Identification and Confirmation of Biomarker-defined Populations in the Personalized Pharmacotherapy

The strong progress in genomic medicine increases the personalization of therapies, especially in oncology, and innovative study designs as adaptive enrichment designs, basket trials and ad hoc methods for the exploration of subgroups are established.

This project aims to identify more quickly those biomarker-based therapy approaches which are not valid or promising and may help to develop faster the very promising approaches. Different aims of the project cover on the one hand aspects on subgroup analyses and on the other hand provide an updated comprehensive classification of biomarkers with focus in oncology in order to estimate the validity of biomarker more precisely regarding clinical outcomes. Additionally, the currently accepted biomarker and the respective scientific advice approaches as well as the relevance of adequate in vitro test systems will be evaluated.

The project started in May 2016 and is supported by the Federal Institute for Drugs and Medical Devices (BfArM).

Aims

Owing to the advances in genomic medicine and a better understanding of individual causes of disease drug targets are identified, which exist mainly in subgroups of a disease. This fact together with a substantial progress in identification of causes of individual characteristics in pharmacokinetic and tolerability are leading to a stratification and personalization of treatments, in particular in oncology. Some oncological therapies are actual in biomarker-defined populations permitted. Innovative study designs such as adaptive enrichment designs and so-called basket trials have been established and a number of statistical methods for the exploration of subgroups have been proposed over the past years.

On this background this project aims to:

1. investigate empirically the evidence on subgroup effects,
2. systematically review and compare exploratory statistical methods for subgroup identification in a clinical scenario evaluation,
3. asses these methods based on criteria relevant to the regulatory process,
4. extend the methods proposed to the setting of more than one study by appropriately modelling between-study heterogeneity and assess the properties of these new methods in simulation studies including
5. an assessment of the regulatory consequences of the between-study heterogeneity, and finally
6. evaluate the possibilities of combining exploratory and confirmatory subgroup identification in clinical development programs, in particular by using adaptive enrichment designs and basket trials.
7. In the complementary molecular medically oriented approach the first aims are to provide an updated comprehensive classification of biomarkers with a focus in oncology. The new classification will be performed according to the drug targets, the molecular types of the biomarkers and several other factors and the new classification should allow a better assessment of the validity of the biomarker in prediction of the clinical outcomes (efficacy and adverse effects). This will also help the regulatory agencies to provide better evidence-based scientific advice in the regulatory process.
8. Another subproject approaches the question by systematic analysis of the German Fachinformationen, the European SmPCs and the FDA drug labels. As far as possible, also the scientific advice procedures from the recent 5 years will be systematically analyzed.
9. Finally, on selected examples we will analyzed the impact of adequate in vitro test systems as a first and essential step in the development and validation of biomarkers.

All this might help to sort out more quickly those biomarker-based therapy approaches which are not valid or promising and to develop faster the very promising approaches.

Publications

1. Jensen O, Ansari S, Gebauer L, Müller SF, Lowjaga KAAT, Geyer J, Tzvetkov MV, Brockmöller J (2020) A double-Flp-in method for stable overexpression of two genes. Scientific Reports. 10:14018.
2. Huber C, Benda N, Friede T (under review) Subgroup identification in individual patient data meta-analysis using model-based recursive partitioning. Advances in Data Analysis and Classification.
3. Huber C, Benda N, Friede T (2019) A comparison of subgroup identification methods in clinical drug development: Simulation study and regulatory considerations. Pharmaceutical Statistics. 1– 27. [Open Access].