Artificial intelligence-enhanced electrocardiography for early assessment of coronavirus disease 2019 severity

Yong-Soo Baek; Yoonsu Jo; Sang-Chul Lee; Wonik Choi; Dae-Hyeok Kim

doi:10.1038/s41598-023-42252-5

Artificial intelligence-enhanced electrocardiography for early assessment of coronavirus disease 2019 severity

Yong-Soo Baek, Yoonsu Jo, Sang-Chul Lee, Wonik Choi^*, Dae-Hyeok Kim

^*Corresponding author for this work

Computer Science

Research output: Contribution to journal › Article › peer-review

18 Downloads (Pure)

Abstract

Despite challenges in severity scoring systems, artificial intelligence-enhanced electrocardiography (AI-ECG) could assist in early coronavirus disease 2019 (COVID-19) severity prediction. Between March 2020 and June 2022, we enrolled 1453 COVID-19 patients (mean age: 59.7 ± 20.1 years; 54.2% male) who underwent ECGs at our emergency department before severity classification. The AI-ECG algorithm was evaluated for severity assessment during admission, compared to the Early Warning Scores (EWSs) using the area under the curve (AUC) of the receiver operating characteristic curve, precision, recall, and F1 score. During the internal and external validation, the AI algorithm demonstrated reasonable outcomes in predicting COVID-19 severity with AUCs of 0.735 (95% CI: 0.662-0.807) and 0.734 (95% CI: 0.688-0.781). Combined with EWSs, it showed reliable performance with an AUC of 0.833 (95% CI: 0.830-0.835), precision of 0.764 (95% CI: 0.757-0.771), recall of 0.747 (95% CI: 0.741-0.753), and F1 score of 0.747 (95% CI: 0.741-0.753). In Cox proportional hazards models, the AI-ECG revealed a significantly higher hazard ratio (HR, 2.019; 95% CI: 1.156-3.525, p = 0.014) for mortality, even after adjusting for relevant parameters. Therefore, application of AI-ECG has the potential to assist in early COVID-19 severity prediction, leading to improved patient management.

Original language	English
Article number	15187
Number of pages	10
Journal	Scientific Reports
Volume	13
Issue number	1
DOIs	https://doi.org/10.1038/s41598-023-42252-5
Publication status	Published - 13 Sept 2023

Bibliographical note

Funding:
This work was supported by an Inha University Research Grant (Y-SB) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number RS-2023-00265440). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.

Keywords

Humans
Male
Adult
Middle Aged
Aged
Female
Artificial Intelligence
COVID-19/diagnosis
Algorithms
Electrocardiography
Area Under Curve

Access to Document

10.1038/s41598-023-42252-5Licence: Creative Commons: Attribution (CC BY)

BaekY2023ArtificialFinal published version, 2.59 MBLicence: Creative Commons: Attribution (CC BY)

Cite this

@article{8a56e9b1d5904e71b5d7408198d69c3f,

title = "Artificial intelligence-enhanced electrocardiography for early assessment of coronavirus disease 2019 severity",

abstract = "Despite challenges in severity scoring systems, artificial intelligence-enhanced electrocardiography (AI-ECG) could assist in early coronavirus disease 2019 (COVID-19) severity prediction. Between March 2020 and June 2022, we enrolled 1453 COVID-19 patients (mean age: 59.7 ± 20.1 years; 54.2% male) who underwent ECGs at our emergency department before severity classification. The AI-ECG algorithm was evaluated for severity assessment during admission, compared to the Early Warning Scores (EWSs) using the area under the curve (AUC) of the receiver operating characteristic curve, precision, recall, and F1 score. During the internal and external validation, the AI algorithm demonstrated reasonable outcomes in predicting COVID-19 severity with AUCs of 0.735 (95% CI: 0.662-0.807) and 0.734 (95% CI: 0.688-0.781). Combined with EWSs, it showed reliable performance with an AUC of 0.833 (95% CI: 0.830-0.835), precision of 0.764 (95% CI: 0.757-0.771), recall of 0.747 (95% CI: 0.741-0.753), and F1 score of 0.747 (95% CI: 0.741-0.753). In Cox proportional hazards models, the AI-ECG revealed a significantly higher hazard ratio (HR, 2.019; 95% CI: 1.156-3.525, p = 0.014) for mortality, even after adjusting for relevant parameters. Therefore, application of AI-ECG has the potential to assist in early COVID-19 severity prediction, leading to improved patient management.",

keywords = "Humans, Male, Adult, Middle Aged, Aged, Female, Artificial Intelligence, COVID-19/diagnosis, Algorithms, Electrocardiography, Area Under Curve",

author = "Yong-Soo Baek and Yoonsu Jo and Sang-Chul Lee and Wonik Choi and Dae-Hyeok Kim",

note = "Funding: This work was supported by an Inha University Research Grant (Y-SB) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number RS-2023-00265440). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.",

year = "2023",

month = sep,

day = "13",

doi = "10.1038/s41598-023-42252-5",

language = "English",

volume = "13",

journal = "Scientific Reports",

issn = "2045-2322",

publisher = "Nature Publishing Group",

number = "1",

}

TY - JOUR

T1 - Artificial intelligence-enhanced electrocardiography for early assessment of coronavirus disease 2019 severity

AU - Baek, Yong-Soo

AU - Jo, Yoonsu

AU - Lee, Sang-Chul

AU - Choi, Wonik

AU - Kim, Dae-Hyeok

N1 - Funding: This work was supported by an Inha University Research Grant (Y-SB) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number RS-2023-00265440). The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation.

PY - 2023/9/13

Y1 - 2023/9/13

N2 - Despite challenges in severity scoring systems, artificial intelligence-enhanced electrocardiography (AI-ECG) could assist in early coronavirus disease 2019 (COVID-19) severity prediction. Between March 2020 and June 2022, we enrolled 1453 COVID-19 patients (mean age: 59.7 ± 20.1 years; 54.2% male) who underwent ECGs at our emergency department before severity classification. The AI-ECG algorithm was evaluated for severity assessment during admission, compared to the Early Warning Scores (EWSs) using the area under the curve (AUC) of the receiver operating characteristic curve, precision, recall, and F1 score. During the internal and external validation, the AI algorithm demonstrated reasonable outcomes in predicting COVID-19 severity with AUCs of 0.735 (95% CI: 0.662-0.807) and 0.734 (95% CI: 0.688-0.781). Combined with EWSs, it showed reliable performance with an AUC of 0.833 (95% CI: 0.830-0.835), precision of 0.764 (95% CI: 0.757-0.771), recall of 0.747 (95% CI: 0.741-0.753), and F1 score of 0.747 (95% CI: 0.741-0.753). In Cox proportional hazards models, the AI-ECG revealed a significantly higher hazard ratio (HR, 2.019; 95% CI: 1.156-3.525, p = 0.014) for mortality, even after adjusting for relevant parameters. Therefore, application of AI-ECG has the potential to assist in early COVID-19 severity prediction, leading to improved patient management.

AB - Despite challenges in severity scoring systems, artificial intelligence-enhanced electrocardiography (AI-ECG) could assist in early coronavirus disease 2019 (COVID-19) severity prediction. Between March 2020 and June 2022, we enrolled 1453 COVID-19 patients (mean age: 59.7 ± 20.1 years; 54.2% male) who underwent ECGs at our emergency department before severity classification. The AI-ECG algorithm was evaluated for severity assessment during admission, compared to the Early Warning Scores (EWSs) using the area under the curve (AUC) of the receiver operating characteristic curve, precision, recall, and F1 score. During the internal and external validation, the AI algorithm demonstrated reasonable outcomes in predicting COVID-19 severity with AUCs of 0.735 (95% CI: 0.662-0.807) and 0.734 (95% CI: 0.688-0.781). Combined with EWSs, it showed reliable performance with an AUC of 0.833 (95% CI: 0.830-0.835), precision of 0.764 (95% CI: 0.757-0.771), recall of 0.747 (95% CI: 0.741-0.753), and F1 score of 0.747 (95% CI: 0.741-0.753). In Cox proportional hazards models, the AI-ECG revealed a significantly higher hazard ratio (HR, 2.019; 95% CI: 1.156-3.525, p = 0.014) for mortality, even after adjusting for relevant parameters. Therefore, application of AI-ECG has the potential to assist in early COVID-19 severity prediction, leading to improved patient management.

KW - Humans

KW - Male

KW - Adult

KW - Middle Aged

KW - Aged

KW - Female

KW - Artificial Intelligence

KW - COVID-19/diagnosis

KW - Algorithms

KW - Electrocardiography

KW - Area Under Curve

U2 - 10.1038/s41598-023-42252-5

DO - 10.1038/s41598-023-42252-5

M3 - Article

C2 - 37704692

SN - 2045-2322

VL - 13

JO - Scientific Reports

JF - Scientific Reports

IS - 1

M1 - 15187

ER -

Artificial intelligence-enhanced electrocardiography for early assessment of coronavirus disease 2019 severity

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this