Laparoscopic resection vs thermal ablation for 3-5cm hepatocellular carcinoma: A multicenter study based on machine learning

Abstract

AbstractBackground Very early recurrence (VER) significantly affects the survival of patients with treated primary hepatocellular carcinoma. Purpose To compare the survival of patients with VER after laparoscopic resection and thermal ablation treatment. To establish a model predicting very early recurrence after laparoscopic resection. Methods 956 primary HCC patients treated with thermal ablation or laparoscopic resection from 12 centers were included in our study. The overall survival (OS) and cancer specific survival (CSS) rate of all patients and patients with VER in two groups were compared. Propensity score matching (PSM) was done to keep baseline balance. Risk factors that may lead to VER occurrence were presented. Different machine learning methods were used to build prediction models of VER occurrence after laparoscopic resection. Results VER significantly affects the OS and CSS of patients in both groups(p < 0.0001). But OS and CSS of patients with VER in two groups has no statistical difference before (p = 0.39 and 0.27, respectively) or after PSM (p = 0.4 and 0.85, respectively). A model predicting the occurrence of VER after laparoscopic resection was established based on machine leaning with 5 factors (AFP, ALT, TBil, DBil and number of tumors) included, which had the highest AUC value of 0.722. Conclusion 3-5cm primary hepatocellular carcinoma patients with VER after laparoscopic resection can benefits more if choose thermal ablation instead. And patients with VER after laparoscopic resection can be reliably predicted by the model presented in this study.