• Performance of Supervised Mach...

*Result*: Performance of Supervised Machine Learning Models for Cardiac Surgery-Associated Acute Kidney Injury in Children: Multicenter Retrospective Cohort Study, 2019-2022.

Title:
Performance of Supervised Machine Learning Models for Cardiac Surgery-Associated Acute Kidney Injury in Children: Multicenter Retrospective Cohort Study, 2019-2022.
Authors:
Baloglu O; Department of Heart, Vascular, and Thoracic, Division of Cardiology and Cardiovascular Medicine, Children's Institute, Cleveland Clinic Children's Center for Artificial Intelligence (C4AI), Cleveland, OH.; Department of Integrated Hospital Care, Division of Pediatric Critical Care, Children's Institute, Cleveland Clinic, Cleveland, OH., Akbasli IT; Department of Heart, Vascular, and Thoracic, Division of Cardiology and Cardiovascular Medicine, Children's Institute, Cleveland Clinic Children's Center for Artificial Intelligence (C4AI), Cleveland, OH., Morca A; Department of Heart, Vascular, and Thoracic, Division of Cardiology and Cardiovascular Medicine, Children's Institute, Cleveland Clinic Children's Center for Artificial Intelligence (C4AI), Cleveland, OH., Latifi SQ; Department of Heart, Vascular, and Thoracic, Division of Cardiology and Cardiovascular Medicine, Children's Institute, Cleveland Clinic Children's Center for Artificial Intelligence (C4AI), Cleveland, OH.; Department of Integrated Hospital Care, Division of Pediatric Critical Care, Children's Institute, Cleveland Clinic, Cleveland, OH., Gist KM; Department of Pediatrics, University of Colorado, Anschutz Medical Campus, Children's Hospital Colorado, Aurora, CO., Penk JS; Division of Pediatric Cardiology, Department of Pediatrics, Ann and Robert H. Lurie Children's Hospital of Chicago, Chicago, IL., Marino BS; Department of Heart, Vascular, and Thoracic, Division of Cardiology and Cardiovascular Medicine, Children's Institute, Cleveland Clinic Children's Center for Artificial Intelligence (C4AI), Cleveland, OH.
Source:
Pediatric critical care medicine : a journal of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies [Pediatr Crit Care Med] 2026 Jan 01; Vol. 27 (1), pp. 3-13. Date of Electronic Publication: 2025 Nov 10.
Publication Type:
Journal Article; Multicenter Study
Language:
English
Journal Info:
Publisher: Lippincott Williams & Wilkins Country of Publication: United States NLM ID: 100954653 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1529-7535 (Print) Linking ISSN: 15297535 NLM ISO Abbreviation: Pediatr Crit Care Med Subsets: MEDLINE
Imprint Name(s):
Original Publication: Baltimore, MD : Lippincott Williams & Wilkins, c2000-
MeSH Terms:
Acute Kidney Injury*/etiology , Acute Kidney Injury*/diagnosis , Acute Kidney Injury*/epidemiology , Cardiac Surgical Procedures*/adverse effects , Postoperative Complications*/etiology , Postoperative Complications*/diagnosis , Postoperative Complications*/epidemiology , Supervised Machine Learning*, Risk Assessment/methods ; Humans ; Retrospective Studies ; Child ; Female ; Male ; Adolescent ; Child, Preschool ; Infant ; Infant, Newborn ; ROC Curve ; Machine Learning
References:
Nadim MK, Forni LG, Bihorac A, et al.: Cardiac and vascular surgery-associated acute kidney injury: The 20th international consensus conference of the ADQI (Acute Disease Quality Initiative) group. J Am Heart Assoc 2018; 7:e008834.
Sharma A, Chakraborty R, Sharma K, et al.: Development of acute kidney injury following pediatric cardiac surgery. Kidney Res Clin Pract 2020; 39:259–268.
Alten JA, Cooper DS, Blinder JJ, et al.; Neonatal and Pediatric Heart and Renal Outcomes Network (NEPHRON) Investigators: Epidemiology of acute kidney injury after neonatal cardiac surgery: A report from the multicenter neonatal and pediatric heart and renal outcomes network. Crit Care Med 2021; 49:e941–e951.
Bertrandt RA, Gist K, Hasson D, et al.; Neonatal and Pediatric Heart and Renal Outcomes Network (NEPHRON) Investigators: Cardiac surgery-associated acute kidney injury in neonates undergoing the Norwood operation: Retrospective analysis of the multicenter neonatal and pediatric heart and renal outcomes network dataset, 2015-2018. Pediatr Crit Care Med 2024; 25:e246–e257.
Van den Eynde J, Rotbi H, Gewillig M, et al.: In-hospital outcomes of acute kidney injury after pediatric cardiac surgery: A meta-analysis. Front Pediatr 2021; 9:733744.
Parikh CR, Devarajan P, Zappitelli M, et al.; TRIBE-AKI Consortium: Postoperative biomarkers predict acute kidney injury and poor outcomes after pediatric cardiac surgery. J Am Soc Nephrol 2011; 22:1737–1747.
Dong L, Ma Q, Bennett M, et al.: Urinary biomarkers of cell cycle arrest are delayed predictors of acute kidney injury after pediatric cardiopulmonary bypass. Pediatr Nephrol 2017; 32:2351–2360.
Brazzelli M, Aucott L, Aceves-Martins M, et al.: Biomarkers for assessing acute kidney injury for people who are being considered for admission to critical care: A systematic review and cost-effectiveness analysis. Health Technol Assess 2022; 26:1–286.
Ramírez M, Chakravarti S, Busovsky-McNeal M, et al.: Elevated levels of urinary biomarkers TIMP-2 and IGFBP-7 predict acute kidney injury in neonates after congenital heart surgery. J Pediatr Intensive Care 2021; 11:153–158.
Lee H, Yoon H, Nam K, et al.: Derivation and validation of machine learning approaches to predict acute kidney injury after cardiac surgery. J Clin Med 2018; 7:322.
Tseng P, Chen Y, Wang C, et al.: Prediction of the development of acute kidney injury following cardiac surgery by machine learning. Crit Care 2020; 24:478.
Luo XQ, Kang YX, Duan SB, et al.: Machine learning-based prediction of acute kidney injury following pediatric cardiac surgery: Model development and validation study. J Med Internet Res 2023; 25:e41142.
Nagy M, Onder AM, Rosen D, et al.: Predicting pediatric cardiac surgery-associated acute kidney injury using machine learning. Pediatr Nephrol 2024; 39:1263–1270.
Khwaja A: KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract 2012; 120:c179–c184.
Collins GS, Moons KGM, Dhiman P, et al.: TRIPOD+AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ 2024; 385:e078378.
Jacobs ML, Jacobs JP, Thibault D, et al.: Updating an empirically based tool for analyzing congenital heart surgery mortality. World J Pediatr Congenit Heart Surg 2021; 12:246–281.
Lundberg SM, Lee S-I: A unified approach to interpreting model predictions. In: Advances in Neural Information Processing Systems 30. Guyon I, Luxburg UV, Bengio S, (Eds). Long Beach, CA, Curran Associates, 2017, pp 4765–4774.
Kim JH: Multicollinearity and misleading statistical results. Korean J Anesthesiol 2019; 72:558–569.
Blinder JJ, Asaro LA, Wypij D, et al.: Acute kidney injury after pediatric cardiac surgery: A secondary analysis of the safe pediatric Euglycemia after cardiac surgery trial. Pediatr Crit Care Med 2017; 18:638–646.
Goldstein SL, Akcan-Arikan A, Alobaidi R, et al.; Pediatric ADQI Collaborative: Consensus-based recommendations on priority activities to address acute kidney injury in children: A modified Delphi consensus statement. JAMA Netw Open 2022; 5:e2229442.
Fuhrman DY, Stanski NL, Krawczeski CD, et al.; ADQI 26 workgroup: A proposed framework for advancing acute kidney injury risk stratification and diagnosis in children: A report from the 26th Acute Disease Quality Initiative (ADQI) conference. Pediatr Nephrol 2024; 39:929–939.
Mathis MR, Mentz GB, Cao J, et al.; MPOG Collaborators: Hospital and clinician practice variation in cardiac surgery and postoperative acute kidney injury. JAMA Netw Open 2025; 8:e258342.
Contributed Indexing:
Keywords: acute kidney injury; cardiac surgery; congenital heart disease; pediatrics; supervised machine learning
Entry Date(s):
Date Created: 20251110 Date Completed: 20260108 Latest Revision: 20260115
Update Code:
20260130
PubMed Central ID:
PMC12771969
DOI:
10.1097/PCC.0000000000003857
PMID:
41212047
Database:
MEDLINE

*Further Information*

*Objectives: To derive and externally validate supervised machine learning (ML) models predictive of cardiac surgery-associated acute kidney injury (CS-AKI).
Design: Retrospective cohort analysis.
Setting: Multicenter (4), cardiac surgical centers from January 2019 to February 2022.
Patients: Seven days to 18 years old who had undergone cardiac surgery.
Interventions: None.
Measurements and Main Results: CS-AKI was defined using Kidney Disease: Improving Global Outcomes criteria, with stages 2/3 classified as severe, during the first 7 postoperative days. Data analysis followed two approaches: 1) combining three centers for derivation and using a fourth for external validation and 2) randomly dividing the entire dataset into derivation and validation cohorts in a 4:1 ratio. Forty ML models were developed across five derivation-validation pairs using four ML algorithms (light gradient-boosting machine, extreme gradient boosting, categorical boosting, and histogram gradient boosting) to predict two outcomes (any and severe CS-AKI) utilizing preoperative, intraoperative, and immediate postoperative variables. SHapley Additive exPlanations was used for input variable importance analysis. A cohort of 1100 patients was analyzed. Any CS-AKI and severe CS-AKI occurred in 49.1% and 23.1% patients, respectively. Wide range of variations in external validation of model performance were observed among all 40 ML models. For any CS-AKI, the range in metrics were: area under the receiver operating characteristic curve (AUROC) 0.64-0.83, sensitivity 0.29-0.86, specificity 0.46-0.95, positive predictive value (PPV) 0.50-0.85, and negative predictive value (NPV) 0.60-0.86. For severe CS-AKI, we found the range in metrics with AUROC 0.65-0.77, sensitivity 0.04-0.58, specificity 0.77-0.99, PPV 0.32-0.75, and NPV 0.78-0.90. Preoperative serum creatinine, cardiopulmonary bypass, aortic cross-clamp duration, weight, and age at surgery were the most important predictors associated with CS-AKI.
Conclusions: This analysis of a retrospective multicenter dataset shows that external performance of ML models vary, highlighting challenges in generalizability, which may be due to center-based differences in practice.
(Copyright © 2025 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.)*

*Dr. Gist’s institution received funding from the Gerber Foundation; she received funding from BioPorto Diagnostics. The remaining authors have disclosed that they do not have any potential conflicts of interest.*