The aim of this study was to evaluate the prognostic significance of the combination of the magnetic resonance spectroscopy (MRS) parameters and systemic immune-inflammation index (SII) in patients with brain metastases (BMs) from non-small cell lung cancer (NSCLC) treated with stereotactic radiotherapy.A total of 118 NSCLC patients with BM who were treated with stereotactic radiotherapy were retrospectively enrolled in this study. All patients underwent MRS and blood samples test for SII analysis before the initiation of stereotactic radiotherapy. The correlation between the parameters of MRS and SII level was assessed using Spearman's correlation coefficient. The cutoff values for the parameters of MRS, SII, and clinical laboratory variables were defined by the receiver operating characteristic (ROC) curve analysis to quantify these predictive values. The prognostic factors of overall survival (OS) and progression-free survival (PFS) curves were assessed using the Kaplan-Meier and Cox proportional hazards models.The median follow-up time was 25 months (range, 12-49 months). The optimal cutoff point for the choline/creatine (Cho/Cr) ratio and SII were 1.50 and 480, respectively. The Cho/Cr ratio was negatively correlated with SII (rs = 0.164, p = 0.075), but there was a trend. The C-SII score was established by combining the Cho/Cr ratio and SII. Patients with both an elevated Cho/Cr ratio (>1.50) and an elevated SII (>480) were given a C-SII score of 2, and patients with one or neither were given a C-SII score of 1 or 0, respectively. The Kaplan-Meier analysis showed that a C-SII score of 2 was significantly linked with poor OS and PFS (p < 0.001 and p < 0.001, respectively). In the Cox proportional hazards model, the C-SII score independently predicted OS [hazard ratio (HR), 1.749; 95% CI, 1.176-2.601; p = 0.006] and PFS (HR, 2.472; 95% CI, 1.624-3.763; p < 0.001).The C-SII score was more accurate for predicting the clinical outcomes of NSCLC patients with BM who underwent stereotactic radiotherapy. The C-SII score, which was superior to either score alone, could be used to identify BM in NSCLC patients with poor outcomes.Copyright © 2022 Guo, Liu, Li, Chen, Zhao, Guo, Zhu and Ji.