Khargi, Indira (2024) Reducing Metal Artifacts of Transcatheter Aortic Valves on CTA with a Novel Deep Learning Metal Artifact Reduction Algorithm (DL-C-MAR): a Retrospective Cohort and Phantom Study. thesis, Medicine.
Full text available on request.Abstract
Objective: To assess the safety and performance of a novel deep learning-based cardiac metal artifact reduction algorithm (DL-C-MAR) in a retrospective comparison with unedited conventional computed tomography angiograms (CTAs) after transcatheter aortic valve implantation (TAVI) and phantom experiments. Methods: DL-C-MAR was trained using simulated metal implants and artifacts in 1000 CTAs. Performance of DL-C-MAR was investigated retrospectively in 50 TAVI patients and compared to unedited conventional CTA scans. To determine image quality, noise, contrast-to-noise ratio (CNR), artifact index (AI), and artifact volume were calculated. DL-C-MAR’s safety was assessed by conducting phantom experiments with metal cylinders of four different sizes. Diameters of the cylinders were measured by two separate readers and compared to their conventional counterparts and the ground truth. Results: In the TAVI CTAs, DL-C-MAR resulted in a higher CNR (9.1±5.8 vs. 7.9±4.8), and lower noise (57.2±33.9 vs. 82.1±54.0), AI (53.0±36.0 vs. 75.7±55.9), and artifact volume (0.02±0.12mL vs. 0.06±0.42mL) compared to unedited conventional CTAs (all p<0.001). In the phantom scans, DL-C-MAR decreased the diameter of the implants in all cases compared to the conventional cases. No significant difference in the shape of the cylinder was found after DL-C-MAR compared to the unedited CTA’s Conclusion: DL-C-MAR does not seem to hallucinate on the phantom images and effectively increases image quality and reduces metal artifacts in CTA scans after TAVI implantation
| Item Type: | Thesis (UNSPECIFIED) |
|---|---|
| Supervisor name: | Boomsma, dr. M.F. and Leiner, Prof. dr. T. |
| Faculty: | Medical Sciences |
| Date Deposited: | 24 Oct 2025 12:51 |
| Last Modified: | 24 Oct 2025 12:51 |
| URI: | https://umcg.studenttheses.ub.rug.nl/id/eprint/3839 |
Actions (login required)
 |
View Item |