In this proof-of-concept study, we demonstrate the feasibility of in vivo MRE of MAT and present preliminary reference values for CD patients and healthy controls. AUC results show diagnostic accuracy in detecting CD to be excellent for SWS of MATCD_Opp (AUC = 0.82) but poor for SWS of MATCD (AUC = 0.52). A statistically significant decrease in mean stiffness of 7% was found for MATCD_Opp vs. MATCTRL. Possibly, this indicates a general softening of MAT in CD with a (statistically nonsignificant) trend toward 8% higher stiffness in areas of focal inflammatory bowel lesions (MATCD). This result could be related to an increased density of network elements by distinctively small adipocytes in CD and enhanced leukocyte infiltration, which contribute to the secondary barrier function of creeping fat. Furthermore, this could represent the biophysical signature that prevents bacterial translocation from bowel walls with CD-related higher permeability, thus reducing the inflammatory response from systemic to localized. Nevertheless, further studies are needed to corroborate this preliminary hypothesis.
As the composition of extracellular matrix components of MAT is constantly changing during inflammation26, a noninvasive imaging modality allowing characterization of these changes would be a valuable new diagnostic tool. Yet only a few imaging studies have investigated MAT in CD and even fewer have used quantitative MRI. For instance, van Schelt et al. investigated seven patients with active CD and seven healthy volunteers using tomoelastography from 30–60 Hz and histopathological assessment in a subset of patients37. Preliminary results show no significant difference for mean SWS (CD: 0.80 ± 0.21 vs. healthy: 0.67 ± 0.07 m/s, p = 0.18), whereas a significant increase was found for mean φ (CD: 0.58 ± 0.15 rad vs. healthy: 0.45 ± 0.08 rad, p = 0.02). While our results are of the same order of magnitude, we found no significant group differences for φ. Possibly, this could be related to the use of mannitol solution as an oral contrast agent in patients but not in healthy controls, and to the increased inflammation in active CD compared with our study with rather fibrotic CD. Desreumaux et al. observed a shift in the ratio of intra-abdominal fat to total body fat on MRI in CD patients compared to controls, indicating MAT hyperplasia in CD patients17. Although creeping fat cannot directly be visualized by noninvasive imaging techniques, Li et al. tackled this challenge by developing a CT-based “mesenteric creeping fat index” which indirectly assesses creeping fat by using vascular findings as surrogate38. The proposed index allowed the authors to predict the extent of creeping fat in surgical specimens in a prospective study population with an AUC of 0.7638. Moreover, in a study using CT-enterography, Sakurai et al. found that mesenteric findings such as hypervascularity (comb sign) and enlarged mesenteric lymph nodes, rather than mural findings, were highly correlated with the endoscopically determined severity of mucosal ulcerations in CD39. On another note, Marticorena Garcia et al. presented a case report showing that MRE-based stiffness and loss angle increased in acute appendicitis (SWS, 2.56 ± 0.12 m/s; φ, 1.37 ± 0.24 rad) and its surrounding MAT (SWS, 2.38 ± 0.16 m/s; φ, 1.32 ± 0.28 rad), followed by a decrease after antibiotic treatment (appendix: SWS, 1.54 ± 0.22 m/s; φ, 0.92 ± 0.22 rad; MAT: 0.74 ± 0.02 m/s; φ, 0.46 ± 0.05 rad) – the latter being comparable to the values for MAT we found in our CD group (SWS, 0.82 ± 0.07 m/s; φ, 0.46 ± 0.07 rad)31. Although appendicitis and CD are different entities, all CD patients in the present study were scheduled for elective surgery, preselecting patients with fibrotic changes rather than active inflammation following treatment with anti-inflammatory medication.
Many MRI studies15,40−43 and recently also a few MRE studies1,2,31 assessed the bowel wall and not MAT. For instance, a recent MRE study conducted in a patient population overlapping with the present study showed significantly increased stiffness and loss angle of the bowel wall in inflammatory bowel disease1. Compared to the present study, there was a strong correlation of φ between the bowel wall and adjacent MAT in CD patients (r = 0.84, p = 0.001), whereas no correlation was found for SWS (r = 0.16, p = 0.64). In another study, Avila et al. also reported increased stiffness in patients with CD as well as a link between advanced fibrosis on MRI and clinical events, e.g., surgical resection2. Ultrasound-based elastography techniques have been used as diagnostic tools for the assessment of the bowel wall for more than two decades, however, without quantitative analysis of MAT44.
Our study has limitations. First, we only investigated a small number of patients in this proof-of-concept study. Second, all CD patients and healthy controls have been reported previously1. However, there is no overlap of MRE data as the earlier study focused on the bowel wall in inflammatory bowel disease. Third, intestinal adhesions and pathologically altered movements in CD might have influenced the comparison with healthy controls. Furthermore, no spasmolytic agents were administered to enable comparison with healthy controls in this explorative setting. Finally, although all patients underwent surgery the day after the MRE scan and CD was histopathologically confirmed, there was no uniform or quantified account available for the extent of creeping fat or for the extent of intestinal inflammation and fibrosis. Nevertheless, we were able to measure MAT directly adjacent to pathological intestinal lesions. Still, other factors such as mesenteric inflammation and fibrosis might have had an influence. Future studies should take this into consideration to further improve the correlation of quantitative parameters. Use of intraoperative ultrasound elastography or tabletop MRE of surgical specimens would allow direct comparison of mechanical properties of creeping fat with histopathology.
In this proof-of-concept study, we demonstrate the feasibility of in vivo MRE of MAT and present preliminary reference values for CD patients and healthy controls. Our results motivate further studies for the biophysical characterization of MAT in inflammatory bowel disease.