Impact of non-obese metabolic dysfunction-associated fatty liver disease on risk factors for the recurrence of esophageal squamous cell carcinoma treated with endoscopic submucosal dissection: A multicenter study.

Abstract

Metabolic dysfunction is a risk factor for esophageal squamous cell carcinoma (ESCC). We investigated the impact of the recently proposed metabolic dysfunction-associated fatty liver disease (MAFLD) and its subtypes on ESCC recurrence after endoscopic treatment. This multicenter observational cohort study enrolled consecutive patients newly diagnosed with ESCC after endoscopic treatment. Patients were classified into MAFLD or non-MAFLD groups. The MAFLD group was further classified into non-obese and obese MAFLD groups with a body mass index cutoff value of 25kg/m2 . The impact of MAFLD on the recurrence of ESCC was evaluated using a decision tree algorithm and random forest analysis. A total of 147 patients (average age 69years; male:female, 127:20; observational period, 2.4years) were enrolled. The 1-, 3-, and 5-year recurrence rates were 2.0%, 21.1%, and 33.7%, respectively. Independent risk factors for the recurrence of ESCC were MAFLD (HR 2.2812; 95% confidence interval 1.0497-4.9571; p=0.0373), drinking status, and smoking status. Metabolic dysfunction-associated fatty liver disease was identified as the second most important classifier for recurrence, followed by drinking status. The cumulative incidence of ESCC recurrence was higher in the MAFLD group than in the non-MAFLD group. In a subanalysis, the cumulative incidence of recurrence was significantly higher in the non-obese than in the obese MAFLD group among abstainers/non-drinkers. Directed acyclic graphs revealed that MAFLD directly contributes to ESCC recurrence. MAFLD was independently and directly associated with ESCC recurrence after endoscopic treatment; a high recurrence rate was observed in patients with non-obese MAFLD. Metabolic dysfunction-associated fatty liver disease may identify patients at high risk for ESCC recurrence.