P633 Concordance between endoscopy and histology scores at baseline and following induction therapy with the JAK1 inhibitor filgotinib in active Crohn’s disease: Results from FITZROY study

Abstract

Background: Mucosal healing (MH) in Crohn's disease is defined predominantly by the absence of ulceration on endoscopy. However, even in patients with MH, inflammation may persist on histologic examination (Neurath, 2014). A greater proportion of moderately to severely active CD patients treated with the JAK1 selective inhibitor filgotinib (200 mg QD for 10 weeks) showed endoscopic remission and histologic improvement compared with placebo (PBO)‐treated patients (FITZROY; Vermeire, 2016). In this post hoc analysis of FITZROY, we explored concordance between endoscopy (SES‐CD) and histology (Total Global Histologic Activity Score, TGHAS; D'Haens, 1998; Geboes, 2005) for a subset of patients in whom baseline (BL) and Week 10 (Week 10) results for all five endoscopically scored intestinal segments (rectum, left‐, transverse‐, right colon, ileum) were available. Methods: Biopsies corresponding to the most affected area in each segment were obtained. Endoscopies and histopathology were read centrally. Histologic healing per segment was defined as TGHAS <2. Agreement between endoscopy and histology was evaluated by kappa and %agreement at BL and Week 10, by segment and all segments combined. Results: Across 300 segments examined, overall there was moderate (kappa) to good (%) agreement between the absence of ulcers on endoscopy (SES‐CD ulcer size and ulcerated surface = 0) and histologic healing, both at BL (75‐80%; K 0.56, p < 0.001) and Week 10 (75‐82%; K 0.56, p < 0.001). There was only minor variation in the degree of agreement between segments, independent of time point (K 0.46‐0.62; Table 1). Focusing on segments with no macroscopic ulceration, two potential sources of discordance are noted. In segments with no histologic evidence of inflammation, 15/113 BL and 20 of 137 Week 10 cases reported evidence of non‐specific lesions (affected surface ≥1). However, in segments with TGHAS ≥ 2, 17 of 25 BL and 20 of 33 Week 10 were assessed as non‐affected (SES‐CD subscores = 0), indicating undetected inflammation. Cases where ulcers were detected macroscopically but not histologically may be due to poor biopsy site choice. (Table presented) Conclusions: There is overall moderate to good agreement between endoscopic and histologic severity across all five intestinal segments. The lack of concordance between non‐ulcerated segments on endoscopy and histologic healing is driven by residual histologic inflammation despite macroscopically normal appearing mucosa or the presence of non‐specific lesions such as erythema. Combination of histology and endoscopy scoring could provide a more robust assessment of intestinal disease activity.