REMARK checklist elaborated to improve tumor prognostic study reporting
Experts have elaborated on a previously published checklist of 20 items -- including descriptions of design, methods, and analysis -- that researchers should address when publishing studies of prognostic markers. These markers are indicators that enable doctors to predict the behavior of a patient’s disease and may aid them in making therapy decisions. The recommendations were originally developed and published simultaneously in five journals in August 2005, to address widespread deficiencies in reporting of such studies, which sometimes has led to inaccurate or irreproducible conclusions. The new article, which offers an extensive explanation of each item on the list of the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK), appeared concurrently in BMC Medicine and PLoS Medicine today and is written by four of the original authors, including Lisa McShane, Ph.D., Division of Cancer Treatment and Diagnosis, NCI. The authors hope that the enhanced checklist will provide a better understanding of how various factors, such as specimen selection, biomarker tests, and statistical study design and analysis, can lead to different study results and interpretations and why it is important to provide information to allow readers to fully evaluate and interpret the results. A “study profile,” developed by the authors, is recommended as a useful format for presenting the key information that fulfills the checklist reporting elements. Examples of good reporting drawn from the published literature accompany the explanations provided for each of the 20 checklist items. Although the original scope of the REMARK recommendations was primarily the reporting of studies that evaluated the prognostic value of a single marker, the new paper explains that most parts of the recommendations are relevant to studies investigating more than one marker, such as multivariable classification functions or indices.
This new article not only informs the reader about good reporting but provides a valuable reference related to many issues that should be considered when designing, conducting, and analyzing tumor marker studies. Many of the recommendations also apply to reporting of studies of predictive markers and to prognostic and predictive studies in other diseases. The authors hope that the overall quality of prognostic studies will ultimately improve if researchers give greater forethought to the REMARK checklist items at the time they originally plan their studies. Improved study quality and more transparent reporting will facilitate comparisons across studies and lead to more informed overviews upon which evidence-based conclusions about the utility of prognostic and predictive markers can be based.