Explicit formulas for a continuous stochastic maturation model. Application to anticancer drug pharmacokinetics/pharmacodynamics

Djalil Chafai,Didier Concordet

DOI: https://doi.org/10.1080/15326340802232244

2008-04-02

Abstract: We present a continuous time model of maturation and survival, obtained as the limit of a compartmental evolution model when the number of compartments tends to infinity. We establish in particular an explicit formula for the law of the system output under inhomogeneous killing and when the input follows a time-inhomogeneous Poisson process. This approach allows the discussion of identifiability issues which are of difficult access for finite compartmental models. The article ends up with an example of application for anticancer drug pharmacokinetics/pharmacodynamics.

Probability,Quantitative Methods

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about how to establish a continuous - time maturation and survival model, and derive explicit formulas in the pharmacokinetics / pharmacodynamics of anti - tumor drugs through this model. Specifically, the authors are concerned with the limiting behavior of the model when the number of compartments tends to infinity in a finite - compartment model. They hope to find a method that can provide explicit formulas for the average occupancy rate of specific compartments while allowing for the discussion of important identifiability issues. These issues are difficult to handle in a finite - compartment model.

### Main contributions of the paper

1. **Continuous - time model**: The authors proposed a continuous - time maturation and survival model, which is obtained through the limit process of the discrete - compartment model when the number of compartments tends to infinity.

2. **Explicit formulas**: Established explicit formulas for the system output law under non - homogeneous killing conditions, especially when the input follows a time - non - homogeneous Poisson process.

3. **Identifiability issues**: This method allows for a more in - depth discussion of identifiability issues, which are difficult to address in a finite - compartment model.

4. **Application example**: The paper finally provides an application example, demonstrating the application of this model in the pharmacokinetics / pharmacodynamics of anti - tumor drugs.

### Key techniques and methods

- **Non - homogeneous Markov process**: The model is based on a time - non - homogeneous Markov process, especially the non - homogeneous M/M/∞ queuing system.
- **Feynman - Kac formula**: Use the Feynman - Kac formula to derive explicit formulas for the output occupancy law.
- **Poisson process**: Consider the input process as a time - non - homogeneous Poisson process to simulate the input and output of drugs.

### Mathematical expressions

- **Explicit formula for the output law**:
\[
\beta(s,t)=\int_{s - \tau}^{t - \tau}\lambda(u)\exp\left(-\int_{0}^{\tau}g(u + w,\rho w)dw\right)\alpha(u + \tau,t)du
\]
where:
- \(\lambda(u)\) is the input rate.
- \(g(u + w,\rho w)\) is the killing rate in the maturation process.
- \(\alpha(u + \tau,t)\) is the survival probability, defined as:
\[
\alpha(s,t)=\exp\left(-\int_{s}^{t}\mu(w)dw\right)
\]
- \(\mu(t)\) is the survival rate after maturation.

- **Expected value**:
\[
E(N_t|N_s = m)=m\alpha(s,t)+\beta(s,t)
\]

### Conclusion

This paper successfully solves the problem of explicit formulas in the pharmacokinetics / pharmacodynamics of anti - tumor drugs under non - homogeneous conditions by establishing a continuous - time maturation and survival model. These formulas not only provide an accurate description of the average occupancy rate of specific compartments but also provide a new perspective for discussing identifiability issues. These results are of great significance in practical applications, especially in the fields of quantitative biology and drug design.