Citations

Background: Integrating machine learning (ML) with population pharmacokinetic (PPK) modeling may improve therapeutic drug monitoring predictions. Methods: Tacrolimus trough concentrations from lung transplant patients were split into a training data set (1151 concentrations in 80 patients) and a testing data set (224 concentrations in 20 patients). A PPK model was developed using NONMEM, followed by the development of 10 ML models to fit individual pharmacokinetic parameters from the PPK model with Bayesian forecasting. The best performing ML model was selected as the final model. Both the final PPK and ML models were compared for prediction performance. A web-based dashboard was established with R-shiny to recommend dosing regimens based on patient data. Results: In the PPK model, postoperative days, hematocrit, aspartate aminotransferase, tacrolimus daily dose, coadministered voriconazole or posaconazole, and the CYP3A5*3 genotype were identified significant covariates on the clearance. The Cubist ML model outperformed the PPK model, showing lower root mean squared error for tacrolimus concentrations in the testing data set. An online web-based precision dosing dashboard was created, accessible at (https://tac-dose-ml.shinyapps.io/shiny/). Conclusions: Integrating ML with PPK modeling could yield superior tacrolimus concentration predictions for lung transplant patients, offering an efficient alternative to Bayesian forecasting. The online dashboard provides rapid, individualized dosing recommendations.

The relationship between selenium and metabolic syndrome (MetS) has been discussed controversially, and limited studies have examined the associations of single nucleotide polymorphisms in selenoproteins genes with MetS. Hence, to examine the associations of plasma selenium concentrations and selenoprotein P rs7579 polymorphism with MetS, a case-control study of 1279 MetS cases and 1279 sex- and age- (±2 years) matched controls was conducted based on the baseline data of the Tongji-Ezhou Cohort study. Plasma selenium concentrations were measured by inductively coupled plasma mass spectrometry. MetS was defined using the definition of the Joint Interim Statement, adjusted for the Chinese population. In addition, the rs7579 polymorphism was genotyped by the Agena MassARRAY System. Plasma selenium concentrations in the MetS group were higher than in the control group (93.88 μg/L (83.17-107.41) vs. 92.66 μg/L (82.36-103.53), P<0.05). Compared with quartile 4 (≥103.53 μg/L), the multivariate-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) associated with MetS were 0.79 (0.59-1.06) for quartile 1 (<82.36 μg/L), 0.75 (0.56-1.01) for quartile 2 (82.37-92.66 μg/L), and 0.61 (0.45-0.83) for quartile 3 (92.67-103.52 μg/L). The cubic spline analyses revealed a U-shaped association between plasma selenium and MetS, with the lowest risk at around 93.69 μg/L. Moreover, in cubic spline analyses, plasma selenium showed U-shaped associations with central obesity and high blood pressure, positive associations with hypertriglyceridemia and hyperglycemia, and a negative association with low high-density lipoprotein cholesterol. Additionally, both the GA and GA+AA genotype carriers were associated with increased ORs of MetS comparing with the GG genotype carriers. Our findings suggested a U-shaped association between plasma selenium and MetS and diverse associations between plasma selenium and components of MetS. Furthermore, our study found that the A allele of rs7579 was associated with higher odds of MetS. Further studies are needed to confirm our findings and elucidate the underlying mechanisms.