Evaluation of automated detection system of three-dimensio nal ultrasonography in differentiation of breast tumors / 中华超声影像学杂志

ABSTRACT

Objective To evaluated a new automate ultra sonographic system for detecting benign and malignant breast tumors in three-dimensions.Methods A magnetic sensor to track the probe in a real time during scanning and a series of sequential cross-sectional images,together with information giving precise location and the direction of movement of the probe,were fed into the computer.Fuzzy logic was used to classify tumor, boundary, and normal breast tissue in order to extract an image of the tumor and its margin through three-dimensional (3D) transformation. Results The difference between the surface roughness of benign and malignant tumors was found by comparing their computer -rendered surface images.The values of the S/V ratio of surface area (S 3) divided by volume (V 2) of malignant tumors were larger than those of benign tumors. Mean ratio was 3.43 in benign tumors, 9.08 in malignant tumors.T-test results showed these differences between benign and malignant tumors to be significant(P