD and 24 of those had RECIST PD occur within one cycle of CA125 PD (PPV: 59%; 95% CI: 43-72). Of the 27 patients without a CA125 PD, 9 patients also did not have a RECIST PD (NPV: 33%; 95% CI: 19-52). Conclusions: GCIG CA125 response is an independent predictor of response to CHK1i in HGSOC prior to the first restaging scans and is associated with a higher CBR and improved PFS and OS. However, GCIG CA125 PD has poor PPV and NPV within a one cycle window, indicating that CA125 alone is not sufficient in monitoring for PD in HGSOC treated with CHK1i.