Introduction

Measuring changes in mucosal damage and inflammation is important for assessing efficacy of experimental therapeutics in celiac disease clinical trials; however, there is no consensus on the optimal histologic endpoints. We previously demonstrated that quantitative endpoints (villous height to crypt depth ratio [Vh:Cd] and intraepithelial lymphocyte [IEL] counts), have better concordance among readers than qualitative measurement of pathology (Marsh-Oberhuber) when using a digital microscopy platform (SlideVantage, Jilab Inc.). The aim of this study was to further compare reader agreement when using composite and multiple component celiac pathology measures.

Aims & Methods

We compared reader agreement with three composite endpoints that combine Vh:Cd and IEL measurements into one score: Quantitative Mucosal Algorithmic Rules for Scoring Histology (Q-MARSH), Vh:Cd IEL (VCIEL), and multiparametric VCIEL (mpVCIEL), using duodenal biopsies from a phase 2 trial in patients with celiac disease (NCT03738475).

Results

We observed a kappa statistic for Q-MARSH of 0.35 and concordance correlation coefficient (CCC) of 0.86 for VCIEL and 0.60 for mpVCIEL. Compared to previous measurements with IEL (CCC = 0.85) or Vh:Cd (CCC = 0.80), VCIEL and mpVCIEL performed comparably and had better reader agreement than the categorical scoring systems, Q-MARSH and Marsh-Oberhuber (kappa = 0.42). We further examined reader agreement when measuring a change from baseline after gluten challenge or after therapeutic treatment and graphically contrasted how changes in each composite endpoint associated with traditional endpoints.

Conclusion

The composite measures assessed in this study capture changes in both architectural (Vh:Cd) and inflammatory (IEL) components of intestinal pathology in celiac disease. Quantitative endpoints, measuring single components or combined into a composite quantitative endpoint, provide the best interobserver agreement and appear most appropriate for assessment of celiac histology in clinical trials.

Disclosure

JM, DR, MZ, QL, VG, DL and GS are employees and shareholders of Takeda. JI is the founder and owner of Jilab, Inc.; is a Full Professor at Tampere University, Medical Faculty; and has a consulting agreement with Takeda. JL and KS are full-time employees of Jilab, Inc. TTW has nothing to disclose. JAM holds grants (to institution) from National Institutes of Health, Immunogenix, Takeda Pharmaceutical, Allakos, ProventionBio, Oberkotter Foundation, and 9Meters, Inc.; contract (to institution) from Kanyos Bio (a wholly owned subsidiary of Anokion); and consultancy fees from Johnson and Johnson, Bristol Myers Squibb, Intrexon Corporation, Dren Bio, Neoleukin, Reistone pharma, GSK, Immunic Therapeutics, Brightseed Bio, Chugai Pharma, Alimentiv, Equillium, Ukko, Medibeacon, Israel Biotech Fund, Precigen, triangle insights group, Sanofi, Zag Bio, Loop bio, and Orbimed and has received royalties from Evelo. MER is a consultant for Takeda, Perspectum and Alimentiv.