Fig. 4

Nomogram predicting lesion depths of SESCC. This model (A) was constructed based on six diagnosis relevant WLI-based features, including Paris classification, lesion length, circumferential extension, nodularity, granularity and spontaneous bleeding. As shown in B, good agreement between prediction and observation was demonstrated. SESCC superficial esophageal squamous cell carcinoma, MM/SM1 tumor invading the muscularis mucosa or superficial submucosa, SM2 tumor invading the submucosa or deep submucosal invasion