Abstract

Background: Most patients with BRAF V600-mutant melanoma progress on BRAF and MEK inhibition. Rechallenge after an interval of other therapy produces responses in a substantial minority, and the mechanism usually invoked is that resistance is partly epigenetic and reverses during time off drug. That mechanism makes a testable prediction. Objectives: To quantify rechallenge efficacy against first-course efficacy in the same patients; to test whether the duration of the drug-free interval predicts rechallenge response as the resensitisation model requires; and to identify what does predict response. Methods: Cohorts reporting outcomes of BRAF-directed rechallenge in advanced melanoma were eligible. Reported response and disease control rates were compared between first course and rechallenge within cohorts, and reported associations between interval duration and outcome were extracted. No new pooling was performed. Analyses were performed in Python 3. Results: Pooled across cohorts totalling 400 patients, rechallenge produced an objective response rate of 34.25% (95% CI 28.50-40.00), disease control in 65.01% (57.31-72.72), median progression-free survival of 5.0 months (4.0-5.9) and median overall survival of 9.8 months (9.3-20.4). Within cohorts, rechallenge retained 38% of the first-course objective response rate (27% against 72%) and 63% to 68% of the disease control rate. Three independent cohorts examined the drugfree interval and none found an association with response, disease control, progression-free survival or overall survival. The single predictive factor identified was depth of response to the first course: all patients achieving complete response initially achieved disease control on rechallenge, against 60% of those achieving partial response (p = 0.002). Among patients who did respond, median duration of response was longer on rechallenge than on first exposure (14.0 against 10.7 months), and toxicity was reported lower. Conclusions: Rechallenge is a worthwhile option in a setting with few alternatives. The mechanism usually invoked to justify it is not supported by the data: the drug-free interval, which the resensitisation model predicts should matter, consistently does not, while prior response depth does. Rarer but longer responses on rechallenge point the same way. These observations are better explained by selection of intrinsically sensitive tumours than by resistance reversing during time off drug, which has implications for whom to rechallenge and for whether scheduling a drug holiday is worthwhile.

Keywords: Melanoma, BRAF inhibitor, MEK inhibitor, Rechallenge, Acquired resistance, Drug holiday