P272 Faecal calprotectin and magnetic resonance imaging are highly accurate to detect endoscopic postoperative recurrence in Crohn's disease
Baillet P.1, Cadiot G.2, Goutte M.3,4, Goutorbe F.3,5, Allimant C.3, Reymond M.3, Obritin H.3, Bommelaer G.3,4, Pereira B.6, Hordonneau C.1, Buisson A.*3,4
1University Hospital Estaing, Radiology Department, Clermont-Ferrand, France 2University hospital of Reims, Gastroenterology, Reims, France 3University Hospital Estaing, Gastroenterology Department, Clermont-Ferrand, France 4UMR 1071 Inserm/Université d'Auvergne; USC-INRA 2018, Microbes, Intestine, Inflammation and Susceptibility of the host, Clermont-Ferrand, France 5Hospital of Bayonne, Gastroenterology Department, Bayonne, France 6University Hospital, Biostatistics Unit, DRCI, Clermont-Ferrand, France
The POCER trial  has recently confirmed that the therapeutic management has to be adapted according to endoscopic findings within the first year following surgery to prevent postoperative recurrence (POR) in Crohn's disease (CD). However, as colonoscopy is burdensome, alternative tools have been developed.
We aimed to compare the performances of MRI and faecal calprotectin to detect endoscopic POR (≥ i2 according to Rutgeerts' score) or severe endoscopic POR (≥ i3) in CD patients.
Adult CD patients from two tertiary centers who underwent ileal or ileocolonic resection were consecutively included in this prospective study. All the patients underwent magnetic resonance enterography including diffusion-weighted sequences with apparent diffusion coefficient (ADC) calculation and evaluation of Clermont score , MaRIA  and MR score , 6 months IQR [5.0–9.3] after surgery or restoration of intestinal continuity. Colonoscopy was performed within a median time of 14 days with stools collection the day before.
Overall, 28 CD patients were enrolled in the study (Table 1). Eight and 6 patients were respectively i0 or i1 according to Rutgeerts' score at 6 months after surgery. Endoscopic POR was observed in 14 patients including 7 patients classified as i2 (3 as i2a and 4 as i2b) and 7 patients with severe endoscopic POR (5 with i3 and 2 with i4). While the mean ADC value was lower in patients with endoscopic POR (2.08 vs 2.38, p=0.02), Clermont score (7.4 vs 10.2, p=0.04)and related contrast enhancement (RCE) (77% vs 129%, 0.05)were higher in patients experiencing endoscopic POR. MaRIA value was not significantly higher in patients with endoscopic POR (5.1 vs 8.1, p=0.18). Using ROC curves, we showed that ADC <2.3mm2/s (Se=0.82, Spe=0.71, NPV=0.83, PPV=0.69), Clermont score >8.4 (Se=1.0, Spe=0.55, NPV=0.74, PPV=1.0) and MaRIA >7 (Se=0.50, Spe=0.82, NPV=0.60, PPV=0.75) demonstrated substantial performances to detect endoscopic POR. Besides, ADC <2.3mm2/s (Se=0.67, Spe=0.86, NPV=0.5, PPV=0.92), Clermont score >12.5 (Se=0.57, Spe=0.78, NPV=0.82, PPV=0.50) and MaRIA >11 (Se=0.57, Spe=0.75, NPV=0.80, PPV=0.50) were highly effective to detect severe endoscopic POR.
Using a cut-off value of 100μg/g, faecal calprotectin was very accurate to detect endoscopic POR (Se=0.62, Spe=1.0, NPV=0.68, PPV=1.0) or severe endoscopic POR (Se=1.0, Spe=0.94, NPV=1.0, PPV=0.88).
Faecal calprotectin and MRI are reliable tools to detect endoscopic POR in CD patients and could be used as non-invasive alternative options to colonoscopy.
 De Cruz et al. (2015), Lancet
 Hordonneau et al. (2014), Am J Gastroenterol
 Rimola et al. (2009), Gut
 Sailer et al. (2008), Eur radiol