Abstract 13564: Novel Electrocardiogram Generating Technique Using Artificial Intelligence: From 2-lead to 12-lead
Abstract only Introduction: Based on the development of AI (artificial intelligence) and big data, an emerging number of methods achieved outstanding performance in myocardial infarction (MI) diagnosis by an electrocardiogram (ECG). However, conventional interpretation methods have low reliability f...
Saved in:
Published in | Circulation (New York, N.Y.) Vol. 146; no. Suppl_1 |
---|---|
Main Authors | , , , , , , |
Format | Journal Article |
Language | English |
Published |
08.11.2022
|
Online Access | Get full text |
Cover
Loading…
Abstract | Abstract only
Introduction:
Based on the development of AI (artificial intelligence) and big data, an emerging number of methods achieved outstanding performance in myocardial infarction (MI) diagnosis by an electrocardiogram (ECG). However, conventional interpretation methods have low reliability for detecting MI and are challenging to apply to 2 leads of wearable devices.
Objectives:
To evaluate whether the novel method can facilitate MI diagnosis by only 2-lead ECG.
Methods:
We propose T2T (from 2-lead to 12-lead), a deep generative model that simulates a standard 12-lead ECG from an input of two asynchronous leads by generating ten leads. (
Figure a
) We used and selected 15,012 ECGs (9,527 normal, 5,485 samples with any MI) from the Physikalisch-Technische Bundesanstalt dataset. This dataset was split into stratified training, validation, and test sets with a ratio of 7:1:2. Three models using 2-lead, T2T, and 12-lead ECG were developed and validated to predict a diagnosis of MI, and their accuracy was compared.
Results:
The area under the receiver operating characteristic curves of the 2-lead, T2T and 12-lead ECG were 0.937, 0.948, and 0.960, respectively.
Figure b
shows representative cases over a two-second window, where the green and pink lines are the signals from the original ECG and T2T. The generated signals have less noise (seen in aVR and aVL in 2a and V4, V5, and V6) and baseline wander. For the V1 and V5 leads generated by T2T, the differences in amplitude are 6.4% and 7.3%, respectively, and the missing positional errors are under 10 ms.
Conclusions:
Novel ECG T2T algorithm demonstrated favorable performance in detecting MI using 12-lead ECG.
Figure a.
ECGT2T model architecture. The model is comprised of style, mapping, generative, and discriminative networks. Each network is built with residual blocks.
Figure b.
Representative samples of ECG T2T. The green line denotes the original signal, while the pink line represents the signals generated by ECGT2T. |
---|---|
AbstractList | Abstract only
Introduction:
Based on the development of AI (artificial intelligence) and big data, an emerging number of methods achieved outstanding performance in myocardial infarction (MI) diagnosis by an electrocardiogram (ECG). However, conventional interpretation methods have low reliability for detecting MI and are challenging to apply to 2 leads of wearable devices.
Objectives:
To evaluate whether the novel method can facilitate MI diagnosis by only 2-lead ECG.
Methods:
We propose T2T (from 2-lead to 12-lead), a deep generative model that simulates a standard 12-lead ECG from an input of two asynchronous leads by generating ten leads. (
Figure a
) We used and selected 15,012 ECGs (9,527 normal, 5,485 samples with any MI) from the Physikalisch-Technische Bundesanstalt dataset. This dataset was split into stratified training, validation, and test sets with a ratio of 7:1:2. Three models using 2-lead, T2T, and 12-lead ECG were developed and validated to predict a diagnosis of MI, and their accuracy was compared.
Results:
The area under the receiver operating characteristic curves of the 2-lead, T2T and 12-lead ECG were 0.937, 0.948, and 0.960, respectively.
Figure b
shows representative cases over a two-second window, where the green and pink lines are the signals from the original ECG and T2T. The generated signals have less noise (seen in aVR and aVL in 2a and V4, V5, and V6) and baseline wander. For the V1 and V5 leads generated by T2T, the differences in amplitude are 6.4% and 7.3%, respectively, and the missing positional errors are under 10 ms.
Conclusions:
Novel ECG T2T algorithm demonstrated favorable performance in detecting MI using 12-lead ECG.
Figure a.
ECGT2T model architecture. The model is comprised of style, mapping, generative, and discriminative networks. Each network is built with residual blocks.
Figure b.
Representative samples of ECG T2T. The green line denotes the original signal, while the pink line represents the signals generated by ECGT2T. |
Author | Kwon, Joon-Myoung Jo, Yong-Yeon Son, Jeong Min Lee, Min Sung Jang, Jong-Hwan Lee, Hak Seung Kang, Sora |
Author_xml | – sequence: 1 givenname: Hak Seung surname: Lee fullname: Lee, Hak Seung organization: Med AI, Seoul – sequence: 2 givenname: Yong-Yeon surname: Jo fullname: Jo, Yong-Yeon organization: Med AI, Seoul – sequence: 3 givenname: Sora surname: Kang fullname: Kang, Sora organization: Med AI, Seoul – sequence: 4 givenname: Jong-Hwan surname: Jang fullname: Jang, Jong-Hwan organization: Med AI, Seoul – sequence: 5 givenname: Jeong Min surname: Son fullname: Son, Jeong Min organization: Med AI, Seoul – sequence: 6 givenname: Min Sung surname: Lee fullname: Lee, Min Sung organization: Med AI, Seoul – sequence: 7 givenname: Joon-Myoung surname: Kwon fullname: Kwon, Joon-Myoung organization: Mediplex Sejong Hosp, Incheon |
BookMark | eNqdj0FOwzAURC1UJFLgCuhfIMGOE1ftrkItdMOqrC3j_gQjxw7fLhK3pyk9AauZWTxp3pzNQgzI2IPglRBKPFpHthKNqtJxHL0WlZCtaq5YIdq6KZtWLmes4Jwvy4Ws6xs2T-nzNJVctAWj9XvKZGyGM7WC1_iNHjYebaZoDR1c7MkM8IwByWQXetij_Qju64jwlqa9puw6Z53xsAsZvXc9Bosr2FIcoC49mgPkCOKv3rHrzviE95e8ZWq72T-9lJZiSoSdHskNhn604Hoy1JOhPhnqi6E-f5X_Bn8BiHFhsQ |
ContentType | Journal Article |
DBID | AAYXX CITATION |
DOI | 10.1161/circ.146.suppl_1.13564 |
DatabaseName | CrossRef |
DatabaseTitle | CrossRef |
DatabaseTitleList | CrossRef |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Medicine Anatomy & Physiology |
EISSN | 1524-4539 |
ExternalDocumentID | 10_1161_circ_146_suppl_1_13564 |
GroupedDBID | --- .-D .3C .XZ .Z2 01R 0R~ 0ZK 18M 1J1 29B 2FS 2WC 354 40H 4Q1 4Q2 4Q3 53G 5GY 5RE 5VS 6PF 71W 77Y 7O~ AAAAV AAAXR AAGIX AAHPQ AAIQE AAJCS AAMOA AAMTA AAQKA AARTV AASCR AASOK AASXQ AAUEB AAWTL AAXQO AAYXX ABASU ABBUW ABDIG ABJNI ABOCM ABPMR ABQRW ABVCZ ABXVJ ABZAD ACDDN ACEWG ACGFO ACGFS ACILI ACOAL ACRKK ACWDW ACWRI ACXJB ACXNZ ADBBV ADCYY ADGGA ADHPY AE6 AEBDS AENEX AFCHL AFDTB AFEXH AFSOK AFUWQ AGINI AHMBA AHOMT AHQNM AHRYX AHVBC AIJEX AINUH AJIOK AJNWD AJZMW AKULP ALMA_UNASSIGNED_HOLDINGS ALMTX AMJPA AMKUR AMNEI AOHHW ASPBG AVWKF AWKKM AYCSE AZFZN BAWUL BOYCO BQLVK BYPQX C45 CITATION CS3 DIK DIWNM DU5 E3Z EBS EEVPB ERAAH EX3 F2K F2L F2M F2N F5P FCALG GNXGY GQDEL GX1 H0~ HLJTE HZ~ IKREB IKYAY IN~ IPNFZ JF9 JG8 JK3 K-A K-F K8S KD2 KMI KQ8 L-C L7B N9A N~7 N~B O9- OAG OAH OBH OCB ODA ODMTH OGEVE OHH OHYEH OJAPA OK1 OL1 OLB OLG OLH OLU OLV OLW OLY OLZ OPUJH OVD OVDNE OVIDH OVLEI OVOZU OWBYB OWU OWV OWW OWX OWY OWZ OXXIT P2P PQQKQ RAH RHF RIG RLZ S4R S4S T8P TEORI TR2 TSPGW UPT V2I VVN W2D W3M W8F WH7 WOQ WOW X3V X3W XXN XYM YFH YOC YSK YYM YZZ ZFV ZY1 ~H1 |
ID | FETCH-crossref_primary_10_1161_circ_146_suppl_1_135643 |
ISSN | 0009-7322 |
IngestDate | Fri Aug 23 01:13:04 EDT 2024 |
IsPeerReviewed | true |
IsScholarly | true |
Issue | Suppl_1 |
Language | English |
LinkModel | OpenURL |
MergedId | FETCHMERGED-crossref_primary_10_1161_circ_146_suppl_1_135643 |
ParticipantIDs | crossref_primary_10_1161_circ_146_suppl_1_13564 |
PublicationCentury | 2000 |
PublicationDate | 2022-11-08 |
PublicationDateYYYYMMDD | 2022-11-08 |
PublicationDate_xml | – month: 11 year: 2022 text: 2022-11-08 day: 08 |
PublicationDecade | 2020 |
PublicationTitle | Circulation (New York, N.Y.) |
PublicationYear | 2022 |
SSID | ssj0006375 |
Score | 4.9128704 |
Snippet | Abstract only
Introduction:
Based on the development of AI (artificial intelligence) and big data, an emerging number of methods achieved outstanding... |
SourceID | crossref |
SourceType | Aggregation Database |
Title | Abstract 13564: Novel Electrocardiogram Generating Technique Using Artificial Intelligence: From 2-lead to 12-lead |
Volume | 146 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV1LS8NAEF5qBfEiPvHNHsSLpDZpHm1vpbZURU8V6ilk41bEmkhIFf09_lBnZjcPtYj1EkJIJtnMx7z221nGjpw7p24LzzLM8dgzbFH3DGFj1ioD0xKy6YwlsS2u3cGNfTFyRpXKR4m1NE1FLXyfua7kP1qFa6BXXCU7h2ZzoXABzkG_cAQNw_FPOu4ILFSE6YnZcFwbk_vr-EVOTnpqb5uQuKZIv9LdpYniPMy7tiq6QCchupBqulH050RpfVx7YhkTwAHGqKY6Lcez3Yck1BuAzdrXp1Rn0IyfQfAI9mmq_SVRd8gLxNG9cStLnICsjh0nueO4CDL6MNw8eNW41iULyHapDPvFDLcMr6EWJNektryWbdiO6myUm2ZdnlQYpH1OfXO24XfR8IcwaLT94Hjo1hr9_sLVZdP73zxgzkukjMg1fZSDSZGv5fgkZ4EtWl7LQeLo2fll7u_dhudk-_XhoPQ6dJBzOvt7SiFQKZYZrrIVnYTwjkLUGqvIaJ1tdKIgjZ_e-DEnWjDNt6yzpSvNvthgSYY3Ti9oc0Ib_4E2XqCN52jjhDZeoI2X0dbmiDWuAMbTmGusbTK33xt2B0Y2FP9ZtUfxf_-JjS1WjeJIbjMuTVNY9cCVEBPZyMOB5KEZNIUrce45FDvsdE7hu3M_sceWC4Tus2qaTOUBhJmpOCQtfwLKqX_j |
link.rule.ids | 315,786,790,27955,27956 |
linkProvider | Flying Publisher |
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=Abstract+13564%3A+Novel+Electrocardiogram+Generating+Technique+Using+Artificial+Intelligence%3A+From+2-lead+to+12-lead&rft.jtitle=Circulation+%28New+York%2C+N.Y.%29&rft.au=Lee%2C+Hak+Seung&rft.au=Jo%2C+Yong-Yeon&rft.au=Kang%2C+Sora&rft.au=Jang%2C+Jong-Hwan&rft.date=2022-11-08&rft.issn=0009-7322&rft.eissn=1524-4539&rft.volume=146&rft.issue=Suppl_1&rft_id=info:doi/10.1161%2Fcirc.146.suppl_1.13564&rft.externalDBID=n%2Fa&rft.externalDocID=10_1161_circ_146_suppl_1_13564 |
thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0009-7322&client=summon |
thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0009-7322&client=summon |
thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0009-7322&client=summon |