Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography
BACKGROUND: Ejection fraction (EF) estimation informs patient plans in the ICU, and low EF can indicate ventricular systolic dysfunction, which increases the risk of adverse events including heart failure. Automated echocardiography models are an attractive solution for high-variance human EF estima...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Elsevier BV
2025
|
| Online Access: | https://hdl.handle.net/1721.1/158123 |
| _version_ | 1824458011319468032 |
|---|---|
| author | Chung, Daniel J Lee, Somin Mindy Kaker, Vasu Zhao, Yongyi Bin, Irbaz Perera, Sudheesha Sasankan, Prabhu Tang, George Kazzi, Brigitte Kuo, Po-Chih Celi, Leo A Kpodonu, Jacques |
| author2 | Sloan School of Management |
| author_facet | Sloan School of Management Chung, Daniel J Lee, Somin Mindy Kaker, Vasu Zhao, Yongyi Bin, Irbaz Perera, Sudheesha Sasankan, Prabhu Tang, George Kazzi, Brigitte Kuo, Po-Chih Celi, Leo A Kpodonu, Jacques |
| author_sort | Chung, Daniel J |
| collection | MIT |
| description | BACKGROUND: Ejection fraction (EF) estimation informs patient plans in the ICU, and low EF can indicate ventricular systolic dysfunction, which increases the risk of adverse events including heart failure. Automated echocardiography models are an attractive solution for high-variance human EF estimation, and key to this goal are echocardiogram vector embeddings, which are a critical resource for computational researchers. OBJECTIVES: The authors aimed to extract the vector embeddings from each echocardiogram in the EchoNet dataset using a classifier trained to classify EF as healthy (>50%) or unhealthy (<= 50%) to create an embeddings dataset for computational researchers. METHODS: We repurposed an R3D transformer to classify whether patient EF is below or above 50%. Training, validation, and testing were done on the EchoNet dataset of 10,030 echocardiograms, and the resulting model generated embeddings for each of these videos. RESULTS: We extracted 400-dimensional vector embeddings for each of the 10,030 EchoNet echocardiograms using the trained R3D model, which achieved a test AUC of 0.916 and 87.5% accuracy, approaching the performance of comparable studies. CONCLUSIONS: We present 10,030 vector embeddings learned by this model as a resource to the cardiology research community, as well as the trained model itself. These vectors enable algorithmic improvements and multimodal applications within automated echocardiography, benefitting the research community and those with ventricular systolic dysfunction (https://github.com/Team-Echo-MIT/r3d-v0-embeddings). |
| first_indexed | 2025-02-19T04:19:06Z |
| format | Article |
| id | mit-1721.1/158123 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2025-02-19T04:19:06Z |
| publishDate | 2025 |
| publisher | Elsevier BV |
| record_format | dspace |
| spelling | mit-1721.1/1581232025-01-28T21:32:22Z Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography Chung, Daniel J Lee, Somin Mindy Kaker, Vasu Zhao, Yongyi Bin, Irbaz Perera, Sudheesha Sasankan, Prabhu Tang, George Kazzi, Brigitte Kuo, Po-Chih Celi, Leo A Kpodonu, Jacques Sloan School of Management Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Harvard--MIT Program in Health Sciences and Technology. Laboratory for Computational Physiology BACKGROUND: Ejection fraction (EF) estimation informs patient plans in the ICU, and low EF can indicate ventricular systolic dysfunction, which increases the risk of adverse events including heart failure. Automated echocardiography models are an attractive solution for high-variance human EF estimation, and key to this goal are echocardiogram vector embeddings, which are a critical resource for computational researchers. OBJECTIVES: The authors aimed to extract the vector embeddings from each echocardiogram in the EchoNet dataset using a classifier trained to classify EF as healthy (>50%) or unhealthy (<= 50%) to create an embeddings dataset for computational researchers. METHODS: We repurposed an R3D transformer to classify whether patient EF is below or above 50%. Training, validation, and testing were done on the EchoNet dataset of 10,030 echocardiograms, and the resulting model generated embeddings for each of these videos. RESULTS: We extracted 400-dimensional vector embeddings for each of the 10,030 EchoNet echocardiograms using the trained R3D model, which achieved a test AUC of 0.916 and 87.5% accuracy, approaching the performance of comparable studies. CONCLUSIONS: We present 10,030 vector embeddings learned by this model as a resource to the cardiology research community, as well as the trained model itself. These vectors enable algorithmic improvements and multimodal applications within automated echocardiography, benefitting the research community and those with ventricular systolic dysfunction (https://github.com/Team-Echo-MIT/r3d-v0-embeddings). 2025-01-28T21:32:20Z 2025-01-28T21:32:20Z 2024-09 2025-01-28T21:24:44Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/158123 Chung, Daniel J, Lee, Somin Mindy, Kaker, Vasu, Zhao, Yongyi, Bin, Irbaz et al. 2024. "Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography." JACC: Advances, 3 (9). en 10.1016/j.jacadv.2024.101196 JACC: Advances Creative Commons Attribution-NonCommercial-NoDerivs License https://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf Elsevier BV Elsevier BV |
| spellingShingle | Chung, Daniel J Lee, Somin Mindy Kaker, Vasu Zhao, Yongyi Bin, Irbaz Perera, Sudheesha Sasankan, Prabhu Tang, George Kazzi, Brigitte Kuo, Po-Chih Celi, Leo A Kpodonu, Jacques Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title | Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title_full | Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title_fullStr | Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title_full_unstemmed | Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title_short | Echocardiogram Vector Embeddings Via R3D Transformer for the Advancement of Automated Echocardiography |
| title_sort | echocardiogram vector embeddings via r3d transformer for the advancement of automated echocardiography |
| url | https://hdl.handle.net/1721.1/158123 |
| work_keys_str_mv | AT chungdanielj echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT leesominmindy echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT kakervasu echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT zhaoyongyi echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT binirbaz echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT pererasudheesha echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT sasankanprabhu echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT tanggeorge echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT kazzibrigitte echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT kuopochih echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT celileoa echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography AT kpodonujacques echocardiogramvectorembeddingsviar3dtransformerfortheadvancementofautomatedechocardiography |