Unipolar Electrogram Eigenvalue Distribution Analysis for the Identification of Atrial Fibrosis
| Autor: | Jennifer Riccio, Juan Pablo Martinez, Sara Rocher, Pablo Laguna, Javier Saiz, Alejandro Alcaine, Laura Martinez-Mateu |
|---|---|
| Jazyk: | angličtina |
| Předmět: |
Clique
Receiver operating characteristic business.industry Orientation (computer vision) medicine.medical_treatment 020206 networking & telecommunications Pattern recognition Atrial fibrillation 02 engineering and technology 030204 cardiovascular system & hematology medicine.disease Ablation Signal TECNOLOGIA ELECTRONICA 03 medical and health sciences 0302 clinical medicine Fibrosis 0202 electrical engineering electronic engineering information engineering medicine Artificial intelligence business Eigenvalues and eigenvectors Mathematics |
| Zdroj: | Zaguán: Repositorio Digital de la Universidad de Zaragoza Universidad de Zaragoza Zaguán. Repositorio Digital de la Universidad de Zaragoza instname 2020 Computing in Cardiology Conference (CinC) CinC RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| ISSN: | 2325-887X |
| DOI: | 10.22489/cinc.2020.434 |
| Popis: | [EN] Atrial fibrosis plays an important role in the pathogenesis of atrial fibrillation (AF). Low bipolar electrograms (b-EGMs) peak-to-peak voltage areas indicate scar tissue and are considered targets for AF substrate ablation. However, this approach ignores the spatiotemporal information embedded in the signal and the dependence of b-EGMs on catheter orientation. This work proposes an approach to detect fibrosis based on the eigenvalue dominance ratio (EIGDR) in an ensemble (clique) of unipolar electrograms (u-EGMs). A 2-D tissue with a central circular patch of fibrosis has been simulated using the Courtemanche cellular model. Maps of EIGDR have been computed using two sizes of electrode cliques, from the original u-EGMs within the ensemble or after a time alignment of these signals. Performance of each map in detecting fibrosis has been evaluated using receiver operating characteristic curves and detection accuracy. Best results achieve an area under the curve (AUC) of 0.98 and an accuracy (ACC) of 1 when we use as marker the gain in eigenvalue dominance produced by the ensemble alignment Funding comes from EU Programme H2020 under the Marie Sklodowska-Curie Grant No 766082 (MY-ATRIA), Gobierno de Aragon (BSICoS Group T39-20R) cofunded by FEDER 2014-2020 "Building Europe from Aragon", fellowship ACIF/2018/174 from Generalitat Valenciana, and PID2019-104881RB-I00 from MICINN, Spain |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|