A Point-based Risk Stratification System for Breast Masses: Potential Alternative to the Existing ACR BIRADS

High variability is observed in the application of the American College of Radiology (ACR) Breast Imaging Reporting and Data System (BIRADS), which can lead to misclassifications. However, limited efforts have been made to develop a more practical point-based BIRADS (P-BIRADS). This study aims to construct a P-BIRADS and to compare it with the ACR BIRADS.

This prospective multicenter study enrolled consecutive adult females with breast masses who underwent breast ultrasound followed by biopsy or surgery from August 2020 to March 2024. Univariable and multivariable logistic regression analyses were conducted to assess the association between qualitative ultrasound features and breast cancer in a development dataset. Features with a p-value of less than 0.05 were utilized to construct the P-BIRADS. The area under the receiver operating characteristic curve (AUC-ROC) was compared between P-BIRADS and ACR BIRADS in the validation dataset.

A total of 518 consecutive patients with 578 masses (benign: mean age 39.5 ± 10.3 y; malignancy: mean age 50.2 ± 12.6 y) were enrolled in the development dataset. Additionally, 69 consecutive patients with 81 masses (benign: mean age 38.8 ± 13.0 y; malignancy: mean age 54.6 ± 9.3 y) were included in the external validation dataset. Independent risk factors associated with breast cancer included mass size, orientation, margin and vascularity. The area under the curve values (AUC) for the ACR BIRADS and P-BIRADS demonstrated no significant difference in the development dataset (0.81 vs. 0.83, p = 0.494) and the validation set (0.92 vs. 0.90, p = 0.618), respectively.

The diagnostic performance of P-BIRADS is comparable to that of ACR BIRADS. However, P-BIRADS offers multimodal predictors and a simpler, more efficient risk stratification system for radiologists in clinical practice.