Limits...
Model-Based Individualized Treatment of Chemotherapeutics: Bayesian Population Modeling and Dose Optimization.

Jayachandran D, Laínez-Aguirre J, Rundell A, Vik T, Hannemann R, Reklaitis G, Ramkrishna D - PLoS ONE (2015)

Bottom Line: Pharmacogenomic variation observed in the TPMT enzyme produces a significant variation in drug response among the patient population.In the face of scarcity of data in clinical settings, a global sensitivity analysis based model reduction approach is used to minimize the parameter space.The model and the control approach can be utilized in the clinical setting to individualize 6-MP dosing based on the patient's ability to metabolize the drug instead of the traditional standard-dose-for-all approach.

View Article: PubMed Central - PubMed

Affiliation: School of Chemical Engineering, Purdue University, 480 Stadium Mall Way, West Lafayette, IN, 47907, United States of America.

ABSTRACT
6-Mercaptopurine (6-MP) is one of the key drugs in the treatment of many pediatric cancers, auto immune diseases and inflammatory bowel disease. 6-MP is a prodrug, converted to an active metabolite 6-thioguanine nucleotide (6-TGN) through enzymatic reaction involving thiopurine methyltransferase (TPMT). Pharmacogenomic variation observed in the TPMT enzyme produces a significant variation in drug response among the patient population. Despite 6-MP's widespread use and observed variation in treatment response, efforts at quantitative optimization of dose regimens for individual patients are limited. In addition, research efforts devoted on pharmacogenomics to predict clinical responses are proving far from ideal. In this work, we present a Bayesian population modeling approach to develop a pharmacological model for 6-MP metabolism in humans. In the face of scarcity of data in clinical settings, a global sensitivity analysis based model reduction approach is used to minimize the parameter space. For accurate estimation of sensitive parameters, robust optimal experimental design based on D-optimality criteria was exploited. With the patient-specific model, a model predictive control algorithm is used to optimize the dose scheduling with the objective of maintaining the 6-TGN concentration within its therapeutic window. More importantly, for the first time, we show how the incorporation of information from different levels of biological chain-of response (i.e. gene expression-enzyme phenotype-drug phenotype) plays a critical role in determining the uncertainty in predicting therapeutic target. The model and the control approach can be utilized in the clinical setting to individualize 6-MP dosing based on the patient's ability to metabolize the drug instead of the traditional standard-dose-for-all approach.

No MeSH data available.


Related in: MedlinePlus

Marginal parameter distribution with correlation matrix estimated through Bayesian approach.The diagonal cells show the marginal distribution for individual parameters. The off-diagonal cells show the pairwise joint distribution of parameters and their corresponding correlation coefficient.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4520687&req=5

pone.0133244.g003: Marginal parameter distribution with correlation matrix estimated through Bayesian approach.The diagonal cells show the marginal distribution for individual parameters. The off-diagonal cells show the pairwise joint distribution of parameters and their corresponding correlation coefficient.

Mentions: With the data sets D1 and D2, model parameters were estimated using the maximum likelihood approach as explained in Section 2.4.1. Feasible ranges of parameters were chosen based on experimental and clinical studies from literature. The statistics of the estimated parameters are given in Table 2. We assumed that the priors, p(θ, ξ), are log-normally distributed with this statistics. Using this prior distribution and individual patients’ 6-TGN data , the posterior joint distribution of model and error parameters were estimated for each patient through the variational-Bayes approach outlined earlier. The posterior distributions for selected patients were verified with MCMC approach as this is a standard approach for performing Bayesian estimations. MCMC was done using Metropolis-Hastings algorithm implemented through R package ‘mcmcpack’ [44,45]. The posterior parameter distribution is essentially an updated form of the prior distribution in light of new information i.e. individual patient’s 6-TGN concentration. Fig 3 shows marginal distribution of model and error parameters for a representative patient. Correlation coefficients for all parameters are acceptable except between kcm and kme. To examine the adequacy of the model in representing drug concentration data for various patients, we employed the global lack-of-fit test described by Blau et al. [26]. This test compares the occurrence of the experimental points within the highest probability density (HPD) regions for concentration predicted by the model. By definition, HPD is the region in which there is a 100(1 − α)% probability that the true value falls within the area under which satisfies . This confidence region (CR) is the smallest interval region among all credible intervals and hence is termed as the highest probability density region [46]. In our simulation, 259 out of 263 experimental points remained within the 95% HPD concentration confidence region, resulting in a confidence level for the lack-of-fit of 0.019. Since this value is less than 0.05, the selected model is adequate as measured by the global lack-of-fit test.


Model-Based Individualized Treatment of Chemotherapeutics: Bayesian Population Modeling and Dose Optimization.

Jayachandran D, Laínez-Aguirre J, Rundell A, Vik T, Hannemann R, Reklaitis G, Ramkrishna D - PLoS ONE (2015)

Marginal parameter distribution with correlation matrix estimated through Bayesian approach.The diagonal cells show the marginal distribution for individual parameters. The off-diagonal cells show the pairwise joint distribution of parameters and their corresponding correlation coefficient.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4520687&req=5

pone.0133244.g003: Marginal parameter distribution with correlation matrix estimated through Bayesian approach.The diagonal cells show the marginal distribution for individual parameters. The off-diagonal cells show the pairwise joint distribution of parameters and their corresponding correlation coefficient.
Mentions: With the data sets D1 and D2, model parameters were estimated using the maximum likelihood approach as explained in Section 2.4.1. Feasible ranges of parameters were chosen based on experimental and clinical studies from literature. The statistics of the estimated parameters are given in Table 2. We assumed that the priors, p(θ, ξ), are log-normally distributed with this statistics. Using this prior distribution and individual patients’ 6-TGN data , the posterior joint distribution of model and error parameters were estimated for each patient through the variational-Bayes approach outlined earlier. The posterior distributions for selected patients were verified with MCMC approach as this is a standard approach for performing Bayesian estimations. MCMC was done using Metropolis-Hastings algorithm implemented through R package ‘mcmcpack’ [44,45]. The posterior parameter distribution is essentially an updated form of the prior distribution in light of new information i.e. individual patient’s 6-TGN concentration. Fig 3 shows marginal distribution of model and error parameters for a representative patient. Correlation coefficients for all parameters are acceptable except between kcm and kme. To examine the adequacy of the model in representing drug concentration data for various patients, we employed the global lack-of-fit test described by Blau et al. [26]. This test compares the occurrence of the experimental points within the highest probability density (HPD) regions for concentration predicted by the model. By definition, HPD is the region in which there is a 100(1 − α)% probability that the true value falls within the area under which satisfies . This confidence region (CR) is the smallest interval region among all credible intervals and hence is termed as the highest probability density region [46]. In our simulation, 259 out of 263 experimental points remained within the 95% HPD concentration confidence region, resulting in a confidence level for the lack-of-fit of 0.019. Since this value is less than 0.05, the selected model is adequate as measured by the global lack-of-fit test.

Bottom Line: Pharmacogenomic variation observed in the TPMT enzyme produces a significant variation in drug response among the patient population.In the face of scarcity of data in clinical settings, a global sensitivity analysis based model reduction approach is used to minimize the parameter space.The model and the control approach can be utilized in the clinical setting to individualize 6-MP dosing based on the patient's ability to metabolize the drug instead of the traditional standard-dose-for-all approach.

View Article: PubMed Central - PubMed

Affiliation: School of Chemical Engineering, Purdue University, 480 Stadium Mall Way, West Lafayette, IN, 47907, United States of America.

ABSTRACT
6-Mercaptopurine (6-MP) is one of the key drugs in the treatment of many pediatric cancers, auto immune diseases and inflammatory bowel disease. 6-MP is a prodrug, converted to an active metabolite 6-thioguanine nucleotide (6-TGN) through enzymatic reaction involving thiopurine methyltransferase (TPMT). Pharmacogenomic variation observed in the TPMT enzyme produces a significant variation in drug response among the patient population. Despite 6-MP's widespread use and observed variation in treatment response, efforts at quantitative optimization of dose regimens for individual patients are limited. In addition, research efforts devoted on pharmacogenomics to predict clinical responses are proving far from ideal. In this work, we present a Bayesian population modeling approach to develop a pharmacological model for 6-MP metabolism in humans. In the face of scarcity of data in clinical settings, a global sensitivity analysis based model reduction approach is used to minimize the parameter space. For accurate estimation of sensitive parameters, robust optimal experimental design based on D-optimality criteria was exploited. With the patient-specific model, a model predictive control algorithm is used to optimize the dose scheduling with the objective of maintaining the 6-TGN concentration within its therapeutic window. More importantly, for the first time, we show how the incorporation of information from different levels of biological chain-of response (i.e. gene expression-enzyme phenotype-drug phenotype) plays a critical role in determining the uncertainty in predicting therapeutic target. The model and the control approach can be utilized in the clinical setting to individualize 6-MP dosing based on the patient's ability to metabolize the drug instead of the traditional standard-dose-for-all approach.

No MeSH data available.


Related in: MedlinePlus