Treatment decisions for complex lung cancer often involve several defensible pathways, whose eligibility, sequencing and safety depend on unresolved information. MedGPT Clinical Explorer (MCE) organizes alternatives, decision‑changing unknowns, safety constraints and fallback options into a conditional strategy for clinician review. The study selected a purposive corpus of 100 cases, established reference standards for 40 of them, and recruited 250 physicians from 98 institutions to produce 2,250 strategies under unaided, retrieval‑reference and MCE‑assisted conditions. MCE‑assisted strategies achieved higher Admissible Pathway Attainment Scores (APAS, 0‑100) than unaided (adjusted difference 12.87, 95% CI 11.18‑14.55) and retrieval‑reference (difference 5.22, 3.52‑6.93). With the same knowledge base, MCE added content on candidate pathways, decision‑critical information and safety constraints. Whole‑strategy acceptability correlated positively with APAS (Spearman rho = 0.671), and a complementary audit verified coherent links among candidates, conditions and subsequent actions. The findings identify two complementary dimensions of open‑ended decision support: coverage of clinically relevant content and coherent connections among pathways, conditions and actions. MCE provides a shared decision object that makes consequential omissions and pathway contingencies visible before execution; prospective studies should assess its impact on workflow and patient outcomes.
Review