P181 Diagnostic performance of Diffusion-Weighted Magnetic Resonance Imaging (DW-MRI) in the evaluation of stenosis and endoscopic recurrence in Crohn's disease
L. Samperi*1, 2, P. Foti3, L. Puzzo4, V. Ficara2, G. Riguccio2, A. Ricciardolo2, G. Inserra2
1Ospedale Morgagni Pierantoni, Dipartimento di medicina specialistica - U.O. Gastroenterologia ed Endoscopia Digestiva, Forlì, Italy, 2University of Catania, Dipartimento di Scienze Mediche e Pediatriche, UO Medicina Interna, Catania, Italy, 3University of Catania, Dipartimento Specialità medico-chirurgiche, Sezione di Scienze Radiologiche, Catania, Italy, 4University of Catania, Dipartimento di Medicina Legale e Tossicologia, Sezione di Anatomia Patologica, Catania, Italy
In active Crohn's disease (CD) the high viscosity and cellularity of inflamed tissue reduce the extracellular space, restricting water diffusion. DWI represents a new application of MRI to study the diffusion of water molecules. ADC (Apparent Diffusion Coefficient) is a quantitative parameter of this phenomenon. We aimed to verify the correlation between DWI findings both with pathological evaluation of fibrosis in surgical specimen (Fig.1) and severity of endoscopic recurrence.
21 pts undergoing ileocolonic resection for fibrostenosing CD were submitted to MRI-DWI; 6 pts within 2 months before surgery; 15 pts before follow-up ileocolonoscopy. Conventional MRI findings of terminal or neo-terminal ileum were recorded together with a semiquantitative evaluation of DWI signal intensity using a 3-point scale. To obtain ADC, images were magnified and a ROI was placed on neo-terminal ileum. For the first aim, Acute Inflammatory and Fibrostenosis Score (AIS) was correlated by Pearson's r to ADC values and DWI grading of matched ileo-cecal segment. For the second purpose, patients were divided in three and two classes of endoscopic recurrence based on Rutgeerts' score (Rs) (Rs i0-i1, Rs i2-i3, Rs i4 and Rs ≤ i2, Rs >i2). The correlation was studied by Spearman's rho or Persons'r. ROC curves analysis was used to find out an ADC cut-off value able to distinguish "low and high grade" severity of post-operative recurrence.
Comparison of DWI findings with surgical specimens pathologic evaluation showed a very good correlation, inverse and statistically significant, between ADC and AIS total (r=-0.91, p=0.013). There was a good correlation, not statistically significant, between the qualitative assessment of DWI and total AIS, tab1. The mean ADC value of the 3 pts with mild/moderate fibrosis was not statistically different from the 3 pts without fibrosis. The comparison between ADC and Rs (i0-i4) showed a good correlation, inverse and statistically significant (rho=-0.73, P=0.002), as well as between ADC and the other criteria of division, tab2.
ROC curves analysis highlighted that a value of ADC ≤ 1.82 x10-3 mm2/s could predict a severe recurrence (Rs >i2) with a sensitivity of 88.9% (51.8 to 99.7, CI 95%) and a specificity of 83.3% (35.9 to 99.6, CI 95%).
The Results of this study showed the ability of DWI sequences to provide quantitative measures, allowing a more objective assessment of the CD, bringing out a new "imaging biomarker" capable to monitor the progress/regress of the disease and the effectiveness of therapies. The limitation of our study is mainly the small number of patients that did not allow us to understand whether ADC evaluation could predict the presence of fibrosis in surgical specimens.