Jeanette Soon1, Nada Sulaiman1, Jong Kwan Park2, Shaw-Hua Anthony Kueh1, Christopher Naoum3, Darra Murphy1, Jennifer Ellis1, Cameron J Hague1, Philipp Blanke1, Jonathon Leipsic4. 1. Dept of Radiology and Centre for Heart Valve Innovation, St Paul's Hospital, Vancouver, Canada. 2. Division of Cardiology, National Health Insurance Service Ilsan Hospital, Goyang, South Korea. 3. Cardiology Department, Concord Hospital, The University of Sydney, Australia. 4. Dept of Radiology and Centre for Heart Valve Innovation, St Paul's Hospital, Vancouver, Canada. Electronic address: JLeipsic@providencehealth.bc.ca.
Abstract
BACKGROUND: Motion correction (MC) algorithms have been shown to improve image quality, interpretability and diagnostic accuracy in coronary CT angiography. We sought to determine whether MC extended to the whole heart would demonstrate improved image quality and reproducibility of aortic annular measurements in pre-TAVR CT. MATERIALS AND METHODS: Twenty-two consecutive contrast enhanced CT data sets acquired for pre TAVR evaluation using retrospective ECG synchronization during a single heart beat were retrospectively identified. Image data sets were obtained from raw data acquired at 35% and 75% of the R-R interval using both standard (STD) and motion corrected (MC) reconstruction algorithms. Four data sets (2 STD, 2 MC) per patient were analyzed by 2 independent, blinded readers for aortic annular area, short and long axis, perimeter and average diameter. Image quality was graded using a 5 point Likert score (1 and 2 non diagnostic, 5 excellent). Statistical analysis was performed using Wilcoxon matched paired tests, Bland-Altman (B-A) plots and Lin's concordance coefficient comparing 35% STD to 35% MC, and 75% STD to 75% MC. RESULTS: Eighty-eight datasets were analyzed (44 STD, 44 MC). At 35%, there was a significant improvement in image quality for MC (Likert score 3.3 ± 0.9 STD vs. 3.9 ± 0.7 MC, p < 0.007). While B-A analysis demonstrated narrower interobserver agreement for aortic annular area (bias 0.03 vs 0.02 cm(2), range -0.32 to 0.39 cm(2) vs -0.50 to 0.55 cm(2)), and perimeter (bias 0.3 vs 0.3 mm, range -3.1 to 3.8 mm vs -4.6 to 5.3 mm), this was not statistically significant by concordance correlation coefficient. At 75%, there was no significant difference in image quality (Likert score 3.3 ± 0.9 vs. 3.5 ± 0.76, p = 0.454) or annular measurement agreement intervals. CONCLUSION: Motion correction algorithms may yield significant improvements of image quality in systolic CT data sets of the heart. Further validation studies are required to determine the effect on annular measurements and translation into clinical practice.
BACKGROUND: Motion correction (MC) algorithms have been shown to improve image quality, interpretability and diagnostic accuracy in coronary CT angiography. We sought to determine whether MC extended to the whole heart would demonstrate improved image quality and reproducibility of aortic annular measurements in pre-TAVR CT. MATERIALS AND METHODS: Twenty-two consecutive contrast enhanced CT data sets acquired for pre TAVR evaluation using retrospective ECG synchronization during a single heart beat were retrospectively identified. Image data sets were obtained from raw data acquired at 35% and 75% of the R-R interval using both standard (STD) and motion corrected (MC) reconstruction algorithms. Four data sets (2 STD, 2 MC) per patient were analyzed by 2 independent, blinded readers for aortic annular area, short and long axis, perimeter and average diameter. Image quality was graded using a 5 point Likert score (1 and 2 non diagnostic, 5 excellent). Statistical analysis was performed using Wilcoxon matched paired tests, Bland-Altman (B-A) plots and Lin's concordance coefficient comparing 35% STD to 35% MC, and 75% STD to 75% MC. RESULTS: Eighty-eight datasets were analyzed (44 STD, 44 MC). At 35%, there was a significant improvement in image quality for MC (Likert score 3.3 ± 0.9 STD vs. 3.9 ± 0.7 MC, p < 0.007). While B-A analysis demonstrated narrower interobserver agreement for aortic annular area (bias 0.03 vs 0.02 cm(2), range -0.32 to 0.39 cm(2) vs -0.50 to 0.55 cm(2)), and perimeter (bias 0.3 vs 0.3 mm, range -3.1 to 3.8 mm vs -4.6 to 5.3 mm), this was not statistically significant by concordance correlation coefficient. At 75%, there was no significant difference in image quality (Likert score 3.3 ± 0.9 vs. 3.5 ± 0.76, p = 0.454) or annular measurement agreement intervals. CONCLUSION: Motion correction algorithms may yield significant improvements of image quality in systolic CT data sets of the heart. Further validation studies are required to determine the effect on annular measurements and translation into clinical practice.
Authors: Jin Young Kim; Young Joo Suh; Suyon Chang; Dong Jin Im; Yoo Jin Hong; Hye-Jeong Lee; Jin Hur; Young Jin Kim; Byoung Wook Choi Journal: Int J Cardiovasc Imaging Date: 2017-08-02 Impact factor: 2.357