Abstract
BACKGROUND: Brain metastases (BMs) from non-small cell lung cancer (NSCLC) remain a major clinical challenge, and existing prognostic tools such as the Graded Prognostic Assessment (GPA) do not incorporate imaging biomarkers or adequately reflect the impact of immunotherapy. Meningeal lymphatic vessels (mLVs), which regulate cerebrospinal fluid drainage and immune surveillance, have been implicated in tumor-immune interactions. We aimed to develop and internally validate a multivariable prognostic model integrating mLV remodeling measured by black-blood magnetic resonance imaging (BB-MRI) with clinical predictors to improve early prediction of treatment response. METHODS: We retrospectively analyzed 130 patients with pathologically confirmed NSCLC (100 with BM, 30 without BM). Among the BM cohort, 56 patients achieved favorable treatment response [stable disease (SD) or partial response (PR)] and 44 experienced progressive disease (PD). Candidate predictors were pre-specified based on clinical relevance, and the final model incorporated total mLV diameter, immunotherapy exposure, sex, and extracranial lesion count. Internal validation was performed with 1,000 bootstrap resamples. Model performance was assessed by discrimination, calibration, and decision curve analysis (DCA). RESULTS: The final multivariable model demonstrated good discrimination [area under the curve (AUC) =0.82; 95% confidence interval (CI): 0.75-0.90], excellent calibration, and consistent net clinical benefit across a range of threshold probabilities. The calibration and decision curves showed promising internal performance, but external validation is required before clinical application. CONCLUSIONS: BB-MRI-derived mLV remodeling may be an early and noninvasive indicator of treatment efficacy in BM. The proposed nomogram enables the individualized prediction of systemic therapy response, supporting precision immunotherapy for patients with intracranial metastases.