Deep Learning–based Automated Coronary Plaque Quantification
This study evaluated a novel DL-based tool with automatic case preparation for quantitative coronary plaque in UHR CCTA data sets. The algorithm was technically robust and reproducible, delivering anatomically consistent outputs not requiring manual correction. Reconstructions with lower temporal resolution (125 ms) systematically overestimated plaque burden compared with higher temporal resolution (66 ms), underscoring that protocol standardization is essential for reliable DL-based plaque quantification.
Authors
- Mehmet Gulsun
- Matthias Eberhard (ORCID: https://orcid.org/0000-0002-9476-3759)
- Konstantin Klambauer (ORCID: https://orcid.org/0000-0002-4114-2043)
- Susanne Burger
- Michael Wels (ORCID: https://orcid.org/0000-0002-1239-1243)
- Tristan T. Demmert
- Thomas Allmendinger (ORCID: https://orcid.org/0000-0002-3133-4017)
- Hatem Alkadhi (ORCID: https://orcid.org/0000-0002-2581-2166)
- Victor Mergen (ORCID: https://orcid.org/0000-0003-4544-4804)
- Bernhard T. Schmidt (ORCID: https://orcid.org/0000-0001-5894-6374)
- Chris Schwemmer (ORCID: https://orcid.org/0000-0002-0700-5053)
- Max Schöbinger
- Lukas J. Moser
Institutions
- Universitätsklinikum Erlangen (DE)
- University Hospital of Zurich (CH)
- Siemens (United States) (US)
- Siemens Healthcare (United States) (US)
- Bayer (Germany) (DE)
Publication Details
- Journal
- Investigative Radiology
- Published
- 2025-08-22
- DOI
- https://doi.org/10.1097/rli.0000000000001233
- Primary Topic
- Advanced X-ray and CT Imaging
- Type
- article
- Field-Weighted Citation Impact
- 0.00