Room: Exhibit Hall | Forum 9
Purpose: to evaluate the performance of CT scanners and clinical protocols across a large installed base using data collected in annual testing by multiple clinical physics groups in different regions of the country. This analysis supports identifying typical protocol parameters and evaluation of scanner and protocol performance in terms of image quality and dose.
Methods: Protocol parameters, measured contrast-to-noise ratio, and measured CTDIvol were collected from reports of 100 annual CT scanner physics surveys in New York and Ohio for adult head and abdomen and pediatric head and abdomen protocols. CNR and dose were compared across vendors as well as across models for specific vendors. Values were compared to ACR accreditation pass/fail criteria and protocol parameters were compared to AAPM-published reference CT protocols.
Results: Typical clinical protocols produce CNR values markedly higher than ACR minimum values at doses markedly lower than ACR reference doses. Some variability in clinical parameters such as reconstructed slice thickness was noted, deviating from AAPM reference protocols. There was some variability in dose efficiency across vendors and scanner models (CNR vs. CTDIvol). Typical CTDIvol for the adult abdomen protocol agreed with published values from the ACR Dose Index Registry.
Conclusion: ACR accreditation CNR and dose criteria for phantom testing are likely not stringent enough to evaluate scanner performance and assess achievable dose reductions or image quality improvements in annual testing. This review suggests additional criteria that could be used to assess the clinical performance of scanners and protocols, based on actual performance of tests on scanners and protocols that are clinically accepted by radiologists in a geographically diverse installed base of hospitals and free-standing imaging centers.