Circulating tumor DNA (ctDNA) monitoring in participants (pts) with ovarian cancer treated with neoadjuvant pembrolizumab (pembro) + chemotherapy (chemo) ± anti–immunoglobulin-like transcript 4 (ILT4) monoclonal antibody MK-4830.

Abstract

5563 Background: ctDNA is a promising biomarker for predicting disease progression, survival, and surgical outcomes in patients with solid tumors. In a phase 1 study (NCT03564691), ILT4 inhibitor MK-4830 + pembro had a manageable safety profile and showed antitumor activity in pts with solid tumors. We present results from a global, randomized phase 2 study (NCT05446870) evaluating quantitative change in ctDNA in pts with high-grade serous ovarian cancer (HGSOC) who received neoadjuvant pembro + chemo ± MK-4830. Methods: Eligible pts were female, aged ≥18 y, with previously untreated histologically confirmed FIGO stage 3 or 4 HGSOC and an ECOG performance status of 0 or 1, and were candidates for interval debulking surgery. Pts were randomly assigned 1:1 to receive neoadjuvant MK-4830 800 mg + pembro 200 mg + chemo (paclitaxel 175 mg/m 2 and carboplatin AUC 5-6) (arm 1) or pembro + chemo (arm 2) IV Q3W for 3 cycles. Pts underwent interval debulking surgery followed by 3 cycles of adjuvant therapy with the neoadjuvant regimen; adjuvant bevacizumab IV Q3W was permitted. ctDNA was assessed at each cycle using the Signatera assay (Natera, Inc.). A constrained longitudinal data analysis model was used to estimate the posterior probability that the coefficient for treatment assignment was <0, evaluating whether the reduction in ctDNA from cycle 1 (C1) was larger in arm 1. The primary end point was change in ctDNA from C1 at C3 in pts with detectable ctDNA; safety was a secondary end point. Results: At data cutoff (Dec 20, 2023), 160 pts were enrolled; 159 pts received treatment (arm 1, n = 79; arm 2, n = 80). Median study follow-up was 8.3 months (range, 1.9-16.4) in arm 1 and 8.3 months (range, 2.1-16.3) in arm 2. Median age was 61.5 y in both arms. 64 pts (80.0%) in arm 1 and 68 pts (85.0%) in arm 2 had available ctDNA data at C1; 51 pts (63.8%) and 63 pts (78.8%), respectively, had available ctDNA data at C3. Median ratio of ctDNA C3 to C1 was 0.02 (range, 0-1.53) in arm 1 and 0.01 (range, 0-0.60) in arm 2. The posterior probability that the coefficient for treatment assignment was <0 in the model was 38.8%, indicating a low posterior certainty of larger ctDNA reduction in arm 1. AEs occurred in 75 pts (94.9%) in arm 1 and all pts (100%) in arm 2. TRAEs occurred in 74 pts (93.7%) in arm 1 and in 79 pts (98.8%) in arm 2; grade 3-5 events occurred in 37 pts (46.8%) and 44 pts (55.0%), respectively. TRAEs led to death in 2 pts in arm 1; no treatment-related deaths occurred in arm 2. Conclusions: In pts with HGSOC, reductions in ctDNA were similar between neoadjuvant/adjuvant MK-4830 + pembro + chemo vs pembro + chemo. Real-time tumor-informed ctDNA testing may be feasibly incorporated into future clinical trials as a surrogate outcome to evaluate response. The safety profile of MK-4830 + pembro + chemo was comparable to pembro + chemo. Clinical trial information: NCT05446870 .