Preoperative Diagnosis of Ipsilateral and Contralateral Breast Cancer: Role of Diffusion-weighted MRI

Background The interpretation of the multiparametric MRI, which combines dynamic contrast-enhanced (DCE) MRI with diffusion-weighted imaging (DWI), has the potential to increase MRI diagnostic accuracy. Purpose To investigate and validate the potential of DWI with an apparent diffusion coefficient (ADC) cutoff in evaluating additional lesions detected at preoperative MRI in patients with breast cancer. Materials and Methods In this retrospective review, data from patients with additional lesions (Breast Imaging Reporting and Data System [BI-RADS] category 3 or higher) who underwent MRI between June 2019 and June 2021 were evaluated. Two breast radiologists independently evaluated additional enhanced lesions and measured the ADC values. The optimal ADC cutoff for downgrading lesions was determined according to the Youden index and was applied to a separate validation cohort. The efficacy of prespecified ADC values (1.53 × 10^-3 mm²/sec and 1.3 × 10^-3 mm²/sec) was investigated. Diagnostic performance was evaluated and compared using generalized estimating equations. Results Data from 219 patients (mean age, 50.8 years ± 10.2 [SD]) with 292 additional lesions were evaluated. The optimal ADC cutoff was 1.0 × 10^-3 mm²/sec, which, compared with that for DCE MRI alone, increased specificity (from 28.7% to 73.1%; P < .001) but decreased sensitivity (from 99.2% to 89.6%; P < .001) of MRI. In the validation cohort (104 patients, 133 additional lesions), 48 of 133 (36.1%) lesions were diagnosed as cancer, and applying the ADC cutoff derived from the development cohort increased specificity (from 21.2% to 68.2%; P < .001) but decreased sensitivity (from 97.9% to 83.3%; P = .01). For ipsilateral lesions, specificity increased (from 8.7% to 65.2%; P < .001), but the sensitivity did not decrease significantly (from 97.1% to 85.3%; P = .052). For contralateral lesions, similar performance improvements were observed for specificity (from 35.9% to 71.8%; P < .001), but a greater decrease in sensitivity was observed (from 100% to 78.6%; P < .001). Application of the prespecified ADC cutoffs achieved higher sensitivity values but smaller improvements in specificity. Conclusion The ADC cutoff based on MRI with DWI improved the performance of preoperative MRI in diagnosing additional lesions in patients with breast cancer. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Honda and Iima in this issue.