Population Pharmacokinetics Of High-Dose Methotrexate In Adults And Children With Non-Hodgkin Lymphoma

High-dose methotrexate?HDMTX?combined with leucovorin?LV?rescue is a common treatment for the non-Hodgkin's lymphoma?NHL?.MTX is an anti-folate antitumor drug,which has large toxicity and side effects,narrow treatment window and large inter-individual variationin pharmacokinetics.Population pharmacokinetics?PPK?can identify and quantitatively study the physiological and pathological factors that affect pharmacokinetics of MTX,and provide a reference for individualized treatment.Objectives:?1?To establish PPK models of HDMTX for adult and children NHL patients respectively,and to obtain the population typical values of corresponding pharmacokinetic parameters.?2?To study factors that willaffect the pharmacokinetic behavior of MTX in NHL patients.?3?To validate the stability and predictive performance of the established PPK model internally and externally,so as to provide a reference for the individualized administration of MTX to NHL patients.Methods:?1?Adult NHL patient PPK study:Demographics and pharmacokinetic data of adult NHL patients who received HDMTX treatment were collected from Sun Yat-sen University Cancer Center from January 1,2016 to September 1,2018.A nonlinear mixed-effects model?NOMEN?was used to establish a population pharmacokinetic model,and potentialcovariates such as height,weight,age,body surface area,laboratory testing indicators,and combination medication were explored.After the final model was established,internal and external verification were performed to evaluate the predictive performance and stability of the final model.Internal verification included scatter plots of predicted concentrations vs measured concentrations,individual weighted residual?IWRES?plots,visual predictive check?VPC?plots,and bootstrap method.External verification used the data set that was not involved in modeling as the verification group,and calculated the average prediction error?MPE?,average absolute prediction error?MAE?,and relative prediction error within±20%and±30%to evaluate the predictive performance of the model.?2?Children NHL patient PPK study:relevant data for children with NHL who received HDMTX treatment from January 1,2014 to September 1,2018 were collected from the Sun Yat-sen University Cancer Center.Modeling and verification methods were similar to adult study.Results:?1?Results for the adult study:A total of 189 adult NHL patients were included.,159 patients were randomly selected as the modeling group and 30 patients were used as the verification group with the aid of SPSS 21.0.The two-compartment model was selected as the structural model,and the exponential model was used to describe the variability between patients and residual variability.A stepwise method was used to select covariates,including forward inclusion and backward elimination,the final PPK model were CL=11.1×?AGE/46?^-0.24×?BSA/1.67?^0.796×?CYSC/0.84?^-0.342L/h;V₁=23.8×?WT/60?^0.698L;Q=0.544 L;V₂=4.96 L.The results showed that age,BSA,and CYSC were covariates that affect ED CL,while weight was a covariate that affected V1.The diagnostic plot results showed that the final model had a good goodness of fit.The bootstrap vaildation results showed that the parameters estimated by the final model were close to the median of the parameters obtained by Bootstrap,and all fell within the 95%confidence interval,indicating that the final model was stable.In the part of external verification,The MPE and MAE values of the final model were 5.8%and 26%respectively,which were better than the basic model?13.6%,40.7%?,indicating that the final model had good prediction performance.?2?Results for the children study:A total of 128 children with NHL were included,with 98children as the modeling group and 30 children as the verification group.The final model for childen were CL=7.31×?WT/24.25?^0.824×?CYSC/0.78?^-0.375L/h;V₁=21.1 L;Q=0.138×?WT/24.25?^0.691L/h;V₂=7.82 L.The results showed that body weight and CYSC were covariates that affected CL,and body weight had an effect on Q.The results of the diagnostic plot showed that the final model fitted poorly in the range of high concentration?>10 mg/L?,while the fitting in low concentration range?<1 mg/L?was good,and the bootstrap results showed that the stability of the final model was good.External verification results showed that the MPE and MAE values of the final model were 10.8%and 30.2%,respectively,which were not better than of the basic model?10.4%,31.5%?.However,the MPE and MAE values of the final model in the lower concentration range were 4.7%and16.3%,respectively,which were better than the basic model?6.3%,31.6%?.Conclusion:?1?Adult PPK study:this study successfully established a population pharmacokinetic model of HDMTX in adult NHL patients.Both internal and external verification results showed that the final model had good predictive performance and stability,which could provide a reference for the individualized administration of HDMTX to NHL patients,improve the efficacy of drugs and reduce the occurrence of adverse reactions.?2?Children PPK study:the results of internal and external verification indicated that the model should be further be optimized by changing the design of the blood collection scheme or adopting a more sensitive method for detecting blood concentration.