For advanced breast cancer with severe local disease (ABC) (stage III/IV), neoadjuvant chemotherapy improves local control and surgical outcome. However, about approximately 20 to 30% of advanced cancers show either no or poor response to chemotherapy. To prevent unnecessary treatment, a capability of predicting clinical response to neoadjuvant chemotherapy of ABC is highly desirable. Vascularity index (VI) of breast cancers was derived from the quantification results in 30 ABC patients by using power Doppler sonography. Power Doppler sonography evaluation was performed every one to two weeks during chemotherapy. The overall response rate for 30 advanced patients tested was 70%, when 50% or more reduction in tumor size was the objective clinical response. Chemotherapy response was unrelated to the original tumor size (p = 0.563) or chemotherapy agents used (p = 0.657). The median VI for all 30 patients was 4.99%. The response rates for hypervascular tumors vs. hypovascular tumors, based on initial median value, were 86.7% and 53.3%, respectively (p = 0.109). The average VIs in responders and nonresponders were 7.67 +/- 4.77% and 4.01 +/- 3.82% (p = 0.052). There was a tendency for responders who have a relatively high initial vascularity. The VI change in responder group shows a pattern of initial increasing in vascularity followed by decreasing in vascularity. All patients (17/17) with a VI increment of >5% during chemotherapy had good chemotherapy response, whereas in patients with a VI increment of <5%, the response rate was 30.8% (4/13) (p < 0.001). For patients with a peak VI of >10% during chemotherapy, the response rate was 94.1% (16/17). However, in patients with a peak VI of <10%, the response rate was 38.5% (5/13) (p = 0.001). This prediction was made mostly within one month (25.47 +/- 12.96 d for VI increments >5% and 25.44 +/- 12.41 d for VI increased to >10%). In the meantime, the differences in size reduction shown in B-mode sonography were insignificant between responders and nonresponders (patient group with VI increment >5%, p = 0.308; patient group with peak VI >10%, p = 0.396). In conclusion, we propose that VI as determined by using power Doppler sonography is a good and inexpensive clinical tool for monitoring vascularity changes during neoadjuvant chemotherapy in ABC patients. Two parameters--VI increment >5% and peak VI >10%--are potential early predictors for good responses to neoadjuvant chemotherapy within one month in patients with ABC.