Prediction of the Malignancy of a Breast Lesion Detected on Breast Ultrasound: Radiomics Applied to Clinical Practice

SIMPLE SUMMARY: Breast cancer is the most frequent cancer among women: early diagnosis and management of breast lesions are crucial to achieve a better prognosis for patients with this diagnosis. Breast ultrasound (US) is one of the main techniques for the management of breast lesions and it is important in doubtful findings on mammography and in the evaluation of dense breasts. Unfortunately, US has a high rate of false positive and has high operator dependence. Ultrasound CAD (computer-aided diagnosis) and radiomics are newly developed tools that can help solve these issues: this study aims to create a radiomics score from breast US to predict malignancy of a breast lesion, and to also combine this score with CAD and sonographer performances. Finally, we would like to create a prediction tool of US radiomics features combined with CAD, clinical parameters, and Breast Imaging Reporting and Data System evaluation for the prediction of malignancy of breast lesions. ABSTRACT: The study aimed to evaluate the performance of radiomics features and one ultrasound CAD (computer-aided diagnosis) in the prediction of the malignancy of a breast lesion detected with ultrasound and to develop a nomogram incorporating radiomic score and available information on CAD performance, conventional Breast Imaging Reporting and Data System evaluation (BI-RADS), and clinical information. Data on 365 breast lesions referred for breast US with subsequent histologic analysis between January 2020 and March 2022 were retrospectively collected. Patients were randomly divided into a training group (n = 255) and a validation test group (n = 110). A radiomics score was generated from the US image. The CAD was performed in a subgroup of 209 cases. The radiomics score included seven radiomics features selected with the LASSO logistic regression model. The multivariable logistic model incorporating CAD performance, BI-RADS evaluation, clinical information, and radiomic score as covariates showed promising results in the prediction of the malignancy of breast lesions: Area under the receiver operating characteristic curve, [AUC]: 0.914; 95% Confidence Interval, [CI]: 0.876–0.951. A nomogram was developed based on these results for possible future applications in clinical practice.