Abstract Title

Population Pharmacokinetics of Pemetrexed in Adult Non-Small Cell Lung Cancer Patients in India

Presenter Name

Meenakshi Srinivasan

RAD Assignment Number

2003

Abstract

Purpose: Pemetrexed (PEM) is indicated for the treatment of non-small cell lung cancer (NSCLC) in combination with cisplatin. The study aimed to characterize pharmacokinetics (PK) and effects of covariates on PK of PEM using a population pharmacokinetic (PPK) approach.

Methods: PPK analysis was performed using plasma samples from 85 patients following 500 mg/m2 IV infusion. The model was developed using NONMEM® (v.7.3.0). Diagnostic plots were generated with packages XPOSE and ggplot2 in R (v.3.4.2). The structural model was a two compartment model parameterized with clearance (CL), central and peripheral volume of distribution (V1 & V2), and inter-compartmental clearance (Q). Exponential error model was assumed for inter-individual variability (IIV) in PK parameters. Residual variability (RV) was modeled as combined additive (ADD) and proportional (PROP) error model. The estimation method was the stochastic approximation expectation maximization (SAEM). Standard errors (SE) and objective function value were calculated using Monte Carlo importance sampling method. Full covariate model approach was utilized by specifying clinically meaningful covariates in the model. Power model was used to describe covariate relationships. Markov chain monte carlo Bayesian analysis (BAYES) was used to compute 95% credible intervals (CI) for the posterior distribution for covariate effects. This CI was used to reduce the full covariate model by eliminating the covariates whose CI included null hypothesis.

Results:Total of 850 PEM levels were used for model development. The final model included weight allometrically scaled to V1, V2 and Q. The fixed effect parameter estimates were CL-3.13 L/h; V1–5.54 L; V2–7.01 L and Q–5.42 L/h. Covariates that retained significance following BAYES CI’s- CrCl on CL, Sex on CL and Albumin on V1 with estimates of 0.52, 1.22 and -1.46 respectively. IIV on CL, V1, V2 and Q were 52%, 51%, 122% and 88% respectively. Relative SE of estimates were in the range of 10 to 43 % and 18 to 128% for fixed and random effect parameters respectively. ADD (SD) was 0.54 µg/mL and PROP RV was 29%. Goodness-of-fit plots showed no major bias in the model. VPC showed agreement between distribution of model simulated and observed data.

Conclusion: PPK model for PEM in Indian subjects was successfully developed using full covariate modeling approach. The covariate relationships identified could be used to individualize dosing based on patient characteristics.

Is your abstract for competition or not for competition?

Competition

Research Area

Pharmacology

Presentation Type

Poster

This document is currently not available here.

Share

COinS
 

Population Pharmacokinetics of Pemetrexed in Adult Non-Small Cell Lung Cancer Patients in India

Purpose: Pemetrexed (PEM) is indicated for the treatment of non-small cell lung cancer (NSCLC) in combination with cisplatin. The study aimed to characterize pharmacokinetics (PK) and effects of covariates on PK of PEM using a population pharmacokinetic (PPK) approach.

Methods: PPK analysis was performed using plasma samples from 85 patients following 500 mg/m2 IV infusion. The model was developed using NONMEM® (v.7.3.0). Diagnostic plots were generated with packages XPOSE and ggplot2 in R (v.3.4.2). The structural model was a two compartment model parameterized with clearance (CL), central and peripheral volume of distribution (V1 & V2), and inter-compartmental clearance (Q). Exponential error model was assumed for inter-individual variability (IIV) in PK parameters. Residual variability (RV) was modeled as combined additive (ADD) and proportional (PROP) error model. The estimation method was the stochastic approximation expectation maximization (SAEM). Standard errors (SE) and objective function value were calculated using Monte Carlo importance sampling method. Full covariate model approach was utilized by specifying clinically meaningful covariates in the model. Power model was used to describe covariate relationships. Markov chain monte carlo Bayesian analysis (BAYES) was used to compute 95% credible intervals (CI) for the posterior distribution for covariate effects. This CI was used to reduce the full covariate model by eliminating the covariates whose CI included null hypothesis.

Results:Total of 850 PEM levels were used for model development. The final model included weight allometrically scaled to V1, V2 and Q. The fixed effect parameter estimates were CL-3.13 L/h; V1–5.54 L; V2–7.01 L and Q–5.42 L/h. Covariates that retained significance following BAYES CI’s- CrCl on CL, Sex on CL and Albumin on V1 with estimates of 0.52, 1.22 and -1.46 respectively. IIV on CL, V1, V2 and Q were 52%, 51%, 122% and 88% respectively. Relative SE of estimates were in the range of 10 to 43 % and 18 to 128% for fixed and random effect parameters respectively. ADD (SD) was 0.54 µg/mL and PROP RV was 29%. Goodness-of-fit plots showed no major bias in the model. VPC showed agreement between distribution of model simulated and observed data.

Conclusion: PPK model for PEM in Indian subjects was successfully developed using full covariate modeling approach. The covariate relationships identified could be used to individualize dosing based on patient characteristics.