The prognostic predictive value of indirect bilirubin-inflammation score in patients with nasopharyngeal carcinoma

Objective: To construct an effective prognostic model based on indirect bilirubin (IBIL) and inflammatory markers, including neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and platelet-to-lymphocyte ratio (PLR), to predict overall survival (OS) in patients with nasopharyngeal carcinoma (NPC). Methods: Hematological parameters, including IBIL and parameters of peripheral blood cells, were retrospectively analyzed in 688 NPC patients before treatment. These patients were randomly divided into a training set (n=481) and a test set (n=207). The IBIL-inflammation (IBI) score was developed using the machine learning. A nomogram was established, and the performance of the prediction model was measured by the concordance index (C-index), time-dependent receiver operating characteristic (ROC) curve, and decision curve analysis (DCA). The interaction and mediation between the biomarkers were further analyzed. Results: By comparing 14 types of machine learning algorithms, the optimal model, oblique random survival forest, was selected to construct IBI score. The C-index of the IBI score was 0.722 in the training set, 0.564 in the test set, and 0.672 in the entire set. The area under the curve of time-dependent ROC at 1, 3, and 5 years was 0.762, 0.712, and 0.705 in the entire set respectively. IBI score was significantly positively correlated with clinical TNM stage (P＜0.05). The nomogram, which integrated age, sex, clinical stage, and IBI score, demonstrated good clinical utility and predictive ability, as evaluated by the DCA. Significant interaction was found between IBIL and PLR, and inflammatory markers did not exhibit any medicating effects on the influence of IBIL on NPC survival. Conclusion: The IBI score, as a potential prognostic factor for NPC patients, offers advantages in convenience and cost-effectiveness for detection. It can provide a basis for personalized prognostic predictions and the formulation of clinical treatment strategies for NPC.