Computerized Analysis and Detection of Missed Cancer in Screening Mammogram

Abstract

This project is to explore an innovative CAD strategy for improving early detection of breast cancer in screening mammograms by focusing on computerized analysis and detection of cancers missed by radiologists. The research scope in past year is on database generation and analysis of missed cancers. Several major progresses have been made including (1) By reviewing more than 1334 cases, a total of 83 missed cancer cases were collected which were used to generate three different datasets including mammograms with missed cancer, mammograms with screening-detected cancer and normal mammograms. (2) Regions-of-interest (ROIs) containing a detected or a missed cancer were extracted, and a ground truth was generated by an experienced radiologist for feature extraction and analysis purpose. (3) With the datasets and the ground truth, a variety of computerized features were extracted and analyzed to explore the difference of detected and missed cancer cases. A set of tests was applied to the extracted features individually from which the significant features distinguishing the missed cancer from detected ones could be identified and applied potentially to the CAD design in next steps.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 2004
Accession Number
ADA428194

Entities

People

  • Lihua Li

Organizations

  • H. Lee Moffitt Cancer Center & Research Institute

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Abstracts
  • Biomedical Research
  • Breast Cancer
  • Classification
  • Computer Vision
  • Data Analysis
  • Data Science
  • Databases
  • Detection
  • Extraction
  • Feature Extraction
  • Information Science
  • Medical Personnel
  • Normality
  • Physicians
  • Statistical Analysis
  • Statistical Tests

Readers

  • Oncology and Biomarker-Based Cancer Detection.

Technology Areas

  • AI & ML