A simplified prediction model for early intrahepatic recurrence after hepatectomy for patients with unilobar hepatocellular carcinoma without macroscopic vascular invasion: An implication for adjuvant therapy and postoperative surveillance

An accurate prediction model of early recurrence of hepatocellular carcinoma (HCC) after hepatectomy is important to ascertain the postoperative adjuvant treatment and surveillance. This is a retrospective cohort study including 1125 patients with HCC underwent curative hepatic resection. They were randomly divided into training (n = 562) and validation (n = 563) sets. Early intrahepatic recurrence within 18 months from surgery is the primary outcome. In the training set, a prediction scoring model (Recurrent Liver Cancer Score RLCS) was developed, which was legitimised in the validation set. RLCS was developed based on four clinicopathologic risk factors (serum alpha fetoprotein, tumor size, multiple tumors or satellite nodules, and microvascular invasion). Low-risk and high-risk groups had statistically significant differences in early recurrence rates (18% vs. 43.8%). The 5-year recurrence-free survival rates of low risk and high risk groups were 52.9% and 27.8%, respectively. This model showed good calibration and discriminatory ability in the validation set (c-index of 0.647). RLCS is a user-friendly prediction scoring model which can accurately predict the occurrence of early intrahepatic recurrence of HCC. It establishes the basis of postoperative adjuvant treatment and surveillance in future studies. •User-friendly prediction model (RLCS) of early recurrence of HCC.•4 risk factors (serum AFP, tumor size, multiplicity, and microvascular invasion).•Low-risk and high-risk groups had different recurrence and survival rates.•Good calibration and discriminatory ability of RLCS.