From Crush Triage to Kidney Rescue: Artificial Intelligence for Early Renal Risk Stratification in Crush Injury, Rhabdomyolysis and Major Trauma A Scoping Review and Evidence-Informed Framework Development Study

Abstract

CRUSH-AKI-AI (From Crush Triage to Kidney Rescue: Artificial Intelligence for Early Renal Risk Stratification in Crush Injury, Rhabdomyolysis and Major Trauma) is a two-phase methodological study combining a scoping review with evidence-informed conceptual framework development. The project aims to systematically map the use of artificial intelligence (AI), machine learning (ML), and explainable AI (XAI) for the early prediction of acute kidney injury (AKI), severe AKI, kidney replacement therapy (KRT), and other clinically relevant renal outcomes in patients with crush injury, rhabdomyolysis, major trauma, polytrauma, or traumatic shock. Phase 1 will be conducted according to JBI methodology for scoping reviews and reported in accordance with PRISMA-ScR and PRISMA-S. The review will map populations, clinical settings, algorithms, predictors, prediction horizons, validation approaches, discrimination and calibration metrics, clinical utility, missing-data handling, explainability, fairness, implementation maturity, and potential transferability to disaster and mass-casualty settings. PROBAST+AI will be used descriptively to assess risk of bias and applicability, while TRIPOD+AI will inform the assessment of reporting transparency. Searches cover MEDLINE/PubMed, Embase, CINAHL, Scopus, Web of Science Core Collection, IEEE Xplore, ACM Digital Library, and Google Scholar, supplemented by backward and forward citation searching. The database-specific search strategies and yields are archived separately; the initial searches identified records across all eight planned information sources. Phase 2 will translate the mapped evidence through a prespecified and traceable evidence-to-framework process. Planned outputs include a Minimum Renal Dataset for Mass-Casualty Care, a Renal AI Operational Readiness Matrix, and the preliminary R3-AKI (Renal Risk Response for Acute Kidney Injury) Framework, structured using a green-yellow-red renal-risk architecture. R3-AKI is intended as an evidence-informed renal-risk overlay and not as a validated triage system, clinical prediction model, or autonomous decision-making tool. It will not replace or downgrade established mass-casualty triage priorities, and no unsupported clinical or probability thresholds will be generated. The study uses exclusively published or publicly accessible evidence and involves no recruitment of human participants or processing of non-public individual-level health data.