Background/Objectives: Brain metastases (BMs) are a major cause of morbidity and mortality in non-small-cell lung cancer (NSCLC). In this setting, EGFR-mutant and ALK-rearranged tumors represent clinically actionable, CNS-relevant oncogenic subgroups for which matched TKIs are essential to management, yet lesion-level molecular profiling is not always feasible or immediately available. We aimed to develop and externally validate DriverNet, a clinical and MRI-based framework for noninvasive pre-treatment molecular triage of EGFR/ALK status in NSCLC-BM. Methods: In this multicenter study, we analyzed pretreatment clinical, T1CE and T2-FLAIR MRI data to develop unimodal radiomics, 2D/2.5D deep learning (DL), and multimodal fusion models. The final model used ImageNet-pretrained CNNs for feature extraction and a Transformer-based architecture for fusion. The primary cohort was split strictly at the patient level before slice extraction and model development, and two independent external cohorts were used for testing. Clinical-only, imaging-only, and clinical-imaging models were compared using discrimination, calibration, Brier score, and decision-curve analyses. Model interpretability and exploratory prognostic stratification were also assessed. Results: A total of 374 patients from three centers were included. Center 1 comprised 224 patients (Chinese) and was divided into training (n = 179, EGFR/ALK+ 55.3%) and internal validation (n = 45, EGFR/ALK+ 57.8%) sets. External cohorts included 54 patients from Center 2 (Chinese, test 1, EGFR/ALK+ 42.6%) and 96 from Center 3 (Western, test 2, EGFR/ALK+ 12.5%). Among all evaluated models, DriverNet achieved the best overall performance, with AUCs of 0.967, 0.947, 0.962, and 0.952 in the training, internal validation, and two external cohorts, respectively, outperforming the clinical-only and imaging-only models. Model-derived labels were also associated with overall survival in exploratory analyses. Conclusions: DriverNet is a clinical and MRI-based framework for noninvasive pre-treatment molecular triage in NSCLC-BM. It may provide complementary information for future molecular triage studies when lesion-level profiling is unavailable or delayed. Prospective validation in larger and more molecularly balanced cohorts remains necessary before any clinical implementation can be considered.

Background: Patients presenting with concurrent lung and brain lesions pose a unique diagnostic and therapeutic challenge. The role of neurosurgical resection in the management of de novo non-small-cell lung cancer brain metastases (NSCLC-BM) remains poorly defined. Methods: Consecutive patients undergoing resection for NSCLC-BM between February 2018 and October 2025 were retrospectively analyzed. De novo NSCLC-BM was defined as radiographic detection of pulmonary and intracranial lesions concurrently or within 1 month during the initial diagnostic work-up. Clinical characteristics and outcomes were compared between de novo and metachronous brain metastases (BM). Multivariable Cox regression identified factors associated with overall survival (OS). Factors associated with long-term postoperative survival in the de novo subgroup were further explored using Firth logistic regression. Results: Among 248 surgically treated patients, 113 (45.6%) presented with de novo BM. Compared with metachronous BM, de novo patients more frequently presented with symptomatic and multifocal intracranial disease. Among 73 de novo patients with evaluable survival, median postoperative OS was 57.0 months (95% CI, 22.9–not reached). Postoperative OS and intracranial progression-free survival (iPFS) did not differ significantly between the two groups, and de novo presentation was not independently associated with postoperative survival. Driver-positive status and lower intracranial tumor burden were associated with durable postoperative survival in the de novo subgroup. Conclusions: Neurosurgical resection represents a reasonable option for appropriately selected patients with de novo NSCLC-BM. These findings provide real-world evidence to inform multidisciplinary, individualized decision-making in patients presenting with concurrent lung and brain lesions.