Improving cardiac computed tomography angiography image quality and diagnostic confidence for atrial fibrillation after left atrial appendage closure using a second-generation whole-heart motion correction algorithm

Introduction

Left atrial appendage (LAA) closure has emerged as an important treatment in the prevention of stroke in patients with non-valvular atrial fibrillation (AF), and cardiac computed tomography angiography (CTA) is a feasible, non-invasive alternative to transesophageal echocardiography (TEE) for follow-up imaging after LAA closure (1-3). Cardiac CTA has good spatial and temporal resolution, a large field of view, and multiplanar reconstruction capability, and is increasingly being used to monitor device complications, including position dislocation, endothelialization, peri-device leaks, device-related thrombosis, and pericardial effusion (4-6). However, the visualization of the LAA itself and its adjacent structures is sometimes hindered by artifacts caused by LAA closure devices, and represents a technical challenge that can degrade image quality and diagnostic confidence in the assessment of associated complications (7-9).

LAA device-related artifacts are complex and have multiple causes (10). Some are caused by closure device-induced beam hardening and scattering, while others are related to metal edges causing additional streak artifacts due to under-sampling and cardiac motion (11). In addition, AF is characterized by a high heart rate (HR) and HR variability (12), and systolic and diastolic phases of different lengths, which can easily lead to motion artifacts and further increase data acquisition inconsistencies, worsening closure-related artifacts. Thus, further improving the temporal resolution of cardiac CTA could improve the image quality of patients with LAA closure.

The first-generation motion correction algorithm, Snapshot Freeze 1 (SSF1; GE HealthCare, Waukesha, WI, USA), was developed to reduce motion artifacts in coronary CTA, with a primary focus on coronary arteries. More recently, a second-generation motion correction algorithm, Snapshot Freeze 2 (SSF2; GE HealthCare), was introduced to extend the correction across the entire heart, including valves and large vessels, thereby enhancing the overall image quality and diagnostic performance in cardiac computed tomography (CT).

SSF2 has been shown to improve image quality and diagnostic accuracy in multiple coronary indications, including in both adult and child patients with increased HRs (13,14), coronary stents (15), and myocardial bridges/mural coronary arteries (16). Further, the utility of SSF2 in non-coronary regions, including mechanical valves (17), and pre-transcatheter aortic valve implantation aortic annulus evaluation (18-20), has also been investigated. However, to date, no research appears to have been conducted on the utility of SSF2 in improving the cardiac CTA image quality of AF patients after LAA closure.

Thus, this study sought to compare the ability of SSF2 to improve image quality and diagnostic confidence for closure devices and adjacent structures of cardiac patients after LAA closure with that of standard reconstruction without motion correction [i.e., the standard reconstruction method (SRM)] and SSF1. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-689/rc).

Methods

Patient selection and study design

This was a retrospective cross-sectional study. Initially, 67 patients with AF after LAA occlusion who underwent cardiac CTA between June 2023 and April 2024 were enrolled in the study. However, 18 patients with a lack of raw data for reconstruction with SSF2 and seven patients with incomplete clinical data were excluded from the study. Thus, ultimately, 42 patients were included in the study. Figure 1 shows the patient selection flowchart. These 42 patients were implanted with five distinct types of occlusion devices; that is, Watchman and Watchman FLX (both from Boston Scientific, Marlborough, MA, USA), Amplatzer Amulet (Abbott Laboratories, Chicago, IL, USA), LAmbre (Lifetech Scientific Corp, Shenzhen, China), and LAMax (Salubris, Shenzhen, China). The Amplatzer Amulet has an anchoring lobe and a proximal occluder, while the other devices are only anchored by occluders with fixation barbs.

Figure 1 Flowchart of study patient selection. AF, atrial fibrillation; CTA, computed tomography angiography; LAA, left atrial appendage; SRM, standard reconstruction method; SSF1, Snapshot Freeze 1; SSF2, Snapshot Freeze 2.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Ethics Board of the First Affiliated Hospital of Xi’an Jiaotong University (No. XJTU1AF2023LSK-168), and the requirement of individual consent for this retrospective analysis was waived.

Image acquisition

The scan and contrast-injection parameters are presented in Table 1. All the patients underwent cardiac CTA scans on a 256-row, 16-cm wide-detector CT scanner (Revolution CT, GE HealthCare) using the prospectively electrocardiographic-triggered scan mode in a single heartbeat. The scanning parameters were as follows: CT detector configuration: 14 or 16 cm, depending on the patient’s heart size; tube potential: 100 kVp; tube current modulation: automatic (300–720 mA) (Smart mA, GE Healthcare); and preset noise index: 21 Hounsfield units (HU). Using the auto-gating technique, the optimal cardiac phase [the R-peak-to-R-peak (R-R) interval] for acquiring projections was selected based on each patient’s HR before scanning as follows: 70–80% of the R-R interval for patients with a HR <65 beats per minute (bpm); 40–80% for those with a HR of 65–80 bpm; and 30–50% for those with a HR >80 bpm.

Under the contrast-injection protocol, 50–60 mL of the contrast agent iopromide (370 mgI/mL, Ultravist, Bayer Schering Pharma, Berlin, Germany) was injected at a flow rate of 4.5–5.0 mL/s, followed by 40 mL of saline solution. The bolus tracking technique was used to trigger the scanning with a 3.1-s delay after the attenuation value in the descending aorta exceeded the threshold of 210 HU. The contrast dose and flow rate were based on each patient’s body mass index and venous conditions.

Image reconstruction

The SmartPhase technique (GE HealthCare) was used to select the cardiac phase with the smallest motion artifacts. To provide baseline images for comparison, the images were first reconstructed using the SRM without any motion correction technique; Additional image sets (SSF1 and SSF2) were reconstructed by applying the motion correction algorithms SSF1 and SSF2, respectively. The other reconstruction parameters were as follows: image matrix: 512×512; and slice thickness: 0.625 mm. All images were reconstructed using the adaptive statistical iterative reconstruction-V (ASIR-V; GE HealthCare) algorithm at the 50% strength level to reduce image noise. Three data sets (SRM, SSF1, and SSF2) were generated from a single cardiac CTA for every patient and transferred to a post-processing workstation (ADW 4.7, GE HealthCare) for analysis.

Image analysis

Quantitative measurement

An experienced technologist quantified the image artifacts by measuring standard deviations (SD_HU) in several regions of interest (ROIs) (21) using the GE Advantage Workstation software (ADW 4.7, GE HealthCare). The ROIs were initially drawn on the axial slice of the SRM image with the visually largest artifacts, at the sites with hypodense artifacts, hyperdense artifacts, and artifacts within the LAA. An artifact-free region in the thoracic aorta at the same level was also recorded as a reference point to document the true background image noise (Figure 2). The ROIs were copied to the exact same locations in the other two image groups for each patient. The inter-observer measurements were taken with an interval of more than 30 days. A consistency test was carried out on the two measurement results, and the average value of the two measurements was used for the analysis.

Figure 2 Examples of ROIs placed over hyperdense artifacts (blue circles), hypodense artifacts (yellow circles), the LAA with artifacts (red circles), and an artifact-free region (green circles) on an axial image. LAA, left atrial appendage; ROIs, regions of interest.

Qualitative measurement

Two experienced radiologists (with 9 and 12 years of experience in cardiac CT imaging, respectively), who were blinded to the reconstruction algorithms, assessed the image quality of the LAA closure and adjacent structures using the multiplanar reformation images. If the two readers disagreed, consensus was reached to obtain a final image quality score. To visualize the LAA closure, the default window setting (window level 300 HU; window width 1,500 HU) for displaying the vertebrae was selected.

The two radiologists performed the qualitative image quality evaluation using a five-point scale (22). The qualitative evaluation included: artifact severity caused by the LAA closure device in three regions (proximal, central, and distal regions of the LAA); the boundary clarity of the LAA closure; the visualization of the LAA closure; and the diagnostic confidence for the LAA closure. The artifact severity scores were inversely scaled (higher = worse), whereas the image visibility, boundary clarity, and diagnostic confidence scores were positively scaled (higher = better). Specifically, the criteria for the per-region artifact severity scores were graded on a 0–4 scale (where 0= none, 1= mild, 2= moderate, 3= pronounced, and 4= massive); the LAA closure boundary clarity was rated on a 1–5 scale (where 1= severely deformed borders, 2= distinct blurred or deformed borders, 3= somewhat blurred borders, 4= slightly blurred borders, and 5= clear borders); the LAA closure visualization was assessed on a 1–5 scale (where 1= not visible, 2= poor visualization, 3= fair visualization, 4= good visualization, and 5= excellent visualization); and the diagnostic confidence was scored on a 1–5 scale (where 1= non diagnostic, 2= suboptimal (limited diagnostic confidence), 3= acceptable, 4= diagnosis with confidence, and 5= highly diagnostic confidence). The images with diagnostic confidence scores ≥4 were considered excellent. The excellent diagnostic confidence rate was calculated as follow: rate = (number of excellent diagnostic confidence patients/total number of patients) × 100 (%) (23).

Statistical analysis

All the statistical analyses were performed using the SPSS 22.0 software (SPSS Inc., Chicago, IL, USA). The continuous variables are presented as the mean ± SD, or as the median and interquartile range. The categorical variables are expressed as the frequency and percentage. Differences in cardiac CT image quantitative measurements (SD) and qualitative score (artifacts, closure boundary, visualization, and diagnostic confidence for LAA closure) for the SRM, SSF1, and SSF2 image sets were analyzed by Friedman’s test and the post-hoc multiple comparison test with Bonferroni correction. Interobserver agreement for the quantitative measurements was determined using intraclass correlation coefficients (ICCs) where 0.76–1.0 indicated excellent agreement; 0.40–0.75 indicated fair to good agreement; and <0.40 indicated poor agreement. The Cohen Kappa test was used to test interobserver agreement in which k values <0.20 indicated the worst agreement; 0.21–0.40 indicated poor agreement; 0.41–0.6 indicated moderate agreement; 0.61–0.80 indicated good agreement; and 0.81–1.00 indicated excellent agreement. The excellent diagnostic confidence rates of the three reconstruction algorithms were compared using the McNemar test. A P value less than 0.05 was considered statistically significant.

Results

Patient and LAA closure characteristics

Of the 42 patients, 29 (69%) were male. The mean age of the patients was 71.38±7.37 years (range, 62–87 years). The average HR was 77.07±18.42 bpm. Of the 42 patients, 14 received Watchman, 18 received Watchman FLX, and 10 received other LAA closure devices. The characteristics of the patients and LAA closures are summarized in Table 2.

Quantitative image analysis

The quantitative artifact measurement results for the three image groups are shown in Table 3 and Figure 3. The ICCs for both measurements were greater than 0.84. In terms of the SD of the reference tissue, there was no significant difference in the SD for the reference ROI (thoracic aorta) among the three groups (P=0.215). Thus, the SD in the artifact-impaired regions was used as a measure of the artifact burden. In relation to the three regions in the plane of the visually largest artifacts (i.e., those with the most hypodense, hyperdense artifacts and artifacts within the LAA), the SSF2 image group had the smallest SD values while the SRM image group had the largest SD values. The SD values (in HU) of the hyperdense artifact regions were 129.89 (99.85–169.74) for the SRM, 117.03 (83.41–152.03) for SSF1, and 68.19 (57.20–91.14) for SSF2. The SD values (in HU) for the hypodense artifact regions were 132.81 (103.38–164.53) for the SRM, 104.80 (79.22–141.60) for SSF1, and 67.93 (53.49–85.36) for SSF2. The SD values (in HU) in the LAA artifact regions were 120.07 (78.36–169.74) for the SRM, 94.21 (74.51–121.86) for SSF1, and 59.08 (47.45–72.59) for SSF2. Based on these results, the quantitative measurements in the three regions were significantly more reduced for SSF2 than for the SRM and SSF1 (all P<0.001). However, the SD comparisons of the three regions revealed no significant differences between the SRM and SSF1 (P=0.108, P=0.082, P=0.232).

Figure 3 Box-plot diagram displaying the SD values in the hypodense and hyperdense artifact regions adjacent to the LAA closure in the SRM, SSF1, and SSF2 image groups. ns, no statistical significance; ***, statistical significance at the levels of P<0.001; ****, statistical significance at the levels of P<0.0001. LAA, left atrial appendage; SD, standard deviation; HU, Hounsfield units; SRM, standard reconstruction method; SSF1, Snapshot Freeze 1; SSF2, Snapshot Freeze 2.

Qualitative image analysis

The interobserver agreement was good or excellent for artifacts (in the proximal, central, and distal regions), boundary, visualization, and diagnostic confidence for LAA closure (kappa =0.689–0.874). The two observers’ independent results for the qualitative scores are shown in Table S1, and the final results are presented in Figure 4. The final scores of the two observers for artifacts (in the proximal, central, and distal regions) were 2 (1–2), 2 (1–3), and 1 (0–3) for the SRM, 1 (1–2), 2 (1–3), and 1 (0–2) for SSF1, and 1 (0–1), 1 (0–1), and 1 (0–1) for SSF2, respectively. The scores for the closure boundaries were 3 (2–4) for the SRM, 4 (3–4) for SSF1, and 4 (4–5) for SSF2. The scores for the LAA visualization were 3 (3–4) for the SRM, 4 (3–4) for SSF1, and 4 (4–5) for SSF2. The scores for diagnostic confidence were 3 (2–4) for the SRM, 4 (3–4) for SSF1, and 5 (4–5) for SSF2. The SSF2 group had the highest scores for the LAA closure boundaries, visualization, and diagnostic confidence, and the lowest artifact scores for the proximal, central, and distal regions compared to the SRM and SSF1 reconstruction groups (all P<0.001). No statistically significant difference was found when comparing the qualitative scores between the SRM and SSF1 image groups. Representative cases are shown in Figure 5.

Figure 4 Final qualitative score results (artifacts, LAA closure borders, LAA visualization, and diagnostic confidence) for the SRM, SSF1, and SSF2 reconstruction algorithms. The artifact severity scores are inversely scaled (higher score = worse). While all the other scores are positively scaled (higher score = better). LAA, left atrial appendage; SRM, standard reconstruction method; SSF1, Snapshot Freeze 1; SSF2, Snapshot Freeze 2.

Figure 5 Example of CT image sets of a 64-year-old-female after percutaneous LAA closure (Lambre2430). The patient’s heart rate during the scan monitoring was 150 bpm (atrial fibrillation). (A-C) Proximal LAA closure reconstructed with the SRM (A), SSF1 (B), and SSF2 (C). (D-F) Central LAA closure reconstructed with the SRM (D), SSF1 (E), and SSF2 (F). (G-I) Distal LAA closure reconstructed with the SRM (G), SSF1 (H), and SSF2 (I). Images with the SRM reconstruction (first column: A,D,G) and with SSF1 (second column: B,E,H) showing the LAA closure had severe deformation or severe artifacts (as indicated by the arrows). The artifacts from the LAA closure were reduced after the application of the SSF2 correction (third column: C,F,I). The LAA structure is more clearly seen in the SSF2 images than the SRM and SSF1 images. bpm, beats per minute; CT, computed tomography; LAA, left atrial appendage; SRM, standard reconstruction method; SSF1, Snapshot Freeze 1; SSF2, Snapshot Freeze 2.

The excellent diagnostic confidence rate was 30.9% for the SRM, 52.4% for SSF1, and 90.5% for SSF2. The excellent diagnostic confidence rate was significantly higher for the SSF2 group than for the other two image groups (all P<0.05); but no statistically significant difference was found between the SRM and SSF1 groups (P=0.357).

Discussion

This study showed that SSF2, the newly introduced second-generation, whole-heart motion correction algorithm, improved the image quality of cardiac CTA and the diagnostic confidence for the LAA in AF patients after LAA closure compared with the SRM without motion correction and SSF1, the first-generation coronary artery motion correction algorithm. In the quantitative analysis, SSF2 achieved significant improvements in the reduction of LAA closure-related artifacts compared with the SRM and SSF1. Similarly, the qualitative analysis showed that the SSF2 was superior to both the SRM and SSF1, with significantly reduced image artifacts and improved visualization of the LAA and LAA closure device. The excellent diagnostic confidence rate for LAA closure was significantly higher for the SSF2 algorithm than the SRM and SSF1 algorithms. The results showed the potential of the SSF2 algorithm to improve LAA patient evaluation after closure.

In studies involving cardiac CTA without metallic components, SSF1 may perform adequately in the imaging of coronary arteries (24-26). However, our study specifically analyzed artifact correction in the presence of LAA closure devices, which introduces a distinct challenge. We found that the application of SSF1 did not significantly improve the image quality of and the diagnostic confidence for the LAA and adjacent tissues after LAA closure. This lack of any substantial benefit can be primarily attributed to the fact that SSF1 was specifically engineered to correct motion-related artifacts arising solely from coronary artery movement rather than to address motion artifacts associated with the entire cardiac structure, including the motion of the LAA and surrounding tissues, which are key sources of imaging interference in post-LAA closure assessments (27).

Conversely, SFF2 extends motion correction to the whole heart beyond the coronary arteries for whole-heart motion correction. SSF2 uses the information from adjacent cardiac phases, available from a single axial rotation, to characterize the whole-heart motion at the prescribed target phase along all three axes. As the whole-heart correction requires motion characterization along all three axes, this also provides greater robustness in the coronary motion correction itself, which is especially helpful for extreme motion scenarios and motion paths predominantly along the z-axis (13,15). Young et al. (17) reported that SSF2 was effective in enhancing the quality of CT scans of mechanical valves in patients compared with the SRM. We confirmed the same in the image quality evaluation after LAA closure. Notably, our study specifically focused on the LAA and its closure, and we found that a significant reduction in both motion artifacts and artifacts caused by LAA closure, which has not been reported previously.

The combination of cardiac motion and metals, such as those in LAA closure devices, exacerbates the sampling inconsistency in CT data acquisition, which worsens the motion/metal artifacts in the images. In our study, we found that the application of SSF2, which reduces the sampling inconsistency in data acquisition, significantly reduced the appearance of such artifacts in cardiac CT images after LAA closure. However, our results also suggested that the ability of SSF2 to reduce the metal artifacts themselves was limited. In our study population, four patients with the Amplatzer Amulet closure (a proximal and distal end screw structure) showed severe beam hardening/metal artifacts, and had an image quality inferior to that of other closure types, even after the application of SSF2. This type of closure may require an additional specific metal artifact removal algorithm to further improve image quality.

This study had several limitations. First, although this study provides preliminary evidence, its single-center retrospective nature may limit the generalizability of the findings due to potential institutional biases and the exclusion of cases with missing data. While exclusions were necessary to ensure data integrity, future prospective multi-center studies with larger cohorts are needed to validate these preliminary findings. Second, this study only evaluated the diagnostic confidence (and not the performance) of SSF2 in assessing LAA device complications. Further studies are needed to evaluate its diagnostic accuracy in comparison with that of TEE. Third, we only evaluated one cardiac phase (the Smart Phase); thus, further evaluation of the effectiveness of the SSF2 algorithm in the assessment of the LAA in a wider range of the cardiac cycles is needed. Fourth, since this was a patient study with inherent cardiac motions, we did not have ground-truth motion-less images for comparison. However, the radiologists or cardiologists did not notice any obvious distorted anatomical structures or artificial details introduced by either SSF1 or SSF2.

Conclusions

Compared to the SRM without motion correction and the first-generation coronary artery motion correction algorithm (SSF1), SSF2 significantly reduced the artifacts caused by LAA closure, and enhanced image quality, visualization, and diagnostic confidence in the cardiac CT imaging of patients who have undergone percutaneous LAA closure. Improved image quality and diagnostic confidence could enhance the detection of device-related complications, such as residual leaks or device-related thrombus, during follow-up CTA, which may directly influence patient management.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-689/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-689/dss

Funding: This study was supported by The Clinical Research Award of the First Affiliated Hospital of Xi’an Jiaotong University, China (No. XJTU1 AF-CRF-2019-013).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-689/coif). All authors report receiving research funding support from the Clinical Research Award of the First Affiliated Hospital of Xi’an Jiaotong University. J.L. is an employee of GE Healthcare, the manufacturer of the CT scanner used in this study. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by institutional ethics board of the First Affiliated Hospital of Xi’an Jiaotong University (No. XJTU1AF2023LSK-168) and the requirement for individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Holmes DR Jr, Doshi SK, Kar S, Price MJ, Sanchez JM, Sievert H, Valderrabano M, Reddy VY. Left Atrial Appendage Closure as an Alternative to Warfarin for Stroke Prevention in Atrial Fibrillation: A Patient-Level Meta-Analysis. J Am Coll Cardiol 2015;65:2614-23. [Crossref] [PubMed]
2. Bajaj NS, Parashar A, Agarwal S, Sodhi N, Poddar KL, Garg A, Tuzcu EM, Kapadia SR. Percutaneous left atrial appendage occlusion for stroke prophylaxis in nonvalvular atrial fibrillation: a systematic review and analysis of observational studies. JACC Cardiovasc Interv 2014;7:296-304. [Crossref] [PubMed]
3. Saw J, Holmes DR, Cavalcante JL, Freeman JV, Goldsweig AM, Kavinsky CJ, Moussa ID, Munger TM, Price MJ, Reisman M, Sherwood MW, Turi ZG, Wang DD, Whisenant BK. SCAI/HRS Expert Consensus Statement on Transcatheter Left Atrial Appendage Closure. JACC Cardiovasc Interv 2023;16:1384-400. [Crossref] [PubMed]
4. Rajiah P, Alkhouli M, Thaden J, Foley T, Williamson E, Ranganath P. Pre- and Postprocedural CT of Transcatheter Left Atrial Appendage Closure Devices. Radiographics 2021;41:680-98. [Crossref] [PubMed]
5. Fauchier L, Cinaud A, Brigadeau F, Lepillier A, Pierre B, Abbey S, Fatemi M, Franceschi F, Guedeney P, Jacon P, Paziaud O, Venier S, Deharo JC, Gras D, Klug D, Mansourati J, Montalescot G, Piot O, Defaye P. Device-Related Thrombosis After Percutaneous Left Atrial Appendage Occlusion for Atrial Fibrillation. J Am Coll Cardiol 2018;71:1528-36. [Crossref] [PubMed]
6. Kramer AD, Korsholm K, Jensen JM, Nørgaard BL, Peelukhana S, Herbst T, Horton R, Kar S, Saw J, Alkhouli M, Nielsen-Kudsk JE. Cardiac computed tomography following Watchman FLX implantation: device-related thrombus or device healing? Eur Heart J Cardiovasc Imaging 2023;24:250-9. [Crossref] [PubMed]
7. Malhotra P. Use of Computed Tomography for Left Atrial Appendage Occlusion Procedure Planning and Post-Procedure Assessment. Interv Cardiol Clin 2024;13:19-28. [Crossref] [PubMed]
8. Li S, Dong J, Luo J, Wang G, Xie D, Zhou L. Comparison of different quantitative evaluation protocols for peri-device leak detection using cardiac computed tomography angiography after left atrial appendage closure. Int J Cardiovasc Imaging 2023;39:659-66. [Crossref] [PubMed]
9. Iriart X, Blanc G, Bouteiller XP, Legghe B, Bouyer B, Sridi-Cheniti S, Bustin A, Vasile C, Thambo JB, Elbaz M, Cochet H. Clinical Implications of CT-detected Hypoattenuation Thickening on Left Atrial Appendage Occlusion Devices. Radiology 2023;308:e230462. [Crossref] [PubMed]
10. Kalisz K, Buethe J, Saboo SS, Abbara S, Halliburton S, Rajiah P. Artifacts at Cardiac CT: Physics and Solutions. Radiographics 2016;36:2064-83. [Crossref] [PubMed]
11. Boas FE, Fleischmann D. CT artifacts: causes and reduction techniques. Imaging in Medicine 2012;4:229-40.
12. Andreini D, Pontone G, Mushtaq S, Mancini ME, Conte E, Guglielmo M, Volpato V, Annoni A, Baggiano A, Formenti A, Ditali V, Perchinunno M, Fiorentini C, Bartorelli AL, Pepi M. Image quality and radiation dose of coronary CT angiography performed with whole-heart coverage CT scanner with intra-cycle motion correction algorithm in patients with atrial fibrillation. Eur Radiol 2018;28:1383-92. [Crossref] [PubMed]
13. Liang J, Sun Y, Ye Z, Sun Y, Xu L, Zhou Z, Thomsen B, Li J, Sun Z, Fan Z. Second-generation motion correction algorithm improves diagnostic accuracy of single-beat coronary CT angiography in patients with increased heart rate. Eur Radiol 2019;29:4215-27. [Crossref] [PubMed]
14. Sun J, Okerlund D, Cao Y, Li H, Zhu Y, Li J, Peng Y. Further Improving Image Quality of Cardiovascular Computed Tomography Angiography for Children With High Heart Rates Using Second-Generation Motion Correction Algorithm. J Comput Assist Tomogr 2020;44:790-5. [Crossref] [PubMed]
15. Wu Z, Han Q, Liang Y, Zheng Z, Wu M, Ai Z, Ma K, Xiang Z. Enhancing diagnostic performance and image quality in coronary CT angiography: Impact of SnapShot Freeze 2 algorithm across varied heart rates in stent patients. J Appl Clin Med Phys 2024;25:e14412. [Crossref] [PubMed]
16. Zhang Z, Liu Z, Hong N, Chen L. Effect of a second-generation motion correction algorithm on image quality and measurement reproducibility of coronary CT angiography in patients with a myocardial bridge and mural coronary artery. Clin Radiol 2024;79:e462-7. [Crossref] [PubMed]
17. Suh YJ, Kim YJ, Kim JY, Chang S, Im DJ, Hong YJ. Choi BW. A whole-heart motion-correction algorithm: Effects on CT image quality and diagnostic accuracy of mechanical valve prosthesis abnormalities. J Cardiovasc Comput Tomogr 2017;11:474-81. [Crossref] [PubMed]
18. Soon J, Sulaiman N, Park JK, Kueh SH, Naoum C, Murphy D, Ellis J, Hague CJ, Blanke P, Leipsic J. The effect of a whole heart motion-correction algorithm on CT image quality and measurement reproducibility in Pre-TAVR aortic annulus evaluation. J Cardiovasc Comput Tomogr 2016;10:386-90. [Crossref] [PubMed]
19. Matsumoto Y, Fujioka C, Yokomachi K, Kitera N, Nishimaru E, Kiguchi M, Higaki T, Kawashita I, Tatsugami F, Nakamura Y, Awai K. Evaluation of the second-generation whole-heart motion correction algorithm (SSF2) used to demonstrate the aortic annulus on cardiac CT. Sci Rep 2023;13:3636. [Crossref] [PubMed]
20. Zhang Y, Liu Z, Cheng Y, Li Z, Wang Z, Peng L, Li J, Shuai T. New Whole-Heart motion correction algorithm enables diagnostic CT of aortic valve and coronary arteries in systolic phase for transcatheter aortic valve implantation candidates. Eur J Radiol 2023;168:111141. [Crossref] [PubMed]
21. Konst B, Ohlsson L, Henriksson L, Sandstedt M, Persson A, Ebbers T. Optimization of photon counting CT for cardiac imaging in patients with left ventricular assist devices: An in-depth assessment of metal artifacts. J Appl Clin Med Phys 2024;25:e14386. [Crossref] [PubMed]
22. Yamaguchi S, Ichikawa Y, Takafuji M, Sakuma H, Kitagawa K. Usefulness of second-generation motion correction algorithm in improving delineation and reducing motion artifact of coronary computed tomography angiography. J Cardiovasc Comput Tomogr 2024;18:281-90. [Crossref] [PubMed]
23. Nagayama Y, Emoto T, Hayashi H, Kidoh M, Oda S, Nakaura T, Sakabe D, Funama Y, Tabata N, Ishii M, Yamanaga K, Fujisue K, Takashio S, Yamamoto E, Tsujita K, Hirai T. Coronary Stent Evaluation by CTA: Image Quality Comparison Between Super-Resolution Deep Learning Reconstruction and Other Reconstruction Algorithms. AJR Am J Roentgenol 2023;221:599-610. [Crossref] [PubMed]
24. Sheta HM, Egstrup K, Husic M, Heinsen LJ, Nieman K, Lambrechtsen J. Impact of a motion correction algorithm on image quality in patients undergoing CT angiography: A randomized controlled trial. Clin Imaging 2017;42:1-6. [Crossref] [PubMed]
25. Leipsic J, Labounty TM, Hague CJ, Mancini GB, O'Brien JM, Wood DA, Taylor CM, Cury RC, Earls JP, Heilbron BG, Ajlan AM, Feuchtner G, Min JK. Effect of a novel vendor-specific motion-correction algorithm on image quality and diagnostic accuracy in persons undergoing coronary CT angiography without rate-control medications. J Cardiovasc Comput Tomogr 2012;6:164-71. [Crossref] [PubMed]
26. Fuchs TA, Stehli J, Dougoud S, Fiechter M, Sah BR, Buechel RR, Bull S, Gaemperli O, Kaufmann PA. Impact of a new motion-correction algorithm on image quality of low-dose coronary CT angiography in patients with insufficient heart rate control. Acad Radiol 2014;21:312-7. [Crossref] [PubMed]
27. Bhagalia R, Pack JD, Miller JV, Iatrou M. Nonrigid registration-based coronary artery motion correction for cardiac computed tomography. Med Phys 2012;39:4245-54. [Crossref] [PubMed]