Preoperative albumin-bilirubin grade combined with aspartate aminotransferase-to-platelet count ratio index predict outcomes of patients with hepatocellular carcinoma within Milan criteria after liver resection

Abstract

There is little information concerning the prognostic significance of combined albumin-bilirubin (ALBI) grade and aspartate aminotransferase-to-platelet count ratio index (APRI) in hepatocellular carcinoma (HCC). Therefore, we performed this study to assess the prognostic utility of combining ALBI and APRI (ALBI-APRI score) for predicting the prognosis of patients with HCC within Milan criteria after liver resection. Two hundred thirty-nine patients were involved in this study. Patients with a high APRI score were allocated a score of 1, whereas patients with a low APRI score were allocated a score of 0. The ALBI-APRI score is the summation of APRI score and ALBI grade. The area under the receiver operating characteristic curve (AUC) was used to estimate the predictive accuracy of different models. During the study period, 132 patients experienced recurrence, and 52 patients died. Multivariate analysis revealed the ALBI-APRI score (HR = 1.753, 95% CI = 1.293-2.377, p < 0.001), presence of microvascular invasion (MVI, HR = 2.693, 95%CI = 1.832-3.960, p < 0.001) and multiple tumors (HR = 1.973, 95% CI = 1.300-2.995, p = 0.001) were all associated with recurrence. In addition, blood transfusion (HR = 3.113, 95% CI = 1.677-5.778, p < 0.001), high preoperative alpha-fetoprotein (AFP, HR = 2.272, 95% CI = 1.298-3.976, p = 0.004), ALBI-APRI score (HR = 2.046, 95% CI = 1.237-3.382, p = 0.005) and presence of MVI (HR = 4.524, 95% CI = 2.514-8.140, p < 0.001) were correlated with postoperative mortality. The AUCs of ALBI-APRI score were significantly higher than either ALBI or APRI alone for predicting both postoperative recurrence and mortality. ALBI-APRI score may be a predictor for the prognosis of patients with HCC within Milan criteria following liver resection. A more well-designed and large-scale study are warranted to prove our findings.