Image retake analysis in digital radiography using DICOM header information.

Area of Science:

• Radiology
• Medical Imaging
• Health Informatics

Background:

• Digital radiology departments often lack robust systems for analyzing rejected images.
• Existing Picture Archiving and Communication Systems (PACS) and workstations are not designed for reject analysis.
• The QCOnline system was adapted for image management and dose monitoring.

Purpose of the Study:

• To develop and evaluate a methodology for automatically detecting potential retakes in digital imaging.
• To utilize Digital Imaging and Communications in Medicine (DICOM) header information for this detection.
• To identify performance deficiencies within a digital radiology department.

Main Methods:

• A system (QCOnline) was employed to identify images with identical patient ID, modality, description, projection, date, cassette orientation, and comments.
• DICOM header information was leveraged for automated comparison of imaging parameters.
• Pilot data from abdomen and chest images were collected and analyzed.

Main Results:

• Pilot experience indicated initial repetition rates of 6.6% for abdomen and 1.9% for chest images.
• Further analysis revealed actual repetition rates of 3.3% for abdomen and 0.9% for chest images.
• Incorrect image identification was identified as a primary cause for discrepancies.

Conclusions:

• The automated methodology using DICOM headers is feasible for detecting potential retakes.
• This approach can identify various departmental performance issues, including identification errors, positioning problems, and technical or equipment malfunctions.
• Automatically collected retake data can serve as a valuable resource for continuous staff training.