Enabling Retrospective Biomarker Studies: Resolving the Conflict between Short and Long Term Goals

In the field of clinical biomarker research, it is common to need to explore new hypotheses after the conclusion of a clinical study (i.e. retrospective studies).  However, if the proper samples are not available, even the best ideas are no more than fantasies.  While this statement might seem trivial, it is surprising to discover that many bio / pharmaceutical companies struggle to implement the proper strategic and tactical steps to enable retrospective biomarker studies. 

In my experience, the most common strategic issue facing bio / pharmaceutical companies in this area is in resolving the conflict between short term and long term corporate goals.  Specifically, in the context of the conduct of clinical studies, the need to meet recruitment quotas and deadlines often clashes with the proposal to acquire supplementary samples that, at the time, have only theoretical value (i.e. potential use in retrospective studies).  Indeed, there is a general consensus among the teams responsible for running clinical trials (i.e. clinical operation) that adding sample collection procedures can complicate approval of protocols by the Institutional Review Boards (responsible for clinical study protocol approval on behalf of the institution and their patients) and can impede patient recruitment.  Therefore, unless there is a strong concrete justification for collecting certain samples, additional sample collections tend to be excluded from clinical protocols.  The solution to this apparent conflict resides in a strong corporate policy in support of biomarker research in general and retrospective biomarker research in particular.  Without the assurance that the logistical constrains imposed by sample acquisition for biomarker research will be fully acknowledged as a factor affecting the conduct of clinical studies, clinical operation will favor the bottom line (i.e. completion of studies in the shortest possible time).

Beyond a biomarker-friendly corporate attitude, the scientists and clinicians responsible for biomarker research need to have a sound understanding of the logistical impact of additional sample collection on clinical studies.  Biomarker researchers need to be able to negotiate intelligently with their colleagues in clinical operation.  Reciprocally, clinical operation staff needs to be with the scientific questions explored by the biomarker researchers.  Therefore, cross-training of biomarker researchers and clinical operation staff is one of the key aspects of a successful clinical biomarker research program.

In some cases, clinical samples that were collected for one purpose (e.g. pharmacokinetics) can be repurposed for biomarker research.  However, if the proper informed consent was not put in place at the time of sample collection, using these samples for retrospective biomarker studies is not acceptable.  Indeed, current ethical and legal standards mandate that all individuals enrolled in a clinical study be fully informed about the use of the biological samples collected in course of the study.  The issue of drafting informed consent forms that adequately inform the patients about future biomarker research can be quite tricky.  While it is impossible to describe all potential future use of clinical samples for biomarker research, it is important to define the overall intent and the limit of this research.  Also, it is often desirable to draft the informed consent form with the option for the patient of opting in (or out) of future biomarker research. 

Finally, assuming that clinical samples exist and are properly consented, efficient retrospective biomarker research requires a solid sample management system.  Beyond the physical inventory of samples, such as system ideally needs to seamlessly integrate anonymized patient medical information, clinical study specific information, consent status (whether patient opted in or out of biomarker research), and prior data obtained from these samples.  Hence, an efficient patient sample management is as much about inventory management as it is about information management.

Thierry Sornasse for Integrated Biomarker Strategy 

Biomarker Studies: samples, hypothesis, and statistics

In a recent post on BiomarkerBlog, David Mosedale highlights a common problem with the design of biomarker discovery studies: reductionist clinical samples selection.  While it is tempting to initially explore for potential new biomarkers in highly contrasted clinical samples (e.g. healthy vs. diseased, benign early cancer vs. advanced metastatic cancer), this approach is almost guaranteed to yield over-optimistic results that do not translate easily to the real, complex world.  As a solution to this common problem, the author proposes that the design of the initial biomarker discovery study should reflect more accurately the intended application of the biomarker by including a spectrum of cases representative of the true complexity of the target patient population.  While I fully agree with David’s point, I would like to suggest an alternative view on this issue.

I would argue that the root-cause of the disconnection between biomarker discovery and their translation to medical use is the application of the right statistics to the wrong questions (i.e. statistical hypothesis).  Based on this premise, is there a fundamental issue with biomarker exploration using highly contrasted clinical samples?  I would argue that this approach can be useful as long as it is recognized for what it is: an initial screening step designed to test the minimalist hypothesis of whether a distinguishing factor (or factors) can be detected under artificially contrasted conditions.  Thus, the strength of the statistical association between the distinguishing factor and the selected sample phenotypes only reflects the pre-defined sample choice, not the true nature of the factor’s statistical association in the real-world population.  Hence, the use of this approach should be limited to the selection of potential biomarker candidates intended to be studied in a representative clinical sample.

Another case of inappropriate hypothesis definition is often encountered in the so-called validation of candidate biomarkers where a subset of the clinical sample used for discovery is used to determine the predictive value of the candidate biomarker using techniques such as Receiver Operator Curve analysis.  Here again, the strength of the statistical predictive value (Positive and Negative Predictive Values) derived from this approach is skewed by the initial sample selection, offering limited information about the predictive value of candidate biomarkers in the real-world.

So what is the solution to this somewhat frustrating trend in biomarker research?  I would argue that biomarker scientists should learn to ask the right questions to statisticians, and that statisticians should learn to challenge biomarker scientists about the actual hypothesis they wish to test.

Thierry Sornasse for Integrated Biomarker Strategy