<img alt="" src="https://secure.lote1otto.com/219869.png" style="display:none;">
Skip to content

Pharmacometrics for Biomarker Driven Clinical Strategy

QPP (sometimes called QP2) remains at the heart of model based drug development. Short for Quantitative Pharmacology & Pharmacometrics, it refers to several types of quantitative modeling including meta-analysis, PK/PD , statistical modeling and the modeling of go-no-go decision rules.

Sometimes referred to as a ‘population approach’ to clinical pharmacology, QPP aims to build mathematical models of pharmacological phenomena that have taken into consideration all published literature about a certain therapeutic’s pharmacological components.

Model-based meta-analysis for biomarker confirmation has become a popular new method within late phase trials. The many uses of population enrichment in late phase trials have been articulated in other Cytel blogs (see [1][2] and [3]). An important consideration for some trials, however, is whether any positive difference will be observed within a certain amount of time.

For example, it may be that the prevalence of a certain biomarker will impact the effectiveness of a new therapy positively – but will it do so in a manner that is observable within six months? This length of time might affect the degree to which population enrichment makes sense as a part of clinical development strategy.

If all the literature to date has published the effects of a biomarker when given standard of care, (but has said nothing about the effects for the new drug in question,) it may not be clear whether a distinction caused by a biomarker will be helpful for your trial. You need to determine whether any observed differences will be observable within a given time. A pharmacometrician can take the existing literature and build a model that can help you decide the optimal development path.

According to a recent presentation by Cytel pharmacometrician Cecilia Fosser, a model-based meta-analysis in such situations can benefit from the following structure:

Step 1: Ask a quantitative question

Example: Is there a quantifiable relationship between the biomarker observed at Week 10 and the endpoint at 6 months?

Step 2: Update Literature Search on Biomarker and Endpoint

Cecilia suggests that the steps to the literature search are themselves multifaceted. They include identifying all relevant publications, updating inclusion criteria for published findings that should inform the model, adding new publications to this list, and then building the model itself. This process can have the structure of the graphic shown.

Step 3: Longitudinal Model Based Meta-analysis for the Biomarker

Identify key relationships between biomarker and standard of care using the widest possible range of relevant studies.

Step 4: Longitudinal Model Based Meta-analysis for the Endpoint

Also use the studies identified in Step 2 to generate model based meta-analysis for the endpoint in question.

Step 5: Determine Relationship Between Biomarker and Endpoint

There are several approaches one could take here. Ultimately the question is whether you can reject the hypothesis of correlation. Luckily most pharmacometricians come well prepared with a statistics background to take on this task. Supplied with informatoin from Step 3 and Step 4, they should be all set to help you create a sound strategy for Phase 3. 

Related Items of Interest

[1] 5 Reasons to Invest in Adaptive Designs for Population Enrichment

[2] Mehta Publishes Article on Adaptive Designs for Biomarker Driven Population Enrichment in Oncology

[3] 2 Methods for Evaluating Biomarker Subpopulations in (Adaptive Enrichment) Time to Event Trials

contact iconSubscribe back to top