Related Experiment Videos
Analysis of clustered microcalcifications by using a single numeric classifier extracted from mammographic digital
S S Buchbinder1, I S Leichter, P N Bamberger
1Department of Radiology, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, NY 10461, USA.
Academic Radiology
|November 11, 1998
Summary
A new computer system accurately analyzes mammogram microcalcifications. This pattern recognition classifier uses quantitative features to distinguish benign from malignant lesions, aiding in diagnosis.
Area of Science:
- Medical imaging analysis
- Computer-aided diagnosis
- Breast cancer screening
Background:
- Clustered microcalcifications are key indicators in mammography.
- Accurate interpretation is crucial for early breast cancer detection.
- Automated analysis can potentially improve diagnostic accuracy.
Purpose of the Study:
- To evaluate a novel numeric classifier for clustered microcalcifications.
- To assess the performance of computer-extracted quantitative features.
- To determine the diagnostic value of a pattern recognition system.
Main Methods:
- Digitization of mammograms with clustered microcalcifications.
- Automatic extraction and quantitative analysis of microcalcification features.
- Development of a discriminant analysis classifier trained on reference cases and tested on prospective cases.
Main Results:
- Thirty-seven parameters showed significant differences between benign and malignant lesions.
- Seven factors were selected for the final classifier.
- Receiver operating characteristic analysis yielded an area under the curve of 0.88 for prospective cases.
Conclusions:
- The pattern recognition classifier demonstrates satisfactory performance.
- Quantitative feature analysis of microcalcifications shows promise.
- The developed software may assist in mammographic interpretation.