Multiparametric MRI characteristics for differentiating primary cancer origin in brain metastases

Abstract Background Accurate identification of primary cancer origin in brain metastases (BMs) is crucial for diagnosis and treatment planning. Identification is challenging, particularly in unknown primary cases. Multiparametric Magnetic Resonance Imaging (MRI) offers non-invasive imaging characteristics as diagnostic clues. This study aimed to investigate the relationship between multiparametric MRI features of BMs and their primary tumor origin. Methods This retrospective study included 125 patients with intra-axial brain metastases, classified as breast cancer, small cell lung cancer (SCLC), non-small cell lung cancer (NSCLC), malignant melanoma, and other cancers. Multiparametric MRI, including T2-FLAIR, contrast-enhanced 3D T1-weighted imaging, apparent diffusion coefficient (ADC) maps, susceptibility-weighted imaging (SWI), and cerebral blood volume (CBV) maps, was performed. Lesion volume, necrotic volume, necrosis ratio, peritumoral edema volume, and edema ratio were measured. 3D Slicer ® software was used for semi-automated volumetric measurements. Quantitative ADC and rCBV values, intratumoral susceptibility signal (ITSS) scores, and metastasis number and location were analyzed to differentiate primary tumor origins. Results 291 metastatic lesions from 125 patients were analyzed. NSCLC metastases showed significantly higher necrosis ratios than breast ( p < 0.001) and SCLC ( p = 0.025) metastases. Malignant melanoma metastases exhibited significantly higher edema ratios than breast ( p = 0.001) and SCLC ( p = 0.004) metastases. NSCLC metastases also showed a significantly higher edema ratio than breast ( p = 0.011) metastases. Mean ADC values were significantly higher in NSCLC metastases compared to all other groups, with an optimal cut-off of ≥ 0.905 mm²/s (AUC: 0.775). Malignant melanoma and SCLC metastases had significantly higher rCBV values than breast cancer metastases. Malignant melanoma metastases consistently showed the highest ITSS scores among all groups, with an optimal cut-off of ≥ 1 (AUC: 0.769). Furthermore, multiple metastases ≤ 0.5 cm³ were significantly associated with breast cancer ( p = 0.041). Infratentorial metastases were more prevalent in breast cancer (OR = 2.490, p = 0.001). Conclusion Brain metastases exhibit distinct multiparametric MRI characteristics by primary cancer origin. Breast cancer metastases tend to be smaller, multiple, and infratentorial. NSCLC metastases show greater necrosis and higher ADC values, while melanoma metastases demonstrate higher intratumoral susceptibility. These features offer valuable diagnostic clues for differentiating primary cancer types in patients with brain metastases. Clinical trial number Not applicable.