Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs
Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection. Within 32 239 individuals of the UK Biobank prospective cohort who underwent...
Gespeichert in:
| Veröffentlicht in: | Circulation. Cardiovascular imaging 2021-06, Vol.14 (6), p.e012281-e012281 |
|---|---|
| Hauptverfasser: | , , , , , , , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | e012281 |
|---|---|
| container_issue | 6 |
| container_start_page | e012281 |
| container_title | Circulation. Cardiovascular imaging |
| container_volume | 14 |
| creator | Khurshid, Shaan Friedman, Samuel Pirruccello, James P. Di Achille, Paolo Diamant, Nathaniel Anderson, Christopher D. Ellinor, Patrick T. Batra, Puneet Ho, Jennifer E. Philippakis, Anthony A. Lubitz, Steven A. |
| description | Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection.
Within 32 239 individuals of the UK Biobank prospective cohort who underwent CMR and 12-lead ECG, we trained a convolutional neural network to predict CMR-derived LV mass using 12-lead ECGs (left ventricular mass-artificial intelligence [LVM-AI]). In independent test sets (UK Biobank [n=4903] and Mass General Brigham [MGB, n=1371]), we assessed correlation between LVM-AI predicted and CMR-derived LV mass and compared LVH discrimination using LVM-AI versus traditional ECG-based rules (ie, Sokolow-Lyon, Cornell, lead aVL rule, or any ECG rule). In the UK Biobank and an ambulatory MGB cohort (MGB outcomes, n=28 612), we assessed associations between LVM-AI predicted LVH and incident cardiovascular outcomes using age- and sex-adjusted Cox regression.
LVM-AI predicted LV mass correlated with CMR-derived LV mass in both test sets, although correlation was greater in the UK Biobank (r=0.79) versus MGB (r=0.60, P |
| doi_str_mv | 10.1161/CIRCIMAGING.120.012281 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2541318963</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2541318963</sourcerecordid><originalsourceid>FETCH-LOGICAL-c4703-81d9ca5795ad57c4b27f60bf36a2ca2f6089643ee84a11ff78e6e8b6812288e83</originalsourceid><addsrcrecordid>eNpNkctO3DAUhq2KqlzaV0BesslwfInjLFGAYaQBKtR2a3mcEyYlkwTbAc2u79A37JPUaACxOmfx_f-RvkPIMYMZY4qdVou7anF9Nl_czGeMwwwY55p9IgeslDwD0LD3Yd8nhyH8BlACcv2F7AvJuCoUPyDTOeJIl2h93_b3NA70u8e6dZFW1tetdfTa3vcYW0fvMAy97R3--_P3HH37hHUKNpH-wj761k2d9YkOgdq-plfbEX30w7je0ks_bCjjWTpT04tqHr6Sz43tAn57nUfk5-XFj-oqW97OF9XZMnOyAJFpVpfO5kWZ2zovnFzxolGwaoSy3Fmedl0qKRC1tIw1TaFRoV4p_SJDoxZH5GTXO_rhccIQzaYNDrvO9jhMwfBcMsFSiUio2qHODyF4bMzo2431W8PAvCg3H5SbpNzslKfg8euNabXB-j325jgBcgc8D11EHx666Rm9WaPt4jq1CFHIUmccOAMFAOljAEL8B_fkjVc</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2541318963</pqid></control><display><type>article</type><title>Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs</title><source>MEDLINE</source><source>American Heart Association Journals</source><source>EZB-FREE-00999 freely available EZB journals</source><creator>Khurshid, Shaan ; Friedman, Samuel ; Pirruccello, James P. ; Di Achille, Paolo ; Diamant, Nathaniel ; Anderson, Christopher D. ; Ellinor, Patrick T. ; Batra, Puneet ; Ho, Jennifer E. ; Philippakis, Anthony A. ; Lubitz, Steven A.</creator><creatorcontrib>Khurshid, Shaan ; Friedman, Samuel ; Pirruccello, James P. ; Di Achille, Paolo ; Diamant, Nathaniel ; Anderson, Christopher D. ; Ellinor, Patrick T. ; Batra, Puneet ; Ho, Jennifer E. ; Philippakis, Anthony A. ; Lubitz, Steven A.</creatorcontrib><description>Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection.
Within 32 239 individuals of the UK Biobank prospective cohort who underwent CMR and 12-lead ECG, we trained a convolutional neural network to predict CMR-derived LV mass using 12-lead ECGs (left ventricular mass-artificial intelligence [LVM-AI]). In independent test sets (UK Biobank [n=4903] and Mass General Brigham [MGB, n=1371]), we assessed correlation between LVM-AI predicted and CMR-derived LV mass and compared LVH discrimination using LVM-AI versus traditional ECG-based rules (ie, Sokolow-Lyon, Cornell, lead aVL rule, or any ECG rule). In the UK Biobank and an ambulatory MGB cohort (MGB outcomes, n=28 612), we assessed associations between LVM-AI predicted LVH and incident cardiovascular outcomes using age- and sex-adjusted Cox regression.
LVM-AI predicted LV mass correlated with CMR-derived LV mass in both test sets, although correlation was greater in the UK Biobank (r=0.79) versus MGB (r=0.60, P<0.001 for both). When compared with any ECG rule, LVM-AI demonstrated similar LVH discrimination in the UK Biobank (LVM-AI c-statistic 0.653 [95% CI, 0.608 -0.698] versus any ECG rule c-statistic 0.618 [95% CI, 0.574 -0.663], P=0.11) and superior discrimination in MGB (0.621; 95% CI, 0.592 -0.649 versus 0.588; 95% CI, 0.564 -0.611, P=0.02). LVM-AI-predicted LVH was associated with incident atrial fibrillation, myocardial infarction, heart failure, and ventricular arrhythmias.
Deep learning-inferred LV mass estimates from 12-lead ECGs correlate with CMR-derived LV mass, associate with incident cardiovascular disease, and may improve LVH discrimination compared to traditional ECG rules.</description><identifier>ISSN: 1942-0080</identifier><identifier>ISSN: 1941-9651</identifier><identifier>EISSN: 1942-0080</identifier><identifier>DOI: 10.1161/CIRCIMAGING.120.012281</identifier><identifier>PMID: 34126762</identifier><language>eng</language><publisher>United States: Lippincott Williams & Wilkins</publisher><subject>Artificial Intelligence ; Deep Learning ; Electrocardiography - methods ; Female ; Follow-Up Studies ; Heart Ventricles - diagnostic imaging ; Humans ; Hypertrophy, Left Ventricular - diagnosis ; Male ; Middle Aged ; Prospective Studies</subject><ispartof>Circulation. Cardiovascular imaging, 2021-06, Vol.14 (6), p.e012281-e012281</ispartof><rights>Lippincott Williams & Wilkins</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c4703-81d9ca5795ad57c4b27f60bf36a2ca2f6089643ee84a11ff78e6e8b6812288e83</citedby><cites>FETCH-LOGICAL-c4703-81d9ca5795ad57c4b27f60bf36a2ca2f6089643ee84a11ff78e6e8b6812288e83</cites><orcidid>0000-0002-7987-4768 ; 0000-0001-9256-0678 ; 0000-0002-0053-2002 ; 0000-0002-9599-4866 ; 0000-0002-2067-0533 ; 0000-0001-6822-0593 ; 0000-0001-6088-4037 ; 0000-0002-2840-4539</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,3685,27923,27924</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/34126762$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Khurshid, Shaan</creatorcontrib><creatorcontrib>Friedman, Samuel</creatorcontrib><creatorcontrib>Pirruccello, James P.</creatorcontrib><creatorcontrib>Di Achille, Paolo</creatorcontrib><creatorcontrib>Diamant, Nathaniel</creatorcontrib><creatorcontrib>Anderson, Christopher D.</creatorcontrib><creatorcontrib>Ellinor, Patrick T.</creatorcontrib><creatorcontrib>Batra, Puneet</creatorcontrib><creatorcontrib>Ho, Jennifer E.</creatorcontrib><creatorcontrib>Philippakis, Anthony A.</creatorcontrib><creatorcontrib>Lubitz, Steven A.</creatorcontrib><title>Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs</title><title>Circulation. Cardiovascular imaging</title><addtitle>Circ Cardiovasc Imaging</addtitle><description>Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection.
Within 32 239 individuals of the UK Biobank prospective cohort who underwent CMR and 12-lead ECG, we trained a convolutional neural network to predict CMR-derived LV mass using 12-lead ECGs (left ventricular mass-artificial intelligence [LVM-AI]). In independent test sets (UK Biobank [n=4903] and Mass General Brigham [MGB, n=1371]), we assessed correlation between LVM-AI predicted and CMR-derived LV mass and compared LVH discrimination using LVM-AI versus traditional ECG-based rules (ie, Sokolow-Lyon, Cornell, lead aVL rule, or any ECG rule). In the UK Biobank and an ambulatory MGB cohort (MGB outcomes, n=28 612), we assessed associations between LVM-AI predicted LVH and incident cardiovascular outcomes using age- and sex-adjusted Cox regression.
LVM-AI predicted LV mass correlated with CMR-derived LV mass in both test sets, although correlation was greater in the UK Biobank (r=0.79) versus MGB (r=0.60, P<0.001 for both). When compared with any ECG rule, LVM-AI demonstrated similar LVH discrimination in the UK Biobank (LVM-AI c-statistic 0.653 [95% CI, 0.608 -0.698] versus any ECG rule c-statistic 0.618 [95% CI, 0.574 -0.663], P=0.11) and superior discrimination in MGB (0.621; 95% CI, 0.592 -0.649 versus 0.588; 95% CI, 0.564 -0.611, P=0.02). LVM-AI-predicted LVH was associated with incident atrial fibrillation, myocardial infarction, heart failure, and ventricular arrhythmias.
Deep learning-inferred LV mass estimates from 12-lead ECGs correlate with CMR-derived LV mass, associate with incident cardiovascular disease, and may improve LVH discrimination compared to traditional ECG rules.</description><subject>Artificial Intelligence</subject><subject>Deep Learning</subject><subject>Electrocardiography - methods</subject><subject>Female</subject><subject>Follow-Up Studies</subject><subject>Heart Ventricles - diagnostic imaging</subject><subject>Humans</subject><subject>Hypertrophy, Left Ventricular - diagnosis</subject><subject>Male</subject><subject>Middle Aged</subject><subject>Prospective Studies</subject><issn>1942-0080</issn><issn>1941-9651</issn><issn>1942-0080</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>EIF</sourceid><recordid>eNpNkctO3DAUhq2KqlzaV0BesslwfInjLFGAYaQBKtR2a3mcEyYlkwTbAc2u79A37JPUaACxOmfx_f-RvkPIMYMZY4qdVou7anF9Nl_czGeMwwwY55p9IgeslDwD0LD3Yd8nhyH8BlACcv2F7AvJuCoUPyDTOeJIl2h93_b3NA70u8e6dZFW1tetdfTa3vcYW0fvMAy97R3--_P3HH37hHUKNpH-wj761k2d9YkOgdq-plfbEX30w7je0ks_bCjjWTpT04tqHr6Sz43tAn57nUfk5-XFj-oqW97OF9XZMnOyAJFpVpfO5kWZ2zovnFzxolGwaoSy3Fmedl0qKRC1tIw1TaFRoV4p_SJDoxZH5GTXO_rhccIQzaYNDrvO9jhMwfBcMsFSiUio2qHODyF4bMzo2431W8PAvCg3H5SbpNzslKfg8euNabXB-j325jgBcgc8D11EHx666Rm9WaPt4jq1CFHIUmccOAMFAOljAEL8B_fkjVc</recordid><startdate>20210601</startdate><enddate>20210601</enddate><creator>Khurshid, Shaan</creator><creator>Friedman, Samuel</creator><creator>Pirruccello, James P.</creator><creator>Di Achille, Paolo</creator><creator>Diamant, Nathaniel</creator><creator>Anderson, Christopher D.</creator><creator>Ellinor, Patrick T.</creator><creator>Batra, Puneet</creator><creator>Ho, Jennifer E.</creator><creator>Philippakis, Anthony A.</creator><creator>Lubitz, Steven A.</creator><general>Lippincott Williams & Wilkins</general><scope>CGR</scope><scope>CUY</scope><scope>CVF</scope><scope>ECM</scope><scope>EIF</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0002-7987-4768</orcidid><orcidid>https://orcid.org/0000-0001-9256-0678</orcidid><orcidid>https://orcid.org/0000-0002-0053-2002</orcidid><orcidid>https://orcid.org/0000-0002-9599-4866</orcidid><orcidid>https://orcid.org/0000-0002-2067-0533</orcidid><orcidid>https://orcid.org/0000-0001-6822-0593</orcidid><orcidid>https://orcid.org/0000-0001-6088-4037</orcidid><orcidid>https://orcid.org/0000-0002-2840-4539</orcidid></search><sort><creationdate>20210601</creationdate><title>Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs</title><author>Khurshid, Shaan ; Friedman, Samuel ; Pirruccello, James P. ; Di Achille, Paolo ; Diamant, Nathaniel ; Anderson, Christopher D. ; Ellinor, Patrick T. ; Batra, Puneet ; Ho, Jennifer E. ; Philippakis, Anthony A. ; Lubitz, Steven A.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c4703-81d9ca5795ad57c4b27f60bf36a2ca2f6089643ee84a11ff78e6e8b6812288e83</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Artificial Intelligence</topic><topic>Deep Learning</topic><topic>Electrocardiography - methods</topic><topic>Female</topic><topic>Follow-Up Studies</topic><topic>Heart Ventricles - diagnostic imaging</topic><topic>Humans</topic><topic>Hypertrophy, Left Ventricular - diagnosis</topic><topic>Male</topic><topic>Middle Aged</topic><topic>Prospective Studies</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Khurshid, Shaan</creatorcontrib><creatorcontrib>Friedman, Samuel</creatorcontrib><creatorcontrib>Pirruccello, James P.</creatorcontrib><creatorcontrib>Di Achille, Paolo</creatorcontrib><creatorcontrib>Diamant, Nathaniel</creatorcontrib><creatorcontrib>Anderson, Christopher D.</creatorcontrib><creatorcontrib>Ellinor, Patrick T.</creatorcontrib><creatorcontrib>Batra, Puneet</creatorcontrib><creatorcontrib>Ho, Jennifer E.</creatorcontrib><creatorcontrib>Philippakis, Anthony A.</creatorcontrib><creatorcontrib>Lubitz, Steven A.</creatorcontrib><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><jtitle>Circulation. Cardiovascular imaging</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Khurshid, Shaan</au><au>Friedman, Samuel</au><au>Pirruccello, James P.</au><au>Di Achille, Paolo</au><au>Diamant, Nathaniel</au><au>Anderson, Christopher D.</au><au>Ellinor, Patrick T.</au><au>Batra, Puneet</au><au>Ho, Jennifer E.</au><au>Philippakis, Anthony A.</au><au>Lubitz, Steven A.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs</atitle><jtitle>Circulation. Cardiovascular imaging</jtitle><addtitle>Circ Cardiovasc Imaging</addtitle><date>2021-06-01</date><risdate>2021</risdate><volume>14</volume><issue>6</issue><spage>e012281</spage><epage>e012281</epage><pages>e012281-e012281</pages><issn>1942-0080</issn><issn>1941-9651</issn><eissn>1942-0080</eissn><abstract>Classical methods for detecting left ventricular (LV) hypertrophy (LVH) using 12-lead ECGs are insensitive. Deep learning models using ECG to infer cardiac magnetic resonance (CMR)-derived LV mass may improve LVH detection.
Within 32 239 individuals of the UK Biobank prospective cohort who underwent CMR and 12-lead ECG, we trained a convolutional neural network to predict CMR-derived LV mass using 12-lead ECGs (left ventricular mass-artificial intelligence [LVM-AI]). In independent test sets (UK Biobank [n=4903] and Mass General Brigham [MGB, n=1371]), we assessed correlation between LVM-AI predicted and CMR-derived LV mass and compared LVH discrimination using LVM-AI versus traditional ECG-based rules (ie, Sokolow-Lyon, Cornell, lead aVL rule, or any ECG rule). In the UK Biobank and an ambulatory MGB cohort (MGB outcomes, n=28 612), we assessed associations between LVM-AI predicted LVH and incident cardiovascular outcomes using age- and sex-adjusted Cox regression.
LVM-AI predicted LV mass correlated with CMR-derived LV mass in both test sets, although correlation was greater in the UK Biobank (r=0.79) versus MGB (r=0.60, P<0.001 for both). When compared with any ECG rule, LVM-AI demonstrated similar LVH discrimination in the UK Biobank (LVM-AI c-statistic 0.653 [95% CI, 0.608 -0.698] versus any ECG rule c-statistic 0.618 [95% CI, 0.574 -0.663], P=0.11) and superior discrimination in MGB (0.621; 95% CI, 0.592 -0.649 versus 0.588; 95% CI, 0.564 -0.611, P=0.02). LVM-AI-predicted LVH was associated with incident atrial fibrillation, myocardial infarction, heart failure, and ventricular arrhythmias.
Deep learning-inferred LV mass estimates from 12-lead ECGs correlate with CMR-derived LV mass, associate with incident cardiovascular disease, and may improve LVH discrimination compared to traditional ECG rules.</abstract><cop>United States</cop><pub>Lippincott Williams & Wilkins</pub><pmid>34126762</pmid><doi>10.1161/CIRCIMAGING.120.012281</doi><orcidid>https://orcid.org/0000-0002-7987-4768</orcidid><orcidid>https://orcid.org/0000-0001-9256-0678</orcidid><orcidid>https://orcid.org/0000-0002-0053-2002</orcidid><orcidid>https://orcid.org/0000-0002-9599-4866</orcidid><orcidid>https://orcid.org/0000-0002-2067-0533</orcidid><orcidid>https://orcid.org/0000-0001-6822-0593</orcidid><orcidid>https://orcid.org/0000-0001-6088-4037</orcidid><orcidid>https://orcid.org/0000-0002-2840-4539</orcidid><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1942-0080 |
| ispartof | Circulation. Cardiovascular imaging, 2021-06, Vol.14 (6), p.e012281-e012281 |
| issn | 1942-0080 1941-9651 1942-0080 |
| language | eng |
| recordid | cdi_proquest_miscellaneous_2541318963 |
| source | MEDLINE; American Heart Association Journals; EZB-FREE-00999 freely available EZB journals |
| subjects | Artificial Intelligence Deep Learning Electrocardiography - methods Female Follow-Up Studies Heart Ventricles - diagnostic imaging Humans Hypertrophy, Left Ventricular - diagnosis Male Middle Aged Prospective Studies |
| title | Deep Learning to Predict Cardiac Magnetic Resonance–Derived Left Ventricular Mass and Hypertrophy From 12-Lead ECGs |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-11T10%3A57%3A42IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Deep%20Learning%20to%20Predict%20Cardiac%20Magnetic%20Resonance%E2%80%93Derived%20Left%20Ventricular%20Mass%20and%20Hypertrophy%20From%2012-Lead%20ECGs&rft.jtitle=Circulation.%20Cardiovascular%20imaging&rft.au=Khurshid,%20Shaan&rft.date=2021-06-01&rft.volume=14&rft.issue=6&rft.spage=e012281&rft.epage=e012281&rft.pages=e012281-e012281&rft.issn=1942-0080&rft.eissn=1942-0080&rft_id=info:doi/10.1161/CIRCIMAGING.120.012281&rft_dat=%3Cproquest_cross%3E2541318963%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2541318963&rft_id=info:pmid/34126762&rfr_iscdi=true |