Jonathan Chauvin (1), Geraldine Ayral (1), Pauline Traynard (1)
(1) Lixoft, Antony, France
Objectives: Therapeutic drug monitoring (TMD) and dose individualization can contribute to increased benefits for patients by augmenting the efficacy and/or decreasing the risk of toxicity. TMD is especially interesting for drugs exhibiting a highly variable exposure between patients and a small therapeutic window. Target therapeutic windows are usually defined for the steady-state through concentration of repeated dosing regimens and a single drug concentration measurement is made.
For a number of drugs, such as Kinase Inhibitors [1] or Animoglycosides [2], clear relationships between exposure and treatment outcome have been established, thus allowing for the definition of target exposure values.
While dose individualization has been more and more advocated over the years, the lack of dedicated, user-friendly and reliable decision-support software hampers its use on a large scale in hospital care. We present dose-recommendation tools for two TKIs (sunitinib and imatinib) and the results of their retrospective application on TMD hospital data.
Methods: The dose adaptation procedure is divided in two steps: first we determine the pharmacokinetic parameters of the patient (such as volume and clearance) and second we use these parameters to perform simulations of alternative doses. For the first step, we integrate the information from a population PK model (usually a literature model revalidated and adjusted on internal data), the patient covariates if relevant (age, weight, …) and the TMD drug concentration measurement(s) to calculate the conditional probability distribution of the individual parameters. For the second step, simulations of the trough concentration after new doses taking into account the operational constraints (such as available tablet doses) are performed using the estimated individual parameters. The dose most likely to reach the target is selected.
This procedure has been implemented within two applications: one for sunitinib and one for imatinib. The applications share the same calculation engine as the MonolixSuite, which is widely used for population PK/PD modeling. They have been developed in collaboration with clinical pharmacologists to fit to the practical needs in hospitals.
Results: The interface is meant to be usable by non-modelers such as clinicians. The interface allows to enter:
- the current treatment (dose, inter-dose interval and treatment start day)
- the date and time of the last dose
- the date and time of the measurement
- the patient covariates (if relevant and available)
- patient ID
- user ID
The application returns a dose recommendation as well as the drug concentration profile (and its uncertainty) with the current dose and with the proposed dose. A report is generated automatically based on a template and the entered information is saved to an audit trail local data base. The two applications are currently experimentally used in routine in 3 hospitals in France.
To evaluate in advance the proportion of the patients that would benefit from sunitinib dose individualization, we have applied our dose-recommendation application to the TMD (without adaptation) data base of the Cochin Hospital (Paris, France). The data base records around 900 PK measures for 233 cancer patients. For only 16% of the patients the application recommended to maintain the standard dose, while for 67% the recommended dose was below the standard and for the remaining 17% above.
Conclusions: The developed dose-recommendation applications permit to use all the available information in a rigorous mathematical framework to suggest the dose most likely to reach the therapeutic target. In addition, estimating the individual parameters gives more flexibility for the logistics: the measurement does not need to be at the trough and the steady-state does not need to be reached, as the steady-state concentration at through can be calculated from the individual parameters.
The retrospective study on past sunitinib data shows the need for TMD and dose adaptation. The routine use of the dose adaptation applications will allow a more precise assessment of the benefits. In addition, the dose adaptation on sunitinib is expected to reduce the overall cost of the treatment as the average recommended dose is smaller than the standard dose.
The presented applications are easy to use and currently tested in several hospitals. Applications for further types of drugs (everolimus, vancomycin, …) are in test phase or under development.
References:
[1] Verheijen, R. B., Yu, H., Schellens, J. H. M., Beijnen, J. H., Steeghs, N., & Huitema, A. D. R. (2017). Practical Recommendations for Therapeutic Drug Monitoring of Kinase Inhibitors in Oncology. Clinical Pharmacology & Therapeutics, 102(5).
[2] Matthews, I., Kirkpatrick, C., & Holford, N. (2004). Quantitative justification for target concentration intervention – Parameter variability and predictive performance using population pharmacokinetic models for aminoglycosides. British Journal of Clinical Pharmacology, 58(1), 8–19.
Reference: PAGE 28 (2019) Abstr 9059 [www.page-meeting.org/?abstract=9059]
Poster: Clinical Applications