Diagnostic merits of the Eosinophilic Esophagitis Diagnostic Panel from a single esophageal biopsy

Abstract

Background: Eosinophilic esophagitis (EoE) is a histologically "patchy" disease with uneven eosinophil distribution along the esophagus, posing a dilemma for histologically analyzing endoscopic biopsy samples, especially when biopsy samples are limited to only the distal esophagus.

Objective: We investigated whether molecular mRNA profiling of a distal esophageal biopsy sample predicts eosinophilia in the proximal esophagus.

Methods: Esophageal biopsy samples (n = 507) from subjects with EoE were collected from multiple institutions, spanning adults and pediatric patients. Subjects were grouped on the basis of distinct distal (D) and proximal (P) eosinophil counts (D⁺P⁺, D⁺P^-, D^-P⁺, and D^-P^-, with + and - defined as ≥15 or <15 eosinophils/hpf, respectively). Molecular profiles were assessed by using the EoE Diagnostic Panel (EDP), a set of 96 esophageal transcripts used to derive the EDP score.

Results: The distal EDP score was correlated with proximal eosinophil levels (r = -0.73; P < .0001). EDP analysis of a histologically negative distal biopsy sample predicted the presence of proximal esophagitis with high sensitivity (85%). In a 2-year follow-up focusing on the cases with discordant histologic and EDP results, histologically negative patients (D^-P^-) had higher rates of EoE relapse when the EDP was positive than when the EDP was negative (odds ratio = 11; P = .003), indicating predictive medicine capacity.

Conclusion: EDP analysis of a single distal esophageal biopsy sample predicts remote inflammation in patients with spatially heterogeneous eosinophilia and disease relapse in patients with histologic remission, providing diagnostic merit and predictive medicine capacity for molecular diagnosis of EoE.

Keywords: EoE transcriptome; Eosinophilic esophagitis; diagnostic panel; eosinophil; eosinophilic gastrointestinal diseases; molecular diagnostics; predictive medicine.

Figure 1.. Schematic overview of the study.

(A) Project schematic with pink ovals representing patchy inflammation. (B) Flow of the study. (C) Correlation graphs of EDP score with proximal or distal eosinophil counts (obtained from histology); Spearman r. Each dot represents an individual subject. CEGIR, Consortium of Eosinophilic Gastrointestinal Disease Researchers; Cincinnati Children’s, Cincinnati Children’s Hospital Medical Center; D+, distal esophageal biopsy histologically positive (≥ 15 eos/hpf) for active EoE; D−, distal esophageal biopsy histologically negative (< 15 eos/hpf) for active EoE; EoE, eosinophilic esophagitis; EDP, EoE Diagnostic Panel. eos or EOS, eosinophils; hpf, high-power field; P+, proximal esophageal biopsy histologically positive (≥ 15 eos/hpf) for active EoE; P−, proximal esophageal biopsy histologically negative (< 15 eos/hpf) for active EoE.

Figure 2.. EDP quantifies relative spatial inflammation along the esophagus.

(A) Heat diagram of hierarchical clustering (red, upregulated; blue, downregulated) of EDP genes; each column indicates distinct individuals in the D−P+ histology group. (B) Nested scatterplot of EDP scores as mean ± SEM; markers represent distinct individuals. Discrepant histology groups (EDP+D−P+ [red], EDP−D−P+ [blue]) and diagnostic cutoff (EDP score <333; dashed line) are shown. (C) Pie chart of discrepant cohorts. (D) Peak tissue eosinophil counts and EDP scores in D+P− patients (n=3) with available distal and proximal biopsy data at the same endoscopy site; eosinophil cutoff (>15 eos/hpf) and EDP diagnostic cutoff <333 marked with dashed lines. *P < .05, t test. NS, not significant. (E) Heat diagram of EDP scores and transcriptional dysregulation of individual histologic features per the Histology Scoring System (HSS) among D−P+ group; more intense coloration indicates greater degree of histologic feature. (F) Negative log10 P value of Spearman correlation between a HSS structural feature and EDP of CEGIR cohort (red, positive correlation; blue, negative correlation). Top 10 genes of HSS-EDP correlation superimposed; upper and lower dashed line indicate absolute Spearman r value of .5 and P value of .05, respectively. CEGIR, Consortium of Eosinophilic Gastrointestinal Disease Researchers; D+, distal esophageal biopsy histologically positive (≥15 eos/hpf) for active EoE; D−, distal esophageal biopsy histologically negative (<15 eos/hpf) for active EoE; EoE, eosinophilic esophagitis; EDP, EoE Diagnostic Panel; EDP+, EDP score <333, indicating active EoE; EDP−, EDP score ≥333, indicating inactive EoE; HPF, high-power microscopic field; P+, proximal esophageal biopsy histologically positive (≥15 eos/hpf) for active EoE; P−, proximal esophageal biopsy histologically negative (<15 eos/hpf) for active EoE.

Figure 3.. A positive EDP predicts future EoE onset in patients with negative histology.

(A) Longitudinal follow-up of 15 patients with a positive EDP (< 333) and negative histology (D−P−) and 15 random, eosinophilia-matched, age-matched patients with a negative EDP (≥ 333) and negative histology (D−P−). The chart shows the presence of EoE episodes within 2 years of the index biopsy. The size of the circle correlates with the severity of eosinophilia. A diamond depicts the index biopsy date used to calculate the EDP score. Red rectangles depict patients who experienced positive episodes within ± 2 years of the index biopsy date. (B) Summary table of positive episodes in both cohorts with statistical analysis results. (C) Diagram of various therapy regimens for patients with EDP+ and EDP− (n=15, each) at various biopsy dates, with shaded box indicating those under treatment at that time point. SGC (swallowed glucocorticoids) and PPI (proton pump inhibitors). CI, confidence interval; D−, distal esophageal biopsy histologically negative (< 15 eos/hpf) for active EoE; EDP, Eosinophilic Esophagitis Diagnostic Panel; OR, odds ratio; P−, proximal esophageal biopsy histologically negative (< 15 eos/hpf) for active EoE.