Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury

Using a large national database of cardiac surgical procedures, we applied machine learning (ML) to risk stratification and profiling for cardiac surgery-associated acute kidney injury. We compared performance of ML to established scoring tools. Four ML algorithms were used, including logistic regre...

Full description

Bibliographic Details
Main Authors: Penny-Dimri, J.C., Bergmeir, C., Reid, Christopher, Williams-Spence, J., Cochrane, A.D., Smith, J.A.
Format: Journal Article
Language:English
Published: ELSEVIER INC 2021
Subjects:
Online Access:http://purl.org/au-research/grants/nhmrc/1136372
http://hdl.handle.net/20.500.11937/93772
_version_ 1848765785819840512
author Penny-Dimri, J.C.
Bergmeir, C.
Reid, Christopher
Williams-Spence, J.
Cochrane, A.D.
Smith, J.A.
author_facet Penny-Dimri, J.C.
Bergmeir, C.
Reid, Christopher
Williams-Spence, J.
Cochrane, A.D.
Smith, J.A.
author_sort Penny-Dimri, J.C.
building Curtin Institutional Repository
collection Online Access
description Using a large national database of cardiac surgical procedures, we applied machine learning (ML) to risk stratification and profiling for cardiac surgery-associated acute kidney injury. We compared performance of ML to established scoring tools. Four ML algorithms were used, including logistic regression (LR), gradient boosted machine (GBM), K-nearest neighbor, and neural networks (NN). These were compared to the Cleveland Clinic score, and a risk score developed on the same database. Five-fold cross-validation repeated 20 times was used to measure the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Risk profiles from GBM and NN were generated using Shapley additive values. A total of 97,964 surgery events in 96,653 patients were included. For predicting postoperative renal replacement therapy using pre- and intraoperative data, LR, GBM, and NN achieved an AUC (standard deviation) of 0.84 (0.01), 0.85 (0.01), 0.84 (0.01) respectively outperforming the highest performing scoring tool with 0.81 (0.004). For predicting cardiac surgery-associated acute kidney injury, LR, GBM, and NN each achieved 0.77 (0.01), 0.78 (0.01), 0.77 (0.01) respectively outperforming the scoring tool with 0.75 (0.004). Compared to scores and LR, shapely additive values analysis of black box model predictions was able to generate patient-level explanations for each prediction. ML algorithms provide state-of-the-art approaches to risk stratification. Explanatory modeling can exploit complex decision boundaries to aid the clinician in understanding the risks specific to individual patients.
first_indexed 2025-11-14T11:40:46Z
format Journal Article
id curtin-20.500.11937-93772
institution Curtin University Malaysia
institution_category Local University
language English
last_indexed 2025-11-14T11:40:46Z
publishDate 2021
publisher ELSEVIER INC
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-937722024-01-09T07:16:32Z Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury Penny-Dimri, J.C. Bergmeir, C. Reid, Christopher Williams-Spence, J. Cochrane, A.D. Smith, J.A. Science & Technology Life Sciences & Biomedicine Cardiac & Cardiovascular Systems Cardiovascular System & Cardiology Artificial intelligence Machine learning Emerging technology Cardiac surgery-associated acute kidney injury ACUTE-RENAL-FAILURE ERYTHROPOIETIN MODELS Artificial intelligence Cardiac surgery-associated acute kidney injury Emerging technology Machine learning Acute Kidney Injury Algorithms Cardiac Surgical Procedures Humans Logistic Models Machine Learning Risk Factors Humans Cardiac Surgical Procedures Logistic Models Risk Factors Algorithms Acute Kidney Injury Machine Learning Using a large national database of cardiac surgical procedures, we applied machine learning (ML) to risk stratification and profiling for cardiac surgery-associated acute kidney injury. We compared performance of ML to established scoring tools. Four ML algorithms were used, including logistic regression (LR), gradient boosted machine (GBM), K-nearest neighbor, and neural networks (NN). These were compared to the Cleveland Clinic score, and a risk score developed on the same database. Five-fold cross-validation repeated 20 times was used to measure the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Risk profiles from GBM and NN were generated using Shapley additive values. A total of 97,964 surgery events in 96,653 patients were included. For predicting postoperative renal replacement therapy using pre- and intraoperative data, LR, GBM, and NN achieved an AUC (standard deviation) of 0.84 (0.01), 0.85 (0.01), 0.84 (0.01) respectively outperforming the highest performing scoring tool with 0.81 (0.004). For predicting cardiac surgery-associated acute kidney injury, LR, GBM, and NN each achieved 0.77 (0.01), 0.78 (0.01), 0.77 (0.01) respectively outperforming the scoring tool with 0.75 (0.004). Compared to scores and LR, shapely additive values analysis of black box model predictions was able to generate patient-level explanations for each prediction. ML algorithms provide state-of-the-art approaches to risk stratification. Explanatory modeling can exploit complex decision boundaries to aid the clinician in understanding the risks specific to individual patients. 2021 Journal Article http://hdl.handle.net/20.500.11937/93772 10.1053/j.semtcvs.2020.09.028 English http://purl.org/au-research/grants/nhmrc/1136372 http://purl.org/au-research/grants/nhmrc/1092642 ELSEVIER INC restricted
spellingShingle Science & Technology
Life Sciences & Biomedicine
Cardiac & Cardiovascular Systems
Cardiovascular System & Cardiology
Artificial intelligence
Machine learning
Emerging technology
Cardiac surgery-associated acute kidney injury
ACUTE-RENAL-FAILURE
ERYTHROPOIETIN
MODELS
Artificial intelligence
Cardiac surgery-associated acute kidney injury
Emerging technology
Machine learning
Acute Kidney Injury
Algorithms
Cardiac Surgical Procedures
Humans
Logistic Models
Machine Learning
Risk Factors
Humans
Cardiac Surgical Procedures
Logistic Models
Risk Factors
Algorithms
Acute Kidney Injury
Machine Learning
Penny-Dimri, J.C.
Bergmeir, C.
Reid, Christopher
Williams-Spence, J.
Cochrane, A.D.
Smith, J.A.
Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title_full Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title_fullStr Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title_full_unstemmed Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title_short Machine Learning Algorithms for Predicting and Risk Profiling of Cardiac Surgery-Associated Acute Kidney Injury
title_sort machine learning algorithms for predicting and risk profiling of cardiac surgery-associated acute kidney injury
topic Science & Technology
Life Sciences & Biomedicine
Cardiac & Cardiovascular Systems
Cardiovascular System & Cardiology
Artificial intelligence
Machine learning
Emerging technology
Cardiac surgery-associated acute kidney injury
ACUTE-RENAL-FAILURE
ERYTHROPOIETIN
MODELS
Artificial intelligence
Cardiac surgery-associated acute kidney injury
Emerging technology
Machine learning
Acute Kidney Injury
Algorithms
Cardiac Surgical Procedures
Humans
Logistic Models
Machine Learning
Risk Factors
Humans
Cardiac Surgical Procedures
Logistic Models
Risk Factors
Algorithms
Acute Kidney Injury
Machine Learning
url http://purl.org/au-research/grants/nhmrc/1136372
http://purl.org/au-research/grants/nhmrc/1136372
http://hdl.handle.net/20.500.11937/93772