National and Kapodistrian University of Athens
Χαρακτηρισμός της υφής του μυοκαρδίου με ct radiomics σε ασθενείς με στένωση αορτικής βαλβίδας
Abstract
dc:descriptionPurpose: Tissue characterization constitutes a cornerstone of medical science, enabling the identification of physiological and pathological phenotypes. Beyond traditional methods such as biopsy, imaging-based tissue characterization has recently emerged, with radiomics analysis being a prominent modality. The objective of this doctoral thesis is the development of a radiomic signature that characterizes the myocardium in patients suffering from aortic valve stenosis. Methods: The study included 400 preoperative CT angiographies from consecutive patients with aortic valve stenosis who underwent transcatheter aortic valve replacement (TAVI). The left ventricle was segmented using a semi-automated approach, and 851 radiomic features were extracted for each patient. Using artificial intelligence models, the features most characteristic of the myocardium in the presence of severe aortic stenosis were selected. Results: The initial set of 851 radiomic features was filtered to select those with the lowest inter-correlation and was adjusted for CT scanner technical parameters. Ultimately, the 71 most robust and stable features were selected to characterize the myocardium, forming the "myocardial radiomic signature". Furthermore, a subset of ten radiomic features per case successfully differentiated patients with coronary artery disease, arterial hypertension, diabetes mellitus, and systolic dysfunction. Additionally, the developed methodology was extended to investigate the diagnostic value of radiomics in detecting transthyretin cardiac amyloidosis (ATTR-CM). This application led to the development of the "AmyloidRS" composite radiomic signature, based on ten selected features, which demonstrated high predictive performance (C-index up to 0.88) for ATTR-CM detection, showcasing the feasibility of transitioning this methodology from theoretical development to clinical practice. Conclusions: Overall, a comprehensive framework is presented for the extraction of a CT-derived radiomic signature for myocardial tissue characterization in patients with aortic stenosis. Further application of this methodology confirmed its utility as a non-invasive quantitative analysis tool for assessing myocardial texture and detecting transthyretin cardiac amyloidosis, a disease traditionally diagnosed via conventional biopsy. This radiomic signature could facilitate the risk stratification of patients with aortic stenosis using only routine CT imaging.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Panagiotopoulos, Ioannis
- Παναγιωτόπουλος, Ιωάννης
Subjects
dc:subject × 16- Ραδιομική
- Υφή μυοκαρδίου
- Στένωση αορτικής βαλβίδας
- Τεχνητή νοημοσύνη
- Ραδιομική υπογραφή
- Radiomics
- Myocardial texture
- Aortic valve stenosis
- Artificial intelligence
- Radiomic signature
- Κλινική Ιατρική
- Ιατρική και Επιστήμες Υγείας
- Καρδιολογία και Καρδιοαγγειακά νοσήματα
- Clinical Medicine
- Medical and Health Sciences
- Cardiology and Cardiovascular Medicine
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/61031