Multi-Center Assessment of Infiltration Rates in FDG-PET/CT scans: Detection, Incidence, and Contributing Factors

Objectives: 18F-FDG dose infiltration impacts PET/CT image quality and quantification accuracy. However, standardized Quality Control (QC) or Quality Assurance (QA) have not generally been implemented for PET radiopharmaceutical administration. The objective of this project is to determine actual infiltration rates in patients undergoing routine FDG PET/CT scans and to identify potential contributing factors. Methods: Imaging centers at MD Anderson Cancer Center, UCLA, University of Tennessee Medical Center, Wake Radiology Services, Carilion New River Mobile, Wake Forest University, and Carilion Memorial Hospital participated in this multi-center quality improvement (QI) project. An IRB for each center determined that such a QI project did not require formal IRB approval. The quality of FDG administration was monitored using an FDA-listed device, Lara (Lucerno Dynamics, LLC). Technologists were trained to use the device and encouraged to routinely monitor all injections for a period of 2-4 months. Sensors were topically applied proximal to the injection site and contralateral (reference) arm for each patient and Time Activity Curves (TACs) acquired from each sensor. The TACs were uploaded to Lucerno for analysis together with additional information related to the injection technique and patient-specific factors that could influence the injection quality. No protected patient health information was collected. The centers were blinded to their results throughout the project. The TACs were evaluated using a classifier developed for the purpose and correlated with the additional, patient-specific information. For each center the utilization and infiltration rates and site-specific factors associated with infiltrations were determined, although the results for a given center were not shared with the other participating centers. Results: Data for 2,431 FDG injections were collected. The mean and median utilization rate was 91% and 93% (range: 30% - 99%). The mean and median patient enrollment was 347 and 343 (range: 52 - 705). Sensors identified a total of 150 infiltrations (6.2%) and classified 44 (1.8%) as potentially moderate or severe. The mean and median infiltration rate was 7.4% and 3.2% (range: 1.9% - 15.7%). Certified and experienced (range: