Nipple Discharge: Imaging Variability Among U.S. Radiologists.

Abstract

The purpose of this study was to assess radiologists' choice of imaging modality for the evaluation of clinical symptoms of physiologic nipple discharge (e.g., bilateral discharge, multiple-duct orifices, and yellow, green, or white color) and pathologic nipple discharge (e.g., unilateral discharge, single-duct orifices, spontaneous and serous discharge, and clear or bloodstained color). An online survey was sent to lead interpreting physicians at mammography facilities accredited by the American College of Radiology (ACR). Statistical analysis was performed using chi-square tests for frequency data and multinomial logistic regression. A total of 849 responses to 8170 distributed surveys were received, for a response rate of 10.4%. For the workup of physiologic nipple discharge, 30% of respondents recommended screening mammography (SM); 24%, diagnostic mammography (DM) only; and 46%, both DM and targeted ultrasound (US) (DM plus US). For the workup of physiologic nipple discharge, practitioners in nonacademic settings and those who read breast images during less than 50% of their practice were significantly more likely to recommend DM (with or without US), compared with SM (the standard recommended by the ACR). Those reading breast images less than 50% of the time were also more likely to recommend MRI after conventional imaging revealed negative results. For the workup of pathologic nipple discharge, 91.0% of respondents recommended DM plus US; 8.5%, DM only; and fewer than 1.0%, SM. Nonacademic providers and those who read breast images less than 50% of the time were significantly less likely to recommend DM plus US (the standard recommended by the ACR), compared with DM only. The present study shows variability in imaging modality selection among U.S. radiologists handling the imaging workflow for benign and pathologic nipple discharge. Radiologists do not uniformly follow ACR practice guidelines, which potentially leads to unnecessary workups and extra health care costs.