A radiomics-immune integrated nomogram for postoperative survival prediction in non-small cell lung cancer

The research aimed to develop and validate an integrative prognostic nomogram combining CT radiomics and tumor immune features to achieve personalized prediction of postoperative survival in non-small cell lung cancer (NSCLC). A total of 146 NSCLC patients who underwent surgical resection between January 2010 and December 2015 were analyzed. Radiomic features were extracted from preoperative CT images, and then a radiomics signature (Rad - score) was constructed using LASSO - Cox regression. Immune markers were assessed via immunohistochemistry to establish immune subtypes. Cox regression analysis was performed to identify independent prognostic factors, including clinicopathological characteristics (microvascular invasion, T stage,N stage, clinical stage, and family history of cancer), imaging features (solid density), and immune markers (immune subtype). These confirmed independent factors were subsequently used to develop four prognostic models: a clinicopathological model, a radiomics model, a clinicopathological - radiomics model, and a comprehensive nomogram. The model performance was evaluated using concordance index (C-index), calibration curves, as well as decision curve analysis (DCA). Seven radiomic features were identified and integrated to construct the Rad-score. Independent prognostic factors, determined through multivariate analysis, included microvascular invasion, clinical stage, lesion solid density, family history of cancer, as well as immune subtype. The integrated nomogram developed from these predictors demonstrated superior predictive performance compared to alternative models (P < 0.05), with C-indices of 0.873 in the training set and 0.841 in the test set. Calibration curves suggested agreement between model predictions and observed outcomes. DCA further confirmed that the model provided enhanced clinical net benefit across a broad range of threshold probabilities. The developed nomogram integrating clinicopathological, immune, as well as radiomic features can accurately predict postoperative survival in NSCLC patients, showing good calibration and clinical utility for individualized prognostic assessment.