Cargando…

A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound

OBJECTIVE: Echocardiography (ECG) is the most common method used to diagnose heart failure (HF). However, its accuracy relies on the experience of the operator. Additionally, the video format of the data makes it challenging for patients to bring them to referrals and reexaminations. Therefore, this...

Descripción completa

Detalles Bibliográficos
Autores principales: Liu, Zeye, Huang, Yuan, Li, Hang, Li, Wenchao, Zhang, Fengwen, Ouyang, Wenbin, Wang, Shouzheng, Luo, Zhiling, Wang, Jinduo, Chen, Yan, Xia, Ruibing, Li, Yakun, Pan, Xiangbin
Formato: Online Artículo Texto
Lenguaje:English
Publicado: De Gruyter 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10680380/
https://www.ncbi.nlm.nih.gov/pubmed/38025953
http://dx.doi.org/10.2478/jtim-2023-0088
_version_ 1785142379151360000
author Liu, Zeye
Huang, Yuan
Li, Hang
Li, Wenchao
Zhang, Fengwen
Ouyang, Wenbin
Wang, Shouzheng
Luo, Zhiling
Wang, Jinduo
Chen, Yan
Xia, Ruibing
Li, Yakun
Pan, Xiangbin
author_facet Liu, Zeye
Huang, Yuan
Li, Hang
Li, Wenchao
Zhang, Fengwen
Ouyang, Wenbin
Wang, Shouzheng
Luo, Zhiling
Wang, Jinduo
Chen, Yan
Xia, Ruibing
Li, Yakun
Pan, Xiangbin
author_sort Liu, Zeye
collection PubMed
description OBJECTIVE: Echocardiography (ECG) is the most common method used to diagnose heart failure (HF). However, its accuracy relies on the experience of the operator. Additionally, the video format of the data makes it challenging for patients to bring them to referrals and reexaminations. Therefore, this study used a deep learning approach to assist physicians in assessing cardiac function to promote the standardization of echocardiographic findings and compatibility of dynamic and static ultrasound data. METHODS: A deep spatio-temporal convolutional model r2plus1d-Pan (trained on dynamic data and applied to static data) was improved and trained using the idea of “regression training combined with classification application,” which can be generalized to dynamic ECG and static cardiac ultrasound views to identify HF with a reduced ejection fraction (EF < 40%). Additionally, three independent datasets containing 8976 cardiac ultrasound views and 10085 cardiac ultrasound videos were established. Subsequently, a multinational, multi-center dataset of EF was labeled. Furthermore, model training and independent validation were performed. Finally, 15 registered ultrasonographers and cardiologists with different working years in three regional hospitals specialized in cardiovascular disease were recruited to compare the results. RESULTS: The proposed deep spatio-temporal convolutional model achieved an area under the receiveroperating characteristic curve (AUC) value of 0.95 (95% confidence interval [CI]: 0.947 to 0.953) on the training set of dynamic ultrasound data and an AUC of 1 (95% CI, 1 to 1) on the independent validation set. Subsequently, the model was applied to the static cardiac ultrasound view (validation set) with simultaneous input of 1, 2, 4, and 8 images of the same heart, with classification accuracies of 85%, 81%, 93%, and 92%, respectively. On the static data, the classification accuracy of the artificial intelligence (AI) model was comparable with the best performance of ultrasonographers and cardiologists with more than 3 working years (P = 0.344), but significantly better than the median level (P = 0.0000008). CONCLUSION: A new deep spatio-temporal convolution model was constructed to identify patients with HF with reduced EF accurately (< 40%) using dynamic and static cardiac ultrasound images. The model outperformed the diagnostic performance of most senior specialists. This may be the first HF-related AI diagnostic model compatible with multi-dimensional cardiac ultrasound data, and may thereby contribute to the improvement of HF diagnosis. Additionally, the model enables patients to carry “on-the-go” static ultrasound reports for referral and reexamination, thus saving healthcare resources.
format Online
Article
Text
id pubmed-10680380
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher De Gruyter
record_format MEDLINE/PubMed
spelling pubmed-106803802023-07-05 A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound Liu, Zeye Huang, Yuan Li, Hang Li, Wenchao Zhang, Fengwen Ouyang, Wenbin Wang, Shouzheng Luo, Zhiling Wang, Jinduo Chen, Yan Xia, Ruibing Li, Yakun Pan, Xiangbin J Transl Int Med Original Article OBJECTIVE: Echocardiography (ECG) is the most common method used to diagnose heart failure (HF). However, its accuracy relies on the experience of the operator. Additionally, the video format of the data makes it challenging for patients to bring them to referrals and reexaminations. Therefore, this study used a deep learning approach to assist physicians in assessing cardiac function to promote the standardization of echocardiographic findings and compatibility of dynamic and static ultrasound data. METHODS: A deep spatio-temporal convolutional model r2plus1d-Pan (trained on dynamic data and applied to static data) was improved and trained using the idea of “regression training combined with classification application,” which can be generalized to dynamic ECG and static cardiac ultrasound views to identify HF with a reduced ejection fraction (EF < 40%). Additionally, three independent datasets containing 8976 cardiac ultrasound views and 10085 cardiac ultrasound videos were established. Subsequently, a multinational, multi-center dataset of EF was labeled. Furthermore, model training and independent validation were performed. Finally, 15 registered ultrasonographers and cardiologists with different working years in three regional hospitals specialized in cardiovascular disease were recruited to compare the results. RESULTS: The proposed deep spatio-temporal convolutional model achieved an area under the receiveroperating characteristic curve (AUC) value of 0.95 (95% confidence interval [CI]: 0.947 to 0.953) on the training set of dynamic ultrasound data and an AUC of 1 (95% CI, 1 to 1) on the independent validation set. Subsequently, the model was applied to the static cardiac ultrasound view (validation set) with simultaneous input of 1, 2, 4, and 8 images of the same heart, with classification accuracies of 85%, 81%, 93%, and 92%, respectively. On the static data, the classification accuracy of the artificial intelligence (AI) model was comparable with the best performance of ultrasonographers and cardiologists with more than 3 working years (P = 0.344), but significantly better than the median level (P = 0.0000008). CONCLUSION: A new deep spatio-temporal convolution model was constructed to identify patients with HF with reduced EF accurately (< 40%) using dynamic and static cardiac ultrasound images. The model outperformed the diagnostic performance of most senior specialists. This may be the first HF-related AI diagnostic model compatible with multi-dimensional cardiac ultrasound data, and may thereby contribute to the improvement of HF diagnosis. Additionally, the model enables patients to carry “on-the-go” static ultrasound reports for referral and reexamination, thus saving healthcare resources. De Gruyter 2023-07-05 /pmc/articles/PMC10680380/ /pubmed/38025953 http://dx.doi.org/10.2478/jtim-2023-0088 Text en © 2023 Zeye Liu, Yuan Huang, Hang Li, Wenchao Li, Fengwen Zhang, Wenbin Ouyang, Shouzheng Wang, Zhiling Luo, Jinduo Wang, Yan Chen, Ruibing Xia, Yakun Li, Xiangbin Pan, published by De Gruyter on behalf of the SMP https://creativecommons.org/licenses/by/4.0/This work is licensed under the Creative Commons Attribution 4.0 International License.
spellingShingle Original Article
Liu, Zeye
Huang, Yuan
Li, Hang
Li, Wenchao
Zhang, Fengwen
Ouyang, Wenbin
Wang, Shouzheng
Luo, Zhiling
Wang, Jinduo
Chen, Yan
Xia, Ruibing
Li, Yakun
Pan, Xiangbin
A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title_full A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title_fullStr A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title_full_unstemmed A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title_short A generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
title_sort generalized deep learning model for heart failure diagnosis using dynamic and static ultrasound
topic Original Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10680380/
https://www.ncbi.nlm.nih.gov/pubmed/38025953
http://dx.doi.org/10.2478/jtim-2023-0088
work_keys_str_mv AT liuzeye ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT huangyuan ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT lihang ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT liwenchao ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT zhangfengwen ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT ouyangwenbin ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT wangshouzheng ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT luozhiling ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT wangjinduo ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT chenyan ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT xiaruibing ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT liyakun ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT panxiangbin ageneralizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT liuzeye generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT huangyuan generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT lihang generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT liwenchao generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT zhangfengwen generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT ouyangwenbin generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT wangshouzheng generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT luozhiling generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT wangjinduo generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT chenyan generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT xiaruibing generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT liyakun generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound
AT panxiangbin generalizeddeeplearningmodelforheartfailurediagnosisusingdynamicandstaticultrasound