Sarabjeet Singh MD, Presenter: Nothing to Disclose

Mannudeep K. S. Kalra MD, Abstract Co-Author: Consultant, General Electric Company Faculty, General Electric Company

Synho Do PhD, Abstract Co-Author: Nothing to Disclose

Owen J. O'Connor, Abstract Co-Author: Nothing to Disclose

Michael Austin Blake MBBCh, Abstract Co-Author: Editor with royalties, Springer Science+Business Media Deutschland GmbH

To evaluate standard and low dose abdominal CT images reconstructed with filtered back projection (FBP), hybrid (hIRT) and pure (pIRT) iterative reconstruction techniques.

In an IRB approved prospective study, 10 patients (mean age 59.9 ± 7.8 years, M: F 4:6) gave written informed consent for acquisition of additional image series through abdomen on 64 slice MDCT. Images were acquired at 200 and then at 50 mAs with all remaining scan parameters held constant, including, 120 kVp, 0.984:1 pitch and standard reconstruction kernel. Scan raw data were exported and reconstructed to generate hIRT, pIRT and FBP images. Six image datasets (FBP, hIRT, pIRT at standard and low doses) were displayed on DICOM compliant 55 inch 2 MP monitor for blinded randomized evaluation by two experienced abdominal radiologists for lesion detection, image noise, artifacts, visibility of small structures as well as diagnostic confidence. Objective noise and HU values were recorded at two different anatomical locations (liver and descending aorta). Noise power spectrum was assessed for FBP, hIRT and pIRT images.

Lesion conspicuity was improved from poorly visualized margins in FBP and hIRT images to well defined margins on pIRT at low dose. Both radiologists found suboptimal noise on low dose FBP and hIRT images, whereas noise was acceptable with pIRT. Except for minor pixilated appearance of images, no significant artifacts were seen. Regardless of the dose level and reconstruction techniques, there was no difference in CT numbers between FBP, hIRT and pIRT images (p=0.87-0.99). pIRT images had lower noise compared to both hIRT (43% (8.9/15.9) at 200 mAs and 61% (12.4/32.1) at 50 mAs) and FBP (58% (8.9/21.4) at 200 mAs and 70% (12.4/42.4) at 50 mAs) (p<0.0001). Noise power spectrum analyses showed hIRT retains the noise spectral signature as FBP, in spite of lowering the noise, whereas pIRT further lowers the noise and has more regularized noise spectral pattern.

Low radiation dose abdominal CT examinations when reconstructed with pIRT are associated with lower image noise as compared to hIRT and FBP, without any significant image artifacts affecting diagnostic interpretation

Ultra low dose abdominal CT scanning at 50 mAs may be possible with pIRT assisted image reconstruction

Singh, S, Kalra, M, Do, S, O'Connor, O, Blake, M, Assessment of Hybrid and Pure Iterative Reconstruction with Filtered Back Projection Technique for Low Dose Abdominal CT.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11013585.html