How will the reporting recommendations for tumor marker prognostic studies affect clinical trials?

Abstract

The direction of modern medicine is changing, focusing more on treatments tailored to the specific characteristics of a patient’s disease. The sequencing of the human genome has enabled significant progress in understanding the underlying biology of a number of diseases; this has been seen particularly in the field of oncology. As pathways are elucidated and the effects of alterations in these pathways are defined, new drugs are being developed to take advantage of this new knowledge. The principles articulated here apply to biological marker studies in other diseases; however, our examples are drawn from the extensive oncology literature. The need to use new and existing therapeutics most effectively has highlighted the necessity for informative biological markers (also called tumor biomarkers in oncology) to help guide clinical decisions and for the development of assays to appropriately measure these biological markers. Over the last few decades, thousands of publications have reported associations between various tumor biomarkers and clinical parameters and outcomes, including prognostication, prediction of response to therapeutics or monitoring disease progression. However, few of these tumor biomarkers have found their way into regular clinical use. Many factors have contributed to the apparent gap between the number of reports of promising biological markers and the number of tumor biomarker tests that are in regular clinical use. Often the relationships discovered in model systems, such as cell culture or animal models, do not translate to the human disease setting. Other problems, including development of an assay that reproducibly measures the biomarker in the appropriate clinical setting, prove to be more challenging and expensive than anticipated. Research on biological markers, and the assays used to measure them, has often been poorly designed, lacking clear hypotheses and appropriate statistical designs [1]. One of the greatest obstacles to determining the clinical utility of a tumor biomarker stems from the lack of sufficient information in published reports to allow informed interpretation of the data or comparison with other reports about the same biological markers and assays. To address this concern, an international collaboration developed the reporting recommendations for tumor marker prognostic studies (REMARK), originally published simultaneously in five international journals and republished in others [2–8]. The REMARK guidelines include a checklist of items that should be fully reported in all publications, including a diagram that transparently illustrates selection and disposition of patients who ultimately comprised the study population, as well as details about the assay, statistical design and analyses of clinical correlations.