Room: AAPM ePoster Library
Purpose: calcium score (CCS) is based on image acquisitions at 120 kVp. The enhanced calcium at lower kVps makes further dose reduction possible but requires adjustments of multiple thresholds for different Houndsfield unit (HU) ranges. We propose a new approach to achieve correct scores at lower kVps without changing the thresholds.
Methods: were first acquired in axial mode at various kVps (120–70) on a Siemens Force CT from a stationary heart phantom (CIRS) with four 3-cm hydroxyapatite inserts (200 and 800 mg/cc). The ratios of HU at lower kVps to that at 120 kVp were obtained. The phantom was then embedded with 1.2- 5 mm HA inserts (50 –400 mg/cc) on the coronary track and the heart module was set in 3D motion with the ECG of 60 bpm. Axial image acquisitions were made at 120 kVp and various CTDIvol (3 –0.78 mGy). To simulate the calcium enhancement at lower kVps, all calcium pixels were selectively scaled up with the HU ratios obtained from the first step. To apply the CCS thresholds of 120 kVp, a new approach was applied to scale down all pixels in the image to maintain the contrast-to-noise ratio. All calcifications were then scored using the SyngoVia CCS software.
Results: compared to the ground-truth CCS results from the original images acquired at 120 kVp and 3 mGy, the scores obtained from the simulated images for lower kVps using the new approach were consistent (error < 1%), except for the very low dose (0.78 mGy). For images acquired at 120 kVp and 0.78 mGy where the noise generated significant false calcifications, the new approach resulted in better scores at lower kVps.
Conclusion: The CCS using the new approach to the simulated images at lower kVps was verified to be consistent with the ground-truth results at 120 kVp.