Contrast-enhanced ultrasound for differentiating benign from malignant Breast Imaging Reporting and Data System 4a breast lesions: a prospective diagnostic accuracy study.

Background: Breast Imaging Reporting and Data System (BI-RADS) 4a breast lesions are considered low suspicion for malignancy (2-10%), yet current guidelines recommend biopsy for definitive diagnosis, leading to a high rate of unnecessary invasive procedures. Contrast-enhanced ultrasound (CEUS) provides real-time microvascular perfusion information that may improve risk stratification in this indeterminate category. This study aimed to evaluate the diagnostic performance of CEUS in differentiating benign from malignant BI-RADS 4a lesions, with a focus on its potential to reduce avoidable biopsies. Methods: In this prospective diagnostic study, consecutive patients with BI-RADS 4a lesions identified on conventional ultrasound underwent CEUS examination using a sulfur-hexafluoride microbubble contrast agent. CEUS cine loops were independently reviewed by two blinded readers. Lesions were classified as CEUS-positive (suspicious for malignancy) if they exhibited ≥2 predefined malignant perfusion features (e.g., heterogeneous hyper-enhancement, centripetal filling, perfusion defects, fast wash-in/wash-out). Histopathology from biopsy or surgery served as the reference standard. Diagnostic accuracy metrics-sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and area under the receiver operating characteristic curve (AUC)-were calculated. Results: A total of 149 BI-RADS 4a lesions were included. Histopathology revealed 129 benign and 20 malignant/high-risk lesions. CEUS demonstrated a sensitivity of 85.0% [95% confidence interval (CI): 62.1-96.8%], specificity of 89.9% (95% CI: 83.3-94.6%), PPV of 47.2% (32.9-62.0%), and NPV of 98.2% (95% CI: 93.8-99.6%). The overall accuracy was 89.3% (95% CI: 83.2-93.7%), with an AUC of 0.88 (95% CI: 0.81-0.94). Malignant lesions showed significantly higher frequencies of heterogeneous hyper-enhancement, centripetal perfusion, perfusion defects, and fast wash-in/wash-out kinetics (all P<0.001). Inter-observer agreement for CEUS classification was substantial (κ=0.76). Conclusions: CEUS exhibits high diagnostic performance, especially a high NPV, in the evaluation of BI-RADS 4a breast lesions. It can reliably identify a subset of lesions with a very low probability of malignancy, suggesting that CEUS may help identify a subset of patients in whom immediate biopsy could be considered for deferral within a shared decision-making framework, though prospective validation in larger cohorts is needed. These findings support the integration of CEUS as a valuable adjunctive triage tool to reduce unnecessary biopsies in low-suspicion breast lesions.