Image Processing and Analysis of Mucosal Calretinin Staining to Define the Transition Zone in Hirschsprung Disease: A Pilot Study

Abstract

Quantification of calretinin-stained mucosal nerve fibers by image processing and analysis (IPA) may objectively define the transition zone (TZ) of Hirschsprung disease (HD). We tested the utility of IPA as an adjunctive tool in HD. Calretinin immunostain was performed on 15 HD pull-through specimens, and multiple images were captured from the proximal aganglionic zone, TZ, and probable normal zone (NZ). Pixel count (PC), defined as the percentage of brown-stained pixels in the mucosa, was quantified and plotted against distance from the rectal distal end. To validate the method, PCs from 45 images were compared with three-tiered visual scoring by five pathologists. Results were correlated against pertinent variables, which were retrieved from the clinical record. The PC gradually increased in the TZ toward the proximal resection margin in 10/13 (77%) cases. The PC variation in the probable NZ and around the circumference was substantial by the coefficient of variation. The mean PC of images with a visual score of 1 was less than scores of 2 and 3 by all five (100%) pathologists (p < 0.01). One patient had possible TZ pull-through that was clinically confirmed. While the mucosal calretinin staining gradually increases in the TZ, for now, the boundaries of the TZ cannot be accurately defined by mucosal biopsies given the substantial variation of staining around the circumference at the same distance and in the NZ. However, the IPA technique does provide a continuous variable and warrants further utility in HD studies.