Related Experiment Videos
Computer-aided diagnosis in chest radiography. Preliminary experience
K Abe1, K Doi, H MacMahon
1Department of Radiology, Iwate Medical College, Japan.
Investigative Radiology
|November 1, 1993
Summary
Computer-aided diagnosis (CAD) shows promise for improving chest radiograph accuracy by detecting lung nodules and other abnormalities. However, high false-positive rates and technical issues need resolution for clinical integration.
Area of Science:
- Radiology
- Medical Imaging
- Artificial Intelligence in Medicine
Background:
- Computer-aided diagnosis (CAD) systems aim to enhance radiologist accuracy in detecting abnormalities on chest radiographs.
- Current CAD schemes require validation on large clinical datasets to assess their real-world utility.
- Intelligent workstations necessitate defined design parameters for effective CAD integration.
Purpose of the Study:
- To evaluate the potential usefulness and limitations of CAD programs for detecting lung nodules, cardiomegaly, and interstitial infiltrates.
- To identify critical design parameters for developing an effective intelligent workstation for radiological diagnosis.
Main Methods:
- Applied CAD programs to 310 consecutive chest radiographs for automated detection of specific abnormalities.
- Evaluated CAD output for accuracy by radiologists and for technical challenges by physicists.
Main Results:
- Approximately 70% of CAD results were deemed potentially acceptable for clinical use.
- A significant number of false-positive findings were observed.
- Technical limitations included missed subtle abnormalities and false detections from normal anatomy.
Conclusions:
- Computer-aided diagnosis holds potential as a valuable tool for radiologists in clinical practice.
- Overcoming technical challenges, such as false positives and missed findings, is crucial.
- Defining optimal operating points for CAD systems is necessary for successful clinical implementation.