Background Iterative reconstruction (IR) is a novel but significant development in CT image acquisition. There have been a number of studies that have reported on the potential of IR in cardiac CT. These retrospectively applied IR in the image domain to images acquired with standard filtered back projection (FBP) techniques. This study was part of an ongoing randomised control trial [ISRCTN52480460] evaluating the cost effectiveness of cardiac CT.

Methods 250 patients were prospectively enrolled to have a cardiac CT for the investigation of stable chest pain. Written and informed consent was obtained. Data acquisition were performed on a Philips Brilliance 64. The patients were divided into two groups. Cohort A underwent standard FBP imaging, and Cohort B underwent IR with Idose® (Philips, Cleveland, Ohio, USA). Within each cohort the scan parameters (kv, mAs, pitch) and reconstruction protocols (prospective or retrospective) were determined by patient characteristics. Images were assessed for noise and signal quality within regions of interest (ROI) on axial images, and subjectively for image quality by two experienced readers. Noise was defined as the SD of the measured HU, and signal as the HU mean attenuation value. The ROIs were in the ascending aorta, interventricular septum and left ventricular cavity. Subjective image quality was rated blindly using a 5-point Likert scale. Effective radiation dose (ED) of each CTCA was estimated by multiplying the dose-length product by a chest-specific conversion coefficient (κ=0.014 mSv×mGy⁻¹×cm⁻¹).

Results Of the 250 patients enrolled 3 withdrew. 146 of the 247 subjects were male with a mean age of 57.93 (SD 9.93). Cohort A consisted of 124 patients, and cohort B 123, with no significant difference in baseline demographics. The mean dose of all FBP was 6.09 mSv, (SD 3.16) compared to an IR mean of 4.23 mSv, (SD 2.01) which was a dose saving of 1.86 mSv (30.54%). This was a significant dose reduction (p value <0.0001.) Mean image quality score obtained from the IR images was 3.67 (SD 1.04) compared to the FBP images of 3.29 (SD 1.17) p value of 0.0067. There was good agreement between the readers—κ coefficient 0.83. Cohort A consisted of 74 retrospective images and 50 prospective. Cohort B had 116 with retrospective and 7 with prospective. The mean ED for a prospective FBP was 3.50 mSv (SD 1.15), with the IR equivalent being 2.00 mSv (0.72), giving a mean dose saving of 1.50 mSv (42.86%). The mean ED for FBP retrospective studies was 7.85 mSv (SD 2.87), with the IR equivalent being 4.36 mSv (SD 1.99), with a mean dose saving of 3.49 mSv (44.46%). There was no statistical difference in noise or mean attenuation between the IR and FBP images in all three areas of interest Abstract 096 table 1.

Conclusions To our knowledge this is the first study to prospectively compare FBP with IR. It suggests that cardiac IR protocols confer a substantial radiation dose reduction without a compromise in diagnostic quality.