Factors predictive of occult nipple‐areolar complex involvement in patients with carcinoma in situ of the breast

Abstract

To investigate predictors of occult nipple-areolar complex (NAC) involvement in patients with carcinoma in situ (CIS) and to validate an online probability calculator (CancerMath; www.lifemath.net/cancer/breastcancer/nipplecalc/index.php). Mastectomized patients with CIS (n = 104) were retrospectively selected. Clinicopathology and preoperative mammography, ultrasound, and magnetic resonance imaging (MRI) findings were analyzed. Histopathological NAC-positivity was confirmed in 20 (19.2%) patients. Short nipple-tumor distance and suspicious extension to the nipple by mammography were significant but ultrasound was not significant to predict NAC involvement. NAC-positive cases had MRI findings of shorter nipple-tumor distance in both the early and delayed phases. Multivariable regression model showed age >50 years and shorter tumor-nipple distance on the delay phase of MRI were statistically significant predictors of NAC involvement. Area under the receiver operating characteristics curve (AUC) was 0.618 when calculated by CancerMath; however, an AUC of 0.954 was achieved when distance and age were applied together as predictor. Mammographic and MRI findings were significant for predicting NAC involvement, with distance of the tumor from the nipple in delay phase MRI the most significant predictor of NAC involvement. Therefore, breast MRI could be beneficial for planning nipple-sparing mastectomy in patients with CIS.