posologyr

Model-Informed Precision Dosing (MIPD) with R

https://github.com/levenc/posologyr

Keywords

Repository

Model-Informed Precision Dosing (MIPD) with R

README.Rmd

---
output: github_document
---



```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
set.seed(1)
```

# posologyr [](https://github.com/levenc/posologyr/)


[![R-CMD-check](https://github.com/levenc/posologyr/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/levenc/posologyr/actions/workflows/R-CMD-check.yaml)
[![CRAN status](https://www.r-pkg.org/badges/version/posologyr)](https://CRAN.R-project.org/package=posologyr)


## Overview

Personalize drug regimens using individual pharmacokinetic (PK) and pharmacokinetic-pharmacodynamic (PK-PD) profiles. By combining therapeutic drug monitoring (TDM) data with a population model, `posologyr` offers accurate posterior estimates and helps compute optimal individualized dosing regimens.

Key dosage optimization functions in `posologyr` include:

+ `poso_dose_conc()` estimates the optimal dose to achieve a target concentration at any given time
+ `poso_dose_auc()` estimates the dose needed to reach a target area under the concentration-time curve (AUC)
+ `poso_time_cmin()` estimates the time required to reach a target trough concentration (Cmin)
+ `poso_inter_cmin()` estimates the optimal dosing interval to consistently achieve a target Cmin 

Individual PK profiles can be estimated with or without TDM data:

+ `poso_estim_map()` computes Maximum A Posteriori Bayesian Estimates (MAP-BE) of individual PK parameters using TDM results
+ `poso_simu_pop()` samples from the the prior distributions of PK parameters

`posologyr` leverages the simulation capabilities of the
[rxode2](https://github.com/nlmixr2/rxode2) package.

## Installation

You can install the released version of `posologyr` from
[CRAN](https://CRAN.R-project.org) with:

``` {r eval = FALSE}
install.packages("posologyr")
``` 

You can install the development version of `posologyr` from 
[GitHub](https://github.com/) with:

```{r eval = FALSE}
# install.packages("remotes")
remotes::install_github("levenc/posologyr")
```

## Bayesian dosing example

To determine the optimal dose of gentamicin for a patient with `posologyr`, you will need:

1. A prior PK model, written in `rxode2` mini-language

In this example, a gentamicin PK from the literature 

```{r}
mod_gentamicin_Xuan2003 <- function() {
  ini({
    THETA_Cl  = 0.047
    THETA_V   = 0.28
    THETA_k12 = 0.092
    THETA_k21 = 0.071
    ETA_Cl  ~ 0.084
    ETA_V   ~ 0.003
    ETA_k12 ~ 0.398
    ETA_k21 ~ 0.342
    add_sd  <- 0.230
    prop_sd <- 0.237
  })
  model({
    TVl   = THETA_Cl*ClCr
    TVV   = THETA_V*WT
    TVk12 = THETA_k12
    TVk21 = THETA_k21
    
    Cl    = TVl*exp(ETA_Cl)
    V     = TVV*exp(ETA_V)
    k12   = TVk12*exp(ETA_k12)
    k21   = TVk21 *exp(ETA_k21)
    
    ke    = Cl/V
    Cp    = centr/V
    
    d/dt(centr)  = - ke*centr - k12*centr + k21*periph
    d/dt(periph) =            + k12*centr - k21*periph

    Cp ~ add(add_sd) + prop(prop_sd) + combined1()
  })
}
```

2. A table of the patient's TDM data, in a format similar to the data for NONMEM

```{r}
patient_data <- data.frame(ID=1,
                           TIME=c(0.0,1.0,11.0),
                           DV=c(NA,9,2),
                           AMT=c(180,0,0),
                           DUR=c(0.5,NA,NA),
                           EVID=c(1,0,0),
                           ClCr=38,
                           WT=63)
patient_data
```


### Individual PK profile

With these two elements, you can estimate and plot and the individual concentrations over time.

```{r}
library("posologyr")
```

```{r map_plot}
#| fig.alt: >
#|   Plot of the individual profile
patient_map <- poso_estim_map(patient_data,mod_gentamicin_Xuan2003)
plot(patient_map$model,Cc)
```

### Dose optimization

We will optimize the gentamicin dosage for this patient to meet two criteria:

+ A peak concentration of 12 mg/L, 30 minutes after a 30-minute infusion.
+ A trough concentration of less than 0.5 mg/L.

The time required to reach a residual concentration of 0.5 mg/L can be estimated as follows:

```{r}
poso_time_cmin(patient_data,mod_gentamicin_Xuan2003,tdm=TRUE,
               target_cmin = 0.5)
```

The dose required to achieve our target concentration can then be determined for an infusion at H48.

```{r}
poso_dose_conc(patient_data,mod_gentamicin_Xuan2003,tdm=TRUE,
               target_conc = 12,duration=0.5,time_dose = 48,time_c = 49)
```

In conclusion a dose of 240 mg 48 h after the first injection would be appropriate to meet our 2 criteria.

More examples can be found at: https://levenc.github.io/posologyr/

## Performance of the MAP-BE algorithm in posologyr
`posologyr` showed comparable performance to NONMEM MAP estimation with option `MAXEVAL=0`:

* Pharmaceutics **2022**, 14(2), 442; [doi:10.3390/pharmaceutics14020442](https://doi.org/10.3390/pharmaceutics14020442)
* Supporting data: [https://github.com/levenc/posologyr-pharmaceutics](https://github.com/levenc/posologyr-pharmaceutics)

Owner

GitHub Events

Issues and Pull Requests

Last synced: 6 months ago

Packages