Introduction: Since it can be challenging to differentiate children with eosinophilic esophagitis (EoE) and gastroesophageal reflux disease (GERD) based on their clinical and histologic features, we investigated if esophageal basal cell nuclear morphology can differentiate them. Methods: Using QuPath v0.2.0, we analyzed H&E sections of esophageal biopsies obtained from children with EoE [n=5; defined per the 2011 Statement] and GERD [n=4; patients without EoE but with esophageal symptoms and < 10 eos/hpf in their esophageal biopsies]. Children with esophageal symptoms and no histologic abnormalities were considered controls (n=6). Basal cells were defined manually at high power (40x). The basal cell nuclear boundaries were automatically selected by the software. Y.C. reviewed all selections manually to ensure accuracy. The basal cell nuclear parameters of area, perimeter, max/min caliper ratio, circularity, and eccentricity were calculated. The Kruskall-Wallis test was used for groupwise comparisons of the median (IQR). Results: The cohort characteristics are presented in Table 1. In all, 4215, 1384, and 1261 basal cell nuclei from EoE, GERD, and controls, respectively were analyzed. The basal cell nuclear max/min caliper ratio was significantly higher in children with EoE [1.75 (1.46-2.17)] compared to GERD [1.68 (1.41-2.00), P< 0.001] and control [1.61 (1.38-1.92), P< 0.001]. GERD patients had greater max/min caliper ratio than controls (P< 0.01). Basal cell nuclei in EoE had greater eccentricity [0.82 (0.72-0.89)] as compared to GERD [0.80 (0.69-0.87), P< 0.001] and controls [0.78 (0.67-0.86), P< 0.001]. GERD patients had greater eccentricity than controls (P< 0.05). Basal cell nuclear circularity in EoE was decreased [0.79 (0.67-0.88)] as compared to GERD [0.83 (0.74-0.90), P< 0.001] and controls [0.84 (0.74-0.91), P< 0.001]. In controls, basal cell nuclear area was decreased [17.5 (13.75-22.25)] as compared to EoE [20.25 (15.00-26.75) mm2, P< 0.001] and GERD [20.50 (16.00-27.50) mm2, P< 0.001]. GERD nuclear area was increased as compared to EoE (P< 0.01). Last, in controls, basal cell nuclear perimeter [16.27 (14.25-18.82)] was significantly lower than EoE [18.10 (15.24-21.56) mm, P< 0.001] and GERD [17.98 (15.57-21.00) mm, P< 0.001] (Figure 1). Conclusion: Esophageal basal cell nuclear dimensions hold potential to distinguish children with EoE from GERD. Future studies with larger sample size are warranted to validate our findings.Figure 1.: A: Representative images of selecting esophageal basal cells and nuclei, B: Comparing nuclear dimensions between Control, EoE, and GERD.Table 1.: Characteristics of the cohort.
Read full abstract