Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated just now
Peer-reviewedCC BY 4.0European Journal of Radiology Open

Impact of super resolution deep learning reconstruction with a 1024 matrix on CAD-RADS classification of intermediate coronary stenosis: A bicentric retrospective study

  • Jean Philippe Dillenseger1,2,3
  • Fuminari Tatsugami4
  • Aissam Labani1
  • Wataru Fukumoto4
  • Soraya El Ghannudi1
  • Akira Taniguchi5
  • Karim Haioun5
  • Catherine Roy1,3
  • Kazuo Awai4
  • Mickaël Ohana1,2,3
5 affiliations
  1. 1Pole d’imagerie médicale, Hôpitaux universitaire de Strasbourg, Strasbourg, France
  2. 2ICube – UMR 7357, CNRS, Université de Strasbourg, Strasbourg, France
  3. 3Faculté de médecine, maïeutique et des sciences de la santé, Université de Strasbourg, Strasbourg, France
  4. 4Department of Diagnostic Radiology, Graduate School of Biomedical Health Science, Hiroshima University, Japan
  5. 5Canon Inc., Shimomaruko, Tokyo, 146-8501, Japan
Read full textPDFDOICite

Abstract

Introduction: Moderate coronary stenosis (CAD-RADS 3) is a diagnostic gray zone in coronary CT angiography (CCTA). We evaluated whether super-resolution deep learning reconstruction (SR-DLR), particularly with a 1024 matrix, modifies CAD-RADS classification and inter-reader agreement. Materials and methods: Thirty-nine CCTA examinations from two centers, each containing at least one 50–69% stenosis identified on the routine normal-resolution deep learning reconstruction (NR-DLR), were reconstructed with NR-DLR and SR-DLR (512 and 1024 matrices). Five experienced readers blindly assigned CAD-RADS categories. Agreement was assessed with Gwet’s AC2 and paired reclassification with exact McNemar tests. Subgroup analyses (plaque composition, initial image quality) were exploratory.

Results: Forty-six lesions were analyzed (19 calcified, 16 mixed, 10 non-calcified, one in-stent). Agreement remained excellent (AC2 0.89 [0.86–0.92], 0.87 [0.83–0.90] and 0.84 [0.81–0.88] for NR-DLR, SR-DLR-512 and SR-DLR-1024). Consensus CAD-RADS 2 rose from 6.5% to 21.7% with SR-DLR-1024 (p = 0.016; 10 lesions downgraded, 2 upgraded), predominantly in lesions with non-calcified components. Per-reader CAD-RADS 2 rates ranged from 13% to 30%, increasing significantly for two of five readers (p = 0.004).

Conclusion: SR-DLR-1024 modified CAD-RADS classification in about one fifth of intermediate lesions while preserving excellent agreement. Without an invasive reference standard, these reclassifications reflect a change in reader interpretation rather than a proven gain in accuracy.

Related

  1. High-Resolution Parallel-Translation CT Reconstruction Based on a Patch-Based Adaptive-Regularization-Guided Diffusion Model
    Xuanheng Li, Mingze Lv, Jiahong Liu +2
    Optics ExpressPeer-reviewedLicence not statedMedical imaging & radiologyCrossref
  2. A hybrid quantum algorithm for multi-resolution arrhythmia classification
    Melvin Yashnil Ramkhelawan, Santoso Wibowo, Srimannarayana Grandhi +1
    Scientific ReportsPeer-reviewedCC BY-NC-ND 4.0Full textCardiovascular medicineCrossref
  3. Evaluation of Delphi high-resolution computed tomography phenotype classification for treatment-naïve patients with pulmonary sarcoidosis
    Ce Wang, Sijia Guo, Shi Shu +6
    Quantitative imaging in medicine and surgeryPeer-reviewedCC BY-NC-NDFull textMedical imaging & radiologyEurope PMC
  4. The frequency and risk factors of deep vein thrombosis among surgical patients receiving thromboprophylaxis at Oshakati Intermediate Hospital: a six-year retrospective cohort study.
    Tulipo Kisha Ipinge, Albertina Mariina Ndinelao Shatri
    The Pan African medical journalPeer-reviewedCC BYFull textGeneral medicineEurope PMCPMC
  5. Independent linear association between flow-mediated dilation and predicted intermediate-to-high 10-year ASCVD risk in patients with type 2 diabetes: a retrospective study
    Hou KJ, Shen SY, Zong CY +4
    Frontiers in EndocrinologyPeer-reviewedCC BY 4.0Full textCardiovascular medicineCrossrefEurope PMCPMC

Cite

BibTeXRIS
BibTeX
@article{dillenseger2026impact,
  title = {{Impact of super resolution deep learning reconstruction with a 1024 matrix on CAD-RADS classification of intermediate coronary stenosis: A bicentric retrospective study}},
  author = {Jean Philippe Dillenseger and Fuminari Tatsugami and Aissam Labani and Wataru Fukumoto and Soraya El Ghannudi and Akira Taniguchi and Karim Haioun and Catherine Roy and Kazuo Awai and Mickaël Ohana},
  journal = {European Journal of Radiology Open},
  year = {2026},
  month = oct,
  doi = {10.1016/j.ejro.2026.100821},
  url = {https://doi.org/10.1016/j.ejro.2026.100821},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Impact of super resolution deep learning reconstruction with a 1024 matrix on CAD-RADS classification of intermediate coronary stenosis: A bicentric retrospective study
AU  - Jean Philippe Dillenseger
AU  - Fuminari Tatsugami
AU  - Aissam Labani
AU  - Wataru Fukumoto
AU  - Soraya El Ghannudi
AU  - Akira Taniguchi
AU  - Karim Haioun
AU  - Catherine Roy
AU  - Kazuo Awai
AU  - Mickaël Ohana
PY  - 2026
DA  - 2026/10/26
JO  - European Journal of Radiology Open
DO  - 10.1016/j.ejro.2026.100821
UR  - https://doi.org/10.1016/j.ejro.2026.100821
AB  - Introduction: Moderate coronary stenosis (CAD-RADS 3) is a diagnostic gray zone in coronary CT angiography (CCTA). We evaluated whether super-resolution deep learning reconstruction (SR-DLR), particularly with a 1024 matrix, modifies CAD-RADS classification and inter-reader agreement. Materials and methods: Thirty-nine CCTA examinations from two centers, each containing at least one 50–69% stenosis identified on the routine normal-resolution deep learning reconstruction (NR-DLR), were reconstructed with NR-DLR and SR-DLR (512 and 1024 matrices). Five experienced readers blindly assigned CAD-RADS categories. Agreement was assessed with Gwet’s AC2 and paired reclassification with exact McNemar tests. Subgroup analyses (plaque composition, initial image quality) were exploratory.  Results: Forty-six lesions were analyzed (19 calcified, 16 mixed, 10 non-calcified, one in-stent). Agreement remained excellent (AC2 0.89 [0.86–0.92], 0.87 [0.83–0.90] and 0.84 [0.81–0.88] for NR-DLR, SR-DLR-512 and SR-DLR-1024). Consensus CAD-RADS 2 rose from 6.5% to 21.7% with SR-DLR-1024 (p = 0.016; 10 lesions downgraded, 2 upgraded), predominantly in lesions with non-calcified components. Per-reader CAD-RADS 2 rates ranged from 13% to 30%, increasing significantly for two of five readers (p = 0.004).  Conclusion: SR-DLR-1024 modified CAD-RADS classification in about one fifth of intermediate lesions while preserving excellent agreement. Without an invasive reference standard, these reclassifications reflect a change in reader interpretation rather than a proven gain in accuracy.
SN  - 2352-0477
ER  -