Biomarker Validation - Clinical Validation

Biomarker Validation

Biomarker classification depends on its context of use, which in turn defines the extent to which the biomarker must be "validated." In terms of requirements for validation, the "holy grail" of biomarker application is one that can serve as a surrogate endpoint in a clinical trial. A surrogate endpoint is a laboratory (or physical) measurement that is used in a clinical trial as a substitute for a clinical endpoint (e.g., survival or functional endpoint). Therefore, a surrogate endpoint must accurately predict the effect of therapeutic intervention. Guidance for Industry: Pharmacogenomic Data Submissions

The use of "companion diagnostics" to drive molecularly targeted drug development also requires that the "diagnostic" be clinically validated through the FDA's PMA or 510(K) process. In both of these instances the biomarker must be analytically validated (inclusive of the test platform, reagents, etc.).

The subtle differences between biomarkers, diagnostics, surrogate endpoints and companion diagnostics in terms of regulatory requirements can be confusing, and nearly any misstep in the development process can lead to failure. Choosing the correct regulatory path for biomarker qualification and/or validation is critical from early biomarker discovery forward; otherwise, the data required to support the chosen context is likely to be insufficient for approval. 

If the chosen regulatory path for a biomarker is as a companion diagnostic or surrogate endpoint, biomarker validation is a requirement. Validation through clinical trials provides evidence that the methods used for a particular analyte support the intended use of the marker with sufficient statistical power to be approved by the FDA (or other international regulatory agencies) before entering the market. In most instances, clinical validation is necessary for reimbursement.

Although requirements will vary depending on context of use for the biomarker, validation of the test requires data that demonstrates selectivity/specificity, accuracy, precision, reproducibility, and analyte stability, to name a few parameters. The sensitivity and specificity of the assay must be demonstrated through robust ROC curves that provide support for the cut points established to identify normal vs. disease state. (ROC, use to set cut points, is essentially a plot that captures true positive rate against false positive rate of an assay.)

Biomarker validation must also be sufficiently robust to achieve a high level of performance beyond laboratory grade samples to include samples encountered in routine clinical use. In other words, to move beyond the biomarker stage to become an approved diagnostic assay requires that it function at a high level using less than ideal clinical samples. Physicians must use assays to make clinical decisions based on clear cut points derived from samples that reflect the target population.

The clinical trial design needed for validation of a biomarker will depend on the stated context of use. Historically, randomized controlled trials have been the mainstay of drug development across a wide range of diseases. Recently, the need for validation of biomarkers as surrogate endpoints or their role as companion diagnostics in drug development has led to a new generation of trial designs ranging from adaptive trials to more inclusive "basket trials" (used primarily in cancer to evaluate a common target across a number of cancers). Achieving successful validation of a biomarker for use as a companion diagnostic, diagnostic, or surrogate endpoint must begin in early discovery where all aspects of development are critical, including asking the right clinical question; acquiring the correct samples at each stage that anticipates context of use clinical validation; incorporates a robust statistical design inclusive of discovery and all analytical validation and clinical trials studies; ensures that technology standards are employed from discovery through clinical validation (if clinical validation is required); ensures that data is of highest quality with sufficient provenance to answer the clinical question and to support regulatory filings; and employs analytics that support all of the biomarker discovery to clinical validation transitions and enables clinical translation to the clinic. 

The Struggle in Validation

 

 

Qualification and Validation