Postoperative α-fetoprotein response predicts tumor recurrence and survival after hepatectomy for hepatocellular carcinoma: A propensity score matching analysis.

Abstract

To investigate the prognostic values of the change of α-fetoprotein within 1 week after resection of hepatocellular carcinoma. We retrospectively analyzed patients with hepatocellular carcinoma who underwent curative hepatectomy as primary therapy at Zhongshan Hospital, Fudan University (Shanghai, China) from 2009 to 2011. We measured serum α-fetoprotein before (α-fetoprotein₀) and 1 week after (α-fetoprotein₇) hepatectomy, calculated change of α-fetoprotein, namely the α-fetoprotein response by the formula: AR = lgAFP₇/lgAFP₀ (lg = log₁₀), analyzed the relationship between patient survival and α-fetoprotein response, and explored the potential clinical implications of the α-fetoprotein response. The results were validated in an independent cohort of patients from the same institute. A total of 841 eligible patients were analyzed. We determined that the optimal cutoff value of the α-fetoprotein response was 0.8135 and subsequently classified patients from the exploration cohort into the α-fetoprotein responder (α-fetoprotein response ≤ 0.8135; n = 452) and α-fetoprotein nonresponder (α-fetoprotein response > 0.8135; n = 146). Multivariate Cox analysis showed that the α-fetoprotein response independently predicted overall survival (OS) and recurrence-free survival (RFS) time after resection (both P < .001). In patients with a higher risk of tumor recurrence (either single tumor with microvascular invasion or multiple tumors), α-fetoprotein responders were associated with better survival than the nonresponders (P < .05). The results were validated by propensity score matched population and another independent cohort. The α-fetoprotein response is a reliable and simple predictive marker for evaluating the oncological effect of surgical resection for hepatocellular carcinoma with positive α-fetoprotein before resection, independent of tumor features.