Myocardial strain analysis of echocardiography based on deep learning
BackgroundStrain analysis provides more thorough spatiotemporal signatures for myocardial contraction, which is helpful for early detection of cardiac insufficiency. The use of deep learning (DL) to automatically measure myocardial strain from echocardiogram videos has garnered recent attention. How...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-12-01
|
| Series: | Frontiers in Cardiovascular Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fcvm.2022.1067760/full |
| _version_ | 1828164907339612160 |
|---|---|
| author | Yinlong Deng Yinlong Deng Peiwei Cai Li Zhang Xiongcheng Cao Xiongcheng Cao Yequn Chen Shiyan Jiang Zhemin Zhuang Bin Wang |
| author_facet | Yinlong Deng Yinlong Deng Peiwei Cai Li Zhang Xiongcheng Cao Xiongcheng Cao Yequn Chen Shiyan Jiang Zhemin Zhuang Bin Wang |
| author_sort | Yinlong Deng |
| collection | DOAJ |
| description | BackgroundStrain analysis provides more thorough spatiotemporal signatures for myocardial contraction, which is helpful for early detection of cardiac insufficiency. The use of deep learning (DL) to automatically measure myocardial strain from echocardiogram videos has garnered recent attention. However, the development of key techniques including segmentation and motion estimation remains a challenge. In this work, we developed a novel DL-based framework for myocardial segmentation and motion estimation to generate strain measures from echocardiogram videos.MethodsThree-dimensional (3D) Convolutional Neural Network (CNN) was developed for myocardial segmentation and optical flow network for motion estimation. The segmentation network was used to define the region of interest (ROI), and the optical flow network was used to estimate the pixel motion in the ROI. We performed a model architecture search to identify the optimal base architecture for motion estimation. The final workflow design and associated hyperparameters are the result of a careful implementation. In addition, we compared the DL model with a traditional speck tracking algorithm on an independent, external clinical data. Each video was double-blind measured by an ultrasound expert and a DL expert using speck tracking echocardiography (STE) and DL method, respectively.ResultsThe DL method successfully performed automatic segmentation, motion estimation, and global longitudinal strain (GLS) measurements in all examinations. The 3D segmentation has better spatio-temporal smoothness, average dice correlation reaches 0.82, and the effect of target frame is better than that of previous 2D networks. The best motion estimation network achieved an average end-point error of 0.05 ± 0.03 mm per frame, better than previously reported state-of-the-art. The DL method showed no significant difference relative to the traditional method in GLS measurement, Spearman correlation of 0.90 (p < 0.001) and mean bias −1.2 ± 1.5%.ConclusionIn conclusion, our method exhibits better segmentation and motion estimation performance and demonstrates the feasibility of DL method for automatic strain analysis. The DL approach helps reduce time consumption and human effort, which holds great promise for translational research and precision medicine efforts. |
| first_indexed | 2024-04-12T01:34:43Z |
| format | Article |
| id | doaj.art-ff3b73cc905c4ea7997450835771f4bd |
| institution | Directory Open Access Journal |
| issn | 2297-055X |
| language | English |
| last_indexed | 2024-04-12T01:34:43Z |
| publishDate | 2022-12-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Cardiovascular Medicine |
| spelling | doaj.art-ff3b73cc905c4ea7997450835771f4bd2022-12-22T03:53:21ZengFrontiers Media S.A.Frontiers in Cardiovascular Medicine2297-055X2022-12-01910.3389/fcvm.2022.10677601067760Myocardial strain analysis of echocardiography based on deep learningYinlong Deng0Yinlong Deng1Peiwei Cai2Li Zhang3Xiongcheng Cao4Xiongcheng Cao5Yequn Chen6Shiyan Jiang7Zhemin Zhuang8Bin Wang9Department of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Preventive Medicine, Shantou University Medical College, Shantou, ChinaUltrasound Division, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Preventive Medicine, Shantou University Medical College, Shantou, ChinaDepartment of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaDepartment of Electronic Information Engineering, College of Engineering, Shantou University, Shantou, ChinaDepartment of Cardiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, ChinaBackgroundStrain analysis provides more thorough spatiotemporal signatures for myocardial contraction, which is helpful for early detection of cardiac insufficiency. The use of deep learning (DL) to automatically measure myocardial strain from echocardiogram videos has garnered recent attention. However, the development of key techniques including segmentation and motion estimation remains a challenge. In this work, we developed a novel DL-based framework for myocardial segmentation and motion estimation to generate strain measures from echocardiogram videos.MethodsThree-dimensional (3D) Convolutional Neural Network (CNN) was developed for myocardial segmentation and optical flow network for motion estimation. The segmentation network was used to define the region of interest (ROI), and the optical flow network was used to estimate the pixel motion in the ROI. We performed a model architecture search to identify the optimal base architecture for motion estimation. The final workflow design and associated hyperparameters are the result of a careful implementation. In addition, we compared the DL model with a traditional speck tracking algorithm on an independent, external clinical data. Each video was double-blind measured by an ultrasound expert and a DL expert using speck tracking echocardiography (STE) and DL method, respectively.ResultsThe DL method successfully performed automatic segmentation, motion estimation, and global longitudinal strain (GLS) measurements in all examinations. The 3D segmentation has better spatio-temporal smoothness, average dice correlation reaches 0.82, and the effect of target frame is better than that of previous 2D networks. The best motion estimation network achieved an average end-point error of 0.05 ± 0.03 mm per frame, better than previously reported state-of-the-art. The DL method showed no significant difference relative to the traditional method in GLS measurement, Spearman correlation of 0.90 (p < 0.001) and mean bias −1.2 ± 1.5%.ConclusionIn conclusion, our method exhibits better segmentation and motion estimation performance and demonstrates the feasibility of DL method for automatic strain analysis. The DL approach helps reduce time consumption and human effort, which holds great promise for translational research and precision medicine efforts.https://www.frontiersin.org/articles/10.3389/fcvm.2022.1067760/fullechocardiographystraindeep learningsegmentationmotion estimation |
| spellingShingle | Yinlong Deng Yinlong Deng Peiwei Cai Li Zhang Xiongcheng Cao Xiongcheng Cao Yequn Chen Shiyan Jiang Zhemin Zhuang Bin Wang Myocardial strain analysis of echocardiography based on deep learning Frontiers in Cardiovascular Medicine echocardiography strain deep learning segmentation motion estimation |
| title | Myocardial strain analysis of echocardiography based on deep learning |
| title_full | Myocardial strain analysis of echocardiography based on deep learning |
| title_fullStr | Myocardial strain analysis of echocardiography based on deep learning |
| title_full_unstemmed | Myocardial strain analysis of echocardiography based on deep learning |
| title_short | Myocardial strain analysis of echocardiography based on deep learning |
| title_sort | myocardial strain analysis of echocardiography based on deep learning |
| topic | echocardiography strain deep learning segmentation motion estimation |
| url | https://www.frontiersin.org/articles/10.3389/fcvm.2022.1067760/full |
| work_keys_str_mv | AT yinlongdeng myocardialstrainanalysisofechocardiographybasedondeeplearning AT yinlongdeng myocardialstrainanalysisofechocardiographybasedondeeplearning AT peiweicai myocardialstrainanalysisofechocardiographybasedondeeplearning AT lizhang myocardialstrainanalysisofechocardiographybasedondeeplearning AT xiongchengcao myocardialstrainanalysisofechocardiographybasedondeeplearning AT xiongchengcao myocardialstrainanalysisofechocardiographybasedondeeplearning AT yequnchen myocardialstrainanalysisofechocardiographybasedondeeplearning AT shiyanjiang myocardialstrainanalysisofechocardiographybasedondeeplearning AT zheminzhuang myocardialstrainanalysisofechocardiographybasedondeeplearning AT binwang myocardialstrainanalysisofechocardiographybasedondeeplearning |