PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 6683019

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

S Finette·A Bleier·W Swindell
View original source ↗Export metadata ↓

Abstract

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.

Explore related subjects

Breast DiseasesDiagnostic ErrorsFemaleHumansPattern Recognition, AutomatedStatistics as TopicUltrasonographyComparative StudyJournal ArticleResearch Support, U.S. Gov't, P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

S Finette, A Bleier, W Swindell. 1983. Breast tissue classification using diagnostic ultrasound and pattern recognition techniques: I. Methods of pattern recognition.. https://doi.org/10.1177/016173468300500106

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2003

Influence of size and depth on accuracy of electrical impedance scanning.

Cancer cells exhibit altered local dielectric properties which can be assessed using electrical impedance scanning (EIS). The study was aimed at clarifying influence of lesion size and depth on EIS performance. From a series of 387 lesions (129 malignant and 258 benign) from 363 patients being sonographically and/or mammographically evaluated, size and depth information was not available in 112 lesions, size was available in 86 lesions and additional depth information was available in 189 lesions, respectively, while performing EIS. Lesions were either histologically verified or had a follow-up of at least 2 years. One hundred three of 129 malignant lesions and 165 of 258 benign lesions were correctly detected (sensitivity 79.8%, specificity 64.0%, accuracy 71.9%). Sensitivity without knowledge of size and depth was 64.6% (10 of 16 malignant lesions detected). This value increased to 76.2% (32 of 42) with knowledge of the size and further increased to 85.9% with knowledge of size and depth (61 of 71). Specificity values in the three subgroups were almost similar: 64.6 (62 of 96), 65.9 (29 of 44), and 62.7% (74 of 118), respectively. Accuracy rises from 63.6% (without knowledge of size/depth) to 71.1 and 74.3% (with size knowledge and with size and depth knowledge, respectively). Accuracy of EIS improved significantly by including sonographical information about depth and size into the analysis. Ultrasound examination should be performed prior to EIS.

Breast Diseases↗