Prognostic value of preoperative systemic immune-inflammation index/albumin for patients with hepatocellular carcinoma undergoing curative resection

Abstract

Background: Hepatocellular carcinoma (HCC) is a major factor for cancer-associated mortality globally. Although the systemic immune-inflammation index (SII) and albumin (ALB) show individual prognostic value for various cancers, their combined significance (SII/ALB) in HCC patients undergoing curative hepatectomy is still unknown. It is hypothesized that a higher SII/ALB ratio correlates with poorer outcomes with regard to overall survival (OS) and recurrence-free survival (RFS).

Aim: To investigate the effect of preoperative SII/ALB in predicting the prognosis of HCC patients undergoing hepatectomy.

Methods: Patients who received curative surgery for HCC at a single institution between 2014 and 2019 were retrospectively analyzed. Cox proportional hazards models and Kaplan-Meier curves were utilized to estimate OS and RFS. A nomogram was created using prognostic factors determined by the least absolute shrinkage and selection operator method and analyzed using multivariate Cox regression. This nomogram was assessed internally through the calibration plots, receiver operating characteristic (ROC) analysis, decision curve analysis (DCA) and the concordance index (C-index).

Results: This study enrolled 1653 HCC patients. Multivariate analyses demonstrated that SII/ALB independently predicted OS [hazard ratio (HR) = 1.22, 95%CI: 1.03-1.46, P = 0.025] and RFS (HR = 1.19, 95%CI: 1.03-1.38, P = 0.022). Age, alpha-fetoprotein, hepatitis B surface antigen, albumin-bilirubin grade, tumor diameter, portal vein tumor thrombus, tumor number, and SII/ALB were incorporated into the nomogram to predict OS. The nomogram had a C-index of 0.73 (95%CI: 0.71-0.76) and 0.71 (95%CI: 0.67-0.74) for the training and validation cohorts, respectively. The area under the ROC curve, DCA and calibration curves demonstrated high accuracy and clinical benefits.

Conclusion: The SII/ALB may independently predict outcomes in HCC patients who receive curative surgical treatment. In addition, the nomogram can be used in HCC treatment decision-making.

Keywords: Hepatocellular carcinoma; Inflammation; Liver resection; Prognosis; Systemic immune-inflammation index/albumin.

Figure 1

Time-dependent receiver operating characteristic curves for systemic immune-inflammation index/albumin, systemic immune-inflammation index, neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio predicting. A: Time dependent area under the curve (AUC) for overall survival; B: Time dependent AUC for recurrence-free survival. SII/ALB: Systemic immune-inflammation index/albumin; ALB: Albumin; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; AUC: Area under the curve.

Figure 2

Relationship between systemic immune-inflammation index/albumin and both overall survival and recurrence-free survival. A: Examination of the dose-response relationship between systemic immune-inflammation index/albumin (SII/ALB) and 5-year overall survival (OS) by the restricted cubic splines model; B: Cut-off value of SII/ALB in patients with hepatocellular carcinoma; C: Kaplan-Meier curves depicting the association between SII/ALB and OS; D: Kaplan-Meier curves depicting the association between SII/ALB and recurrence-free survival. SII/ALB: Systemic immune-inflammation index/albumin.

Figure 3

Subgroup analysis of systemic immune-inflammation index/albumin in predicting overall survival. HR: Hazard ratio; HbsAg: Hepatitis B surface antigen; AFP: Alpha-fetoprotein; ALBI: Albumin-bilirubin; BCLC: Barcelona Clinic Liver Cancer; MVI: Microvascular invasion; PVTT: Portal vein tumor thrombus.

Figure 4

Subgroup analysis of systemic immune-inflammation index/albumin in predicting recurrence-free survival. HR: Hazard ratio; HbsAg: Hepatitis B surface antigen; AFP: Alpha-fetoprotein; ALBI: Albumin-bilirubin; BCLC: Barcelona Clinic Liver Cancer; MVI: Microvascular invasion; PVTT: Portal vein tumor thrombus.

Figure 5

Least absolute shrinkage and selection operator regression analysis for variable selection. A: Cross-validation graph; B: Least absolute shrinkage and selection operator regression analysis coefficients; C: Construction of a nomogram incorporating systemic immune-inflammation index/albumin and clinical parameters. ALBI: Albumin-bilirubin; λ: Lambda; AFP: Alpha-fetoprotein; AST: Aspartate transaminase; HbsAg: Hepatitis B surface antigen; SII/ALB: Systemic immune-inflammation index/albumin; PVTT: Portal vein tumor thrombus; OS: Overall survival.

Figure 6

Nomogram calibration curves for different time intervals and cohorts. A: The 1-year, training cohort; B: The 3-year, training cohort; C: The 5-year, training cohort; D: The 1-year, validation cohort; E: The 3-year, validation cohort; F: The 5-year, validation cohort.

Figure 7

Time-dependent receiver operating characteristic curves and areas under the curve in different models and time intervals. A: 1-year, 3-year, and 5-year in the training set; B: 1-year, 3-year, and 5-year in the validation set; C: The 1-year, different models in the training set; D: The 1-year, different models in the validation set; E: The 3-year, different models in the training set; F: The 3-year, different models in the validation set; G: The 5-year, different models in the training set; H: The 5-year, different models in the validation set. AUC: Area under the curve; ROC: Receiver operating characteristic; BCLC: Barcelona Clinic Liver Cancer; CNLC: China Liver Cancer Staging.

Figure 8

Decision curve analysis for comparing various models. A: Decision curve analysis (DCA) at 1-year, 3-year, and 5-year intervals in the training set; B: DCA at 1-year, 3-year, and 5-year intervals in the validation set; C: DCA at the 1-year interval comparing models in the training set; D: DCA at the 1-year interval comparing models in the validation set; E: DCA at the 3-year interval comparing models in the training set; F: DCA at the 3-year interval comparing models in the validation set; G: DCA at the 5-year interval comparing models in the training set; H: DCA at the 5-year interval comparing models in the validation set. DCA: Decision curve analysis; BCLC: Barcelona Clinic Liver Cancer; CNLC: China Liver Cancer Staging.