Objective To explore the best cutoff point of the automated breast volume scanner (ABVS) combined with breast BI.RADS classification for differentiating malignant and benign breast lesions and to assess the diagnostic value using scoring in ABVS+ BI.RADS classification. Methods The ABVS images + BI.RADS classification of 656 breast lesions (477 cases) proved by pathology were reviewed. 5 characteristics of the side, edge, burr, entanglement and calcification + BI.RADS classification were used in scring. A receiver operating characteristic (ROC) curve was used to explore diagnostic value. Results The ratio of scring in ABVS images+ BI.RADS in all of the lesions is 6, and is the best cutoff point for differentiating malignant from benign breast lesions. The Az, sensitivity, specificity, accuracy misdiagnosis rate and missed diagnosis rate were 0.965, 90.4%、94.2%、93.4%、9.6%、3.8%. Conclusion The scring in ABVS image+ BI.RADS is a useful quantitative index for differentiating breast malignant lesion from benign lesion.