Prognostic Value of Serum α‐Fetoprotein Level as an Important Characteristic of Tumor Biology for Patients Undergoing Liver Resection of Early-Stage Hepatocellular Carcinoma (BCLC Stage 0/A): A Large Multicenter Analysis

Background and Objective: According to the Barcelona Clinic Liver Cancer (BCLC) algorithm, tumor burden and liver function, but not tumor biology, are the key factors in determining tumor staging and treatment modality, and evaluating treatment prognosis. The serum α‐fetoprotein (AFP) level is an important characteristic of hepatocellular carcinoma (HCC) biology, and we aimed to evaluate its prognostic value for patients undergoing liver resection of early-stage HCC. Methods: Patients who underwent curative liver resection for early-stage HCC were identified from a multi‐institutional database. Patients were divided into three groups according to preoperative AFP levels: low (< 400 ng/mL), high (400–999 ng/mL), and extremely-high (≥ 1000 ng/mL) AFP groups. Overall survival (OS) and recurrence rates were compared among these three groups. Results: Among 1284 patients, 720 (56.1%), 262 (20.4%), and 302 (23.5%) patients had preoperative low, high, and extremely-high AFP levels, respectively. The cumulative 5-year OS and recurrence rates were 71.3 and 38.9% among patients in the low AFP group, 66.3 and 48.5% in the high AFP group, and 45.7 and 67.2% in the extremely-high AFP group, respectively (both p < 0.001). Multivariate Cox regression analysis identified both high and extremely-high AFP levels to be independent risk factors of OS (hazard ratio [HR] 1.275 and 1.978, 95% confidence interval [CI] 1.004–1.620 and 1.588–2.464, respectively; p = 0.047 and p < 0.001, respectively) and recurrence (HR 1.290 and 2.050, 95% CI 1.047–1.588 and 1.692–2.484, respectively; p = 0.017 and p < 0.001, respectively). Conclusions: This study demonstrated the important prognostic value of preoperative AFP levels among patients undergoing resection for early-stage HCC. Incorporating AFP to prognostic estimation of the BCLC algorithm can help guide individualized risk stratification and identify neoadjuvant/adjuvant treatment necessity.