Observer variability in the histopathologic diagnosis of microscopic colitis and subgroups

Research output: Contribution to journalJournal articlepeer-review

The diagnosis of microscopic colitis (MC) is based on histologic findings and includes collagenous colitis (CC) and lymphocytic colitis (LC). Incomplete MC (MCi) denotes patients with chronic diarrhea and a normal endoscopy and morphological changes that do not completely meet the histologic criteria of LC or CC. The aim of this study was to investigate the intraobserver and interobserver agreement on the MC subtypes of CC, LC, and MCi and the ability to discriminate MCi from normal and inflammatory bowel disease/nonspecific reactive changes. A single hematoxylin and eosin-stained specimen from biopsies of the following 5 groups were randomly selected and blinded: CC, LC, MCi, inflammatory bowel disease, and normal. Three pathologists independently reviewed the specimens. The specimens were relabeled and reinterpreted 4 months later. Intraobserver and interobserver agreement was evaluated by κ statistics. κ values for intraobserver agreement were good for 5 diagnostic groups varying from 0.70 to 0.83 and very good when simplifying to only 3 diagnostic groups varying from 0.88 to 0.96, separating MC/MCi from non-MC. κ values for interobserver agreement varied from 0.60 to 0.75 for 5 diagnostic groups and 0.81 to 0.89 for 3 diagnostic groups. The study shows that the intraobserver and interobserver agreement is high for discriminating between MC/MCi and non-MC, whereas the ability to discriminate MCi from CC and LC is lower. A revision and consensus on the histologic criteria of the MC subtypes seem warranted.

Original languageEnglish
JournalHuman Pathology
Volume44
Issue number11
Pages (from-to)2461-6
Number of pages6
ISSN0046-8177
DOIs
Publication statusPublished - Nov 2013

    Research areas

  • Adult, Aged, Biopsy, Colitis, Collagenous, Colitis, Lymphocytic, Colitis, Microscopic, Colonoscopy, Diagnosis, Differential, Female, Humans, Male, Middle Aged, Observer Variation, Reproducibility of Results

ID: 135497140