The added value of contrast-enhanced ultrasonography to BI-RADS in the evaluation of breast lesions presenting with pathological nipple discharge.

BACKGROUND

Pathological nipple discharge (PND) is a frequent presentation in breast clinics, yet its accurate differentiation between benign and malignant lesions remains difficult due to the inherent limitations of current imaging tools. There is still a lack of a practical, non‑invasive approach that incorporates contrast-enhanced ultrasound (CEUS) features with the conventional Breast Imaging Reporting and Data System (BI-RADS) to improve the diagnostic yield. This study aimed to evaluate the feasibility and diagnostic performance of a rerated BI-RADS system combining ultrasound (US) and CEUS versus US BI-RADS alone for breast lesions causing PND.

METHODS

A total of 217 consecutive patients with PND underwent conventional US and CEUS between January 2009 and August 2025. Each lesion was evaluated with US BI-RADS categories, then CEUS BI-RADS categories were re-rated based on US BI-RADS categories and CEUS scores. All responsible lesions were confirmed by pathological results through surgery or biopsy. To account for the potential correlation among multiple lesions from the same patient, we used generalized estimating equations with an exchangeable correlation structure for multivariate analysis to identify independent risk factors. The diagnostic performance of US BI-RADS and CEUS BI-RADS was evaluated using receiver operating characteristic curve analysis.

RESULTS

There were 258 breast lesions from 217 patients consecutively evaluated in this retrospective study. A total of 57 lesions were pathologically confirmed to be malignant, and 201 were confirmed to be benign. There was a statistically significant difference in multiple clinical and ultrasonic features between malignant and benign lesions (P<0.05). Univariate and multivariate logistic regression analysis based on generalized estimating equations identified age, size, enhancement range, enhancement order and spiculate sign as independent predictors of malignant breast lesions with discharge (all P<0.02). The BI-RADS categories based on CEUS showed higher sensitivity, negative predictive value (NPV) and the area under the receiver operating characteristic curve (AUROC) compared to US BI-RADS, which were 0.930 [95% confidence interval (CI): 0.822-0.977], 0.945 (95% CI: 0.858-0.982) and 0.637 (95% CI: 0.575-0.695), respectively.

CONCLUSIONS

In conclusion, CEUS BI-RADS demonstrates high sensitivity and NPV in evaluating breast lesions associated with PND. Although its overall discriminatory ability, as measured by AUROC, was not significantly different from that of conventional US BI-RADS when BI-RADS 4a lesions were considered test-positive, CEUS provides complementary functional information that aids in risk stratification. These findings position CEUS as a valuable adjunct to conventional US, particularly in refining the assessment of low-suspicion lesions and guiding biopsy decisions in patients with PND.