Immuno-inflammatory signature for predicting therapeutic response and survival after stereotactic radiosurgery in NSCLC patients with brain metastases: a retrospective cohort study.

Purpose: This study aimed to delineate critical factors, particularly immune-inflammatory biomarkers, that predict therapeutic response and overall survival (OS) in non-small cell lung cancer (NSCLC) patients with brain metastases (BM) undergoing stereotactic radiosurgery (SRS), and to develop novel decision-tree and nomogram models for prognostication. Patients and methods: In this retrospective study, we analyzed data from 464 NSCLC patients with BM treated with SRS between February 2016 and November 2022. The cohort was randomly split into training and validation sets (7:3 ratio). A C5.0 algorithm was employed to build a decision tree model for treatment response. Prognostic factors for OS were identified via univariate and multivariate Cox regression, and subsequently used to construct graphical and online nomograms. Model performance was assessed with calibration curves and the C-index. Results: The median OS for the entire cohort was 15.8 months (95% confidence interval [CI]: 14.6 to 17.0 months). The decision tree model for treatment response identified NLR as a key predictor, alongside volume of brain metastases, Score Index for Radiosurgery (SIR), edema index (EI), and maximum diameter. Multivariate Cox analysis identified age, volume of brain metastases, EI, and SIR as independent prognostic factors for OS. Graphical and dynamic nomograms were developed based on these factors (available at: https://helloshinyweb.shinyapps.io/brain_metastasis_from_NSCLC/). The calibration curves demonstrated good consistency between predicted and actual survival, and the C-index indicated a moderate discriminative ability. Conclusions: We identified that immune-inflammatory profiles and radiological-clinical factors are significant predictors for treatment response and OS in NSCLC patients with BM undergoing SRS. The developed decision tree and nomogram models, which incorporate immune-inflammatory profiles, provide user-friendly tools to assist clinicians in optimizing personalized management for this patient population.