Guidelines for reporting histopathology studies

With the increasing amount of published papers and the need for evidence-based guidelines for diagnostics and treatment, it becomes of utmost importance to assess the quality of publications. The highest degrees of evidence in medicine are based on prospectively randomised clinical trials. In histopathology, trials are virtually non-existing and prospective studies are still relatively rare. The majority of our practice is based on retrospective studies, quite often from single centres.

Higher levels of evidence can be reached by systematic reviews of the existing literature and meta-analyses, that are increasingly present in the literature (figure 1).1 These meta-analyses are also important to provide information about the prognostic value of traditional factors, against which new diagnostic tools can be compared. However, the reporting of meta-analyses varies,2–4 limiting the possibility to assess strength and weaknesses of the reviews. Therefore, the PRISMA guidelines have been implemented for the reporting of systematic reviews and meta-analyses.5 However the PRISMA guidelines focus on the reporting of a meta-analysis, not on the reporting of the individual studies included in the meta-analyses. The establishment of quality of publications and judgement of the risk of bias is a key element in executing a systematic review, but is considered a subjective measurement. Quality assessment scales and reporting checklists for studies have been developed; among others Quadas,6 Newcastle-Ottawa scale7 and REMARK.8 For most histopathology studies these cannot be applied. The Quadas scale compares diagnostic interventions, not diagnostic criteria. The Newcastle-Ottawa scale has been developed for case-control and cohort studies. The REMARK guidelines are specifically designed for the reporting of tumour marker prognostic studies. The latter initiative comes close to what is needed for histopathology studies, however, some checklist items are only applicable on prognostic studies which assess biological molecules. Sample size calculations are rare in retrospective histopathology studies, moreover prognostic model-building, checking model assumptions, model validation and internal validation is impossible. Therefore we have adjusted the REMARK checklist to make it more suitable for retrospective histopathology studies (box 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Number of publications of systematic reviews and meta-analysis per year. Source: pubmed.1

Journal editors should consider the endorsement of guidelines and standardised checklists for the reporting of histopathology studies since these studies are inherently different from clinical trials and prognostic biomarker research.