Diagnostic test of conventional ultrasonography combined with contrast-enhanced ultrasound in the subcategorization of suspicious Breast Imaging-Reporting and Data System (BI-RADS) 4 breast lesions
Yindi Zhu, Dian Zhang, Xiaona Wang, Yue-Nan Chen, Mei-Fang Pan, Susanna Guerrini, Eugene Ong, Xinxian Gu, Li Jiang
Abstract
Background: Although conventional ultrasonography (CUS) and contrast-enhanced ultrasound (CEUS) play a critical role in cancer detection, diagnosis, and image-guided biopsies, there is no standardized diagnostic approach for the clinical evaluation of suspected Breast Imaging-Reporting and Data System (BI-RADS) category 4 breast lesions. This diagnostic test evaluates the complementary roles of CUS and CEUS in addressing limitations of conventional imaging, such as microvascular visualization. This study aimed to evaluate the diagnostic value of combining CUS with CEUS in subcategorizing suspicious breast lesions classified as BI-RADS for ultrasound (US-BI-RADS) category 4. Methods: The data of 131 patients with BI-RADS category 4 breast lesions, examined between February 2017 and March 2023, were retrospectively analyzed. All lesions underwent pathological examination following surgery and served as the gold standard for diagnosis. Key features such as lesion margins, echogenicity, size, microcalcification, blood flow distribution via color Doppler flow imaging (CDFI), and CEUS characteristics were assessed. CEUS scores were calculated using a five-point scoring system. Stepwise logistic regression was applied to evaluate the odds ratios (ORs) of the lesion characteristics on US and CEUS. The combination of the US-BI-RADS and CEUS scores (termed the CEUS-BI-RADS) was compared to the US-BI-RADS alone, and a receiver operating characteristic (ROC) curve analysis was conducted to determine the diagnostic performance of these methods. Results: Of the 131 lesions, 62 (47.3%) were benign, and 69 (52.7%) were malignant. The multivariate logistic regression identified the primary indicators of malignancy as calcification [OR =1.58, 95% confidence interval (CI): 0.25-2.91, P=0.02], suspicious or abnormal axillary lymph nodes (OR =2.51, 95% CI: 0.59-4.44, P=0.01), obscure margins after enhancement (OR =2.67, 95% CI: 0.35 to 4.99, P=0.02), and increased lesion size (OR =4.89, 95% CI: 1.45-8.33, P=0.005). The sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) of the US-BI-RADS were 73.9%, 74.2%, 74.0%, 71.9%, and 76.1%, respectively, while those of the CEUS-BI-RADS were 92.8%, 79.0%, 86.3%, 90.7%, and 83.1%, respectively. The areas under the ROC curves for the US-BI-RADS and CEUS-BI-RADS were 0.741 and 0.859, respectively. Conclusions: The CEUS-BI-RADS significantly enhances diagnostic efficacy for BI-RADS category 4 breast lesions, outperforming the US-BI-RADS and could reduce unnecessary biopsies.