The effect of iterative image reconstruction algorithms on the feasibility of automated plaque assessment in coronary CT angiography

To evaluate the effect of adaptive statistical (ASIR) and model based (MBIR) iterative reconstruction algorithms on the feasibility of automated plaque assessment in coronary computed tomography angiography (CCTA) compared to filtered back projection reconstruction (FBPR) algorithm. Three ex vivo human donor hearts were imaged by CCTA and reconstructed with FBPR, ASIR and MBIR. Commercial plaque assessment software was applied for the automated delineation of the outer and inner vessel-wall boundaries. Manually corrections were performed where necessary and the percentages were compared between the reconstruction algorithms. In total 2,295 CCTA cross-sections with 0.5 mm increments were assessed (765 co-registered FBPR/ASIR/MBIR triplets). Any boundary corrections were performed in 31.0 % of all cross-sections (N = 712). The percentage of corrected crosssections was lower for MBIR (24.1 %) as compared to ASIR (32.4 %, p = 0.0003) and FBPR (36.6 %, p <0.0001), and marginal between ASIR/FBPR (p = 0.09). The benefit of MBIR over FBPR was associated with the presence of moderate and severe calcification (OR 2.9 and 5.7, p <0.0001; respectively). Using MBIR significantly reduced the need for vessel-wall boundary corrections compared to other reconstruction algorithms, particular at the site of calcifications. Thus, MBIR may improve the feasibility of automated plaque assessment in CCTA and potentially its clinical applicability.