Search in the Abstract Database

Search Abstracts 2016

* = Presenting author

P178 Magnetic resonance imaging characteristics of proctitis in Crohn’s disease in correlation with endoscopy

C. Tutein Nolthenius*1, S. Bipat1, B. Mearadji1, A. Spijkerboer1, A. Montauban van Swijndregt2, C. Ponsioen3, J. Stoker1

1Academic Medical Centre (AMC), Radiology, Amsterdam, Netherlands, 2OLVG, Radiology, Amsterdam, Netherlands, 3Academic Medical Centre (AMC), Gastroenterology, Amsterdam, Netherlands


Pelvic magnetic resonance imaging (MRI) is widely used in perianal Crohn’s disease (CD). Although diagnosing proctitis is important for management, MRI criteria are lacking. In our retrospective study, we studied which MRI characteristics are valuable in assessing proctitis.


Hospital patient database was searched for CD patients (≥ 18 years) who underwent proctoscopy, sigmoidoscopy, or colonoscopy and perianal fistula MRI (T2wTSE, fat sat T2wTSE, or fat sat T1w TSE after gadolinium) within 8 weeks of the endoscopy. Patients were divided into proctitis and no proctitis, based on endoscopy findings (reference standard). All predefined 15 MRI features were blindly scored by 3 observers and correlated to endoscopy using regression analyses. Reproducibility (multirater kappa, intraclass correlation coefficient) was determined for all 3 observer-pairs. MRI features were considered relevant if there was a significant correlation with endoscopy for at least 2 observers, and reproducibility was ≥0.40 for at least 2 of 3 observer pairs.


Of 58 CD patients, 26 (45%) had proctitis and 32 (55%) had no proctitis. Rectal wall thickness, perimural T2 signal, creeping fat, and size of mesorectal lymph nodes showed a significant correlation with endoscopy for all 3 observers (all p-values ≤ 0.05) and kappa/intraclass correlation coefficient were ≥ 0.40 for all 3 observer-pairs. Perimural enhancement and the presence of rectal mural fat showed a significant correlation with endoscopy (p = 0.000 and p = 0.011, 0.025 and 0.172, respectively) for at least 2 of 3 observers, reproducibility was ≥ 0.40 for 2 of 3 observer pairs.


Wall thickness, mural fat, and mesorectal tissue features (perimural T2 signal, perimural enhancement, creeping fat, and size of mesorectal lymph nodes) had the strongest correlation to endoscopy and were reproducible in diagnosing proctitis.