Novel inflammation-based model for postoperative early recurrence prediction of hepatitis B virus-related hepatocellular carcinoma ≤ 5 cm

BACKGROUND Despite surgical treatment, hepatocellular carcinoma (HCC) remains highly recurrent, particularly in hepatitis B virus (HBV)-related cases. The mechanisms of early (≤ 2 years) vs late recurrence differ substantially; however, the inflammation-related divers of early recurrence in HBV-related HCC are poorly defined. AIM To construct an inflammation-related score (IRS) model to predict recurrence in HBV-related HCC with tumors ≤ 5 cm. METHODS A total of 650 eligible patients with HCC were retrospectively enrolled and randomly divided into the training (n = 390) and validation (n = 260) groups. Univariate, least absolute shrinkage and selection operator Cox analyses were used to identify early recurrence-related inflammation indices and construct an IRS model in the training cohort. An IRS-based nomogram was established and assessed for both cohorts. RESULTS Early recurrence accounted for 74.7% (222/297) of all recurrences. Four critical inflammation indices were screened for the IRS model construction. Patients with a high IRS had significantly lower early recurrence-free survival (RFS) than those with a low IRS in both the training and validation cohorts (all P < 0.01). In combination with tumor diameter, number, and microvascular invasion, a novel IRS-based nomogram for predicting early RFS was established. The concordance index was 0.658 in both training and validation cohorts. A high concordance was observed in the calibration curves. The nomogram revealed a better capacity for risk discrimination and clinical usefulness than other staging systems. Finally, patients were classified into different risk groups based on the total points of the nomogram. CONCLUSION The integrated IRS model and corresponding nomogram provide good clinical implications for identifying high-risk populations with early recurrence.