The Role of Magnetic Resonance Imaging in Distinguishing Intraductal Papillary Lesions

Objective: The Third International B3 Lesions Consensus Conference recommends vacuum-assisted excision (VAE) followed by radiological surveillance for intraductal papillary lesions (IPLs). However, VAE is available in a limited number of centers, and the optimal imaging strategy for follow-up remains uncertain. This study aimed to evaluate the diagnostic performance of magnetic resonance imaging (MRI) findings in differentiating benign from malignant IPLs.

Methods: Magnetic resonance imaging records of patients diagnosed with IPL on core needle biopsy were retrospectively reviewed. Magnetic resonance imaging features were assessed, and apparent diffusion coefficient (ADC) values were measured. Histopathologically, lesions were classified into papilloma (Group 1) and papilloma with atypia/malignancy (Group 2).

Results: A total of 83 lesions met the inclusion criteria, comprising 42 in Group 1 and 41 in Group 2. Compared with Group 1, Group 2 lesions more frequently demonstrated nipple retraction (P = .028), larger size (P < .001), irregular margins (P = .018), and peripheral location (P = .036). Both mean (P = .004) and minimum ADC values (P = .001) were lower in Group 2, and more lesions showed a type 3 pattern (P = .008). Receiver-operating characteristic analysis identified lesion size >10.5 mm as the strongest discriminator (area under the curve = 0.743). Mean ADC provided the highest sensitivity (74%), whereas contour irregularity offered the greatest specificity (87%).

Conclusion: Following core biopsy, MRI features, including lesion size >10.5 mm, irregular contours, low ADC values, and type 3 kinetic curves, may help guide clinical decision-making between VAE and surgical excision in patients with IPL.

Cite this article as: Ergin.z E, Kahveci A, Kurt SA, Öztürk T, Kayadibi Y. The role of magnetic resonance imaging in distinguishing intraductal papillary lesions. Cerrahpaşa Med J 2026, 50, 0105, doi: 10.5152/cjm.2026.25105.