Tumor Involvement of the Nipple at Preoperative Contrast-enhanced Mammography: A Multicenter Diagnostic Accuracy...

Background Assessing nipple involvement in breast cancer at preoperative imaging is crucial for nipple-preserving surgery, but no studies have investigated the utility of contrast-enhanced mammography (CEM) in this assessment. Purpose To evaluate the diagnostic accuracy of CEM in ruling in pathologic nipple involvement and identify associated CEM features. Materials and Methods This retrospective multicenter study included patients with biopsy-proven breast cancer treated with primary surgery who underwent preoperative CEM at five referral centers between December 2016 and December 2023. After local image review by board-certified breast radiologists (with ≥5 years of experience), CEM diagnostic accuracy indexes to rule in nipple involvement (specificity [ie, primary end point], positive predictive value, and positive likelihood ratio) were calculated with surgical pathologic findings serving as the reference standard. Associations of CEM features with pathologic nipple involvement and the minimum enhancement-to-nipple distance to rule in pathologic nipple involvement were evaluated by using multivariable logistic regression analysis (adjusted odds ratios [ORs]) and the Youden index. Results Pathologic nipple involvement was present in 147 of 558 (26.3%) cancers from 530 included women (median age, 63.7 years; IQR, 50.4-73.6 years). CEM specificity was 95.1% (391 of 411; 95% CI: 92.6, 97.0), with a positive predictive value of 83.5% (101 of 121; 95% CI: 76.5, 88.7) and a positive likelihood ratio of 14.1 (95% CI: 9.1, 22.0). CEM features associated with pathologic nipple involvement were as follows: periareolar skin thickening (OR, 3.8; 95% CI: 1.7, 8.4; P < .001), disrupted superficial linear enhancement (OR, 3.0; 95% CI: 1.3, 7.0; P = .01), nipple retraction (OR, 2.2; 95% CI: 1.1, 4.6; P = .03), and (with lower odds) 1-mm increases in the minimum enhancement-to-nipple distance (OR, 0.94; 95% CI: 0.93, 0.96; P < .001). For the minimum enhancement-to-nipple distance, a 9.5-mm cutoff best predicted pathologic nipple involvement (Youden index, 0.53). Conclusion CEM was able to help accurately rule in pathologic nipple involvement in the preoperative setting. Supplemental material is available for this article. © RSNA, 2025.