Computer Aid for the Decision to Biopsy Breast Lesions

Abstract

A case based reasoning (CBR) computer algorithm is under development as a computer aid for the decision to biopsy a breast lesion that is suspected to be cancer. The CBR stores the results of many previous cases along with the results of their biopsies. New cases are evaluated by searching these previous cases and selecting those that are similar to the new cases. A prediction of the results of biopsy is based on the results of the similar known cases. In this report, the algorithm has been implemented on computer and a search has been conducted to optimize the set of image features that are used to compare the new images to the known images. An accuracy of better than 83% is reported and the best performing set of features is described.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2000
Accession Number
ADA388047

Entities

People

  • Carey Floyd

Organizations

  • Duke University

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Artificial Intelligence
  • Biomedical Research
  • Breast Cancer
  • Cancer
  • Computers
  • Databases
  • Diagnostic Imaging
  • Health Care
  • Health Services
  • Information Science
  • Information Systems
  • Materials
  • Medical Personnel
  • Neural Networks
  • Operating Systems

Fields of Study

  • Computer science
  • Medicine

Readers

  • Molecular and genetic basis of cancer.
  • Operations Research
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.