Room: AAPM ePoster Library
Purpose: To quantify image artifacts at ultra-low dose levels for three major CT manufacturers and determine when the detected signal approaches the detector electronic noise floor.
Methods: A 30cm diameter cylindrical water phantom was scanned on two models of CT scanners from each of the three manufacturers at multiple CTDIvol values (4, 2, 1, 0.5mGy & lowest achievable). Scanners included: Canon Aquilion 64; Canon Aquilion ONE Genesis Edition; GE Lightspeed VCT; GE Discovery CT750 HD; Siemens Sensation 64; and Siemens Force. At each CTDIvol value, the scan was performed at 80kV and 120kV and a helical pitch of 1.2, 0.9, 0.6 or the closest pitch value available. Automatic exposure control was turned off. All data were reconstructed with a medium smooth kernel (B30 on Sensation, Br40 on Force, FC18 on Canon scanners, Standard on GE scanners) without IR to reflect the inherent detector properties. Images were visually inspected for the presence of photon starvation artifacts. CT number variation as a function of distance from isocenter was used to quantify artifact severity.
Results: For all manufacturers, higher-end scanners had superior performance at ultra-low-dose levels compared to lower-end ones. The higher-end scanner of Manufacturer A had the best overall performance at ultra-low dose, with visible photon starvation artifacts only becoming evident at 0.25mGy, followed by Manufacturer C and then B both of which had obvious artifacts at a dose level of 0.5mGy. Visual impressions were confirmed by quantitative measurements. For a given CTDIvol value, use of lower helical pitch values allows for use of lower tube current values, and led to more severe photon starvation artifacts due to detection of lower signal amplitudes per detector reading.
Conclusion: At the same ultra-low dose levels, different CT systems demonstrate remarkably different levels of electronic noise artifacts due to underlying variations in detector performance.
Funding Support, Disclosures, and Conflict of Interest: Cynthia H. McCollough receives industry funding from Siemens AG.