Refinement and reproducibility of histologic criteria for the assessment of microscopic lesions in patients with gastroesophageal reflux disease: the Esohisto Project.

BACKGROUND: Standardized criteria for assessing microscopic esophageal lesions are required to test their utility as markers of gastroesophageal reflux disease (GERD). AIMS: To finalize draft criteria for assessing microscopic esophageal lesions associated with gastroesophageal reflux and to test them for interobserver agreement.
METHODS: An international group of gastrointestinal pathologists was convened to finalize, using a consensus-based approach, draft criteria for recognizing microscopic esophageal lesions. Finalized criteria were retested for interobserver variability by four of the pathologists using 120 digitized esophageal biopsy slides from patients with GERD.
RESULTS: The finalized criteria included further clarification on lesion definitions and new guidance on how to select the area for assessing each lesion. This latter refinement was guided by the high interobserver agreement observed when draft criteria were previously applied to biopsies where the assessment area was preselected. When finalized criteria were applied in the current study to digitized biopsies without a preselected assessment area, the pairwise agreement was 73-97% for basal cell hyperplasia, papillary elongation, intraepithelial eosinophil, neutrophil and mononuclear cell numbers, and active/healed erosions, with slightly lower agreement (64%) for dilated intercellular spaces (DIS). When a combined severity score was applied, the level of agreement was 77%. The mean kappa ranged from fair to high (0.26-0.77) for individual lesions and was high for the combined score (0.64).
CONCLUSIONS: These levels of agreement are comparable with or higher than those for other accepted histologic definitions. Further steps include clinical validation of these criteria by correlating microscopic lesions with clinical variables such as esophageal acid exposure.