Program Information
Unified Database for Rejected Image Analysis Across Multiple Vendors in Radiography
K Little*, L Liu , K Haas , A Sanchez , I Reiser , Z Lu , The University of Chicago, Chicago, IL
Presentations
SA-B-BRD-11 (Saturday, March 5, 2016) 10:30 AM - 12:30 PM Room: Grand Ballroom D
Purpose: Rejected image analysis is an important part of a QA program to minimize patient radiation dose and maintain quality in radiography. A clinical reject is an x-ray acquisition of a patient that is discarded without being reviewed by a radiologist. However, variations in acquisition reporting by different equipment vendors make reject analysis difficult. Our adult imaging areas use six DR and CR models from three vendors. A centralized reject database and reporting dashboard were developed to allow for consistent clinical reject analysis across all radiographic equipment at our facility.
Methods: Acquisition reports were retrieved monthly for each unit and imported into an SQL database. The database was interfaced with RIS to associate each record with the performing technologist. We developed counting rules for each model to ensure that multiple records for a single acquisition (as in the case of dual-energy acquisitions) were not duplicated. The various vendor-provided clinical reject reasons were mapped into one of five categories: incorrect technique, positioning/collimation, artifact/obstruction, patient motion, and other. Each vendor’s varying anatomy descriptors were mapped to a unified terminology.
Results: Data were analyzed based on reject reason, anatomy/view, clinical area, equipment, and staff. Interventions were developed to target the most frequent reject reason and the procedure with the highest rejects. Reject rates for each technologist were reported to the area manager to identify the need for individualized training. Initial clinical reject rates were found to be higher than expected based on available literature and generally higher for DR than for CR.
Conclusion: The monitoring of rejected images in radiography can be a complicated task when multiple manufacturers and models are used in the same department. A centralized database, repeat reason categorization, and anatomy/view mapping allow clinical reject data to be consistently analyzed and used for process improvement.
Funding Support, Disclosures, and Conflict of Interest: This work was supported in part by the RSNA/AAPM Imaging Physics Residency Program Grant.
Contact Email: