Preoperative gamma-glutamyl transpeptidase for predicting postoperative recurrence of hepatocellular carcinoma with wide resection margins: a multi-institutional study.

Wide resection for hepatocellular carcinoma (HCC) typically enhances tumor clearance and improves prognosis. However, owing to its complex pathogenesis, postoperative recurrence remains common even with wide margins (> 1 cm). This study explored the prognostic value of GGT in the development of nomograms for predicting recurrence in HCC patients with wide resection margin. We retrospectively analyzed data from 750 HCC patients who underwent radical hepatic resection at three medical centers in China between January 2009 and December 2015. Patients were randomly split into a development cohort (n = 525) and a validation cohort (n = 225). Independent risk factors were identified through Cox multivariate regression, leading to the construction of nomogram. The models' accuracy was evaluated using time-dependent ROC and calibration curves, while decision curve analysis (DCA) assessed their clinical utility. Kaplan-Meier survival analysis and log-rank tests were applied to examine the correlation between risk stratification and recurrence-free survival (RFS). The nomogram of WRM model contains Gamma-glutamyl transpeptidase (GGT), tumor size, tumor number, and microvascular invasion (MVI). The model demonstrated good discrimination, with C-indexes of 0.75, 0.71, 0.69 at 1, 3, 5 years, and AUCs of 0.78, 0.76, 0.74 for RFS in the validation cohort. The DCA of the WRM model showed significantly better predictive performances than other models. The model could stratify patients into two different risk groups. The web-based tools are convenient for clinical practice. GGT was identified as an independent risk factor for RFS in HCC patients with wide resection margin. The GGT-based nomograms demonstrated solid predictive performance for recurrence.