Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety

James P. Howard, Christopher M. Cook, Tim P. van de Hoef, Martijn Meuwissen, Guus A. de Waard, Martijn A. van Lavieren, Mauro Echavarria-Pinto, Ibrahim Danad, Jan J. Piek, Matthias Götberg, Rasha K. Al-Lamee, Sayan Sen, Sukhjinder S. Nijjer, Henry Seligman, Niels van Royen, Paul Knaapen, Javier Escaned, Darrel P. Francis, Ricardo Petraco, Justin E. Davies

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Objectives: This study developed a neural network to perform automated pressure waveform analysis and allow real-time accurate identification of damping. Background: Damping of aortic pressure during coronary angiography must be identified to avoid serious complications and make accurate coronary physiology measurements. There are currently no automated methods to do this, and so identification of damping requires constant monitoring, which is prone to human error. Methods: The neural network was trained and tested versus core laboratory expert opinions derived from 2 separate datasets. A total of 5,709 aortic pressure waveforms of individual heart beats were extracted and classified. The study developed a recurrent convolutional neural network to classify beats as either normal, showing damping, or artifactual. Accuracies were reported using the opinions of 2 independent core laboratories. Results: The neural network was 99.4% accurate (95% confidence interval: 98.8% to 99.6%) at classifying beats from the testing dataset when judged against the opinions of the internal core laboratory. It was 98.7% accurate (95% confidence interval: 98.0% to 99.2%) when judged against the opinions of an external core laboratory not involved in neural network training. The neural network was 100% sensitive, with no beats classified as damped misclassified, with a specificity of 99.8%. The positive predictive and negative predictive values were 98.1% and 99.5%. The 2 core laboratories agreed more closely with the neural network than with each other. Conclusions: Arterial waveform analysis using neural networks allows rapid and accurate identification of damping. This demonstrates how machine learning can assist with patient safety and the quality control of procedures.
Original languageEnglish
Pages (from-to)2093-2101
JournalJACC: Cardiovascular Interventions
Volume12
Issue number20
DOIs
Publication statusPublished - 2019

Cite this

Howard, James P. ; Cook, Christopher M. ; van de Hoef, Tim P. ; Meuwissen, Martijn ; de Waard, Guus A. ; van Lavieren, Martijn A. ; Echavarria-Pinto, Mauro ; Danad, Ibrahim ; Piek, Jan J. ; Götberg, Matthias ; Al-Lamee, Rasha K. ; Sen, Sayan ; Nijjer, Sukhjinder S. ; Seligman, Henry ; van Royen, Niels ; Knaapen, Paul ; Escaned, Javier ; Francis, Darrel P. ; Petraco, Ricardo ; Davies, Justin E. / Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety. In: JACC: Cardiovascular Interventions. 2019 ; Vol. 12, No. 20. pp. 2093-2101.
@article{446728f6721343ca9b5cbff24bf5770d,
title = "Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety",
abstract = "Objectives: This study developed a neural network to perform automated pressure waveform analysis and allow real-time accurate identification of damping. Background: Damping of aortic pressure during coronary angiography must be identified to avoid serious complications and make accurate coronary physiology measurements. There are currently no automated methods to do this, and so identification of damping requires constant monitoring, which is prone to human error. Methods: The neural network was trained and tested versus core laboratory expert opinions derived from 2 separate datasets. A total of 5,709 aortic pressure waveforms of individual heart beats were extracted and classified. The study developed a recurrent convolutional neural network to classify beats as either normal, showing damping, or artifactual. Accuracies were reported using the opinions of 2 independent core laboratories. Results: The neural network was 99.4{\%} accurate (95{\%} confidence interval: 98.8{\%} to 99.6{\%}) at classifying beats from the testing dataset when judged against the opinions of the internal core laboratory. It was 98.7{\%} accurate (95{\%} confidence interval: 98.0{\%} to 99.2{\%}) when judged against the opinions of an external core laboratory not involved in neural network training. The neural network was 100{\%} sensitive, with no beats classified as damped misclassified, with a specificity of 99.8{\%}. The positive predictive and negative predictive values were 98.1{\%} and 99.5{\%}. The 2 core laboratories agreed more closely with the neural network than with each other. Conclusions: Arterial waveform analysis using neural networks allows rapid and accurate identification of damping. This demonstrates how machine learning can assist with patient safety and the quality control of procedures.",
author = "Howard, {James P.} and Cook, {Christopher M.} and {van de Hoef}, {Tim P.} and Martijn Meuwissen and {de Waard}, {Guus A.} and {van Lavieren}, {Martijn A.} and Mauro Echavarria-Pinto and Ibrahim Danad and Piek, {Jan J.} and Matthias G{\"o}tberg and Al-Lamee, {Rasha K.} and Sayan Sen and Nijjer, {Sukhjinder S.} and Henry Seligman and {van Royen}, Niels and Paul Knaapen and Javier Escaned and Francis, {Darrel P.} and Ricardo Petraco and Davies, {Justin E.}",
year = "2019",
doi = "10.1016/j.jcin.2019.06.036",
language = "English",
volume = "12",
pages = "2093--2101",
journal = "JACC Cardiovascular Interventions",
issn = "1936-8798",
publisher = "Elsevier Inc.",
number = "20",

}

Howard, JP, Cook, CM, van de Hoef, TP, Meuwissen, M, de Waard, GA, van Lavieren, MA, Echavarria-Pinto, M, Danad, I, Piek, JJ, Götberg, M, Al-Lamee, RK, Sen, S, Nijjer, SS, Seligman, H, van Royen, N, Knaapen, P, Escaned, J, Francis, DP, Petraco, R & Davies, JE 2019, 'Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety' JACC: Cardiovascular Interventions, vol. 12, no. 20, pp. 2093-2101. https://doi.org/10.1016/j.jcin.2019.06.036

Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety. / Howard, James P.; Cook, Christopher M.; van de Hoef, Tim P.; Meuwissen, Martijn; de Waard, Guus A.; van Lavieren, Martijn A.; Echavarria-Pinto, Mauro; Danad, Ibrahim; Piek, Jan J.; Götberg, Matthias; Al-Lamee, Rasha K.; Sen, Sayan; Nijjer, Sukhjinder S.; Seligman, Henry; van Royen, Niels; Knaapen, Paul; Escaned, Javier; Francis, Darrel P.; Petraco, Ricardo; Davies, Justin E.

In: JACC: Cardiovascular Interventions, Vol. 12, No. 20, 2019, p. 2093-2101.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Artificial Intelligence for Aortic Pressure Waveform Analysis During Coronary Angiography: Machine Learning for Patient Safety

AU - Howard, James P.

AU - Cook, Christopher M.

AU - van de Hoef, Tim P.

AU - Meuwissen, Martijn

AU - de Waard, Guus A.

AU - van Lavieren, Martijn A.

AU - Echavarria-Pinto, Mauro

AU - Danad, Ibrahim

AU - Piek, Jan J.

AU - Götberg, Matthias

AU - Al-Lamee, Rasha K.

AU - Sen, Sayan

AU - Nijjer, Sukhjinder S.

AU - Seligman, Henry

AU - van Royen, Niels

AU - Knaapen, Paul

AU - Escaned, Javier

AU - Francis, Darrel P.

AU - Petraco, Ricardo

AU - Davies, Justin E.

PY - 2019

Y1 - 2019

N2 - Objectives: This study developed a neural network to perform automated pressure waveform analysis and allow real-time accurate identification of damping. Background: Damping of aortic pressure during coronary angiography must be identified to avoid serious complications and make accurate coronary physiology measurements. There are currently no automated methods to do this, and so identification of damping requires constant monitoring, which is prone to human error. Methods: The neural network was trained and tested versus core laboratory expert opinions derived from 2 separate datasets. A total of 5,709 aortic pressure waveforms of individual heart beats were extracted and classified. The study developed a recurrent convolutional neural network to classify beats as either normal, showing damping, or artifactual. Accuracies were reported using the opinions of 2 independent core laboratories. Results: The neural network was 99.4% accurate (95% confidence interval: 98.8% to 99.6%) at classifying beats from the testing dataset when judged against the opinions of the internal core laboratory. It was 98.7% accurate (95% confidence interval: 98.0% to 99.2%) when judged against the opinions of an external core laboratory not involved in neural network training. The neural network was 100% sensitive, with no beats classified as damped misclassified, with a specificity of 99.8%. The positive predictive and negative predictive values were 98.1% and 99.5%. The 2 core laboratories agreed more closely with the neural network than with each other. Conclusions: Arterial waveform analysis using neural networks allows rapid and accurate identification of damping. This demonstrates how machine learning can assist with patient safety and the quality control of procedures.

AB - Objectives: This study developed a neural network to perform automated pressure waveform analysis and allow real-time accurate identification of damping. Background: Damping of aortic pressure during coronary angiography must be identified to avoid serious complications and make accurate coronary physiology measurements. There are currently no automated methods to do this, and so identification of damping requires constant monitoring, which is prone to human error. Methods: The neural network was trained and tested versus core laboratory expert opinions derived from 2 separate datasets. A total of 5,709 aortic pressure waveforms of individual heart beats were extracted and classified. The study developed a recurrent convolutional neural network to classify beats as either normal, showing damping, or artifactual. Accuracies were reported using the opinions of 2 independent core laboratories. Results: The neural network was 99.4% accurate (95% confidence interval: 98.8% to 99.6%) at classifying beats from the testing dataset when judged against the opinions of the internal core laboratory. It was 98.7% accurate (95% confidence interval: 98.0% to 99.2%) when judged against the opinions of an external core laboratory not involved in neural network training. The neural network was 100% sensitive, with no beats classified as damped misclassified, with a specificity of 99.8%. The positive predictive and negative predictive values were 98.1% and 99.5%. The 2 core laboratories agreed more closely with the neural network than with each other. Conclusions: Arterial waveform analysis using neural networks allows rapid and accurate identification of damping. This demonstrates how machine learning can assist with patient safety and the quality control of procedures.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85073068914&origin=inward

UR - https://www.ncbi.nlm.nih.gov/pubmed/31563678

U2 - 10.1016/j.jcin.2019.06.036

DO - 10.1016/j.jcin.2019.06.036

M3 - Article

VL - 12

SP - 2093

EP - 2101

JO - JACC Cardiovascular Interventions

JF - JACC Cardiovascular Interventions

SN - 1936-8798

IS - 20

ER -