Imaging and blood-based biomarkers are moving toward earlier decision points. A meta-analysis reported that radiomics models using CT-image features plus clinical variables can predict high-risk pathological features in gastric cancer, but it flagged the need for multicenter external validation. In another blood-based direction, a large meta-analysis found circulating tumor DNA after neoadjuvant therapy strongly predicts worse survival across most solid tumors—supporting ctDNA as a risk stratification input before surgery. These data points reinforce ongoing emphasis on measurable surrogate endpoints that can refine escalation and de-escalation strategies—particularly when tissue-based biomarkers are difficult to repeat frequently.