Purpose: The authors assessed the characteristics of benign and malignant solid breast tumors in harmonic three-dimensional (3D) power Doppler imaging and proposed decision models to classify benign and malignant breast tumors.
Materials and methods: A total of 86 malignant and 97 benign harmonic 3D power Doppler US images were analyzed. All the harmonic 3D power Doppler images were obtained using a Voluson730 US system (GE, Zipf, Austria) equipped with a RSP 6-12 transducer and tissue harmonic imaging modalities. Imaging analysis was performed using the Virtual Organ Computer-aided Analysis (VOCAL)-imaging program. Histogram indices, the vascularization index (VI), flow index (FI) and vascularization-flow index (VFI), were calculated for the intra-tumor and for shells with an outside thickness of 3mm surrounding the breast tumors. The receiver operating characteristic (ROC) curves were calculated to estimate the diagnostic performances.
Results: The results revealed that the choice of decision model comprised the parameters of patient age, intra-tumor VI, and tumor volume to classify benign and malignant breast tumors. The area under the ROC curve (Az) was 0.910, accuracy was 81.4%, and sensitivity and specificity were 81.4% and 81.4%, respectively. The parameter intra-tumor VI was the choice for all of the histogram indices in differentiating between malignant and benign lesions.
Conclusion: The decision model, which was composed of patient age, tumor volume and intra-tumor VI, and a cut-off value for intra-tumor VI at the upper end of patient age and tumor volume, was recommended in clinical application.