Patient population
5 adults (23-57 y.o.) and 12 children (11-17 y.o.) with active Crohn’s disease in the terminal ileal loop were involved in this prospective cross-sectional study. The inclusion criteria were: proven non-stricturizing terminal ileitis with MRI signs of active inflammation: 1) bowel thickness > 3 mm, 2) presence of mural oedema [4], 3) restricted diffusion in both conventional DWI and DWIBS tracking images of b=800 s/mm2, along with low SI in ADC map, 4) early mucosal hyperenhancement in the Gd-enhanced T1 post-contrast series [5]. The exclusion criteria were: 1) locations of active inflammation other than terminal ileum, 2) bowel thickness < 3 mm, 3) artefacts from dynamic blurring in either of series of conventional DWI, DWIBS or T1 post-contrast images, 4) inability to locate active bowel wall inflammation in either of images of conventional DWI, DWIBS sequences as well as in Gd-enhanced T1 post-contrast series within one and the same location.
MRI technique
The MR enterography (MRE) examination was performed with 1.5T MRI system (Ingenia, Philips Medical Systems, Best, the Netherlands) in patients lying in prone position. MRE protocol included SPIR-based conventional DWI sequence and STIR-based DWI sequence (DWIBS) using diffusion factors b equal to 0, 600 and 800 s/mm2, along with the corresponding ADC maps. All patients had been fasting 6 hours before the MRE examination, and no bowel cleansing was performed. The bowel distention was maintained with 1-1.5 l of 2.5% mannitol solution intaken 45 min prior to the examination followed by additional 20 minutes of lying in the right decubitus position and intaking another 0.25 l of mannitol solution. For reduction of bowel peristalsis, buscopan (10 mg in patients < 50 kg and 20 mg in patients ≥ 50 kg, diluted in 20 ml of saline solution) was intravenously administered prior to scanning of DWI and DWIBS sequences as well as T1 dynamic post-contrast series.
MRI image analysis
The inflamed bowel area were divided into 3 cm long segments (n=32 in adult patients, n= 46 in children). In all of the segments, wall thickness was measured, presence or absence of ulcers was estimated, 6 measurements of ADC values of both DWI and DWIBS sequences were performed, and 6 measurements of the wall signal intensity - WSI before (WSI-preGd) and after (WSI-postGD) administration of gadolinium contrast medium in the location with the highest contrast enhancement were performed. Six measurements of standard deviation (SD) representing the image noise were made outside the patient's body before (SD-preGd) and after (SD-post-gad) administration of contrast media. In all calculations, the mean of the six measurement values was used.
MaRIA score was calculated per formula: MaRIA = 1.5 x wall thickness (mm) + 0.02 x RCE + 5 x oedema + 10 x ulcers. RCE was calculated using the formula RCE=(WSI-postGd-WSI-pre-Gd)/(WSI-preGD) x 100 x (SD-preGd/SD-postGd) [4].
The corresponding Clermont score, or DWI-MaRIA, for both 'conventional' DWI and DWIBS sequences was calculated per formula DWI-MaRIA = 1.646 x bowel thikness (mm) - 1.321 x ADC + 5.613 x oedema + 8.306 x ulceration + 5.039 [6].
The measurements of ADC, WSI and SD were performed by one radiologist of 19 years experience in abdominal radiology using 4-9 mm2 oval region of interest (ROI). The image were reviewed and ADC values analyzed with post-processing server Philips Intellispace Portal v. 5.0 (Philips Medical System, Best, The netherlands). WSI and image noise measurements were perfomed with Clear Canvas DICOM Viewer, v. 13.2 (Synaptive Medical, Toronto, Canada, 2019).
Statistical analysis
The analysis was performed using software SPSS 20.0 (IBM Corporation, USA, 2011). The mean values with standard deviations for ADC of SPIR-based conventional DWI and of STIR-based DWIBS, as well as for MaRIA score, DWI-based Clermont score, and DWIBS-based Clermont were calculated. 95% CI were estimated for mean differences. Spearman’s correlation coefficient was used to estimate the correlations between quantitative parameters.