QUANTITATIVE COMPUTER-ASSISTED IMAGE-ANALYSIS OF SUCTION BIOPSY IN PEDIATRIC GASTROESOPHAGEAL REFLUX

Gastroesophageal reflux is a disorder with well-characterized histopathological features in the adult, but the incidence and pathogenesis of epithelial injury in children are poorly understood. Esophageal suction biopsies from 80 infants exhibiting symptoms of reflux were studied by computer-assisted image analysis and the results compared to routine histological scoring. The histological features evaluated were the number of intraepithelial lymphocytes and eosinophils, papillary height, interpapillary basal cell height, and a novel measure ''integrated basal cell height.'' We quantitatively evaluated these histological criteria by computer-assisted image analysis and compared these results to four subjective grades of esophagitis: low, mild, moderate and high. In this report we now describe this quantitative histopathologic method for the evaluation of pediatric esophageal biopsies. Utilizing this method we demonstrate that both inter- and intra-observer variability were sufficiently low to stratify the mucosal changes reliably into at least four categories. The reliability of this objective analytic technique will permit studies into the disease progression and regression in pediatric reflux esophagitis.