David Knipp MD, Abstract Co-Author: Nothing to Disclose

Barton Frederick Lane MD, Abstract Co-Author: Nothing to Disclose

Seth Jay Kligerman MD, Abstract Co-Author: Author, Reed Elsevier

Amar Dhanantwari, Abstract Co-Author: Employee, Koninklijke Philips NV

Barry David Daly MD, Presenter: Research Grant, Koninklijke Philips NV

To assess quantitative and qualitative measures of image quality from cadaver data with progressively reduced-dose abdominopelvic CT scans using a knowledge-based iterative reconstruction algorithm compared to standard dose filtered back projection (FBP).

Two human cadavers were scanned at 120kVp on a 256-slice CT scanner (Philips Medical, Cleveland) at standard dose (300 mAs) and reconstructed with FBP. Subsequent reduced dose scans were performed at 60%, 70%, 80%, 90%, and 95% dose reduction (DR), (as low as 15mAs) and reconstructed with iterative model reconstruction (IMR) technology, (Philips Medical) at both 3mm and 1mm slice thickness. Sample images with focal pathology in the liver, mid abdomen and pelvis were selected for review. Three experienced radiologists graded scans for image quality (IQ), perceived noise, and presence of artifacts using a 1-5 Likert scale. All scans were compared directly to the standard dose FBP scan for reference. ROIs were placed in the liver, mid abdomen and pelvic tissues wall to calculate noise, attenuation, and contrast to noise ratio (CNR).

At full dose FBP, average noise (21.2) and CNR (2.3), was not significantly different than noise at 90% and CNR at 95% DR respectively (p>0.05 for all). There was no significant difference in noise or CNR between 3mm and 1mm slice thickness for IMR. At 60%, 70%, and 80% DR had equal aggregate IQ compared to standard dose FBP (p>0.05 for all). Perceived noise was improved with IMR compared to FBP for all levels of DR up to 80% (p<0.01 for all) and equal to FBP for 90-95% DR. Artifacts were not different between FBP and IMR at up to 70% DR. For IMR, there was no significant difference in scoring of IQ, perceived noise or artifact presence between 3mm and 1mm slice thickness.

In this cadaveric study, quantitative data and subjective reader evaluations suggest that IMR can allow 80% dose reduction compared to standard dose FBP in abdominopelvic CT without loss of image quality. No noise penalty was seen with thinner image slice reconstruction.

The introduction of IMR into clinical use should allow major reductions in radiation dose for abdominopelvic CT while maintaining diagnostic image quality.

Knipp, D, Lane, B, Kligerman, S, Dhanantwari, A, Daly, B, Investigation of Iterative Model Reconstruction to Determine Maximum Obtainable CT Radiation Dose Reduction with Preserved Diagnostic Quality in a Cadaver Study.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14007481.html