Information about shape, provided by a breast tumor contour is important to physicians in making diagnostic decisions. To avoid needless biopsy and enhance the diagnostic accuracy, a computer-aided diagnosis (CAD) system can provide a second beneficial support reference. This paper aimed to evaluate the potential role of the CAD with automatic contouring and morphologic analysis in the differential of breast tumors for ultrasound (US) images. This study evaluated 118 breast lesions. The suspicious tumor contour in the digitized US image was automatically extracted by the proposed contouring algorithm. Then 20 practical morphologic features from the extracted contour were calculated and a support vector machine (SVM) classifier identified the breast tumor as benign or malignant. The area (Az) under the receiver operating characteristic (ROC) curve for the proposed CAD system was 0.91 ? 0.03. This system differentiates benign from malignant breast tumors with relative high accuracy and is therefore clinically useful to reduce patients needed for dispensable breast biopsy.
Relation:
Computer-Assisted Radiology and Surgery Volume 2, Issue 1 SUPPL., June 2007, Pages S344-S346
