International Society for History, Philosophy, and Social Studies of Biology

twitter 2015
     facebook 2015


WEDNESDAY, JULY 8  /  09:00 - 10:30  /  DS-M440
Individual papers
Biomarkers, Individualized medicine and Etiological theories

Multi-modal discordance and the ontology of biomarker research

Spencer Hey (McGill University, Canada); Brianna Barsanti-Innes (McGill University, Canada)

Research in personalized cancer medicine is centered around the discovery and validation of "biomarkers"- testable properties of a patient specimen that can be assayed to prospectively inform a treatment decision. Successful examples of such biomarkers include HER2 overexpression for treating breast cancer and mutated KRAS for metastatic colorectal cancer. Unfortunately, these examples of successful biomarkers are the exceptions. As many commentators have emphasized, biomarker research is currently plagued by disappointment and difficulty. Some of the difficulty can be explained by social factors. For example, in contrast to drug development, there is no central financial actor that stands to benefit from a biomarker's validation, and therefore, there is often little financial support for rigorous biomarker diagnostic development. Other difficulties are methodological: Biomarker studies are often retrospective studies of convenience, notorious for their poor reporting and quality. However, the basic ontology of biomarker research also generates a novel set of philosophical challenges. In this presentation, we analyze a case-study in biomarker development in order to illuminate these challenges. The excision repair cross complement group 1 (ERCC1) gene is a biomarker hypothesized to predict patient responses to platinum-based therapies for non-small cell lung cancer. Although this marker has been extensively tested, its clinical utility remains uncertain. We argue that this continuing uncertainty is due to a fundamental confusion about what is the biomarker. In particular, there is discordance among the alternative modalities for assaying ERCC1 - that is, whether the marker is tested at the genetic, transcription, or protein level. We take this to show that "ERCC1" does not refer to a single ontological entity. Rather, it refers to a set of entities and processes in a complex causal system. Therefore, contrary to the assumptions of translational researchers, "What is the predictive value of 'ERCC1'?" (without further specification) is not a well-formed scientific question.

Evidence and individualized medicine

Megan Delehanty (University of Calgary, Canada)

Individualized medicine (IM) combines the molecular subtyping of complex diseases with an understanding of how an individual’s genomic, transcriptomic, proteomic, and metabolomic profile affect drug response, with the goal of optimizing treatment regimens for each individual. At the core of individualized medicine is the recognition of the extent to which heterogeneity must be accounted for. This is the case not only at the level of individuals displaying the same phenotype, but genotypically identical individuals who may be heterogeneous at other “-omic” levels, as well as cell heterogeneity that is often present within, for instance, a tumour. Each distinct population may show a different pharmacokinetic/pharmacodynamic response to a given drug. The methods used to accommodate this heterogeneity at multiple levels involve complex models and simulations both to identify potentially relevant variants that might be relevant to drug response, as well as to predict the nature of that drug response. In this paper, I will argue that evidential and explanatory bases of IM contribute a strong additional line of argument against the privileging of randomized controlled trials by evidence based medicine (EBM). In particular, IM demonstrates the severity of the problems EBM encounters due to the recognized difficulties with the exportability of results obtained via randomized controlled trials (e.g. Cartwright 2007, 2010; LaCaze 2008, 2009, 2011), as well as providing a case where evidence from models and simulations together with in vitro assays can be expected to produce higher quality evidence.

Re-evaluating concepts of biological function/dysfunction in clinical medicine: How are etiological theories useful?

Benjamin Chin-Yee (University of Toronto, Canada)

Theories of biological function can be categorized into etiological/historical theories, wherein the function is the causal explanation for the existence of the trait in question, or ahistorical theories, wherein the function is the current consequence the trait produces, irrespective of its cause. The existence of numerous understandings of biological function makes it unclear which concepts best apply to clinical medicine. I propose to re-evaluate the concepts of biological function/dysfunction in medicine to determine which might best inform medical discourse, particularly concerning the definition of disease. Objectivist theories of disease, notably Boorse’s Biostatistical Theory (BST) and Wakefield’s Harmful-Dysfunction (HD) account, base their definitions in biological dysfunction, assuming that biology provides a straightforward concept applicable to medicine. I begin by clarifying the concepts of function employed by BST and HD analysis, according to the ‘Taxonomy of Functions’ proposed by Walsh and Ariew (1996). Both theories use evolutionary notions of function, either ahistorical function (BST) or historical function (HD account). I highlight problems that these concepts pose in medicine, using the example of Sickle Cell Trait (SCT), which challenges the function/dysfunction distinction offered by these objectivist accounts. The pathological status of SCT is contentious and raises complex ethical issues. The case of SCT also shows how etiological theories may offer a useful heuristic for understanding illness, but alone are insufficient for defining disease. The concept of ‘Relational Function’ proposed by Walsh (1996), which recognizes function as a contextual property, offers a more pluralistic account of function for medicine, which captures useful aspects of both historical and ahistorical function. Relational Function better handles the case of SCT and captures the pertinent medical considerations at issue. Furthermore, this contextual concept of function may be better suited to contemporary medicine, which increasingly considers the roles of physical and social environments in shaping disease.