Investigation of Three-Group Classifiers to Fully Automate Detection and Classification of Breast Lesions in an Intelligent CAD Mammography Workstation

Abstract

The investigators' goal is to develop a fully automated classification scheme for computer-aided diagnosis (CAD) in mammography. Traditional CAD classification schemes and performance measurement tools, such as receiver operating characteristic (ROC) analysis, are based on the premise that the observations are classified into two groups, most commonly malignant and benign. Such classification schemes are difficult to fully automate, as they analyze radiologist-identified lesions. The difficulty is that many false-positive (FP) detections produced by a computerized detection scheme cannot reasonably be classified as benign or malignant lesions. The proposed scheme would classify computer detections into three groups: malignant lesions, benign lesions, and false-positive computer detections (non-lesions). During the past year, the authors have collected data on 134 mammography cases with clustered microcalcification lesions. They have shown that three decision boundary lines used by the three-group ideal observer are intricately related to one another. They also have analyzed several recently proposed three-group classification methods in terms of the three-group ideal observer. Finally, they have developed principled theoretical motivations for various proposed three-group classification methods, given the selections of restricted or simplified three-group evaluation methods. A three-group classifier could potentially allow radiologists to detect more malignant breast lesions without increasing their false-positive biopsy rates.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 2006
Accession Number
ADA456989

Entities

People

  • Darrin C. Edwards

Organizations

  • University of Chicago

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Biomedical Research
  • Boundaries
  • Breast Cancer
  • Classification
  • Computer-Aided Diagnosis
  • Computers
  • Databases
  • Detection
  • Diagnostic Imaging
  • Electronic Mail
  • Image Processing
  • Machine Learning
  • Medical Personnel
  • Observers
  • Physicians
  • Probability
  • Probability Density Functions

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Oncology and Biomarker-Based Cancer Detection.