Early Prediction of Tacrolimus-Induced Tubular Toxicity in Pediatric Refractory Nephrotic Syndrome Using Machine Learning

Early prediction of growth patterns after pediatric kidney transplantation based on height-related single-nucleotide polymorphisms

BACKGROUND

Growth retardation is a common complication of chronic kidney disease in children, which can be partially relieved after renal transplantation. This study aimed to develop and validate a predictive model for growth patterns of children with end-stage renal disease (ESRD) after kidney transplantation using machine learning algorithms based on genomic and clinical variables.

METHODS

A retrospective cohort of 110 children who received kidney transplants between May 2013 and September 2021 at the First Affiliated Hospital of Zhengzhou University were recruited for whole-exome sequencing (WES), and another 39 children who underwent transplant from October 2021 to March 2022 were enrolled for external validation. Based on previous studies, we comprehensively collected 729 height-related single-nucleotide polymorphisms (SNPs) in exon regions. Seven machine learning algorithms and 10-fold cross-validation analysis were employed for model construction.

RESULTS

The 110 children were divided into two groups according to change in height-for-age Z -score. After univariate analysis, age and 19 SNPs were incorporated into the model and validated. The random forest model showed the best prediction efficacy with an accuracy of 0.8125 and an area under curve (AUC) of 0.924, and also performed well in the external validation cohort (accuracy, 0.7949; AUC, 0.796).

CONCLUSIONS

A model with good performance for predicting post-transplant growth patterns in children based on SNPs and clinical variables was constructed and validated using machine learning algorithms. The model is expected to guide clinicians in the management of children after renal transplantation, including the use of growth hormone, glucocorticoid withdrawal, and nutritional supplementation, to alleviate growth retardation in children with ESRD.

Development of a predictive model for nephrotoxicity during tacrolimus treatment using machine learning methods

AIM

When administering tacrolimus, therapeutic drug monitoring is recommended because nephrotoxicity, an adverse event, occurs at supra-therapeutic whole-blood concentrations of tacrolimus. However, some patients exhibit nephrotoxicity even at the recommended concentrations, therefore establishing a therapeutic range of tacrolimus concentration for the individual patient is necessary to avoid nephrotoxicity. This study aimed to develop a model for individualized prediction of nephrotoxicity in patients administered tacrolimus.

METHODS

We collected data, such as laboratory test data at tacrolimus initiation, concomitant drugs and tacrolimus whole-blood concentration, from medical records of patients who received oral tacrolimus. Nephrotoxicity was defined as an increase in serum creatinine levels within 60 days of tacrolimus initiation. We built 13 prediction models based on different machine learning algorithms: logistic regression, support vector machine, gradient-boosting trees, random forest and neural networks. The best performing model was compared with the conventional model, which classifies patients according to the tacrolimus concentration alone.

RESULTS

Data from 163 and 41 patients were used to construct models and evaluate the best performing one, respectively. Most of the patients were diagnosed with inflammatory or autoimmune diseases. The best performing model was built using a support vector machine; it showed a high F2 score of 0.750 and outperformed the conventional model (0.500).

CONCLUSIONS

A machine learning model to predict nephrotoxicity in patients during tacrolimus treatment was developed using tacrolimus whole-blood concentration and other patient data. This model could potentially assist in identifying high-risk patients who require individualized target therapeutic concentrations of tacrolimus prior to treatment initiation to prevent nephrotoxicity.

Optimizing the initial tacrolimus dosage in Chinese children with lung transplantation within normal hematocrit levels

Background

The appropriate initial dosage of tacrolimus is undefined in Chinese pediatric lung transplant patients with normal hematocrit values. The purpose of this study is to optimize the initial dose of tacrolimus in Chinese children who are undergoing lung transplantation and have normal hematocrit levels.

Methods

The present study is based on a published population pharmacokinetic model of tacrolimus in lung transplant patients and uses the Monte Carlo simulation to optimize the initial tacrolimus dosage in Chinese children with lung transplantation within normal hematocrit levels.

Results

Within normal hematocrit levels, for children with lung transplantation who do not carry the CYP3A5*1 gene and have no coadministration with voriconazole, it is recommended to administer tacrolimus at a dosage of 0.02 mg/kg/day, divided into two doses, for children weighing 10-32 kg, and a dosage of 0.03 mg/kg/day, also divided into two doses, for children weighing 32-40 kg. For children with lung transplantation who carry the CYP3A5*1 gene and have no coadministration with voriconazole, tacrolimus dosages of 0.02, 0.03, and 0.04 mg/kg/day split into two doses are recommended for children weighing 10-15, 15-32, and 32-40 kg, respectively. For children with lung transplantation who do not carry the CYP3A5*1 gene and have coadministration with voriconazole, tacrolimus dosages of 0.01 and 0.02 mg/kg/day split into two doses are recommended for children weighing 10-17 and 17-40 kg, respectively. For children with lung transplantation who carry the CYP3A5*1 gene and have coadministration with voriconazole, a tacrolimus dosage of 0.02 mg/kg/day split into two doses is recommended for children weighing 10-40 kg.

Conclusions

It is the first time to optimize the initial dosage of tacrolimus in Chinese children undergoing lung transplantation within normal hematocrit.

Physiologically based pharmacokinetic model combined with reverse dose method to study the nephrotoxic tolerance dose of tacrolimus

Nephrotoxicity is the most common side effect that severely limits the clinical application of tacrolimus (TAC), an immunosuppressive agent used in kidney transplant patients. This study aimed to explore the tolerated dose of nephrotoxicity of TAC in individuals with different CYP3A5 genotypes and liver conditions. We established a human whole-body physiological pharmacokinetic (WB-PBPK) model and validated it using data from previous clinical studies. Following the injection of 1 mg/kg TAC into the tail veins of male rats, we developed a rat PBPK model utilizing the drug concentration-time curve obtained by LC-MS/MS. Next, we converted the established rat PBPK model into the human kidney PBPK model. To establish renal concentrations, the BMCL5 of the in vitro CCK-8 toxicity response curve (drug concentration range: 2-80 mol/L) was extrapolated. To further investigate the acceptable levels of nephrotoxicity for several distinct CYP3A5 genotypes and varied hepatic function populations, oral dosing regimens were extrapolated utilizing in vitro-in vivo extrapolation (IVIVE). The PBPK model indicated the tolerated doses of nephrotoxicity were 0.14-0.185 mg/kg (CYP3A5 expressors) and 0.13-0.155 mg/kg (CYP3A5 non-expressors) in normal healthy subjects and 0.07-0.09 mg/kg (CYP3A5 expressors) and 0.06-0.08 mg/kg (CYP3A5 non-expressors) in patients with mild hepatic insufficiency. Further, patients with moderate hepatic insufficiency tolerated doses of 0.045-0.06 mg/kg (CYP3A5 expressors) and 0.04-0.05 mg/kg (CYP3A5 non-expressors), while in patients with moderate hepatic insufficiency, doses of 0.028-0.04 mg/kg (CYP3A5 expressors) and 0.022-0.03 mg/kg (CYP3A5 non-expressors) were tolerated. Overall, our study highlights the combined usage of the PBPK model and the IVIVE approach as a valuable tool for predicting toxicity tolerated doses of a drug in a specific group.

Comparison and development of machine learning for thalidomide-induced peripheral neuropathy prediction of refractory Crohn’s disease in Chinese population

BACKGROUND

Thalidomide is an effective treatment for refractory Crohn’s disease (CD). However, thalidomide-induced peripheral neuropathy (TiPN), which has a large individual variation, is a major cause of treatment failure. TiPN is rarely predictable and recognized, especially in CD. It is necessary to develop a risk model to predict TiPN occurrence.

AIM

To develop and compare a predictive model of TiPN using machine learning based on comprehensive clinical and genetic variables.

METHODS

A retrospective cohort of 164 CD patients from January 2016 to June 2022 was used to establish the model. The National Cancer Institute Common Toxicity Criteria Sensory Scale (version 4.0) was used to assess TiPN. With 18 clinical features and 150 genetic variables, five predictive models were established and evaluated by the confusion matrix receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), specificity, sensitivity (recall rate), precision, accuracy, and F1 score.

RESULTS

The top-ranking five risk variables associated with TiPN were interleukin-12 rs1353248 [P = 0.0004, odds ratio (OR): 8.983, 95% confidence interval (CI): 2.497-30.90], dose (mg/d, P = 0.002), brain-derived neurotrophic factor (BDNF) rs2030324 (P = 0.001, OR: 3.164, 95%CI: 1.561-6.434), BDNF rs6265 (P = 0.001, OR: 3.150, 95%CI: 1.546-6.073) and BDNF rs11030104 (P = 0.001, OR: 3.091, 95%CI: 1.525-5.960). In the training set, gradient boosting decision tree (GBDT), extremely random trees (ET), random forest, logistic regression and extreme gradient boosting (XGBoost) obtained AUROC values > 0.90 and AUPRC > 0.87. Among these models, XGBoost and GBDT obtained the first two highest AUROC (0.90 and 1), AUPRC (0.98 and 1), accuracy (0.96 and 0.98), precision (0.90 and 0.95), F1 score (0.95 and 0.98), specificity (0.94 and 0.97), and sensitivity (1). In the validation set, XGBoost algorithm exhibited the best predictive performance with the highest specificity (0.857), accuracy (0.818), AUPRC (0.86) and AUROC (0.89). ET and GBDT obtained the highest sensitivity (1) and F1 score (0.8). Overall, compared with other state-of-the-art classifiers such as ET, GBDT and RF, XGBoost algorithm not only showed a more stable performance, but also yielded higher ROC-AUC and PRC-AUC scores, demonstrating its high accuracy in prediction of TiPN occurrence.

CONCLUSION

The powerful XGBoost algorithm accurately predicts TiPN using 18 clinical features and 14 genetic variables. With the ability to identify high-risk patients using single nucleotide polymorphisms, it offers a feasible option for improving thalidomide efficacy in CD patients.

Predicting the effect of sirolimus on disease activity in patients with systemic lupus erythematosus using machine learning

WHAT IS KNOWN AND OBJECTIVES

The present study aimed to predict the effect of sirolimus on disease activity in patients with systemic lupus erythematosus (SLE) using machine learning and to recommend appropriate sirolimus dosage regimen for patients with SLE.

METHODS

The E_max model was selected for machine learning, where the evaluation indicator was the change rate of systemic lupus erythematosus disease activity index from baseline value.

RESULTS

A total 103 patients with SLE were included for modelling, where the E_max , ET₅₀ were -53.9%, 1.53 months in the final model respectively, and the evaluation of the final model was good. Further simulation found that the follow-up time to achieve 25%, 50%, 75% and 80% (plateau) E_max of sirolimus effecting on disease activity in patients with SLE were 0.51, 1.53, 4.59 and 6.12 months, respectively. In addition, the sirolimus dosage was flexible and adjusted according to drug concentration, where the intersection of sirolimus concentration range included in this study was about 8-10 ng/ml.

WHAT IS NEW AND CONCLUSIONS

This study was the first time to predict the effect of sirolimus on disease activity in patients with SLE and in order to achieve better therapeutic effect maintaining a concentration of 8-10 ng/ml sirolimus for at least 6.12 months was necessary.