Angiogenesis provides blood supply for tumor expansion and also increases the opportunity for tumor cells to enter the blood or lymph circulation. Several proangiogenic factors as well as the contribution of the microenvironment to tumor-induced angiogenesis have been identified. Among these, vascular endothelial growth factor (VEGF) and the angiopoietin (Ang) family play a predominant role involved in the growth for endothelial cells. Tumor vessels are structurally and functionally abnormal because of an imbalance of these angiogenic regulators. In contrast to normal vessels, tumor vasculature is highly disorganized, tortuous and dilated, with uneven diameter and excessive branching. In other words, the morphologic features are likely to carry additional clues that, when used in conjunction with more established parameters, can improve the present diagnostic approaches. In our study, we present a new method that helps to capture the morphologic features from three-dimensional (3-D) power Doppler ultrasound (PDUS) images. After narrowing down the vessels into their skeletons using a 3-D thinning algorithm, we extracted seven features including vessel-to-volume ratio, number of vascular trees, number of bifurcation, mean of radius and three tortuosity measures, from the skeleton and applied a neural network to classify the tumors by using these features. In investigations into 221 solid breast tumors, including 110 benign and 111 malignant cases, the p values using the Student’s t-test for all features were less than 0.05, indicating that the proposed features were deemed statistically significant. The A_Z values for these seven features were 0.84, 0.87, 0.84, 0.75, 0.77, 0.79 and 0.69, respectively. The accuracy, sensitivity, specificity, and positive and negative predictive values were 80.09% (177 of 221), 80.18% (89 of 111), 80% (88 of 110), 80.18% (89 of 111) and 80% (88 of 110), respectively, with an A_Z value of 0.89. The preliminary results show that the proposed method is feasible and has a good agreement with the diagnosis of the pathologists. (E-mail: rfchang@cs.ccu.edu.tw)