Machine learning algorithm using publicly available echo database for simplified "visual estimation" of left ventricular ejection fraction
Left ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of endocardial borders. This significantly limits usability in bedside clinical applications, where ultrasound automation is needed most. To create a...
Saved in:
| Published in: | World journal of experimental medicine Vol. 12; no. 2; p. 16 |
|---|---|
| Main Authors: | , |
| Format: | Journal Article |
| Language: | English |
| Published: |
United States
20.03.2022
|
| Subjects: | |
| ISSN: | 2220-315X, 2220-315X |
| Online Access: | Get more information |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Left ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of endocardial borders. This significantly limits usability in bedside clinical applications, where ultrasound automation is needed most.
To create a simple deep learning (DL) regression-type algorithm to visually estimate left ventricular (LV) ejection fraction (EF) from a public database of actual patient echo examinations and compare results to echocardiography laboratory EF calculations.
A simple DL architecture previously proven to perform well on ultrasound image analysis, VGG16, was utilized as a base architecture running within a long short term memory algorithm for sequential image (video) analysis. After obtaining permission to use the Stanford EchoNet-Dynamic database, researchers randomly removed approximately 15% of the approximately 10036 echo apical 4-chamber videos for later performance testing. All database echo examinations were read as part of comprehensive echocardiography study performance and were coupled with EF, end systolic and diastolic volumes, key frames and coordinates for LV endocardial tracing in csv file. To better reflect point-of-care ultrasound (POCUS) clinical settings and time pressure, the algorithm was trained on echo video correlated with calculated ejection fraction without incorporating additional volume, measurement and coordinate data. Seventy percent of the original data was used for algorithm training and 15% for validation during training. The previously randomly separated 15% (1263 echo videos) was used for algorithm performance testing after training completion. Given the inherent variability of echo EF measurement and field standards for evaluating algorithm accuracy, mean absolute error (MAE) and root mean square error (RMSE) calculations were made on algorithm EF results compared to Echo Lab calculated EF. Bland-Atlman calculation was also performed. MAE for skilled echocardiographers has been established to range from 4% to 5%.
The DL algorithm visually estimated EF had a MAE of 8.08% (95%CI 7.60 to 8.55) suggesting good performance compared to highly skill humans. The RMSE was 11.98 and correlation of 0.348.
This experimental simplified DL algorithm showed promise and proved reasonably accurate at visually estimating LV EF from short real time echo video clips. Less burdensome than complex DL approaches used for EF calculation, such an approach may be more optimal for POCUS settings once improved upon by future research and development. |
|---|---|
| AbstractList | Left ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of endocardial borders. This significantly limits usability in bedside clinical applications, where ultrasound automation is needed most.
To create a simple deep learning (DL) regression-type algorithm to visually estimate left ventricular (LV) ejection fraction (EF) from a public database of actual patient echo examinations and compare results to echocardiography laboratory EF calculations.
A simple DL architecture previously proven to perform well on ultrasound image analysis, VGG16, was utilized as a base architecture running within a long short term memory algorithm for sequential image (video) analysis. After obtaining permission to use the Stanford EchoNet-Dynamic database, researchers randomly removed approximately 15% of the approximately 10036 echo apical 4-chamber videos for later performance testing. All database echo examinations were read as part of comprehensive echocardiography study performance and were coupled with EF, end systolic and diastolic volumes, key frames and coordinates for LV endocardial tracing in csv file. To better reflect point-of-care ultrasound (POCUS) clinical settings and time pressure, the algorithm was trained on echo video correlated with calculated ejection fraction without incorporating additional volume, measurement and coordinate data. Seventy percent of the original data was used for algorithm training and 15% for validation during training. The previously randomly separated 15% (1263 echo videos) was used for algorithm performance testing after training completion. Given the inherent variability of echo EF measurement and field standards for evaluating algorithm accuracy, mean absolute error (MAE) and root mean square error (RMSE) calculations were made on algorithm EF results compared to Echo Lab calculated EF. Bland-Atlman calculation was also performed. MAE for skilled echocardiographers has been established to range from 4% to 5%.
The DL algorithm visually estimated EF had a MAE of 8.08% (95%CI 7.60 to 8.55) suggesting good performance compared to highly skill humans. The RMSE was 11.98 and correlation of 0.348.
This experimental simplified DL algorithm showed promise and proved reasonably accurate at visually estimating LV EF from short real time echo video clips. Less burdensome than complex DL approaches used for EF calculation, such an approach may be more optimal for POCUS settings once improved upon by future research and development. Left ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of endocardial borders. This significantly limits usability in bedside clinical applications, where ultrasound automation is needed most.BACKGROUNDLeft ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of endocardial borders. This significantly limits usability in bedside clinical applications, where ultrasound automation is needed most.To create a simple deep learning (DL) regression-type algorithm to visually estimate left ventricular (LV) ejection fraction (EF) from a public database of actual patient echo examinations and compare results to echocardiography laboratory EF calculations.AIMTo create a simple deep learning (DL) regression-type algorithm to visually estimate left ventricular (LV) ejection fraction (EF) from a public database of actual patient echo examinations and compare results to echocardiography laboratory EF calculations.A simple DL architecture previously proven to perform well on ultrasound image analysis, VGG16, was utilized as a base architecture running within a long short term memory algorithm for sequential image (video) analysis. After obtaining permission to use the Stanford EchoNet-Dynamic database, researchers randomly removed approximately 15% of the approximately 10036 echo apical 4-chamber videos for later performance testing. All database echo examinations were read as part of comprehensive echocardiography study performance and were coupled with EF, end systolic and diastolic volumes, key frames and coordinates for LV endocardial tracing in csv file. To better reflect point-of-care ultrasound (POCUS) clinical settings and time pressure, the algorithm was trained on echo video correlated with calculated ejection fraction without incorporating additional volume, measurement and coordinate data. Seventy percent of the original data was used for algorithm training and 15% for validation during training. The previously randomly separated 15% (1263 echo videos) was used for algorithm performance testing after training completion. Given the inherent variability of echo EF measurement and field standards for evaluating algorithm accuracy, mean absolute error (MAE) and root mean square error (RMSE) calculations were made on algorithm EF results compared to Echo Lab calculated EF. Bland-Atlman calculation was also performed. MAE for skilled echocardiographers has been established to range from 4% to 5%.METHODSA simple DL architecture previously proven to perform well on ultrasound image analysis, VGG16, was utilized as a base architecture running within a long short term memory algorithm for sequential image (video) analysis. After obtaining permission to use the Stanford EchoNet-Dynamic database, researchers randomly removed approximately 15% of the approximately 10036 echo apical 4-chamber videos for later performance testing. All database echo examinations were read as part of comprehensive echocardiography study performance and were coupled with EF, end systolic and diastolic volumes, key frames and coordinates for LV endocardial tracing in csv file. To better reflect point-of-care ultrasound (POCUS) clinical settings and time pressure, the algorithm was trained on echo video correlated with calculated ejection fraction without incorporating additional volume, measurement and coordinate data. Seventy percent of the original data was used for algorithm training and 15% for validation during training. The previously randomly separated 15% (1263 echo videos) was used for algorithm performance testing after training completion. Given the inherent variability of echo EF measurement and field standards for evaluating algorithm accuracy, mean absolute error (MAE) and root mean square error (RMSE) calculations were made on algorithm EF results compared to Echo Lab calculated EF. Bland-Atlman calculation was also performed. MAE for skilled echocardiographers has been established to range from 4% to 5%.The DL algorithm visually estimated EF had a MAE of 8.08% (95%CI 7.60 to 8.55) suggesting good performance compared to highly skill humans. The RMSE was 11.98 and correlation of 0.348.RESULTSThe DL algorithm visually estimated EF had a MAE of 8.08% (95%CI 7.60 to 8.55) suggesting good performance compared to highly skill humans. The RMSE was 11.98 and correlation of 0.348.This experimental simplified DL algorithm showed promise and proved reasonably accurate at visually estimating LV EF from short real time echo video clips. Less burdensome than complex DL approaches used for EF calculation, such an approach may be more optimal for POCUS settings once improved upon by future research and development.CONCLUSIONThis experimental simplified DL algorithm showed promise and proved reasonably accurate at visually estimating LV EF from short real time echo video clips. Less burdensome than complex DL approaches used for EF calculation, such an approach may be more optimal for POCUS settings once improved upon by future research and development. |
| Author | Blaivas, Michael Blaivas, Laura |
| Author_xml | – sequence: 1 givenname: Michael surname: Blaivas fullname: Blaivas, Michael email: mike@blaivas.org organization: Department of Medicine, University of South Carolina School of Medicine, Roswell, GA 30076, United States. mike@blaivas.org – sequence: 2 givenname: Laura surname: Blaivas fullname: Blaivas, Laura organization: Department of Environmental Science, Michigan State University, Roswell, Georgia 30076, United States |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/35433318$$D View this record in MEDLINE/PubMed |
| BookMark | eNpNUE1PAjEUbAxGELl6NA0nL-D2tV2WoyF-JRgvmngjb7uvUNLdxXYXw1_wVwuKiac3LzOZzMw561R1RYxdimSs1VTefK6pHG8FjB2MRXrCegCQjKTQ751_uMsGMa6TJBECphqyM9aVWkkpRdZjX89oVq4i7glD5aolR7-sg2tWJW_j4d-0uXfG7zhu0XnMPXEyq5oX2GCOkbitA4-u3HhnHRV8uHWxRc8pNq7ExtXVkNd2728bvqWqCc60HgOnNZkDy23AH3DBTi36SIPj7bO3-7vX2eNo_vLwNLudj4xMdDpCSElLbTRQqkxm7ISEIgUwmaKa2kztWSOxAJ1rlSko0IpCqQwRVG4LCX12_eu7CfVHu4-5KF005D1WVLdxAamGRAJkB-nVUdrmJRWLTdhXCrvF337wDTSVeEI |
| CitedBy_id | crossref_primary_10_1016_j_ultrasmedbio_2025_03_015 crossref_primary_10_1109_TIM_2025_3553885 crossref_primary_10_1016_j_ccc_2025_02_008 crossref_primary_10_3390_diagnostics13132155 crossref_primary_10_7717_peerj_cs_2506 crossref_primary_10_3390_biomedicines12102324 crossref_primary_10_1016_j_ejmp_2024_104505 crossref_primary_10_1007_s10278_024_01336_y |
| ContentType | Journal Article |
| Copyright | The Author(s) 2022. Published by Baishideng Publishing Group Inc. All rights reserved. |
| Copyright_xml | – notice: The Author(s) 2022. Published by Baishideng Publishing Group Inc. All rights reserved. |
| DBID | NPM 7X8 |
| DOI | 10.5493/wjem.v12.i2.16 |
| DatabaseName | PubMed MEDLINE - Academic |
| DatabaseTitle | PubMed MEDLINE - Academic |
| DatabaseTitleList | PubMed MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Medicine |
| EISSN | 2220-315X |
| ExternalDocumentID | 35433318 |
| Genre | Journal Article |
| GroupedDBID | 5VR 8WL ADBBV ALMA_UNASSIGNED_HOLDINGS AOIJS CCEZO CHBEP CIEJG GX1 HYE NPM OK1 RPM 7X8 |
| ID | FETCH-LOGICAL-c3056-a26e535c52e64c8cf7e14e42279a49f84e53c3ad25b54842daf1d448aa24bfd32 |
| IEDL.DBID | 7X8 |
| ISSN | 2220-315X |
| IngestDate | Fri Jul 11 07:17:41 EDT 2025 Mon Jul 21 05:15:12 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | false |
| IsScholarly | true |
| Issue | 2 |
| Keywords | Deep learning Echocardiography Artificial intelligence Cardiac Point-of-care-ultrasound Ejection fraction |
| Language | English |
| License | The Author(s) 2022. Published by Baishideng Publishing Group Inc. All rights reserved. |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c3056-a26e535c52e64c8cf7e14e42279a49f84e53c3ad25b54842daf1d448aa24bfd32 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| OpenAccessLink | https://doi.org/10.5493/wjem.v12.i2.16 |
| PMID | 35433318 |
| PQID | 2652032283 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_2652032283 pubmed_primary_35433318 |
| PublicationCentury | 2000 |
| PublicationDate | 2022-Mar-20 20220320 |
| PublicationDateYYYYMMDD | 2022-03-20 |
| PublicationDate_xml | – month: 03 year: 2022 text: 2022-Mar-20 day: 20 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | World journal of experimental medicine |
| PublicationTitleAlternate | World J Exp Med |
| PublicationYear | 2022 |
| SSID | ssj0001129528 |
| Score | 2.218489 |
| Snippet | Left ventricular ejection fraction calculation automation typically requires complex algorithms and is dependent of optimal visualization and tracing of... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| StartPage | 16 |
| Title | Machine learning algorithm using publicly available echo database for simplified "visual estimation" of left ventricular ejection fraction |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/35433318 https://www.proquest.com/docview/2652032283 |
| Volume | 12 |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1JS8NAFB7cEC_uS90YxWtqMkuSnkTE4sEWDwq5hcksNaVNtakV_4K_2veSiCdB8JJLyCTMe3nbfO99hFw4EGsWCOMFToWeiJzwMmasFzvBFMsi5ZuKteQ-6vfjJOk8NAW3soFVftvEylCbicYa-SULJZJ9gze8enn1kDUKT1cbCo1FsswhlEFIV5TEPzUWcGayolcFL4jmRib13EZIivjl-9CO2_OAtXPWDsLfI8zK03Q3_vuNm2S9iTHpda0UW2TBFttktdecou-Qz16FoLS0oYwYUDUawDqz5zFFHPyA1sOvRx9UzVU-wu4qasFOUsSTot-jEOrSMkc0uoMYlp7P8_INXokjO-peyHM6cbC-m1EEVFZVRjWldlghvwrqpnVDxS556t4-3tx5DSeDpzHZ8BQLreRSS2ZDoWPtIhsIK3AOoRIdFwu4q7kyTGaQCwlmlAsMpIBKMZE5w9keWSomhT0g1NdSaWxRYYILA3JyXMnQdDKRMT_MXIucfe9yCjqPBxmqsJO3Mv3Z5xbZr0WVvtTDOVIuBedgqA7_8PQRWWPYzeBzMBbHZNnBH29PyIqez_JyelopE1z7D70v5wDXQw |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Machine+learning+algorithm+using+publicly+available+echo+database+for+simplified+%22visual+estimation%22+of+left+ventricular+ejection+fraction&rft.jtitle=World+journal+of+experimental+medicine&rft.au=Blaivas%2C+Michael&rft.au=Blaivas%2C+Laura&rft.date=2022-03-20&rft.issn=2220-315X&rft.eissn=2220-315X&rft.volume=12&rft.issue=2&rft.spage=16&rft_id=info:doi/10.5493%2Fwjem.v12.i2.16&rft_id=info%3Apmid%2F35433318&rft_id=info%3Apmid%2F35433318&rft.externalDocID=35433318 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2220-315X&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2220-315X&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2220-315X&client=summon |