Application of pattern recognition techniques to breast cancer detection: ultrasonic analysis of 100 pathologically confirmed tissue areas

Ultrasound texture-based CAD system for detecting neuromuscular diseases

PURPOSE

Diagnosis of neuromuscular diseases in ultrasonography is a challenging task since experts are often unable to discriminate between healthy and pathological cases. A computer-aided diagnosis (CAD) system for skeletal muscle ultrasonography was developed and tested for myositis detection in ultrasound images of biceps brachii.

METHODS

Several types of features were extracted from rectangular and polygonal image regions-of-interest (ROIs), including first-order statistics, wavelet-based features, and Haralick's features. Features were chosen that are sensitive to the change in contrast and structure for pathological ultrasound images of neuromuscular diseases. The number of features was reduced by applying different sequential feature selection strategies followed by a supervised principal component analysis. For classification, two linear approaches were investigated: Fisher's classifier and the linear support vector machine (SVM) as well as the nonlinear [Formula: see text]-nearest neighbor approach. The CAD system was benchmarked on datasets of 18 subjects, seven of which were healthy, while 11 were affected by myositis. Three expert radiologists provided pre-classification and testing interpretations.

RESULTS

Leave-one-out cross-validation on the training data revealed that the linear SVM was best suited for discriminating healthy and pathological muscle tissue, achieving 85/87 % accuracy, 90 % sensitivity, and 83/85 % specificity, depending on the radiologist.

CONCLUSION

A muscle ultrasonography CAD system was developed, allowing a classification of an ultrasound image by one-click positioning of rectangular ROIs with minimal user effort. The applicability of the system was demonstrated with the challenging example of myositis detection, showing highly accurate results that were robust to imprecise user input.

Breast tissue classification using diagnostic ultrasound and pattern recognition techniques: II. Experimental results

The methods of statistical pattern recognition have been applied to the problem of in vivo ultrasonic characterization of breast disease in humans. Backscattered A-mode signals obtained from a commercial pulse imaging system were used to generate a large set of potentially useful features. Using statistical tests, a small subset of discriminatory features was selected to design a Bayes decision rule for each of two tissue classification schemes: malignant disease vs. benign disease, and malignant disease vs. (benign disease + normal tissue). Classification results obtained by the rotation method included sensitivities of 88 percent and 76 percent for the two schemes, based on data obtained from 32 women. These results are encouraging, though a definitive statement concerning the extrapolation of these numbers to the general population should only be made after obtaining results with a large data base.

Breast tissue classification using diagnostic ultrasound and pattern recognition techniques: I. Methods of pattern recognition

This paper discusses the application of statistical pattern recognition techniques to problems in diagnostic ultrasound. Using our own system as an example, we describe the concepts and specific methods that we have applied to a problem involving the computer-aided classification of breast tissue in vivo. Topics include feature generation, feature selection and classification, as well as a method which estimates the probability of error on classifying future data. An accompanying paper applies these methods to the classification of backscattered RF signals from normal and diseased breast tissue.