Challenges in Biomarker Development

While many challenges remain, the important role of biomarkers in diagnosis, monitoring, and predicting disease is unquestionable.65,73,86 Technology has advanced but the identification of novel biomarkers remains a challenge due to the samples being analyzed containing an abundance of analytes (known/unknown genes, protein/protein fragments, metabolic factors) expressed in a wide dynamic range.102 Moreover, differences between normal and patient populations at this level may not be readily apparent due to lack of comprehensive data. It can still be difficult to correlate a genotypic, proteomic, metabolomic, or biochemical profile to a distinct phenotype. Factors such as variable pene-trance, variations in disease phenotype, and/or environmental factors resulting in varied clinical manifestations make biomarkers less likely to be fully predictive of disease outcome. Therefore, extensive characterization and validation of the biomarker are crucial.

Rigorous evaluation of biomarkers in preclinical models is important whenever possible, but it has limitations. Physiological differences between humans and animal models can hinder translation to the clinic. Though costly, incorporation of biomarkers in clinical trials is necessary to achieve full validation of a particular biomarker. The consideration of scale and population studied (including the proper controls) must also be carefully considered. In the initial Cardiac Arrhythmic Suppression Trial (CAST), the surrogate marker used (electrocardiograms) did not demonstrate a similar benefit when a large, multiyear follow-up study with proper controls was conducted.79

Expectedly, technical and regulatory hurdles also pose significant challenges for the biomarker field.64'65'73 The measures necessary (reproducibility, statistical performance, and predictive value) prior to classification as a bona fide biomarker are not completely defined. Apart from assay validation (sensitivity and specificity), reagent availability, standardization, and information management (medical records privacy, intellectual property) across institutions have yet to be universally implemented.86 Rectifying this problem will require a concerted effort from academic, government, research, and pharmaceutical institutions. A consortium has recently been established by the NCI (and several programs by the National Institutes of Health) to accelerate biomarker discovery and clinical application.65

The need not only to diagnose and monitor disease, but also to predict clinical benefit/outcome has become increasingly important in drug development. Clinical outcome is most often assessed by clinical endpoints as well as surrogate endpoints that may only be apparent later in development (clinical trials).64,74 Attempts have been made to co-develop biomarkers early in drug discovery/development rather than investigate them retrospectively, to gain maximum measurable benefit to the patient as well as streamline drug development, minimizing costs by early identification of nonviable drug candidates. The predictive value of a validated biomarker may also reduce sample size by identification of patients most likely to have a positive response to therapy (Her2/neu and Herceptin) and abbreviate the duration of studies accelerating go/no-go decisions.

Despite existing challenges that may preclude attempts to incorporate biomarkers in clinical practice, remarkable strides to date have demonstrated proof of principle. Altogether, biomarkers have immense potential to impact not only the initial phases of drug discovery/later development but also in patient care, monitoring, and the early evaluation of successful therapies.

0 0

Post a comment