PD Simulation Platform CAMPSIS

library(campsis)

This vignette intends to demonstrate that CAMPSIS can be used to implement almost any PK/PD model, including complex ones.

Filgrastim PK/PD model

Load the filgrastim PK/PD model from the model library as follows. Please note that this model was translated from NONMEM code. The original model file can be found here on the DDMORE repository. Eventually, this model was updated with the final parameters from the corresponding publication (Krzyzanski et al., 2010).

pkpd <- model_suite$literature$filgrastim_pkpd_krzyzanski
pkpd

Let's create a simple demonstration dataset of 250 subjects:

baseDataset <- Dataset(250) %>% 
  add(Covariate("BAS", 0.02)) %>%
  add(Covariate("WT", UniformDistribution(50, 100))) %>%
  add(DoseAdaptation("WT*AMT")) %>% # per kilo dosing
  add(Observations(0:216)) %>%
  add(Covariate("ROUT", 0)) # subcutaneous route (SC)

Assume we want to compare the following subcutaneous administrations of filgrastim:

2.5 μg/kg QD for a week
5 μg/kg QD for a week
10 μg/kg QD for a week

We define the following scenarios:

scenarios <- Scenarios() %>% 
  add(Scenario("2.5 μg/kg SC", dataset=~.x %>% add(Bolus(time=0, amount=2.5, compartment=1, ii=24, addl=6)))) %>%
  add(Scenario("5 μg/kg SC", dataset=~.x %>% add(Bolus(time=0, amount=5, compartment=1, ii=24, addl=6)))) %>%
  add(Scenario("10 μg/kg SC", dataset=~.x %>% add(Bolus(time=0, amount=10, compartment=1, ii=24, addl=6))))

A quick simulation gives us the plasma concentration of filgrastim, as well as the absolute neutrophil count (ANC):

library(ggplot2)

results <- pkpd %>% simulate(dataset=baseDataset, scenarios=scenarios, seed=1)
results <- results %>% dplyr::mutate(SCENARIO=factor(SCENARIO, levels=unique(SCENARIO)))

p1 <- shadedPlot(results, "CP", "SCENARIO") + facet_wrap(~SCENARIO) +
  scale_y_log10(breaks=c(.01,.1,1,10,100)) + ylab("G-CSF Serum Concentration (ng/mL)")
p2 <- shadedPlot(results, "A_14", "SCENARIO") + facet_wrap(~SCENARIO) + 
  ylab("ANC (10^3 cells/μL)")

gridExtra::grid.arrange(p1, p2, nrow=2)