Reliable deep learning for coronary artery disease detection: a patient-level, statistically validated MRI study
Romanian Journal of Morphology and Embryology · 10.03.2026 · 66 (4) · 711-717
DOI: 10.47162/rjme.66.4.09 ↗Kurz gefasst
Fachartikel in der Zeitschrift Romanian Journal of Morphology and Embryology, online erschienen am 03.03.2026. Verfasst von Christiana Raluca Dănciulescu und 5 weiteren Autor:innen. Beteiligte Einrichtung: Doctoral School, University of Medicine and Pharmacy of Craiova, Romania; daniel.stanescu@umfcv.ro. Hauptthemen laut MEDLINE: Deep Learning, Coronary Artery Disease, Magnetic Resonance Imaging. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
Background
Accurate detection of coronary artery disease (CAD) from cardiac magnetic resonance (CMR) imaging can support earlier diagnosis and streamlined clinical decision-making.
Objective
This study evaluated the performance and statistical robustness of two deep learning architectures – DenseNet121 and ResNet50 – for automated CAD classification using multiparametric CMR imaging.
Methods
Images were preprocessed using a valid pipeline and partitioned strictly at the patient level. Model performance was quantified through average accuracy, area under the receiver operating characteristic (ROC) curve (AUC–ROC), precision recall, while distributional assumptions were assessed using Shapiro–Wilk tests and variance homogeneity was explored with Brown–Forsythe test.
Results
ResNet50 demonstrated the strongest performance, achieving an average accuracy of 90.43%, AUC–ROC of 0.862, and area under the precision recall curve (PR–AUC) of 0.891. DenseNet121 showed lower accuracy (81.72%). Statistical analysis revealed non-normal performance distributions and significant variance differences between models.
Conclusions
The findings indicate that ResNet50 offers a reliable and statistically validated solution for CAD detection from CMR imaging. The combined use of realistic preprocessing and comprehensive inferential testing supports the generation of reproducible and clinically meaningful performance estimates.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Doctoral School, University of Medicine and Pharmacy of Craiova, Romania; daniel.stanescu@umfcv.ro.
Zitieren
APA
Dănciulescu, C. R., Ivănescu, C. R., Stănescu, D. R., Băiaşu, A. F., Alexandru, D. O., & Şerbănescu, M. S. (2026). Reliable deep learning for coronary artery disease detection: a patient-level, statistically validated MRI study. Romanian Journal of Morphology and Embryology, 66(4), 711-717. https://doi.org/10.47162/rjme.66.4.09
Vancouver
Dănciulescu CR, Ivănescu CR, Stănescu DR, Băiaşu AF, Alexandru DO, Şerbănescu MS. Reliable deep learning for coronary artery disease detection: a patient-level, statistically validated MRI study. Romanian Journal of Morphology and Embryology. 2026;66(4):711-717. doi:10.47162/rjme.66.4.09
BibTeX
@article{Dnciulescu2026,
author = {Dănciulescu, Christiana Raluca and Ivănescu, Constantin Renato and Stănescu, Daniel Robert and Băiaşu, Andrei Florentin and Alexandru, Dragoş Ovidiu and Şerbănescu, Mircea Sebastian},
title = {Reliable deep learning for coronary artery disease detection: a patient-level, statistically validated MRI study},
journal = {Romanian Journal of Morphology and Embryology},
year = {2026},
volume = {66},
number = {4},
pages = {711--717},
doi = {10.47162/rjme.66.4.09}
}
EUVIMED