The aim of this study was to assess tumor vascularity through three dimensional (3D) power Doppler breast ultrasound (US) and propose a decision model for the classification of benign and malignant breast tumors. Patient recruitment for this study was performed consecutively during a 13-mo period (January 2003 to February 2004). A total of 102 benign and 93 malignant solid breast images were analyzed. Three-dimensional power Doppler US imaging was performed using a Voluson730 ultrasound system equipped with a relative stopping power index (RSP) 6 to 12 transducer. The Virtual Organ Computer-aided Analysis (VOCAL)-imaging program (version 2.1) was used to analyze the stored volume. Histogram indices of the vascularization index (VI), flow index (FI) and vascularization-flow index (VFI) for the intra-tumor and for shells with a thickness of 3 mm surrounding the breast lesion were calculated and showed that for both, malignancy had a higher VI, FI and VFI than benignancy, with statistical significance. Multivariate and stepwise logistic regression revealed the model (including patient age, volume and intra-tumor FI in 3D power Doppler vascularity) to be the best choice for malignant breast tumor characterization. The receiver operating characteristics (ROC) index for the performance of the model was 0.926. Histogram indices for the intra-tumor FI in the 3D power Doppler scan are a good choice of parameter for differentiating between malignant and benign tumors with respect to the power of sensitivity, no matter whether one index is suggested or the patients' age and volume are considered.