Increasingly biomarkers are viewed as indicators of abnormal biology that is upstream of a disease symptom, versus measuring the presence or absence of disease. In that regard, every successful clinically relevant and ultimately FDA approved biomarker begins with relevant clinical question. An effective biomarker that is “fit for purpose” must possess clinical utility for some diagnostic, prognostic, or preventive purpose. These biomarker applications, plus the recent use of biomarkers to stratify patients into clinical trials, must enable a physician to make an informed evidence-based decision relative to a patient's treatment plan. To develop the best clinical question possible, it is critical to consider the following: will use of the biomarker improve patient outcome; is it cost effective; and does the assay to fit into the clinical laboratory diagnostic workflow. Beyond these questions, biomarkers destined for clinical use must be FDA approved. (Note: Certain biomarkers are currently used in clinical settings without FDA approval. FDA has indicated its intent to apply some regulatory requirements to these tests, laboratory developed tests (LDTs) but these actions are pending.)
Study design must be robust in terms of sample sizes and repeat testing processes and must, to the extent possible, remove all forms of bias. Biomarkers must be confirmed in a true comparison between matched cases and controls and then repeated with a fully independent set of cases and controls. Sample sets must be sufficiently sized to statistically power the study, and the data should reflect a robust system capable of generating meaningful results in a variety of samples.
The foundation of a quality biomarker is a quality sample. Bio specimens should be collected using rigorous standard operating procedures (SOPs). All biomarker developers should be prepared to fully disclose the SOP used to obtain the samples, as well as release all of the metadata that was obtained for any single sample. Case and control samples must come from the same source, be thoroughly annotated, and evaluated by an appropriately qualified pathologist.
Currently, technology standards for biomarker discovery have been the subject of a number of local and national efforts. However, the affected communities have not reached consensus on the appropriate standards for several of the most commonly used platforms. The NBDA will pursue bringing the groups together to achieve such consensus. Labs use various technology platforms; each of which has unique performance characteristics and all of which require sets of standards for use in identifying biomarkers across different classes. The NBDA will publish NBDA Standards* for the most relevant technologies in the context of a uniform language for biomarker development, discovery, qualification, and validation. This promises to be a major advance in reducing or eliminating methodological error.
Biomarker data is increasing exponentially. However, real knowledge development from these myriad efforts is almost universally lacking, with little hope for change, unless communities come together to drive needed changes. As a starting context for biomarker development, data should be collected in a manner that guarantees both liquidity and transparency. To enable the development of the biomarker field, data must be able to move among technologies and investigators, while retaining a record of its origins through metadata. Data collection should be standardized for collection, processing, and storage (including the associated metadata that indicates the clinical application or decision process that the data will support).
Assuming the generation of the highest quality data with requisite transparency and liquidity, the next question is how to analyze it. This element of biomarker discovery is in rapid evolution. If and when it is possible to develop a national collection of biomarkers derived from quality bio specimens, using standards-based technologies and derivative high quality data sets, it will be possible to perform cross-validation and reproducibility studies using robust analytical and statistical methods. However, the development of responsive algorithms represents what is perhaps the major barrier in biomarker R&D. The NBDA’s goal is to create this new ecosystem for collecting, analyzing, and analyzing data from biomarker discovery that will underpin an end-to-end process and accelerate the adoption of more effective biomarkers into clinical practice.
*NBDA Standards includes but is not limited to: “official existing standards”, guidelines, principals, standard operating procedures (SOP), and best practices.