P699 Radiologic response reduces the risk of small-bowel surgery in Crohn’s disease

Abstract Background The aim of the present study was (1) to identify factors associated with radiologic response and (2) to assess if radiologic response was associated with better long-term outcomes in Crohn’s disease (CD) patients. Methods We performed a prospective study from 2011 to 2017 in the tertiary centre of Lille Hospital, including all patients with small-bowel (SB) CD who underwent two magnetic resonance imaging (MRI) within 3–6 months delay, with a follow-up for at least one year after the second MRI. Signs of radiologic inflammation were identified by an expert radiologist in CD. At second MRI, complete responders had all improved lesions, non-responders equal or worsening lesions, and partial responders other scenarios. Radiologic response included patients with complete and partial response. Factors associated with radiologic response were studied using multivariable Cox model. Rate of pejorative CD-related outcomes, i.e. hospitalisation, medical therapeutic optimisation, SB surgery, and endoscopic dilations were assessed using multivariable Cox models and were compared between radiologic responders and non-responders. Results A total of 115 SB CD patients were included: 55% had previous intestinal resection, 43% previous failure to immunosuppressant, 33% failure to at least one biologic. At the first MRI, 32% had stricturing disease and 25% penetrating disease. Median time between the two MRIs was 266 days (IQR 211–324), and median follow-up duration after the second MRI 17 months (IQR 11,6–28.3). At the second MRI, 18% had normal MRIs, 3% were complete radiologic responders, 26% were partial responders, 36% had stable abnormal MRIs and 17% presented worsening lesions. At the end, 54 (47%) patients were considered as responders and 61 (53%) as non-responders. In univariate analysis, use of anti-TNF treatment and a CRP ≥5 mg/l at the first MRI were associated with a decreased risk of radiologic response. In multivariate analysis, only a CRP ≥5 mg/l was associated with a decreased risk of radiologic response (OR 0.31; IC 1.10–0.96). At 1 year, 15 (13%) patients were hospitalised, 54 (47%) had medical therapy optimisation, 14 (12%) underwent surgery or endoscopic dilation. In univariate analysis, presence of a fistula or a stricture at the first MRI were associated with an increased risk of surgical or endoscopic procedure. On the opposite, radiologic response was associated with a decreased risk of surgery or endoscopic procedure in univaria