A new SVM-based machine-learning model accurately predicts postoperative recurrence in ESCC.
Using factors like age, ECOG, NLR, CPR, CY211, TNM stage & complications, the SVM6+8 model reached 94% sensitivity.
A nomogram predicts 1-, 3-, 5-yr DFS, helping identify high- vs low-risk patients.
Enables personalized follow-up & adjuvant therapy.