Computer-Aided Interval Change Analysis of Microcalcifications on Mammograms for Breast Cancer Detection

Abstract

The goal of this project is to develop a computer-aided diagnosis (CAD) system for automatic interval change analysis of microcalcification clusters on mammograms. Based on our regional registration method a local area on the prior that may contain the corresponding cluster is determined. A search program is used to detect cluster candidates within the local area. The cluster on the current image is then paired with the candidates to form true (TP-TP) or false (TP-FP) pairs. A correspondence classifier (CC) using automatically extracted features is designed to reduce the false pairs. A temporal classifier (TC) based on current and prior information is used if a cluster is detected in the prior, and a current classifier (CurC) based on current information alone is used if no prior cluster is detected. 175 temporal pairs of mammograms were used for evaluation. The search program detected 90.2% of the clusters on the priors with an average of 0.43 FPs/image. The CC identified 85% (l49/175) of the TP-TP pairs with 15 false matches within the 164 image pairs that had detected clusters. The TC achieved a test Az of 0.83 for the 164 pairs for classifying the clusters as malignant or benign. For the 11 clusters without detection on the prior, the test Az by the CurC was 0.72. The radiologist achieved an Az of 0.72 for both the 175 and the 164 temporal pairs.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2005
Accession Number
ADA443710

Entities

People

  • Lubomir M. Hadjiiski

Organizations

  • University of Michigan

Tags

Communities of Interest

  • Engineered Resilient Systems

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Breast Cancer
  • Computer Vision
  • Computers
  • Data Sets
  • Databases
  • Detection
  • Electronic Mail
  • Feature Extraction
  • Feature Selection
  • Identification
  • Machine Learning
  • Neoplasms
  • Neural Networks
  • North America
  • Pattern Recognition

Fields of Study

  • Physics

Readers

  • Oncology and Biomarker-Based Cancer Detection.
  • Quantum Chemistry