Truace
Subscribe

Machine learning-based nomogram for acute kidney injury after liver transplantation

Acute kidney injury (AKI) after liver transplantation (LT) is a frequent complication associated with adverse outcomes, yet practical early postoperative prediction tools remain limited. This retrospective study developed and internally validated an interpretable model for AKI after LT using perioperative variables available by the time of the first urgent postoperative blood test, generally obtained within 1 h after surgery. Adult LT recipients treated from June 2023 to January 2026 were screened. AKI was defin…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1221· cite & verify
Machine learning-based nomogram for acute kidney injury after liver transplantation
"Bill on dialysis" by fireflythegreat is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/.

In brief

Researchers retrospectively developed an early postoperative interpretable nomogram to predict acute kidney injury after liver transplantation using four variables available within about one hour of surgery. In 127 patients split into training and validation cohorts, the model achieved AUCs of 0.861 and 0.761, with a support vector machine variant reaching 0.846 on validation.

Early AKI prediction matters because AKI after liver transplantation is frequent and linked to adverse outcomes, and a tool using routine early labs could trigger intensified renal monitoring. Uncertainty remains because the model has only internal validation in a single-center cohort and the authors state external multicenter validation is required before clinical implementation.

Main points

  1. Retrospective study of 127 adult liver transplant recipients from June 2023 to January 2026, split 3:1 training and validation, defined AKI by KDIGO criteria.
  2. LASSO selected eGFR by CKD-EPI, anhepatic phase time, postoperative natural logarithm-transformed D-dimer, and postoperative alanine aminotransferase; lower eGFR and longer anhepatic time were independently associated with AKI.
  3. Model used variables available by first urgent postoperative blood test generally within 1 h after surgery, with SHAP identifying eGFR as dominant predictor and internal validation by five-fold cross-validation and bootstrap.

The gain

An interpretable machine learning nomogram using eGFR, anhepatic phase time, postoperative log D-dimer and ALT available within 1 hour after liver transplantation predicted AKI with AUC 0.861 in training and 0.761 in validation, with SVM reaching 0.846, to enable early identification for intensified renal monitoring.

The rundown

The study screened 146 adult LT recipients and included 127, using perioperative variables available by the first urgent postoperative blood test within 1 hour. LASSO selected four predictors and multivariable logistic regression linked lower eGFR and longer anhepatic phase time to AKI defined by KDIGO.

Performance was reported as AUC 0.861 training and 0.761 validation for the nomogram, with SVM highest at 0.846 among five models, supported by five-fold cross-validation, bootstrap internal validation, and SHAP analysis highlighting eGFR as dominant.

Sources

  1. Peer-reviewedRenal Failure2026-09-27

The debate