Study limitation reported by the research below: “The study design is a mix of a prospective RCT (RTOG-0617) and two retrospective real-world cohorts — the level of evidence reflects the strongest component, but the retrospective cohorts introduce selection bias risk.”