Comparison of Iterative and Filtered Back Projection Image Reconstruction Techniques for the Evaluation of Heavily Calcified Vessels with Coronary CT Angiography

Purpose
Heavy calcifications remain a limitation of coronary CT angiography (cCTA) and decrease the accuracy of this test for stenosis detection. We prospectively compared traditional filtered back projection (FBP) and iterative image reconstruction techniques for the cCTA evaluation of heavily calcified vessels.

Method and materials
Dual-source cCTA studies of 25 patients (16 men; 60±7 years) with Agatston scores of >400 were reconstructed with traditional FBP and with iterative reconstruction in image space (IRIS) techniques using corresponding vascular algorithms. Image noise was measured as the HU standard deviation in a region of interest in the aortic root. To assess the effect of the two reconstruction techniques on “blooming” artifacts, the volume of heavy coronary artery calcifications was measured using a threshold based volumetry tool. Two experienced, blinded observers subjectively rated image quality on a 5-point scale and interpreted both, FBP and IRIS series for significant (>50%) stenosis using the AHA 17-segment model. Results were compared with coronary catheter angiography in all patients. Analysis used  Χ2 statistics for categorical variables and Wilcoxon test for continuous variables.

Results
Image quality of IRIS studies was rated significantly (p=0.039) higher than that of FBP reconstructions. Mean image noise was significantly lower with IRIS (27.4±13.5) than with FBP (35.7±14.2, p=0.0001). Heavy calcifications measured significantly lower in studies that were reconstructed with IRIS compared to those that were reconstructed with FBP (34.1±25.0mm3 versus 39.0±25.9mm3, p=0.04). There was an incremental improvement in the detection of significant stenosis with IRIS (accuracy 96%, sensitivity 90%, specificity 96%, NPV 98%, PPV 76%) over FBP (accuracy 93%, sensitivity 85%, specificity 95%, NPV 97%, PPV 75%) when compared with catheter angiography.

Conclusion
Use of IRIS instead of FBP increases the accuracy of stenosis detection with cCTA in heavily calcified vessels. This is likely explained by decreased image noise and “blooming” artifacts from heavy calcifications, as indicated by the lower calcium volume measurements with IRIS, enabling more accurate delineation of calcified vessel segments.

Clinical relevance/application

Iterative image reconstruction techniques appear superior to FBP for the evaluation of heavily calcified vessels with cCTA and should be preferentially used in patients with high calcium scores.

Submission Type: Scientific Presentations RSNA 2010

Authors: M Renker, BSC, Charleston, SC; U J Schoepf, MD; M Weininger, MD; M Kerl, Resident; R Bauer;
T Henzler, MD; et al.