Is MRI useful for the prediction of nipple involvement?

The purpose of this retrospective study was to examine the usefulness of MRI for predicting nipple involvement.

A retrospective review was performed of 81 breasts in 78 breast cancer patients who had undergone nipple or nipple-areolar complex resections during surgery. The enhancement patterns of the nipple on MRI were compared to the contralateral side and were categorized as no enhancement, bilateral enhancement, and unilateral enhancement. The physical examination results, MRI findings, and pathological findings were assessed.

Nipple involvement was observed in 17 of 81 nipples (21%), of which one exhibited infiltration of invasive ductal carcinoma, 12 exhibited ductal carcinoma in situ, and four exhibited Paget's disease. The nipples with no enhancement or with bilateral enhancement on MRI showed a lower rate of nipple involvement than the nipples with unilateral enhancement (0/48 vs. 17/33, p < 0.001). Among the unilaterally enhanced nipples, those with continuous enhancement from the index lesion showed a higher rate of nipple involvement than those without continuous enhancement (16/26 vs. 1/7, p = 0.03). When the enhancement characteristics of the unilaterally enhanced nipples were categorized, the findings with relatively high positive predictive values for nipple involvement were diffuse enhancement (100%, 2/2), unilateral skin enhancement (100%, 2/2), rim enhancement (86%, 6/7), and periductal enhancement (80%, 4/5).

MRI was useful for predicting nipple involvement. Unilateral enhancement of the nipple with continuous enhancement from the index lesion may suggest the possibility of nipple involvement.