Visualizing the vascular-cellular microenvironment in lung cancer brain metastasis via multiparametric fusion of DCE-MRI and DWI.

Background: The heterogeneity of the lung cancer brain metastasis (LCBM) microenvironment limits therapeutic efficacy, while invasive pathological biopsies fail to dynamically assess brain metastasis (BM) comprehensively. Non-invasive imaging techniques thus hold clinical value for visualizing the LCBM microenvironment. This study aimed to achieve non-invasive quantitative analysis of vascular function and cellular structures in LCBM using multiparametric Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) and Diffusion-Weighted Imaging (DWI). Methods: A prospective cohort of 114 LCBM patients (63 lung adenocarcinoma [LUAD]-BM, 28 lung squamous cell carcinoma [LUSC]-BM, 23 small cell lung cancer [SCLC]-BM) underwent DCE-MRI and DWI on a 3.0T MRI scanner. Parameters including volume transfer constant (Ktrans), rate constant (Kep), extravascular extracellular volume (Ve), and plasma volume fraction (Vp) were derived using the Extended Tofts model. Group differences were analyzed via Mann-Whitney U test, diagnostic efficacy via ROC curves, and parameter interactions via multivariate logistic and linear regression. Results: ADC distinguished SCLC-BM from NSCLC-BM with AUC=0.891 (specificity=95.05% at 752.4×10-6mm²/s), while Ktrans differentiated LUAD-BM from LUSC-BM with AUC=0.998 (sensitivity=98.48%, specificity=97.79% at 157 min-1/1000). Microenvironmental profiles: LUAD-BM showed high Vp (51.50/1000) and Ktrans (424.8 min-1/1000); LUSC-BM had low Ktrans (61.15 min-1/1000) and medium ADC (1163×10-6mm²/s); SCLC-BM exhibited high cellular density (ADC=661×10-6mm²/s) and abnormal contrast kinetics (high Kep, low Vp). Vp and ADC were identified as independent predictors for LUAD-BM and SCLC-BM, respectively. Parameter interactions varied by subtype: ADC in LUAD-BM correlated with Kep and Delay Time; in LUSC-BM, with Ktrans and Ve; and in SCLC-BM, showed weaker vascular associations. Ktrans regulation involved distinct parameter contributions across subtypes. Conclusion: A DCE-MRI-DWI "Vascular-Cellular Microenvironment Visualization Model" was established, revealing distinct profiles: high microvascular density/permeability in LUAD-BM, low permeability/medium cellularity in LUSC-BM, and high cellularity/abnormal contrast kinetics in SCLC-BM. This validates multimodal imaging for characterizing LCBM heterogeneity and provides insights into tumor angiogenesis, cellular density, and BBB regulation, supporting microenvironment-targeted therapy.