Sources of Bias in Specimens for Research About Molecular Markers for Cancer

University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA.
Journal of Clinical Oncology (Impact Factor: 18.43). 02/2010; 28(4):698-704. DOI: 10.1200/JCO.2009.25.6065
Source: PubMed

ABSTRACT Claims about the diagnostic or prognostic accuracy of markers often prove disappointing when "discrimination" found between cancers versus normals is due to bias, a systematic difference between compared groups. This article describes a framework to help simplify and organize current problems in marker research by focusing on the role of specimens as a source of bias in observational research and using that focus to address problems and improve reliability. The central idea is that the "fundamental comparison" in research about markers (ie, the comparison done to assess whether a marker discriminates) involves two distinct processes that are "connected" by specimens. If subject selection (first process) creates baseline inequality between groups being compared, then laboratory analysis of specimens (second process) may erroneously find positive results. Although both processes are important, subject selection more fundamentally influences the quality of marker research, because it can hardwire bias into all comparisons in a way that cannot be corrected by any refinement in laboratory analysis. An appreciation of the separateness of these two processes-and placing investigators with appropriate expertise in charge of each-may increase the reliability of research about cancer biomarkers.

Download full-text


Available from: Margaret Lee Gourlay, Jul 22, 2014
  • Source
    • "an antibody, which are useful as the capture and detection reagents in proteomics (Ransohoff and Gourlay, 2010). As an alternative, affinity reagents, e.g. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Cancer is the leading cause of the death, accounts for about 13% of all annual deaths worldwide. Many different fields of science are collaborating together studying cancer to improve our knowledge of this lethal disease, and find better solutions for diagnosis and treatment. Proteomics is one of the most recent and rapidly growing areas in molecular biology that helps understanding cancer from an omics data analysis point of view. The human proteome project was officially initiated in 2008. Proteomics enables the scientists to interrogate a variety of biospecimens for their protein contents and measure the concentrations of these proteins. Current necessary equipment and technologies for cancer proteomics are mass spectrometry, protein microarrays, nanotechnology and bioinformatics. In this paper, we provide a brief review on proteomics and its application in cancer research. After a brief introduction including its definition, we summarize the history of major previous work conducted by researchers, followed by an overview on the role of proteomics in cancer studies. We also provide a list of different utilities in cancer proteomics and investigate their advantages and shortcomings from theoretical and practical angles. Finally, we explore some of the main challenges and conclude the paper with future directions in this field.
    Asian Pacific journal of cancer prevention: APJCP 03/2014; 15(6):2433-8. DOI:10.7314/APJCP.2014.15.6.2433 · 2.51 Impact Factor
  • Source
    • "While it is usually difficult to prove in individual cases, investigator bias, including inadvertent and unrecognized bias (Ransohoff et al., 2010), is widely believed to contribute to the reproducibility problem. To overcome investigator bias, leaders in neuroscience recently argued for more rigorous standards in study design and reporting (Landis et al., 2012). "
    Neural Regeneration Research 01/2014; 9(1):6-7. DOI:10.4103/1673-5374.125322 · 0.23 Impact Factor
  • Source
    • "At the workshop, Dr Shai Silberberg (NINDS) reviewed recently published meta-analyses showing that unintended biases in animal studies tend to substantially inflate effect sizes, or lead to Type 1 errors (false positives) and overestimations of potential therapeutic efficacy. These unintended biases include the lack of allocation concealment [i.e. the investigators have no knowledge of the experimental group to which an animal belongs (Fisher et al., 2009)], blinded assessment of outcome and random allocation of subjects to experimental groups (see Bebarta et al., 2003; Macleod et al., 2005; Ransohoff and Gourlay, 2010; Sena et al., 2007; Landis et al., 2012). Furthermore, in recent efforts to reproduce original findings of potential drug targets in animal models of amyotrophic lateral sclerosis (ALS) (Scott et al., 2008), women's health and cardiovascular disease (Prinz et al., 2011), cancer (Begley and Ellis, 2012; Prinz et al., 2011) or spinal cord injury (Steward et al., 2012), the authors of follow-up studies were unable to replicate most data published by others, including those published in high-profile journals. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In September of 2011, the National Institute of Neurological Disorders and Stroke (NINDS), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), the International Rett Syndrome Foundation (IRSF) and the Rett Syndrome Research Trust (RSRT) convened a workshop involving a broad cross-section of basic scientists, clinicians and representatives from the National Institutes of Health (NIH), the US Food and Drug Administration (FDA), the pharmaceutical industry and private foundations to assess the state of the art in animal studies of Rett syndrome (RTT). The aim of the workshop was to identify crucial knowledge gaps and to suggest scientific priorities and best practices for the use of animal models in preclinical evaluation of potential new RTT therapeutics. This review summarizes outcomes from the workshop and extensive follow-up discussions among participants, and includes: (1) a comprehensive summary of the physiological and behavioral phenotypes of RTT mouse models to date, and areas in which further phenotypic analyses are required to enhance the utility of these models for translational studies; (2) discussion of the impact of genetic differences among mouse models, and methodological differences among laboratories, on the expression and analysis, respectively, of phenotypic traits; and (3) definitions of the standards that the community of RTT researchers can implement for rigorous preclinical study design and transparent reporting to ensure that decisions to initiate costly clinical trials are grounded in reliable preclinical data.
    Disease Models and Mechanisms 11/2012; 5(6):733-45. DOI:10.1242/dmm.011007 · 5.54 Impact Factor
Show more