Artificial intelligence algorithm for predicting mortality of patients with acute heart failure

Joon Myoung Kwon, Kyung Hee Kim, Ki Hyun Jeon, Sang Eun Lee, Hae Young Lee, Hyun Jai Cho, Jin Oh Choi, Eun Seok Jeon, Min Seok Kim, Jae Joong Kim, Kyung Kuk Hwang, Shung Chull Chae, Sang Hong Baek, Seok Min Kang, Dong Ju Choi, Byung Su Yoo, Kye Hun Kim, Hyun Young Park, Myeong Chan Cho, Byung Hee Oh

Research output: Contribution to journalArticle

Abstract

Aims This study aimed to develop and validate deep-learning-based artificial intelligence algorithm for predicting mortality of AHF (DAHF). Methods and results 12,654 dataset from 2165 patients with AHF in two hospitals were used as train data for DAHF development, and 4759 dataset from 4759 patients with AHF in 10 hospitals enrolled to the Korean AHF registry were used as performance test data. The endpoints were in-hospital, 12-month, and 36-month mortality. We compared the DAHF performance with the Get with the Guidelines-Heart Failure (GWTG-HF) score, Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and other machine-learning models by using the test data. Area under the receiver operating characteristic curve of the DAHF were 0.880 (95% confidence interval, 0.876-0.884) for predicting in-hospital mortality; these results significantly outperformed those of the GWTG-HF (0.728 [0.720-0.737]) and other machinelearning models. For predicting 12- and 36-month endpoints, DAHF (0.782 and 0.813) significantly outperformed MAGGIC score (0.718 and 0.729). During the 36-month follow-up, the high-risk group, defined by the DAHF, had a significantly higher mortality rate than the low-risk group(p<0.001). Conclusion DAHF predicted the in-hospital and long-term mortality of patients with AHF more accurately than the existing risk scores and other machine-learning models.

Original languageEnglish
Article numbere0219302
JournalPloS one
Volume14
Issue number7
DOIs
Publication statusPublished - 2019 Jan 1

Fingerprint

artificial intelligence
Artificial Intelligence
heart failure
Artificial intelligence
Heart Failure
Mortality
risk groups
Meta-Analysis
endpoints
meta-analysis
Guidelines
Learning systems
Hospital Mortality
ROC Curve
Registries
Learning
Confidence Intervals
confidence interval
learning
testing

All Science Journal Classification (ASJC) codes

  • Biochemistry, Genetics and Molecular Biology(all)
  • Agricultural and Biological Sciences(all)
  • General

Cite this

Kwon, J. M., Kim, K. H., Jeon, K. H., Lee, S. E., Lee, H. Y., Cho, H. J., ... Oh, B. H. (2019). Artificial intelligence algorithm for predicting mortality of patients with acute heart failure. PloS one, 14(7), [e0219302]. https://doi.org/10.1371/journal.pone.0219302
Kwon, Joon Myoung ; Kim, Kyung Hee ; Jeon, Ki Hyun ; Lee, Sang Eun ; Lee, Hae Young ; Cho, Hyun Jai ; Choi, Jin Oh ; Jeon, Eun Seok ; Kim, Min Seok ; Kim, Jae Joong ; Hwang, Kyung Kuk ; Chae, Shung Chull ; Baek, Sang Hong ; Kang, Seok Min ; Choi, Dong Ju ; Yoo, Byung Su ; Kim, Kye Hun ; Park, Hyun Young ; Cho, Myeong Chan ; Oh, Byung Hee. / Artificial intelligence algorithm for predicting mortality of patients with acute heart failure. In: PloS one. 2019 ; Vol. 14, No. 7.
@article{86341c13a7a74d53a8f70d654df5c06c,
title = "Artificial intelligence algorithm for predicting mortality of patients with acute heart failure",
abstract = "Aims This study aimed to develop and validate deep-learning-based artificial intelligence algorithm for predicting mortality of AHF (DAHF). Methods and results 12,654 dataset from 2165 patients with AHF in two hospitals were used as train data for DAHF development, and 4759 dataset from 4759 patients with AHF in 10 hospitals enrolled to the Korean AHF registry were used as performance test data. The endpoints were in-hospital, 12-month, and 36-month mortality. We compared the DAHF performance with the Get with the Guidelines-Heart Failure (GWTG-HF) score, Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and other machine-learning models by using the test data. Area under the receiver operating characteristic curve of the DAHF were 0.880 (95{\%} confidence interval, 0.876-0.884) for predicting in-hospital mortality; these results significantly outperformed those of the GWTG-HF (0.728 [0.720-0.737]) and other machinelearning models. For predicting 12- and 36-month endpoints, DAHF (0.782 and 0.813) significantly outperformed MAGGIC score (0.718 and 0.729). During the 36-month follow-up, the high-risk group, defined by the DAHF, had a significantly higher mortality rate than the low-risk group(p<0.001). Conclusion DAHF predicted the in-hospital and long-term mortality of patients with AHF more accurately than the existing risk scores and other machine-learning models.",
author = "Kwon, {Joon Myoung} and Kim, {Kyung Hee} and Jeon, {Ki Hyun} and Lee, {Sang Eun} and Lee, {Hae Young} and Cho, {Hyun Jai} and Choi, {Jin Oh} and Jeon, {Eun Seok} and Kim, {Min Seok} and Kim, {Jae Joong} and Hwang, {Kyung Kuk} and Chae, {Shung Chull} and Baek, {Sang Hong} and Kang, {Seok Min} and Choi, {Dong Ju} and Yoo, {Byung Su} and Kim, {Kye Hun} and Park, {Hyun Young} and Cho, {Myeong Chan} and Oh, {Byung Hee}",
year = "2019",
month = "1",
day = "1",
doi = "10.1371/journal.pone.0219302",
language = "English",
volume = "14",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "7",

}

Kwon, JM, Kim, KH, Jeon, KH, Lee, SE, Lee, HY, Cho, HJ, Choi, JO, Jeon, ES, Kim, MS, Kim, JJ, Hwang, KK, Chae, SC, Baek, SH, Kang, SM, Choi, DJ, Yoo, BS, Kim, KH, Park, HY, Cho, MC & Oh, BH 2019, 'Artificial intelligence algorithm for predicting mortality of patients with acute heart failure', PloS one, vol. 14, no. 7, e0219302. https://doi.org/10.1371/journal.pone.0219302

Artificial intelligence algorithm for predicting mortality of patients with acute heart failure. / Kwon, Joon Myoung; Kim, Kyung Hee; Jeon, Ki Hyun; Lee, Sang Eun; Lee, Hae Young; Cho, Hyun Jai; Choi, Jin Oh; Jeon, Eun Seok; Kim, Min Seok; Kim, Jae Joong; Hwang, Kyung Kuk; Chae, Shung Chull; Baek, Sang Hong; Kang, Seok Min; Choi, Dong Ju; Yoo, Byung Su; Kim, Kye Hun; Park, Hyun Young; Cho, Myeong Chan; Oh, Byung Hee.

In: PloS one, Vol. 14, No. 7, e0219302, 01.01.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Artificial intelligence algorithm for predicting mortality of patients with acute heart failure

AU - Kwon, Joon Myoung

AU - Kim, Kyung Hee

AU - Jeon, Ki Hyun

AU - Lee, Sang Eun

AU - Lee, Hae Young

AU - Cho, Hyun Jai

AU - Choi, Jin Oh

AU - Jeon, Eun Seok

AU - Kim, Min Seok

AU - Kim, Jae Joong

AU - Hwang, Kyung Kuk

AU - Chae, Shung Chull

AU - Baek, Sang Hong

AU - Kang, Seok Min

AU - Choi, Dong Ju

AU - Yoo, Byung Su

AU - Kim, Kye Hun

AU - Park, Hyun Young

AU - Cho, Myeong Chan

AU - Oh, Byung Hee

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Aims This study aimed to develop and validate deep-learning-based artificial intelligence algorithm for predicting mortality of AHF (DAHF). Methods and results 12,654 dataset from 2165 patients with AHF in two hospitals were used as train data for DAHF development, and 4759 dataset from 4759 patients with AHF in 10 hospitals enrolled to the Korean AHF registry were used as performance test data. The endpoints were in-hospital, 12-month, and 36-month mortality. We compared the DAHF performance with the Get with the Guidelines-Heart Failure (GWTG-HF) score, Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and other machine-learning models by using the test data. Area under the receiver operating characteristic curve of the DAHF were 0.880 (95% confidence interval, 0.876-0.884) for predicting in-hospital mortality; these results significantly outperformed those of the GWTG-HF (0.728 [0.720-0.737]) and other machinelearning models. For predicting 12- and 36-month endpoints, DAHF (0.782 and 0.813) significantly outperformed MAGGIC score (0.718 and 0.729). During the 36-month follow-up, the high-risk group, defined by the DAHF, had a significantly higher mortality rate than the low-risk group(p<0.001). Conclusion DAHF predicted the in-hospital and long-term mortality of patients with AHF more accurately than the existing risk scores and other machine-learning models.

AB - Aims This study aimed to develop and validate deep-learning-based artificial intelligence algorithm for predicting mortality of AHF (DAHF). Methods and results 12,654 dataset from 2165 patients with AHF in two hospitals were used as train data for DAHF development, and 4759 dataset from 4759 patients with AHF in 10 hospitals enrolled to the Korean AHF registry were used as performance test data. The endpoints were in-hospital, 12-month, and 36-month mortality. We compared the DAHF performance with the Get with the Guidelines-Heart Failure (GWTG-HF) score, Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and other machine-learning models by using the test data. Area under the receiver operating characteristic curve of the DAHF were 0.880 (95% confidence interval, 0.876-0.884) for predicting in-hospital mortality; these results significantly outperformed those of the GWTG-HF (0.728 [0.720-0.737]) and other machinelearning models. For predicting 12- and 36-month endpoints, DAHF (0.782 and 0.813) significantly outperformed MAGGIC score (0.718 and 0.729). During the 36-month follow-up, the high-risk group, defined by the DAHF, had a significantly higher mortality rate than the low-risk group(p<0.001). Conclusion DAHF predicted the in-hospital and long-term mortality of patients with AHF more accurately than the existing risk scores and other machine-learning models.

UR - http://www.scopus.com/inward/record.url?scp=85069302009&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069302009&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0219302

DO - 10.1371/journal.pone.0219302

M3 - Article

C2 - 31283783

AN - SCOPUS:85069302009

VL - 14

JO - PLoS One

JF - PLoS One

SN - 1932-6203

IS - 7

M1 - e0219302

ER -