Development and Assessment of Novel Predictive Nomograms Based on APRI for Hepatitis B Virus-associated Small Solitary Hepatocellular Carcinoma with Stereotactic Body Radiotherapy.

The correlation between serum inflammatory marker before treatment and the survival of patients with hepatitis B virus (HBV)-associated small solitary hepatocellular carcinoma (HCC) after stereotactic body radiotherapy (SBRT) remains unclear. The objective of our study is to estimate survival in such patients using multivariable prediction models and investigate the prognostic value of aspartate aminotransferase-to-platelet index (APRI), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and lymphocyte-to-monocyte ratio (LMR) for HBV-associated small solitary HCC patients treated with SBRT. Patients with HBV-associated small solitary HCC who were newly treated with SBRT were retrospectively analysed in our hospital from 2009 to 2016. We counted the APRI, NLR, PLR, and LMR before treatment and calculated their cut-off values for predicting overall survival (OS) and progression-free survival (PFS) by receiver operating characteristic (ROC) analysis. The random forest model combined with least absolute shrinkage and selection operator (LASSO) regression model for OS and PFS were used to screen potentially prognostic factors from serum inflammatory markers, demographic data, and clinical characteristics. Predictive models for OS and PFS were developed by multivariable COX regression and nomograms were constructed. Discrimination was assessed using the C-index. Internal validation was assessed using the Bootstrap method. Survival analysis was carried out to assess the prognostic value of serum inflammatory markers, and OS and PFS curves were compared by Kaplan-Meier analysis and Log-Rank test, respectively. A total of 72 patients with HBV-associated small solitary HCC were recruited for the study. The median follow-up time was 2015 days (range, 232-3823 days). Age, tumor size, NLR, PLR, and APRI were used to construct nomogram for OS, while gender, age, TNM stage, portal hypertension, AFP, APRI were for PFS. The two models displayed good discriminations with C-indexes of 0.738 (95% CI: 0.632-0.844) and 0.657 (95% CI: 0.538-0.777), and their C-indexes in the internal validation cohort reached 0.790 (95% CI: 0.684-0.896) and 0.739 (95% CI: 0.619-0.859). The multivariable cox analysis indicated that APRI<0.47 was favourable independent prognostic factors for OS and PFS. Compared to APRI≥0.47, APRI<0.47 predicts better OS (=0.003) and PFS (=0.003). Nomograms based on APRI are superior in predicting OS and PFS in HBV-associated small solitary HCC patients who have received SBRT. APRI before treatment is a feasible and convenient prognostic indicator for OS and PFS, which helpfully determines the beneficial population of SBRT for HBV-associated small solitary HCC.