JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Integration of urine retinol-binding protein and genetic markers for early prediction of tacrolimus nephrotoxicity using machine learning

  • 0Department of Pharmacy, Guangzhou Women and Children's Medical Center, Guangzhou Medical University, Guangzhou, China.
Translational Pediatrics +

|

September 15, 2025

|

September 15, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

A new machine learning model predicts tacrolimus (TAC) nephrotoxicity in children with nephrotic syndrome (NS) using genetic markers and urine retinol-binding protein (RP). This tool aids physicians in preventing kidney injury by assessing individual risks.

Area Of Science

  • Pharmacogenomics
  • Nephrology
  • Computational Biology

Background

  • Tacrolimus (TAC) use is limited by variable nephrotoxicity.
  • Urine retinol-binding protein (RP) shows potential for early detection of TAC-induced renal tubular injury.
  • Nephrotic syndrome (NS) patients exhibit significant individual differences in TAC response.

Purpose Of The Study

  • To develop and validate a machine learning model for predicting TAC nephrotoxicity in pediatric NS patients.
  • To integrate clinical features and genetic markers for enhanced prediction accuracy.
  • To utilize urine RP as a key indicator of renal tubular damage.

Main Methods

  • Retrospective cohort of 203 children with NS for model development; 12 for external validation.
  • Inclusion of 38 clinical features and 80 genetic variables.
  • Evaluation of five machine learning algorithms: Extra Random Trees, Gradient Boosting Decision Tree, random forests, eXtreme Gradient Boosting, and logistic regression (LR).

Main Results

  • The logistic regression (LR) model demonstrated the best performance.
  • Key predictors included six genetic markers: CYP3A5*3, NFATC1 rs1660144, NFKB1 rs230526, NFKBIA rs696, CD2AP rs12664637, and PLCE1 rs2274223.
  • The LR model achieved a sensitivity of 78.6%, specificity of 63.8%, accuracy of 67.2%, and AUC of 76.1%.

Conclusions

  • A validated machine learning model incorporating genetic factors predicts TAC renal tubular toxicity in NS patients.
  • The model utilizes urine RP as a marker for renal toxicity.
  • This tool empowers physicians to personalize TAC treatment plans and mitigate nephrotoxicity risk.

Keywords: