Zobraziť v EDS

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

Uložené v:
Podrobná bibliografia
Názov: Explainable AI identifies key biomarkers for acute kidney injury prediction in the ICU
Autori: Koozi, Hazem, Engström, Jonas, Friberg, Hans, Frigyesi, Attila
Prispievatelia: Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Originator, Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Center for cardiac arrest, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Centrum för hjärtstopp, Originator, Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), eSSENCE: The e-Science Collaboration, Lunds universitet, Profilområden och andra starka forskningsmiljöer, Strategiska forskningsområden (SFO), eSSENCE: The e-Science Collaboration, Originator, Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), EpiHealth: Epidemiology for Health, Lunds universitet, Profilområden och andra starka forskningsmiljöer, Strategiska forskningsområden (SFO), EpiHealth: Epidemiology for Health, Originator, Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Intensive Care Epidemiology, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Intensivvårdsepidemiologi, Originator
Zdroj: Intensive Care Medicine Experimental. 13:1-12
Predmety: Medical and Health Sciences, Clinical Medicine, Anesthesiology and Intensive Care, Medicin och hälsovetenskap, Klinisk medicin, Anestesi och intensivvård, Urology, Urologi
Popis: 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.
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: Thestudy 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.
Prístupová URL adresa: https://doi.org/10.1186/s40635-025-00816-x
Databáza: SwePub
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://doi.org/10.1186/s40635-025-00816-x#
    Name: EDS - SwePub (s4221598)
    Category: fullText
    Text: View record in SwePub
  – Url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db=pmc&term=2197-425X[TA]+AND+1[PG]+AND+2025[PDAT]
    Name: FREE - PubMed Central (ISSN based link)
    Category: fullText
    Text: Full Text
    Icon: https://imageserver.ebscohost.com/NetImages/iconPdf.gif
    MouseOverText: Check this PubMed for the article full text.
  – Url: https://resolver.ebscohost.com/openurl?sid=EBSCO:edsswe&genre=article&issn=2197425X&ISBN=&volume=13&issue=&date=20251022&spage=1&pages=1-12&title=Intensive Care Medicine Experimental&atitle=Explainable%20AI%20identifies%20key%20biomarkers%20for%20acute%20kidney%20injury%20prediction%20in%20the%20ICU&aulast=Koozi%2C%20Hazem&id=DOI:10.1186/s40635-025-00816-x
    Name: Full Text Finder
    Category: fullText
    Text: Full Text Finder
    Icon: https://imageserver.ebscohost.com/branding/images/FTF.gif
    MouseOverText: Full Text Finder
  – Url: https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=EBSCO&SrcAuth=EBSCO&DestApp=WOS&ServiceName=TransferToWoS&DestLinkType=GeneralSearchSummary&Func=Links&author=Koozi%20H
    Name: ISI
    Category: fullText
    Text: Nájsť tento článok vo Web of Science
    Icon: https://imagesrvr.epnet.com/ls/20docs.gif
    MouseOverText: Nájsť tento článok vo Web of Science
Header DbId: edsswe
DbLabel: SwePub
An: edsswe.oai.portal.research.lu.se.publications.3a6855d9.bc8b.4cc4.9b0b.e3ab952db56e
RelevancyScore: 1174
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 1173.8935546875
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Explainable AI identifies key biomarkers for acute kidney injury prediction in the ICU
– Name: Author
  Label: Authors
  Group: Au
  Data: &lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Koozi%2C+Hazem%22&quot;&gt;Koozi, Hazem&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Engstr&#246;m%2C+Jonas%22&quot;&gt;Engstr&#246;m, Jonas&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Friberg%2C+Hans%22&quot;&gt;Friberg, Hans&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;AR&quot; term=&quot;%22Frigyesi%2C+Attila%22&quot;&gt;Frigyesi, Attila&lt;/searchLink&gt;
– Name: Author
  Label: Contributors
  Group: Au
  Data: Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Lunds universitet, Medicinska fakulteten, Institutionen f&#246;r kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivv&#229;rd, Originator&lt;br /&gt;Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Center for cardiac arrest, Lunds universitet, Medicinska fakulteten, Institutionen f&#246;r kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivv&#229;rd, Centrum f&#246;r hj&#228;rtstopp, Originator&lt;br /&gt;Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), eSSENCE: The e-Science Collaboration, Lunds universitet, Profilomr&#229;den och andra starka forskningsmilj&#246;er, Strategiska forskningsomr&#229;den (SFO), eSSENCE: The e-Science Collaboration, Originator&lt;br /&gt;Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), EpiHealth: Epidemiology for Health, Lunds universitet, Profilomr&#229;den och andra starka forskningsmilj&#246;er, Strategiska forskningsomr&#229;den (SFO), EpiHealth: Epidemiology for Health, Originator&lt;br /&gt;Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Intensive Care Epidemiology, Lunds universitet, Medicinska fakulteten, Institutionen f&#246;r kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivv&#229;rd, Intensivv&#229;rdsepidemiologi, Originator
– Name: TitleSource
  Label: Source
  Group: Src
  Data: &lt;i&gt;Intensive Care Medicine Experimental&lt;/i&gt;. 13:1-12
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: &lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Medical+and+Health+Sciences%22&quot;&gt;Medical and Health Sciences&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Clinical+Medicine%22&quot;&gt;Clinical Medicine&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Anesthesiology+and+Intensive+Care%22&quot;&gt;Anesthesiology and Intensive Care&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Medicin+och+h&#228;lsovetenskap%22&quot;&gt;Medicin och h&#228;lsovetenskap&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Klinisk+medicin%22&quot;&gt;Klinisk medicin&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Anestesi+och+intensivv&#229;rd%22&quot;&gt;Anestesi och intensivv&#229;rd&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Urology%22&quot;&gt;Urology&lt;/searchLink&gt;&lt;br /&gt;&lt;searchLink fieldCode=&quot;DE&quot; term=&quot;%22Urologi%22&quot;&gt;Urologi&lt;/searchLink&gt;
– Name: Abstract
  Label: Description
  Group: Ab
  Data: 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 &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). 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.&lt;br /&gt;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: Thestudy 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). 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.
– Name: URL
  Label: Access URL
  Group: URL
  Data: &lt;link linkTarget=&quot;URL&quot; linkTerm=&quot;https://doi.org/10.1186/s40635-025-00816-x&quot; linkWindow=&quot;_blank&quot;&gt;https://doi.org/10.1186/s40635-025-00816-x&lt;/link&gt;
PLink https://erproxy.cvtisr.sk/sfx/access?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsswe&AN=edsswe.oai.portal.research.lu.se.publications.3a6855d9.bc8b.4cc4.9b0b.e3ab952db56e
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1186/s40635-025-00816-x
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 12
        StartPage: 1
    Subjects:
      – SubjectFull: Medical and Health Sciences
        Type: general
      – SubjectFull: Clinical Medicine
        Type: general
      – SubjectFull: Anesthesiology and Intensive Care
        Type: general
      – SubjectFull: Medicin och hälsovetenskap
        Type: general
      – SubjectFull: Klinisk medicin
        Type: general
      – SubjectFull: Anestesi och intensivvård
        Type: general
      – SubjectFull: Urology
        Type: general
      – SubjectFull: Urologi
        Type: general
    Titles:
      – TitleFull: Explainable AI identifies key biomarkers for acute kidney injury prediction in the ICU
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Koozi, Hazem
      – PersonEntity:
          Name:
            NameFull: Engström, Jonas
      – PersonEntity:
          Name:
            NameFull: Friberg, Hans
      – PersonEntity:
          Name:
            NameFull: Frigyesi, Attila
      – PersonEntity:
          Name:
            NameFull: Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Originator
      – PersonEntity:
          Name:
            NameFull: Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Center for cardiac arrest, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Centrum för hjärtstopp, Originator
      – PersonEntity:
          Name:
            NameFull: Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), eSSENCE: The e-Science Collaboration, Lunds universitet, Profilområden och andra starka forskningsmiljöer, Strategiska forskningsområden (SFO), eSSENCE: The e-Science Collaboration, Originator
      – PersonEntity:
          Name:
            NameFull: Lund University, Profile areas and other strong research environments, Strategic research areas (SRA), EpiHealth: Epidemiology for Health, Lunds universitet, Profilområden och andra starka forskningsmiljöer, Strategiska forskningsområden (SFO), EpiHealth: Epidemiology for Health, Originator
      – PersonEntity:
          Name:
            NameFull: Lund University, Faculty of Medicine, Department of Clinical Sciences, Lund, Section II, Anesthesiology and Intensive Care, Intensive Care Epidemiology, Lunds universitet, Medicinska fakulteten, Institutionen för kliniska vetenskaper, Lund, Sektion II, Anestesiologi och intensivvård, Intensivvårdsepidemiologi, Originator
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 22
              M: 10
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 2197425X
            – Type: issn-locals
              Value: SWEPUB_FREE
            – Type: issn-locals
              Value: LU_SWEPUB
          Numbering:
            – Type: volume
              Value: 13
          Titles:
            – TitleFull: Intensive Care Medicine Experimental
              Type: main
ResultId 1