Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study

Jussi A. Hernesniemi (Corresponding Author), Shadi Mahdiani, Juho A. Tynkkynen, Leo-Pekka Lyytikäinen, Pashupati P. Mishra, Terho Lehtimäki, Markku Eskola, Kjell Nikus, Kari Antila, Niku Oksala

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

Abstract

Objective: Investigation of the clinical potential of extensive phenotype data and machine learning (ML) in the prediction of mortality in acute coronary syndrome (ACS). Methods: The value of ML and extensive clinical data was analyzed in a retrospective registry study of 9066 consecutive ACS patients (January 2007 to October 2017). Main outcome was six-month mortality. Prediction models were developed using two ML methods, logistic regression and extreme gradient boosting (xgboost). The models were fitted in training set of patients treated in 2007–2014 and 2017 (81%, n = 7344) and validated in a separate validation set of patients treated in 2015–2016 with full GRACE score data available for comparison of model accuracy (19%, n = 1722). Results: Overall, six-month mortality was 7.3% (n = 660). Several variables were found to be significantly associated with six-month mortality by both ML methods. The xgboost scored the best performance: AUC 0.890 (0.864–0.916). The AUC values for logistic regression and GRACE score were 0.867(0.837–0.897) and 0.822 (0.785–0.859), respectively. The AUC value of xgboost was better when compared to logistic regression (p = .012) and GRACE score (p < .00001). Conclusions: The use of extensive phenotype data and novel machine learning improves prediction of mortality in ACS over traditional GRACE score.
Original languageEnglish
Pages (from-to)156-163
Number of pages8
JournalAnnals of Medicine
Volume51
Issue number2
DOIs
Publication statusPublished - 17 Feb 2019
MoE publication typeA1 Journal article-refereed

Fingerprint

Acute Coronary Syndrome
Phenotype
Mortality
Area Under Curve
Logistic Models
Registries
Retrospective Studies
Machine Learning

Keywords

  • machine learning
  • risk factors
  • mortality
  • acute coronary syndrome

Cite this

Hernesniemi, J. A., Mahdiani, S., Tynkkynen, J. A., Lyytikäinen, L-P., Mishra, P. P., Lehtimäki, T., ... Oksala, N. (2019). Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study. Annals of Medicine, 51(2), 156-163. https://doi.org/10.1080/07853890.2019.1596302
Hernesniemi, Jussi A. ; Mahdiani, Shadi ; Tynkkynen, Juho A. ; Lyytikäinen, Leo-Pekka ; Mishra, Pashupati P. ; Lehtimäki, Terho ; Eskola, Markku ; Nikus, Kjell ; Antila, Kari ; Oksala, Niku. / Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study. In: Annals of Medicine. 2019 ; Vol. 51, No. 2. pp. 156-163.
@article{032f40fb1286499286256206330a0c23,
title = "Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study",
abstract = "Objective: Investigation of the clinical potential of extensive phenotype data and machine learning (ML) in the prediction of mortality in acute coronary syndrome (ACS). Methods: The value of ML and extensive clinical data was analyzed in a retrospective registry study of 9066 consecutive ACS patients (January 2007 to October 2017). Main outcome was six-month mortality. Prediction models were developed using two ML methods, logistic regression and extreme gradient boosting (xgboost). The models were fitted in training set of patients treated in 2007–2014 and 2017 (81{\%}, n = 7344) and validated in a separate validation set of patients treated in 2015–2016 with full GRACE score data available for comparison of model accuracy (19{\%}, n = 1722). Results: Overall, six-month mortality was 7.3{\%} (n = 660). Several variables were found to be significantly associated with six-month mortality by both ML methods. The xgboost scored the best performance: AUC 0.890 (0.864–0.916). The AUC values for logistic regression and GRACE score were 0.867(0.837–0.897) and 0.822 (0.785–0.859), respectively. The AUC value of xgboost was better when compared to logistic regression (p = .012) and GRACE score (p < .00001). Conclusions: The use of extensive phenotype data and novel machine learning improves prediction of mortality in ACS over traditional GRACE score.",
keywords = "machine learning, risk factors, mortality, acute coronary syndrome",
author = "Hernesniemi, {Jussi A.} and Shadi Mahdiani and Tynkkynen, {Juho A.} and Leo-Pekka Lyytik{\"a}inen and Mishra, {Pashupati P.} and Terho Lehtim{\"a}ki and Markku Eskola and Kjell Nikus and Kari Antila and Niku Oksala",
year = "2019",
month = "2",
day = "17",
doi = "10.1080/07853890.2019.1596302",
language = "English",
volume = "51",
pages = "156--163",
journal = "Annals of Medicine",
issn = "0785-3890",
publisher = "Informa Healthcare",
number = "2",

}

Hernesniemi, JA, Mahdiani, S, Tynkkynen, JA, Lyytikäinen, L-P, Mishra, PP, Lehtimäki, T, Eskola, M, Nikus, K, Antila, K & Oksala, N 2019, 'Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study', Annals of Medicine, vol. 51, no. 2, pp. 156-163. https://doi.org/10.1080/07853890.2019.1596302

Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study. / Hernesniemi, Jussi A. (Corresponding Author); Mahdiani, Shadi; Tynkkynen, Juho A.; Lyytikäinen, Leo-Pekka ; Mishra, Pashupati P. ; Lehtimäki, Terho; Eskola, Markku; Nikus, Kjell; Antila, Kari; Oksala, Niku.

In: Annals of Medicine, Vol. 51, No. 2, 17.02.2019, p. 156-163.

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

TY - JOUR

T1 - Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study

AU - Hernesniemi, Jussi A.

AU - Mahdiani, Shadi

AU - Tynkkynen, Juho A.

AU - Lyytikäinen, Leo-Pekka

AU - Mishra, Pashupati P.

AU - Lehtimäki, Terho

AU - Eskola, Markku

AU - Nikus, Kjell

AU - Antila, Kari

AU - Oksala, Niku

PY - 2019/2/17

Y1 - 2019/2/17

N2 - Objective: Investigation of the clinical potential of extensive phenotype data and machine learning (ML) in the prediction of mortality in acute coronary syndrome (ACS). Methods: The value of ML and extensive clinical data was analyzed in a retrospective registry study of 9066 consecutive ACS patients (January 2007 to October 2017). Main outcome was six-month mortality. Prediction models were developed using two ML methods, logistic regression and extreme gradient boosting (xgboost). The models were fitted in training set of patients treated in 2007–2014 and 2017 (81%, n = 7344) and validated in a separate validation set of patients treated in 2015–2016 with full GRACE score data available for comparison of model accuracy (19%, n = 1722). Results: Overall, six-month mortality was 7.3% (n = 660). Several variables were found to be significantly associated with six-month mortality by both ML methods. The xgboost scored the best performance: AUC 0.890 (0.864–0.916). The AUC values for logistic regression and GRACE score were 0.867(0.837–0.897) and 0.822 (0.785–0.859), respectively. The AUC value of xgboost was better when compared to logistic regression (p = .012) and GRACE score (p < .00001). Conclusions: The use of extensive phenotype data and novel machine learning improves prediction of mortality in ACS over traditional GRACE score.

AB - Objective: Investigation of the clinical potential of extensive phenotype data and machine learning (ML) in the prediction of mortality in acute coronary syndrome (ACS). Methods: The value of ML and extensive clinical data was analyzed in a retrospective registry study of 9066 consecutive ACS patients (January 2007 to October 2017). Main outcome was six-month mortality. Prediction models were developed using two ML methods, logistic regression and extreme gradient boosting (xgboost). The models were fitted in training set of patients treated in 2007–2014 and 2017 (81%, n = 7344) and validated in a separate validation set of patients treated in 2015–2016 with full GRACE score data available for comparison of model accuracy (19%, n = 1722). Results: Overall, six-month mortality was 7.3% (n = 660). Several variables were found to be significantly associated with six-month mortality by both ML methods. The xgboost scored the best performance: AUC 0.890 (0.864–0.916). The AUC values for logistic regression and GRACE score were 0.867(0.837–0.897) and 0.822 (0.785–0.859), respectively. The AUC value of xgboost was better when compared to logistic regression (p = .012) and GRACE score (p < .00001). Conclusions: The use of extensive phenotype data and novel machine learning improves prediction of mortality in ACS over traditional GRACE score.

KW - machine learning

KW - risk factors

KW - mortality

KW - acute coronary syndrome

UR - http://www.scopus.com/inward/record.url?scp=85065169656&partnerID=8YFLogxK

U2 - 10.1080/07853890.2019.1596302

DO - 10.1080/07853890.2019.1596302

M3 - Article

VL - 51

SP - 156

EP - 163

JO - Annals of Medicine

JF - Annals of Medicine

SN - 0785-3890

IS - 2

ER -

Hernesniemi JA, Mahdiani S, Tynkkynen JA, Lyytikäinen L-P, Mishra PP, Lehtimäki T et al. Extensive phenotype data and machine learning in prediction of mortality in acute coronary syndrome – the MADDEC study. Annals of Medicine. 2019 Feb 17;51(2):156-163. https://doi.org/10.1080/07853890.2019.1596302

Access to Document