FIGURE 1.

(A) Nomogram was constructed by independent recurrence factors, including PNI, EMVI, vascular tumor thrombus, perineural invasion, and tumor size. (B) The calibration curve suggested good consistency between predicted and actual recurrence probabilities. (C) DCA indicated that the risk model provides greater clinical benefit than each individual recurrent factor. (D) The model's AUC value of 0.763 indicated high specificity and sensitivity than each individual recurrent factor. DCA, decision curve analysis; EMVI, extramural vascular invasion; PNI, prognostic nutritional index.