Probe-based confocal laser microscopy identifies criteria predictive of active celiac sprue.

BACKGROUND: Celiac sprue is an underdiagnosed chronic intestinal inflammatory disease. Probe-based confocal laser microscopy (CLM) is a novel endoscopic technique for in vivo inspection of the intestinal mucosa that has not been evaluated in celiac sprue yet. AIMS: To develop CLM criteria most predictive of celiac pathology in a prospective pilot study. METHODS: Twenty-one patients (male n = 5, f = 16, mean age 52 years) with established or suspected celiac sprue, seven of whom had confirmed active disease (Marsh III) and 14 duodenal normal mucosa. CLM images from 91 duodenal sites were assessed. CLM recordings were obtained next to Argon beamer labeled areas. Biopsies were taken from the same spots for precise histological matching. After establishing histology-correlated criteria on one sample per patient, the remaining CLM recordings from the same patients were randomized and blindly evaluated. RESULTS: Villous atrophy and irregular appearing villi were most predictive of celiac pathology. Although the presence of crypts was diagnostic for celiac pathology, it was only recognized in 26.7% of celiac pathology sites. Using these criteria in the blinded assessment, the overall endoscopist's prediction of celiac sprue was accurate in 89.8% of all biopsy sites in 85.7% of all patients. Preliminary interobserver agreement testing villous atrophy, irregular villi, and crypts was poor (kappa 0.05 to 0.26). CONCLUSIONS: Probe-based CLM criteria developed in this pilot trial appear promising for the detection of active celiac sprue. Preliminary interobserver variability was high, indicating a learning curve effect. Our criteria need validation in an independent patient population.

Endocytoscopy for the detection of microstructural features in adult patients with celiac sprue: a prospective, blinded endocytoscopy-conventional histology correlation study.

BACKGROUND: Endocytoscopy (EC) is a novel technique that allows magnified live inspection of the intestinal mucosa. OBJECTIVE: To evaluate EC for the detection of key pathological findings in patients with celiac sprue. DESIGN: A total of 166 EC recordings were prospectively acquired. Matched videos, images, and biopsy specimens were obtained by duodenal argon beamer labeling of the respective sites. SETTING: Academic tertiary referral center. PATIENTS: Forty patients (mean age 51.5 years, 70% women) with established (n = 32) or suspected (n = 8) celiac disease (CD). INTERVENTIONS: A validated scoring system (Marsh classification) was used to assess disease activity. EC criteria were independently evaluated by 2 gastroenterologists and 1 pathologist. MAIN OUTCOME MEASUREMENTS: The primary endpoint was to examine EC correlation with conventional CD histology. RESULTS: Of 166 duodenal biopsy sites, 23% were classified as Marsh III (moderate to severe), 10% as Marsh I (mild), and 67% as Marsh 0 (normal). Using the 450x magnification, we found that identification of crypts was diagnostic for celiac pathology. Four criteria were significant predictors of Marsh III pathology when adjusted by multivariate analysis: low number of villi per visual field (<3; odds ratio [OR] 9.1; 95% CI, 1.3-62.0), confluence of villi (OR 37.1; 95% CI, 1.3-1021.2), irregular epithelial lining (OR 10.9; 95% CI, 2.5-46.7), and inability to delineate loop capillaries (OR 14.9; 95% CI, 3.3-67.0). None was a good predictor of Marsh I pathology. LIMITATIONS: Single-center experience. No prospective validation of the criteria in an independent patient population. CONCLUSIONS: EC at 450x magnification accurately identifies mucosal histopathology of advanced CD, but not early morphological changes.