Randomized controlled trials (RCT) represent the cornerstone of evidence-based medicine but are resource intensive. We propose and evaluate a novel machine learning (ML) strategy of adaptive predictive enrichment through computational trial phenomaps to optimize RCT enrollment. In simulated group sequential analyses of two large cardiovascular outcomes RCTs of (1) a therapeutic drug (pioglitazone versus placebo; Insulin Resistance Intervention after Stroke or IRIS), and (2) a disease management strategy (intensive versus standard systolic blood pressure reduction in SPRINT), we constructed dynamic phenotypic representations to infer profiles of patients benefiting from the intervention versus control during interim trial analyses and examined their association with study outcomes. Across three interim analyses, our strategy learned dynamic phenotypic signatures predictive of individualized cardiovascular benefit in each arm. By conditioning a prospective candidate’s probability of enrollment on their predicted benefit, we estimate that our approach would have enabled a reduction in the final trial size across five simulations (IRIS: -18 ± 4.7%,p=0.008; SPRINT: -27.4 ± 3.4%,p=0.002), while preserving the original average treatment effect (IRIS: hazard ratio of 0.71 ± 0.01 for pioglitazone vs placebo, vs 0.76 in the original trial; SPRINT: hazard ratio of 0.72 ± 0.01 for intensive vs standard systolic blood pressure, vs 0.75 in the original trial; all comparisons withp<0.01). This adaptive framework has the potential to maximize RCT enrollment efficiency.