Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Explainable AI identifies key biomarkers for acute kidney injury prediction in the ICU

Koozi, Hazem LU ; Engström, Jonas LU orcid ; Friberg, Hans LU and Frigyesi, Attila LU (2025) In Intensive Care Medicine Experimental 13. p.1-12
Abstract

BACKGROUND: Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.

METHODS: A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic... (More)

BACKGROUND: Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.

METHODS: A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic regression models, incorporating emerging biomarkers and routine clinical data at ICU admission. SHapley Additive exPlanations (SHAP) were used to identify key predictors. Discrimination was assessed using the mean area under the receiver operating characteristic curve (AUC).

RESULTS: The study included 4732 admissions, with 2603 analysed for new-onset AKI and 4716 for RRT. Top predictors of new-onset AKI were urine output, endostatin, baseline creatinine, lactate, and albumin. Top predictors of RRT were creatinine, urine output, endostatin, neutrophil gelatinase-associated lipocalin (NGAL), and the Simplified Acute Physiology Score (SAPS) 3. Several clinically relevant non-linear relationships were revealed. XGBoost outperformed logistic regression for both new-onset AKI (mean AUC 0.76, 95% CI 0.70-0.81 vs. 0.74, 95% CI 0.68-0.81; p < 0.001) and RRT (mean AUC 0.92, 95% CI 0.89-0.95 vs. 0.90, 95% CI 0.87-0.94; p < 0.001).

CONCLUSION: XGBoost identified key predictors of early new-onset AKI and RRT need in the ICU, highlighting both emerging (endostatin, NGAL) and established biomarkers (lactate, albumin), alongside known clinical predictors. It also improved predictive accuracy for both outcomes. Further clinical evaluation of these biomarkers and XAI models is warranted.

(Less)
Abstract (Swedish)
Background
Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.

Methods
A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic... (More)
Background
Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.

Methods
A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic regression models, incorporating emerging biomarkers and routine clinical data at ICU admission. SHapley Additive exPlanations (SHAP) were used to identify key predictors. Discrimination was assessed using the mean area under the receiver operating characteristic curve (AUC).

Results
The study included 4732 admissions, with 2603 analysed for new-onset AKI and 4716 for RRT. Top predictors of new-onset AKI were urine output, endostatin, baseline creatinine, lactate, and albumin. Top predictors of RRT were creatinine, urine output, endostatin, neutrophil gelatinase-associated lipocalin (NGAL), and the Simplified Acute Physiology Score (SAPS) 3. Several clinically relevant non-linear relationships were revealed. XGBoost outperformed logistic regression for both new-onset AKI (mean AUC 0.76, 95% CI 0.70–0.81 vs. 0.74, 95% CI 0.68–0.81; p < 0.001) and RRT (mean AUC 0.92, 95% CI 0.89–0.95 vs. 0.90, 95% CI 0.87–0.94; p < 0.001).

Conclusion
XGBoost identified key predictors of early new-onset AKI and RRT need in the ICU, highlighting both emerging (endostatin, NGAL) and established biomarkers (lactate, albumin), alongside known clinical predictors. It also improved predictive accuracy for both outcomes. Further clinical evaluation of these biomarkers and XAI models is warranted. (Less)
Please use this url to cite or link to this publication:
author
; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
Intensive Care Medicine Experimental
volume
13
article number
106
pages
1 - 12
publisher
Springer Nature
external identifiers
  • pmid:41123857
ISSN
2197-425X
DOI
10.1186/s40635-025-00816-x
project
SweCrit, a critical care biobank
language
English
LU publication?
yes
additional info
© 2025. The Author(s).
id
3a6855d9-bc8b-4cc4-9b0b-e3ab952db56e
date added to LUP
2025-10-23 02:42:58
date last changed
2025-10-23 09:08:04
@article{3a6855d9-bc8b-4cc4-9b0b-e3ab952db56e,
  abstract     = {{<p>BACKGROUND: Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.</p><p>METHODS: A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic regression models, incorporating emerging biomarkers and routine clinical data at ICU admission. SHapley Additive exPlanations (SHAP) were used to identify key predictors. Discrimination was assessed using the mean area under the receiver operating characteristic curve (AUC).</p><p>RESULTS: The study included 4732 admissions, with 2603 analysed for new-onset AKI and 4716 for RRT. Top predictors of new-onset AKI were urine output, endostatin, baseline creatinine, lactate, and albumin. Top predictors of RRT were creatinine, urine output, endostatin, neutrophil gelatinase-associated lipocalin (NGAL), and the Simplified Acute Physiology Score (SAPS) 3. Several clinically relevant non-linear relationships were revealed. XGBoost outperformed logistic regression for both new-onset AKI (mean AUC 0.76, 95% CI 0.70-0.81 vs. 0.74, 95% CI 0.68-0.81; p &lt; 0.001) and RRT (mean AUC 0.92, 95% CI 0.89-0.95 vs. 0.90, 95% CI 0.87-0.94; p &lt; 0.001).</p><p>CONCLUSION: XGBoost identified key predictors of early new-onset AKI and RRT need in the ICU, highlighting both emerging (endostatin, NGAL) and established biomarkers (lactate, albumin), alongside known clinical predictors. It also improved predictive accuracy for both outcomes. Further clinical evaluation of these biomarkers and XAI models is warranted.</p>}},
  author       = {{Koozi, Hazem and Engström, Jonas and Friberg, Hans and Frigyesi, Attila}},
  issn         = {{2197-425X}},
  language     = {{eng}},
  month        = {{10}},
  pages        = {{1--12}},
  publisher    = {{Springer Nature}},
  series       = {{Intensive Care Medicine Experimental}},
  title        = {{Explainable AI identifies key biomarkers for acute kidney injury prediction in the ICU}},
  url          = {{http://dx.doi.org/10.1186/s40635-025-00816-x}},
  doi          = {{10.1186/s40635-025-00816-x}},
  volume       = {{13}},
  year         = {{2025}},
}