Microstructural Characterization of Breast Lesion Subtypes Using Time-Dependent Diffusion MRI: A Multicenter Study

To evaluate the diagnostic performance of microstructural parameters derived from time-dependent diffusion MRI (Td-dMRI) with the oscillating gradient spin-echo (OGSE) in differentiating breast lesion subtypes.

Patients with clinically diagnosed breast lesions were enrolled from two centers between March 2024 and May 2025. All patients underwent conventional MRI and Td-dMRI examinations. The IMPULSED model was applied to fit Td-dMRI data for extracting microstructural parameters. Lesions were divided into two groups (benign vs malignant) and three groups (benign lesions, in situ carcinoma, invasive carcinoma) for comparative analysis. The diagnostic performance of these parameters was assessed using the area under the ROC curve. Univariate and multivariate logistic regression analyses were performed to evaluate the associations between microstructural parameters and lesion subtypes.

A total of 84 subjects (mean age: 53 ± 14 years) with 89 breast lesions were included. Univariate analysis revealed that microstructural parameters such as cell density were significantly associated with the differentiation of benign and malignant lesions. Multivariate logistic regression analysis identified a higher Vin (P < .001) and a lower ADC_OGSEN1 (P = .006) as independent predictors of malignant lesions. The combined model exhibited favorable diagnostic performance with an AUC of 0.86. Additionally, statistically significant differences in cell density, Vin, ADC_PGSE were observed between benign lesions and carcinoma in situ (P = .004 to .013).

Td-dMRI combined with OGSE technology shows potential for effectively differentiating benign from malignant breast lesions. Moreover, the derived microstructural parameters may help facilitate the distinction between benign lesions and carcinoma in situ.