Clinical evaluation of a nomogram model incorporating multimodal ultrasound features for breast cancer diagnosis.

Background: Breast cancer ranks as the second most common malignancy worldwide and remains the leading cause of cancer-related mortality among women. The aim of this study is to evaluate multimodal ultrasound features of breast lesions categorized as Breast Imaging Reporting and Data System (BI-RADS) 3 to 5 using logistic regression analysis, to identify independent imaging predictors of malignancy, to develop a corresponding nomogram model, and to assess its diagnostic performance. Methods: A retrospective analysis was conducted on 157 breast lesions from 141 patients with histopathologically confirmed diagnoses. The sample was divided into a training cohort (n=116) and a validation cohort (n=41). All lesions underwent B-mode ultrasound, color Doppler flow imaging (CDFI), ultrasound elastography (UE), and contrast-enhanced ultrasound (CEUS). Imaging features from each modality were systematically recorded. In the training cohort, univariate and multivariate logistic regression analyses were performed to identify independent predictors of malignancy, which were then used to construct the nomogram model. Model performance was evaluated using receiver operating characteristic (ROC) curve analysis and calibration curves. Decision curve analysis (DCA) was applied to assess clinical utility, followed by external validation using the independent cohort. Results: (I) The following imaging characteristics were identified as independent predictors of breast cancer: CEUS-based lesion size expansion, irregular lesion margins, presence of penetrating vessels or a radial enhancement pattern, heterogeneous contrast agent distribution, CDFI grades II or III, and UE scores of 4 or 5. (II) The area under the curve (AUC) for the nomogram model in the training cohort was 0.966 [95% confidence interval (CI): 0.939-0.992], with the calibration curve demonstrating strong agreement between predicted and observed probabilities. DCA indicated favorable clinical applicability. In the validation cohort, the model yielded an AUC of 0.819 (95% CI: 0.687-0.952), supporting its reliable diagnostic accuracy. Conclusions: The nomogram model incorporating multimodal ultrasound features provided effective differentiation between benign and malignant breast lesions within BI-RADS categories 3 to 5. This tool offers reliable imaging-based support for clinical decision-making in breast cancer diagnostics.