Regulatory authorities in the US and Europe have identified criteria for a good safety biomarker, resulting in changes to biomarker approaches and programs.
A technically good biomarker doesn’t necessarily mean a “qualified” biomarker. In the current issue of Nature Biotechnology, the editors investigate how regulatory authorities can be convinced of a biomarker’s utility. A first set of studies by the Predictive Safety Testing Consortium (PSTC) provides data supporting the utility of seven renal biomarkers in safety testing in the preclinical setting. The exciting news is that the results have also been accepted by regulatory authorities in the US and Europe. According to the article, a good safety biomarker meets the following criteria:
- The marker must be present in peripheral body tissue and/or fluid (e.g., blood, urine, saliva, breath or cerebrospinal fluid)
- It must be easy to detect or quantify in assays that are both affordable and robust
- Its appearance must be associated as specifically as possible with damage of a particular tissue, preferably in a quantifiable manner
But several key questions remain – what does it take to convince a regulatory authority of a biomarker’s utility and what are the data requirements?
The Critical Path Institute (PSTC), which is was formed in 2006 and includes about 190 industry and government scientists, has made significant headway with regards to a standardized regulatory review process for the qualification of biomarkers. With appropriate data support, a biomarker can be qualified by the respective regulatory authorities.
In the case of the PSTC’s nephrotoxicity biomarkers, the respective regulatory bodies (FDA and EMEA) regard their tests as ‘fit for purpose’ in preclinical research only because the data presented are from animal toxicity testing. Under the new ‘rolling’ qualification process, some of these urinary biomarkers could be qualified for nephrotoxicity once supportive human data are submitted.
In the past, much of the supporting data has been lacking: poor translation into drug development, inadequate description of the sampling, data generation, or statistical analyses. Now that there is a clearer direction for what a good biomarker is, and what types of supporting data are required, it will be interesting to see how the data generation processes adapt.
Read the article at http://www.nature.com/nbt/journal/v28/n5/full/nbt0510-431.html and let us know your thoughts. How will the outcome from efforts by industry/FDA collaborations such as the PSTC affect your biomarker programs? What approaches are you taking to distinguish between good and qualified biomarkers?