Pharma R&D Today
Ideas and Insight supporting all stages of Drug Discovery & Development
Qualification, Validation, and Fit-for-Purpose Biomarkers
Posted on July 22nd, 2016 by Jeffrey Paul, PhD in Pharma R&D
As drug developers we rely greatly on biomarkers for providing information to our drug development programs and supporting submission to regulators. Frustration follows when a drug company provides data on biomarkers to support its claim and FDA responds with a “sorry, the biomarker is not proven.” In 2011, FDA issued a guidance and roadmap for validation and qualification of biomarkers (ref 1). Unfortunately, there have been only a hand full of biomarker submissions and successful qualifications although biomarker validation remains a high interest in industry and FDA. (ref 2).
A few definitions are helpful when discussing biomarker development:
- Validation: Refers to the data demonstrating assay performance. That is, what are the assay methods, calibration standards, accuracy, precision, reproducibility, linearity, quality specifications, etc. These validation elements are well-known among bioanalytical, radiologic, and diagnostic laboratory professionals and are described in the FDA guidance on in Vitro Laboratory Developed Testing (ref 3)
- Qualification: This is a new concept where the submitter explains how results of the biomarker will be used in decision-making. In other words, what is the interpretation or action taken with a given outcome in the biomarker. A Context of Use (COU) statement is the outcome of qualification.
- Fit for Purpose: Means the biomarker has “good enough” validation/qualification information to be used internally or possibly for publication. This requires no regulatory submission.
Drug developers have been frustrated with FDA with regards to biomarkers, although FDA has a high interest in them as tools in drug development. As a Pharma participant I attended several meetings, hosted by PhRMA, with FDA and diagnostic companies, all of whom have a big stake in biomarkers. In one memorable session, an outspoken colleague described developing biomarkers with FDA was like playing football for the first time on a very foggy day. Run down the field, the ball drops from the sky, continue to run in the fog until you see an upright. Oh, BTW- don’t get tackled. Similarly, “we’ll tell you when you have sent enough data.” Understandably, FDA has a strong reluctance to rely on unproven biomarkers in their review because of the risk of false results. In my meetings with FDA (both public and private) they repeatedly cite the CAST study as great example of a false positive. CAST was a drug comparison study in fatal cardiac arrhythmia where one of the test drugs normalized the ECG although mortality increased. Thus, the biomarker (i.e. ECG) looked great, yet more patients died.
In many cases collaboration between industry and academia is crucial to provide the large amount of patient data to qualify a biomarker. A current example is the ADNI collaboration in Alzheimer’s Disease where this group relates brain imaging, amyloid and tau testing to disease progression (ref 4). The CPath Institute has also been active in biomarker qualifications (ref 5).
Now for Context of Use–
The COU elements, are clearly laid out in the FDA guidance and provide an informative, structured exercise for identifying the biomarker’s utility. A few COU elements are key: 1) What aspect of the biomarker is measured and in what form; 2) What is its purpose in drug development; 3) What is its interpretation, and decision/action based on a finding.
A simple example is the COU for renal tubular degeneration from nonclinical toxicology studies:
“KIM-1, Albumin, Clusterin and Trefoil Factor-3 can be included as biomarkers of drug- induced acute kidney tubular alterations in Good Laboratory Practice (GLP) rat studies used to support clinical trials.” N.B- these markers are not qualified for clinical renal safety.
Another example using imaging as a prognostic/stratification of Parkinson’s Disease patients: “…molecular neuroimaging of the dopamine transporter (DAT) can be used as an exploratory prognostic biomarker for enrichment in Parkinson’s disease (PD) trials.” A successful FDA review of a biomarker qualification package results in a Letter of Support to the submitter.
Finally, let’s discuss Fit for Purpose. This is the most common way to use a biomarker. A full validation and qualification using standards for submission may not be possible or of interest. However, development teams may still use the biomarker in ongoing programs. Fit for purpose identifies the most important elements of a validation/qualification program that support its use in internal decision making. Thus, a fit for purpose validation requires judgement and should pass the “red face test.” My experience with fit for purpose is that development teams and management need to be together, on the same page, because there is a known risk for a false negative or false positive. No drug product claims can be made from these data, however, in some circumstances biomarker information can be described in a product insert. Fit for purpose data may be acceptable to be published in a peer reviewed journal. An example is measurement of certain amyloid fragments in the CSF of humans with neurodegenerative disease. Although it will take a lot more than a village to obtain regulatory approval and qualification for the use of amyloid fragments in CSF, is there is enough data to support its use in a research setting? Fit for purpose is the most often used since it allows one to use the biomarker in early drug studies; and if the drug works, a full validation/qualification can be completed during the later phases.
Overall, development teams should be mindful of the FDA validation/qualification program or alternatively, a fit-for-purpose use of biomarkers as a drug development tool.
Industry Perspectives on Biomarker Qualification; Lavezzari, G.,Womack, AW.; Clinical Pharmacology & Therapeutics; 99(2), 2016
R&D Solutions for Pharma & Life SciencesWe're happy to discuss your needs and show you how Elsevier's Solution can help.
Jeffrey Paul, PhD
Principal at JPharm Consulting
- Emulating start-up success with purpose-built data tools
- Finding Answers Where You Least Expect Them
- Machine Learning Could Make Drug Discovery Faster, Cheaper, Better
- AI Can Work Wonders, But Can It Create Miracle Cures?
- Helping Startups in the Pursuit of Drug Discovery Success