RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography

IntroductionAortic stenosis (AS) is a valvular heart disease that obstructs normal blood flow from the left ventricle to the aorta due to pathological changes in the valve, leading to impaired cardiac function. Echocardiography is a key diagnostic tool for AS; however, its accuracy is influenced by...

Full description

Saved in:
Bibliographic Details
Main Authors: Yejia Gan, Wanzhong Huang, Yan Deng, Xiaoying Xie, Yuanyuan Gu, Yaozhuang Zhou, Qian Zhang, Maosheng Zhang, Yangchun Liu
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-04-01
Series:Frontiers in Medicine
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fmed.2025.1587307/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849310417321459712
author Yejia Gan
Wanzhong Huang
Yan Deng
Xiaoying Xie
Yuanyuan Gu
Yaozhuang Zhou
Qian Zhang
Maosheng Zhang
Yangchun Liu
author_facet Yejia Gan
Wanzhong Huang
Yan Deng
Xiaoying Xie
Yuanyuan Gu
Yaozhuang Zhou
Qian Zhang
Maosheng Zhang
Yangchun Liu
author_sort Yejia Gan
collection DOAJ
description IntroductionAortic stenosis (AS) is a valvular heart disease that obstructs normal blood flow from the left ventricle to the aorta due to pathological changes in the valve, leading to impaired cardiac function. Echocardiography is a key diagnostic tool for AS; however, its accuracy is influenced by inter-observer variability, operator experience, and image quality, which can result in misdiagnosis. Therefore, alternative methods are needed to assist healthcare professionals in achieving more accurate diagnoses.MethodsWe proposed a deep learning model, RSMAS-Net, for the automated identification and diagnosis of AS using echocardiography. The model enhanced the ResNet50 backbone by replacing Stage 4 with Spatial and Channel Reconstruction Convolution (SCConv) and Multi-Dconv Head Transposed Attention (MDTA) modules, aiming to reduce redundant computations and improve feature extraction capabilities.ResultsThe proposed method was evaluated on the TMED-2 echocardiography dataset, achieving an accuracy of 94.67%, an F1-score of 94.37%, and an AUC of 0.95 for AS identification. Additionally, the model achieved an AUC of 0.93 for AS severity classification on TMED-2. RSMAS-Net outperformed multiple baseline models in recall, precision, parameter efficiency, and inference time. It also achieved an AUC of 0.91 on the TMED-1 dataset.ConclusionRSMAS-Net effectively diagnoses and classifies the severity of AS in echocardiographic images. The integration of SCConv and MDTA modules enhances diagnostic accuracy while reducing model complexity compared to the original ResNet50 architecture. These results highlight the potential of RSMAS-Net in improving AS assessment and supporting clinical decision-making.
format Article
id doaj-art-22af5a3be7554d4c8fce56fe23417cde
institution Kabale University
issn 2296-858X
language English
publishDate 2025-04-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Medicine
spelling doaj-art-22af5a3be7554d4c8fce56fe23417cde2025-08-20T03:53:43ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2025-04-011210.3389/fmed.2025.15873071587307RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiographyYejia Gan0Wanzhong Huang1Yan Deng2Xiaoying Xie3Yuanyuan Gu4Yaozhuang Zhou5Qian Zhang6Maosheng Zhang7Yangchun Liu8Department of Information and Management, Guangxi Medical University, Nanning, ChinaDepartment of Cardiology, Jiangbin Hospital of Guangxi Zhuang Autonomous Region, Nanning, ChinaDepartment of Ultrasound, The First Affiliated Hospital of Guangxi Medical University, Nanning, ChinaDepartment of Cardiovascular Medicine, The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, ChinaDepartment of Information and Management, Guangxi Medical University, Nanning, ChinaDepartment of Information and Management, Guangxi Medical University, Nanning, ChinaDepartment of Information and Management, Guangxi Medical University, Nanning, ChinaDepartment of Information and Management, Guangxi Medical University, Nanning, ChinaCardiothoracic Surgery Intensive Care Unit, The First Affiliated Hospital of Guangxi Medical University, Nanning, ChinaIntroductionAortic stenosis (AS) is a valvular heart disease that obstructs normal blood flow from the left ventricle to the aorta due to pathological changes in the valve, leading to impaired cardiac function. Echocardiography is a key diagnostic tool for AS; however, its accuracy is influenced by inter-observer variability, operator experience, and image quality, which can result in misdiagnosis. Therefore, alternative methods are needed to assist healthcare professionals in achieving more accurate diagnoses.MethodsWe proposed a deep learning model, RSMAS-Net, for the automated identification and diagnosis of AS using echocardiography. The model enhanced the ResNet50 backbone by replacing Stage 4 with Spatial and Channel Reconstruction Convolution (SCConv) and Multi-Dconv Head Transposed Attention (MDTA) modules, aiming to reduce redundant computations and improve feature extraction capabilities.ResultsThe proposed method was evaluated on the TMED-2 echocardiography dataset, achieving an accuracy of 94.67%, an F1-score of 94.37%, and an AUC of 0.95 for AS identification. Additionally, the model achieved an AUC of 0.93 for AS severity classification on TMED-2. RSMAS-Net outperformed multiple baseline models in recall, precision, parameter efficiency, and inference time. It also achieved an AUC of 0.91 on the TMED-1 dataset.ConclusionRSMAS-Net effectively diagnoses and classifies the severity of AS in echocardiographic images. The integration of SCConv and MDTA modules enhances diagnostic accuracy while reducing model complexity compared to the original ResNet50 architecture. These results highlight the potential of RSMAS-Net in improving AS assessment and supporting clinical decision-making.https://www.frontiersin.org/articles/10.3389/fmed.2025.1587307/fulldeep learningaortic valve stenosiscardiovascular diseaseechocardiographytransthoracic-echocardiographyartificial intelligence
spellingShingle Yejia Gan
Wanzhong Huang
Yan Deng
Xiaoying Xie
Yuanyuan Gu
Yaozhuang Zhou
Qian Zhang
Maosheng Zhang
Yangchun Liu
RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
Frontiers in Medicine
deep learning
aortic valve stenosis
cardiovascular disease
echocardiography
transthoracic-echocardiography
artificial intelligence
title RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
title_full RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
title_fullStr RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
title_full_unstemmed RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
title_short RAMAS-Net: a module-optimized convolutional network model for aortic valve stenosis recognition in echocardiography
title_sort ramas net a module optimized convolutional network model for aortic valve stenosis recognition in echocardiography
topic deep learning
aortic valve stenosis
cardiovascular disease
echocardiography
transthoracic-echocardiography
artificial intelligence
url https://www.frontiersin.org/articles/10.3389/fmed.2025.1587307/full
work_keys_str_mv AT yejiagan ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT wanzhonghuang ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT yandeng ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT xiaoyingxie ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT yuanyuangu ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT yaozhuangzhou ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT qianzhang ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT maoshengzhang ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography
AT yangchunliu ramasnetamoduleoptimizedconvolutionalnetworkmodelforaorticvalvestenosisrecognitioninechocardiography