Clinical Value of Inflammatory Indexes in Predicting Post-Transplantation Prognosis in Hepatocellular Carcinoma: A
Han-Xuan Wang1,2, Xiao-Yong Ye1,2, Tao Jiang1,2
1Division of Hepatobiliary and Pancreaticosplenic Surgery, Department of General Surgery, Beijing Chao-Yang Hospital, Capital Medical University, Beijing, People's Republic of China.
Background:
Liver transplantation (LT) is an effective treatment for hepatocellular carcinoma (HCC), and inflammatory indexes are accessible, non-invasive prognostic biomarkers for HCC patients. This study aims to systematically evaluate the clinical value of various preoperative inflammatory indexes in predicting post-LT outcomes to optimize the stratification of suitable HCC patients that can benefit from LT.
Methods:
This retrospective study included 239 HCC patients who underwent LT from April 2013 to July 2023 and their inflammatory biomarkers were calculated. Univariate and multivariate analysis were applied to identify independent risk factors for postoperative prognosis. Subgroup analysis was used to evaluate the predictive value in different sub-populations. Prediction models based on preoperative indexes were established using multiple machine-learning model. Mediating effect analysis was utilized to elucidate potential mediating factors.
Results:
Inflammation biomarkers were significantly different in patients with different postoperative prognosis. Lymphocyte-to-monocyte ratio (LMR, RR: 0.806, 95% CI: 0.675-0.962, P=0.017) served as an independent predictor for early tumor recurrence, with an optimal cutoff of 2.03. The systemic immune-inflammation index served as an independent risk factor for both overall survival (HR: 3.472, 95% CI: 1.322-9.119, P=0.012) and disease-free survival (HR: 2.604, 95% CI: 1.012-6.698, P=0.047), with an optimal cutoff value of 155.19. The prognostic value of SII was robust across nearly all clinical subgroups, while the predictive value of LMR was limited to certain subgroups. The random survival forest model, incorporating alpha-fetoprotein and SII as its most important variables, achieved apparent and bootstrap-corrected C-indexes of 0.814 and 0.754 and apparent and bootstrap-corrected area under receiver characteristic curve of 0.854 and 0.788, demonstrating the optimal predictive accuracy in all established models. Mediating analysis indicated that tumor T stage significantly mediated the correlation between SII and postoperative survival outcomes, while liver function and tumor burden (T stage and tumor number) exerted significantly suppression and mediating effect on correlation between LMR and early tumor recurrence.
Conclusion:
Preoperative LMR and SII are associated with early tumor recurrence and long-term prognosis in HCC patients after LT. A random survival forest model that integrated these inflammatory indexes with other preoperative variables may serve as a potential complementary tool for early risk stratification in the future after verified by further external validation.