Improving a Deep Learning Model to Accurately Diagnose LVNC

Baron Yusti, Jaime Rafael; Bernabé García, Gregorio; González Férez, Pilar; García Carrasco, José Manuel; Casas, Guillem; González-Carrillo, Josefa

Publication:
Improving a Deep Learning Model to Accurately Diagnose LVNC

dc.contributor.author	Baron Yusti, Jaime Rafael
dc.contributor.author	Bernabé García, Gregorio
dc.contributor.author	González Férez, Pilar
dc.contributor.author	García Carrasco, José Manuel
dc.contributor.author	Casas, Guillem
dc.contributor.author	González-Carrillo, Josefa
dc.contributor.department	Ingeniería y Tecnología de Computadores
dc.date.accessioned	2024-01-30T11:26:01Z
dc.date.available	2024-01-30T11:26:01Z
dc.date.created	2023-10-30
dc.date.issued	2023-12-12
dc.description	©2023. This manuscript version is made available under the CC-BY 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ This document is the Published Manuscript version of a Published Work that appeared in final form in Journal of Clinical Medicine. To access the final edited and published work see https://doi.org/10.3390/jcm12247633	es
dc.description.abstract	Accurate diagnosis of Left Ventricular Noncompaction Cardiomyopathy (LVNC) is critical for proper patient treatment but remains challenging. This work improves LVNC detection by improving left ventricle segmentation in cardiac MR images. Trabeculated left ventricle indicates LVNC, but automatic segmentation is difficult. We present techniques to improve segmentation and evaluate their impact on LVNC diagnosis. Three main methods are introduced: (1) using full 800 × 800 MR images rather than 512 × 512; (2) a clustering algorithm to eliminate neural network hallucinations; (3) advanced network architectures including Attention U-Net, MSA-UNet, and U-Net++.Experiments utilize cardiac MR datasets from three different hospitals. U-Net++ achieves the best segmentation performance using 800 × 800 images, and it improves the mean segmentation Dice score by 0.02 over the baseline U-Net, the clustering algorithm improves the mean Dice score by 0.06 on the images it affected, and the U-Net++ provides an additional 0.02 mean Dice score over the baseline U-Net. For LVNC diagnosis, U-Net++ achieves 0.896 accuracy, 0.907 precision, and 0.912 F1-score outperforming the baseline U-Net. Proposed techniques enhance LVNC detection, but differences between hospitals reveal problems in improving generalization. This work provides validated methods for precise LVNC diagnosis.	es
dc.format	application/pdf	es
dc.format.extent	15	es
dc.identifier.citation	Journal of Clinical Medicine, volumen 12, número 24, año 2023
dc.identifier.doi	https://doi.org/10.3390/jcm12247633
dc.identifier.uri	http://hdl.handle.net/10201/138134
dc.language	eng	es
dc.publisher	Federico Guerra	es
dc.relation	TED2021-129221B-I00 APLICACIÓN DE LA COMPUTACIÓN EFICIENTE DE ALTO RENDIMIENTO CON TÉCNICAS AVANZADAS DE INTELIGENCIA ARTIFICIAL PARA EL DIAGNÓSTICO DE ENFERMEDADES EN SISTEMAS HETEROGÉNEOS ENTIDAD: Ministerio de Ciencia e Inovación/AGENCIA ESTATAL DE INVESTIGACIÓN y “European Union NextGenerationEU/PRTR” COMIENZO: 01/12/2022, FIN: 30/11/2024	es
dc.relation.publisherversion	https://www.mdpi.com/2077-0383/12/24/7633/pdf?version=1702432108	es
dc.rights	Attribution-NonCommercial-NoDerivatives 4.0 Internacional	es
dc.rights	info:eu-repo/semantics/openAccess
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/4.0/	*
dc.subject	Left ventricular non-compaction diagnosis	es
dc.subject	Cardiomyopathies	es
dc.subject	Convolutional neural networks	es
dc.subject	MRI Image segmentation	es
dc.subject.other	CDU::0 - Generalidades.::00 - Ciencia y conocimiento. Investigación. Cultura. Humanidades.::004 - Ciencia y tecnología de los ordenadores. Informática.	es
dc.title	Improving a Deep Learning Model to Accurately Diagnose LVNC	es
dc.type	info:eu-repo/semantics/article	es
dspace.entity.type	Publication	es