Optimisation of neoadjuvant pembrolizumab therapy for locally advanced MSI-H/dMMR colorectal cancer using data-driven delay integro-differential equations

The paper defines and solves the constrained optimization with a TCR threshold at t = 126 days (Eq. 4.6), yielding Treatments 3, 4, and 5 as the multi-dose optima for thresholds 86.25%, 86%, and 85%, respectively, and Treatment 6 as the optimal one-shot at 84% (Table 12). The candidate solution mistakenly enforces the threshold on efficacy (not TCR), so it selects Treatment 5 for all thresholds in (a), which conflicts with the paper’s results. Parts (b) and (c) of the candidate solution match the paper’s conclusions about the one-shot regimen’s efficiency and comparability to FDA regimens. The paper’s occasional wording mix-up around “threshold efficacies” likely caused the confusion, but the analysis and table clearly use a TCR threshold.