Feasibility of TG299 Implementation for Clinical Photon-Counting CT
Abstract
Purpose
In order to initiate a quality control (QC) process for clinical photon-counting computed tomography (PCCT), this work evaluates the applicability of AAPM Task Group Report 299 (TG299) for multi-energy CT (MECT) systems to PCCT, highlighting gaps in and recommendations for implementation.
Methods
Following TG299 methodology, we used a clinical PCCT system (Siemens Healthineers Alpha.Prime) to image an ACR CT phantom and a tissue-equivalent MECT phantom containing iodine and calcium inserts of various concentrations (Sun Nuclear). We utilized various imaging protocols and image reconstructions. We used image viewing software (NilRead) and Siemens’ syngo.via platform to perform standard CT QC tests as well as those specific to MECT: CT number accuracy for virtual monoenergetic images (VMI), virtual non-contrast (VNC) performance, material differentiation quantification accuracy, and differentiation task accuracy. We identified pitfalls in implementation of TG299 methodology.
Results
We compiled a summary of recommended QC tests that includes appropriate scan protocols, reconstructions, and software used for each task. We found most tests were implementable with resulting performance acceptable based on TG299 thresholds. However, material differentiation accuracy and material density quantification accuracy tasks proved unfeasible with the currently available syngo.via software version (VB80F). Attempts at performing material quantification yielded inaccurate results (up to 550% error) for low iodine concentrations. Furthermore, iodine-basis images did not discriminate between iodine and calcium.
Conclusion
While most TG299-recommended QC tests are actionable, it is difficult to accurately perform all recommended tests with common currently available clinical tools. Material quantification and differentiation tasks should be carefully verified using alternative methods prior to clinical utilization.