Click here to


Are you sure ?

Yes, do it No, cancel

A Method for Automated Repeat/reject Rate Analysis in CT

S Rose*, B Viggiano , T Szczykutowicz , University Wisconsin-Madison, Madison, WI


(Thursday, 7/18/2019) 7:30 AM - 9:30 AM

Room: 221AB

Purpose: Repeat/reject rate analysis is federally mandated in mammography and recommended by multiple organizations in radiography. Here we extend repeat/reject rate analysis to CT and evaluate both the rate and excess dose of repeat examinations.

Methods: An algorithm was developed using dose monitoring data to identify repeats due to motion, technologist mistakes, etc. The algorithm does not require manual specification of protocols. A pattern recognition method first identifies a site’s gold standard exams. Repeats are identified via comparison of individual exams to the gold standards. The algorithm estimated repeat/reject rates from high-volume protocols at CT scanners from two sites using data collected over 3 years. We only considered repeats consisting of additional overlapping helical/axial scans. We calculated effective doses from all exams for the ten highest repeat-rate protocols.

Results: Repeat rates were 1.4% at both sites with 95% confidence intervals of [1.2,1.6] and [1.3,1.5]. Among the ten highest repeat rate protocols, the median percent increase in mean effective dose between normal and repeat-containing exams was 107.5% (interquartile range [89.9,130.2]) for site A and 64.6% (interquartile range [44.4,88.8]) for site B. Four of the ten highest repeat rate protocols at site B (zero at site A) have more than one helical/axial scan, lessening a repeat’s relative dose contribution. The ten highest repeat rate protocols differed between sites. This is partially attributable to the emergency department at site B, where physically/mentally compromised patients increase motion repeats for certain protocols. Increased repeat rates for pulmonary embolism protocols at site A have prompted further investigation.

Conclusion: We were able to successfully measure repeat rates for CT. Historically, this has been challenging due to the complex nature of CT ordering sets, which frequently involve multiple irradiation events and fulfilling multiple orders in a single exam. Monitoring repeats could motivate intervention for individual technologists and protocols.

Funding Support, Disclosures, and Conflict of Interest: TPS supplies CT protocols to GE Healthcare under a licensing agreement. TPS is also a consultant for GE and iMALOGIX and is on the CAB of iMALOGIX. He is also the founder of


Quality Control, CT, Dose


IM/TH- Formal quality management tools: Failure modes and effects analysis

Contact Email