ABSTRACT: To develop and evaluate a computer-aided diagnosis (CAD) system with automatic contouring and morphological analysis to aid in the classification of breast tumors using ultrasound.
We evaluated 118 breast lesions (34 malignant and 84 benign tumors). Each tumor contour was automatically extracted from the digitized ultrasound image. Nineteen practical morphological features from the extracted contour were calculated and principal component analysis (PCA) was applied to find independent features. A support vector machine (SVM) classifier utilized the selected principal vectors to identify the breast tumor as benign or malignant. In this study, all the cases were sampled with k-fold cross-validation (k = 10) to evaluate the performance by receiver-operating characteristics (ROC) curve analysis.
The areas under the ROC curves for the proposed CAD systems using all morphological features and the lower-dimensional principal vector were 0.91 and 0.90, respectively. The classification ability for breast tumors using morphological information was good.
This system differentiates benign from malignant breast tumors well and therefore provides a clinically useful second opinion. Moreover, the morphological features are nearly setting-independent and thus available to various ultrasound machines.
Ultrasound in Obstetrics and Gynecology 04/2008; 32(4):565-72. · 3.01 Impact Factor
ABSTRACT: The appearance of cluster of microcalcifications in mammography or sonography is an important indicator for malignancy. Microcalcifications are calcium deposits, which can be identified as tiny areas that are slightly brighter than surrounding tissue. Detection of mammographic microcalcification has been proposed in many studies. Since a microcalcification cluster is a three-dimensional (3-D) entity, its projection onto a two-dimensional (2-D) image results in a loss of spatial information and may also cause superimposition of individual calcifications within the cluster. This paper aims to use the 3-D ultrasound to determine microcalcifications. In each slice, the proposed method adopts the top-hat filter to find bright spots, and employs four 2-D criteria to select the spots as candidate microcalcifications. Finally, spots appearing in sequent slices at the same position are considered as a microcalcification. We suggest using a computer automatically to detect the microcalcification being feasible and microcalcifications being a very important criterion of malignancy on future developing the computer-aided diagnosis for ultrasound. In the future, this technique can be adopted in a computer-aided diagnosis system combined with other diagnosis features for improving the diagnosis performance
Engineering in Medicine and Biology Society, 2005. IEEE-EMBS 2005. 27th Annual International Conference of the; 02/2006
ABSTRACT: We present a computer-aided diagnostic (CAD) system with textural features and image retrieval strategies for classifying benign and malignant breast tumors on various ultrasonic systems. Effective applications of CAD have used different types of texture analysis. Nevertheless, most approaches performed in a specific ultrasonic machine do not indicate whether the technique functions satisfactorily for other ultrasonic systems. This study evaluated a series of pathologically proven breast tumors using various ultrasonic systems.
Altogether, 600 ultrasound images of solid breast nodules comprising 230 malignant and 370 benign tumors were investigated. All ultrasound images were acquired from four diverse ultrasonic systems. The suspicious tumor area in the ultrasound image was manually chosen as the region-of-interest (ROI) subimage. Textural features extracted from the ROI subimage are supported in classifying the breast tumor as benign or malignant. However, the textural feature always behaves as a high-dimensional vector. In practice, high-dimensional vectors are unsatisfactory at differentiating breast tumors. This study applied the principal component analysis (PCA) to project the original textural features into a lower dimensional principal vector that summarized the original textural information. The image retrieval techniques were employed to differentiate breast tumors, according to the similarities of the principal vectors. The query ROI subimages were identified as malignant or benign tumors according to characteristics of retrieved images from the ultrasound image database.
Using the proposed CAD system, historical cases could be directly added into the database without a retraining program. The area under the receiver-operating characteristics curve for the system was 0.970+/-0.006.
The CAD system identified solid breast nodules with comparatively high accuracy in the different ultrasound systems investigated.
Ultrasound in Obstetrics and Gynecology 11/2005; 26(5):558-66. · 3.01 Impact Factor