Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning
Abstract Echocardiography, a simple and noninvasive tool, is the first choice for screening pulmonary hypertension (PH). However, accurate assessment of PH, incorporating both the pulmonary artery pressures and additional signs for PH remained unsatisfied. Thus, this study aimed to develop a machine...
Main Authors: | , , , , , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2023-07-01
|
Series: | Pulmonary Circulation |
Subjects: | |
Online Access: | https://doi.org/10.1002/pul2.12272 |
_version_ | 1797672378777468928 |
---|---|
author | Zuwei Liao Kaikai Liu Shangwei Ding Qinhua Zhao Yong Jiang Lan Wang Taoran Huang LiFang Yang Dongling Luo Erlei Zhang Yu Zhang Caojin Zhang Xiaowei Xu Hongwen Fei |
author_facet | Zuwei Liao Kaikai Liu Shangwei Ding Qinhua Zhao Yong Jiang Lan Wang Taoran Huang LiFang Yang Dongling Luo Erlei Zhang Yu Zhang Caojin Zhang Xiaowei Xu Hongwen Fei |
author_sort | Zuwei Liao |
collection | DOAJ |
description | Abstract Echocardiography, a simple and noninvasive tool, is the first choice for screening pulmonary hypertension (PH). However, accurate assessment of PH, incorporating both the pulmonary artery pressures and additional signs for PH remained unsatisfied. Thus, this study aimed to develop a machine learning (ML) model that can automatically evaluate the probability of PH. This cohort included data from 346 (275 for training set and internal validation set and 71 for external validation set) patients with suspected PH patients and receiving right heart catheterization. Echocardiographic images on parasternal short axis‐papillary muscle level (PSAX‐PML) view from all patients were collected, labeled, and preprocessed. Local features from each image were extracted and subsequently integrated to build a ML model. By adjusting the parameters of the model, the model with the best prediction effect is finally constructed. We used receiver‐operating characteristic analysis to evaluate model performance and compared the ML model with the traditional methods. The accuracy of the ML model for diagnosis of PH was significantly higher than the traditional method (0.945 vs. 0.892, p = 0.027 [area under the curve [AUC]]). Similar findings were observed in subgroup analysis and validated in the external validation set (AUC = 0.950 [95% CI: 0.897−1.000]). In summary, ML methods could automatically extract features from traditional PSAX‐PML view and automatically assess the probability of PH, which were found to outperform traditional echocardiographic assessments. |
first_indexed | 2024-03-11T21:29:11Z |
format | Article |
id | doaj.art-c40b43749429468586f2d1ee70ea2aea |
institution | Directory Open Access Journal |
issn | 2045-8940 |
language | English |
last_indexed | 2024-03-11T21:29:11Z |
publishDate | 2023-07-01 |
publisher | Wiley |
record_format | Article |
series | Pulmonary Circulation |
spelling | doaj.art-c40b43749429468586f2d1ee70ea2aea2023-09-27T13:33:25ZengWileyPulmonary Circulation2045-89402023-07-01133n/an/a10.1002/pul2.12272Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learningZuwei Liao0Kaikai Liu1Shangwei Ding2Qinhua Zhao3Yong Jiang4Lan Wang5Taoran Huang6LiFang Yang7Dongling Luo8Erlei Zhang9Yu Zhang10Caojin Zhang11Xiaowei Xu12Hongwen Fei13Shantou University Medical College Shantou Guangdong ChinaSchool of Information Engineering Northwest A&F University Yangling Shanxi ChinaDepartment of Ultrasound The First Affiliated Hospital of Guangzhou Medical University Guangzhou Guangdong ChinaDepartment of Pulmonary Circulation Shanghai Pulmonary Hospital, Tongji University School of Medicine Shanghai ChinaState Key Laboratory of Cardiovascular Disease, Department of Echocardiography National Center for Cardiovascular Diseases, Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College Beijing ChinaDepartment of Pulmonary Circulation Shanghai Pulmonary Hospital, Tongji University School of Medicine Shanghai ChinaShantou University Medical College Shantou Guangdong ChinaGuangdong Cardiovascular Institute, Guangdong Provincial People's Hospital (Guangdong Academy of Medical Sciences) Southern Medical University Guangzhou Guangdong ChinaGuangdong Cardiovascular Institute, Guangdong Provincial People's Hospital (Guangdong Academy of Medical Sciences) Southern Medical University Guangzhou Guangdong ChinaSchool of Information Engineering Northwest A&F University Yangling Shanxi ChinaSchool of Information Engineering Northwest A&F University Yangling Shanxi ChinaGuangdong Cardiovascular Institute, Guangdong Provincial People's Hospital (Guangdong Academy of Medical Sciences) Southern Medical University Guangzhou Guangdong ChinaGuangdong Cardiovascular Institute, Guangdong Provincial People's Hospital (Guangdong Academy of Medical Sciences) Southern Medical University Guangzhou Guangdong ChinaShantou University Medical College Shantou Guangdong ChinaAbstract Echocardiography, a simple and noninvasive tool, is the first choice for screening pulmonary hypertension (PH). However, accurate assessment of PH, incorporating both the pulmonary artery pressures and additional signs for PH remained unsatisfied. Thus, this study aimed to develop a machine learning (ML) model that can automatically evaluate the probability of PH. This cohort included data from 346 (275 for training set and internal validation set and 71 for external validation set) patients with suspected PH patients and receiving right heart catheterization. Echocardiographic images on parasternal short axis‐papillary muscle level (PSAX‐PML) view from all patients were collected, labeled, and preprocessed. Local features from each image were extracted and subsequently integrated to build a ML model. By adjusting the parameters of the model, the model with the best prediction effect is finally constructed. We used receiver‐operating characteristic analysis to evaluate model performance and compared the ML model with the traditional methods. The accuracy of the ML model for diagnosis of PH was significantly higher than the traditional method (0.945 vs. 0.892, p = 0.027 [area under the curve [AUC]]). Similar findings were observed in subgroup analysis and validated in the external validation set (AUC = 0.950 [95% CI: 0.897−1.000]). In summary, ML methods could automatically extract features from traditional PSAX‐PML view and automatically assess the probability of PH, which were found to outperform traditional echocardiographic assessments.https://doi.org/10.1002/pul2.12272echocardiographymachine learningmean pulmonary artery pressurepulmonary hypertension |
spellingShingle | Zuwei Liao Kaikai Liu Shangwei Ding Qinhua Zhao Yong Jiang Lan Wang Taoran Huang LiFang Yang Dongling Luo Erlei Zhang Yu Zhang Caojin Zhang Xiaowei Xu Hongwen Fei Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning Pulmonary Circulation echocardiography machine learning mean pulmonary artery pressure pulmonary hypertension |
title | Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
title_full | Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
title_fullStr | Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
title_full_unstemmed | Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
title_short | Automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
title_sort | automatic echocardiographic evaluation of the probability of pulmonary hypertension using machine learning |
topic | echocardiography machine learning mean pulmonary artery pressure pulmonary hypertension |
url | https://doi.org/10.1002/pul2.12272 |
work_keys_str_mv | AT zuweiliao automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT kaikailiu automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT shangweiding automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT qinhuazhao automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT yongjiang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT lanwang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT taoranhuang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT lifangyang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT donglingluo automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT erleizhang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT yuzhang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT caojinzhang automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT xiaoweixu automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning AT hongwenfei automaticechocardiographicevaluationoftheprobabilityofpulmonaryhypertensionusingmachinelearning |