Background

• Structural model - Two compartment model

• Route of administration - IV-infusion (with an infusion pump)

• Dosage Regimen - 100 mg/m² of Cisplatin for 1-h at time=0 considering a patient with 1.7m²

• Number of Subjects - 1

Learning Outcome

• By the application of the present model, we will learn how to simulate model for reversible metabolism.

• Simulate data for two IV-Infusion with two different rates of infusion regimen and how to develop a better fit for the available data.

Objectives

In this exercise you will learn how to

• Simulate an IV-infusion two compartment for reversible metabolism.

(Here certain assumptions to be considered:-

• A fraction of the dose (2.3%) is present as the monohydrated complex in the infusion solution,

• That there is a reversible reactions between cisplatin (p) and its monohydrated complex (m),

• The input rate can be split into cisplatin infusion rate (Inp) and monohydrate infusion rate (Inm).

• Write a differential equation for a two-compartment model in terms of Clearance and Volume for simulating reversible metabolism.

Libraries

call the "necessary" libraries to get start.

using Pumas
using Plots
using CSV
using StatsPlots
using Random

Model - Microconstant Model

In this two compartment model, we administer the mentioned dose on Central compartment as well as on Metabolite compartment at 'time= 0'. Also, K12 and K21 are the rate constants for the conversion of cisplatin into monohydrated complex and the monohydrated complex into cisplatin, respectively.

pk_34             = @model begin
    @param begin
      tvvc        ∈ RealDomain(lower=0)
      tvclm       ∈ RealDomain(lower=0)
      tvvm        ∈ RealDomain(lower=0)
      tvclp       ∈ RealDomain(lower=0)
      tvk12       ∈ RealDomain(lower=0)
      tvk21       ∈ RealDomain(lower=0)
      Ω           ∈ PDiagDomain(6)
      σ²_prop     ∈ RealDomain(lower=0)
    end

    @random begin
      η           ~ MvNormal(Ω)
    end

    @pre begin
      Vc          = tvvc * exp(η[1])
      CLm         = tvclm * exp(η[2])
      Vm          = tvvm * exp(η[3])
      CLp         = tvclp * exp(η[4])
      K12         = tvk12 * exp(η[5])
      K21         = tvk21 * exp(η[6])
    end

    @dynamics begin
      Central'    = -(CLp/Vc)*Central - K12*Central + K21*Metabolite*Vc/Vm
      Metabolite' = -(CLm/Vm)*Metabolite - K21*Metabolite + K12*Central*Vm/Vc
    end

    @derived begin
      cp          = @. Central/Vc
      dv_cp       ~ @. Normal(cp, sqrt(cp^2*σ²_prop))
      met         = @. Metabolite/Vm
      dv_met      ~ @. Normal(met, sqrt(cp^2*σ²_prop))
    end
end

PumasModel
  Parameters: tvvc, tvclm, tvvm, tvclp, tvk12, tvk21, Ω, σ²_prop
  Random effects: η
  Covariates: 
  Dynamical variables: Central, Metabolite
  Derived: cp, dv_cp, met, dv_met
  Observed: cp, dv_cp, met, dv_met

Parameters

Parameters provided for simulation are as below. tv represents the typical value for parameters.

• Vc - Volume of central compartment (L)

• CLm - Clearance of metabolite (L/min)

• Vm - Volume of metabolite compartment (μg/L)

• CLp - Clearance of parent (L/min)

• K12 - Rate constant for the conversion of cisplatin into monohydrated complex (min⁻¹)

• K21 - Rate constant for the conversion of monohydrated complex into cisplatin (min⁻¹)

• Ω - Between Subject Variability

• σ - Residual error

param = ( tvvc    = 14.1175,
          tvclm   = 0.00832616,
          tvvm    = 2.96699,
          tvclp   = 0.445716,
          tvk12   = 0.00021865,
          tvk21   = 0.021313,
          Ω       = Diagonal([0.0,0.0,0.0,0.0,0.0,0.0]),
          σ²_prop = 0.001)

(tvvc = 14.1175, tvclm = 0.00832616, tvvm = 2.96699, tvclp = 0.445716, tvk1
2 = 0.00021865, tvk21 = 0.021313, Ω = [0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0
; … ; 0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0], σ²_prop = 0.001)

Dosage Regimen

In this section the Dosage regimen is mentioned for:

• Cisplatin Infusion - A total dose of 170mg (100mg/m² * 1.7m²) i split as Cisplatin 166.09 and Monohydrate 3.91.

• Monohydrate Infusion - A total dose of 10 mg/L is given as Monohydrate

ev1  = DosageRegimen([166.09,3.91], time=0, cmt=[1,2], duration=[60,60])
sub1 = Subject(id=1, events=ev1, time=20:0.1:180)
ev2  = DosageRegimen(10, time=0, cmt=2, duration=2)
sub2 = Subject(id=2, events=ev2, time=5:0.1:180)
pop2_sub = [sub1, sub2]

Population
  Subjects: 2
  Covariates:

Simulation

We will simulate the plasma concentration at the pre specified time points.

Random.seed!(123)
sim_sub1 = simobs(pk_34, pop2_sub, param)
df1      = DataFrame(sim_sub1)

Dataframe and Plot

Convert the simulation to a dataframe and use the dataframe to make your necessary plots

df1_id1    = filter(x -> x.id == "1", df1)
df1_id1_dv = filter(x -> x.time in [20,30,40,60,65,80,95,120,150,180], df1_id1)
df2_id2    = filter(x -> x.id == "2", df1)
df2_id2_dv = filter(x -> x.time in [5,10,15,20,30,45,60,120,150,180], df2_id2)


@df df1_id1 plot(:time, :cp,
              title="Reversible Metabolism" , label="Pred - Cisplatin (Inf-Cisplatin)",
              xlabel="Time (min)", ylabel="Concentration (ug/ml)", linewidth=3,
              xticks=[0,20,40,60,80,100,120,140,160,180], xlims=(-2,182), yticks=[0,1,2,3,4,5,6], ylims=(-0.1,6))
@df df1_id1 plot!(:time, :met, label="Pred - Monohydrate (Inf-Cisplatin)", linewidth=3)
@df df1_id1_dv scatter!(:time, :dv_cp, label="Obs - Conc (Inf-Cisplatin)")
@df df1_id1_dv scatter!(:time, :dv_met, label=false)

@df df2_id2  plot!(:time, :cp, label="Pred - Cisplatin (Inf-Monohydrate)", linewidth=3)
@df df2_id2  plot!(:time, :met, label="Pred - Monohydrate (Inf-Monohydrate)", linewidth=3)
@df df2_id2_dv scatter!(:time, :dv_cp, label="Obs - Conc (Inf-Monohydrate)")
@df df2_id2_dv scatter!(:time, :dv_met, label=false)

---