In patients with Crohn's disease (CD) submitted to ileal resection the evolution of post-operative recurrent CD lesions can be prospectively assessed since their onset. Patients with diffuse recurrent lesions in the neo-terminal ileum within 1 year of resection present symptoms earlier and are more prone to have complications than patients with no or very mild lesions. However, even minimal recurrent CD lesions such as aphtae have the tendency to progress into more severe involvement such as ulcerations and strictures . Based on these observations it has been proposed that patients with CD have endoscopic evaluation of the neo-terminal ileum 6 to 12 months after surgery to guide therapeutic management . Because of its invasiveness and need of intestinal preparation patients' compliance to undergo colonoscopy is poor . Furthermore the endoscopy cannot explore the neo-terminal ileum in patients with score 4 and stricture of ileo-colonic anastomosis or of neo-terminal ileum. Finally ileo-colonoscopy cannot exclude other possible localizations in more proximal parts of small bowel requiring additional ad hoc investigations, the most widely used of which is small bowel standard radiology. However, because of radiation exposure radiology cannot be used at will and should be minimized  particularly in young people, child bearing age women, and in those who may require repetitive assessments for the follow-up of CD as in patients submitted to surgery.
A non-invasive method which visualizes the entire small bowel such as MRI [16–18] or US performed after the ingestion of oral contrast [4–7] is likely to improve patient's compliance to undergo follow-up controls that can be planned early after surgery and timely adjusted at will. A limitation of MRI is the inability to prolong the observation so to observe properly distended and not contracting small bowel loops even when performed with enteroclysis [16, 17]. A limitation of oral contrast US could be the patient's refusal to ingest large amount of solution . Following a previous SICUS study in which variable volumes (125 ml-1,000 ml) of macrogol solution were tested, the amount of 375 ml was identified as the one properly distending the entire small bowel  and acceptable by virtually all patients [6, 7] as demonstrated in the present study.
Normal intestinal wall thickness of ≤3 mm has been identified at SICUS with repeated measurements at different levels of the small bowel in healthy subjects [5, 20]. Based on this cut-off normal value of ≤3 mm it has been shown that the diagnostic accuracy of SICUS is comparable to that of small bowel follow-through [6, 7] in detecting suspected small bowel pathology and of intraoperative findings in assessing site and extension of CD lesions .
Up to now the normal value of wall thickness of ICA in patients submitted to intestinal resection for CD complications is not known. In the present study the US value of 3.5 mm or more of wall thickness at the level of the ileo-colonic anastomosis predicted the presence of Crohn's disease recurrence in 100% of the patients. Increasing values of ICA wall thickness are significantly associated with increasing Rutgeerts severity score (figure 2A). Based on these findings the wall thickness of 3.5 mm appears to be the cut-off value to differentiate normal from abnormal ileo-colonic anastomosis in CD patients. To confirm this observation further studies should be performed in a larger CD population and comparatively in patients undergoing ileo-colonic anastomosis for non inflammatory disease.
Previous studies have assessed wall thickness after curative ileal resection in CD patients at the level of neo-terminal ileum with MR and US [17, 20]. Both MR and US did not provide sufficient resolution to differentiate initial wall thickening in patients with endoscopic scores 1 and 2. Recently two ultrasound studies [8, 9] reported that wall thickness >4 mm at the level of neo-terminal ileum had a high sensitivity in detecting severe endoscopic CD recurrence (i.e score 3 and score 4) as opposed to a low sensitivity in detecting mild lesions (score 1 and score 2). Differently from these studies limited to the US assessment of wall thickness of the neo-terminal ileum with no assessment of its extension, in the present one the wall thickness of the ileo-colonic anastomosis, expressed as the sum of the colonic and ileal limbs, and of the extension of intramural lesions of neo-terminal ileum were used in a model in order to differentiate patients with no recurrent CD lesions from those with recurrence and, more importantly, from those with low grade CD recurrence (i.e. score 1). The ROC curve analysis shows that the two combined variables represent an almost perfect tool in discriminating patients with score 0 from those with score 1-4 and a good tool in discriminating patients with score 0 from those with score 1. Therefore a relevant finding of the present study is that both variables, i.e., the increased wall thickness at level of the ileo-colonic anastomosis and the intramural extension of the lesion at the level of the neo-terminal ileum, should be taken into account to differentiate patients with endoscopic score 0 from score 1 and score 1 from score 2 and the latter from the more severe scores. At the first observation performed 6 months after surgery ICA wall thickening and normal wall thickness of the neo-terminal ileum was detected in 31% of the patients, indicating that a minimal involvement of the ICA precedes extension of the CD recurrent inflammation in the neo-terminal ileum. It is thus of note that assessment of ICA wall thickness at SICUS is relevant to detect postoperative recurrent lesion earlier than the usual US assessment of wall thickness limited to the neo-terminal ileum.
Differently from endoscopic scoring system, this study finds an association between the US grading of intramural lesion at the level of ICA and its extension along the neo-terminal ileum. The observation that in patients with Rutgeerts score 1 the lesions were confined to ICA in 17/31 (54%) at SICUS and in 29/31 (93%) at endoscopy suggests that a substantial intramural involvement may occur in the presence of few aphtae not accompanied by gross mucosal alterations. In addition, results of the present study indicate that while mucosal lesions were associated with intramural lesions in all patients with Rutgeerts score 1-4, in 4 patients intramural lesions, confirmed at MRI and histology, were not associated with mucosal lesions. This finding may indicate that CD inflammation does not necessarily imply the damage of the epithelial lining as also supported by the observation that in 3 patients with an early intramural recurrence at 6 months, mucosal lesions became apparent at ileo-colonoscopy 1 year later. Indeed it has been shown  that in the presence of inflammation, and occasionally even of granulomas at histology of the mucosal biopsy, endoscopic appearance can be completely normal. Thus recurrence can be detected as an intramural lesion without any accompanying mucosal alterations.
Similar to previous report , the degree of endoscopic severity of mucosal lesion in this study was not associated with the extension of wall thickening which was extremely variable, irrespective of the endoscopic score 1 (0-30 cm), 2 (0-40 cm), 3 (1-60 cm) and 4 (1-55 cm). Notably, in this study the presence of stenoses did not allow a proper assessment of CD recurrence at endoscopy in 78% of patients with score 4. In such circumstances, knowledge of the length of the stenosis, that can be assessed at SICUS and not at endoscopy, appears relevant for the possible indication to perform in these patients a perendoscopic dilatation. In the study by Rutgeerts et al patients with no (score 0) or very mild (score 1) and those with severe (score 3, and score 4) lesions at endoscopy were grouped together as they had, respectively, nearly asymptomatic or aggressive disease 1 year after surgery. Patients with intermediate severity of lesions (score 2) had no clear clinical prognosis as they progressed with either mild or aggressive disease.
By assessing both wall thickening at the level of the ileo-colonic anastomosis and its proximal extension in the neo-terminal ileum, SICUS could be proposed to grade the severity of intramural involvement of recurrent CD lesions in patients submitted to ileal resection. The proposed model predicts the absence of recurrence in more than 80%, and the presence of very mild CD recurrence (i.e. score 1) in more than 65%, of patients independently from any other variables.
Based on our results neither the CD disease activity as assessed with CDAI nor symptoms are associated with endoscopic score nor with severity and extension of intramural lesions, while the duration of CD before surgery appears to be marginally associated. This result may be explained by the low reliability of the CDAI system in the assessment of disease activity in patients with previous extensive ileo-colonic resection . Identification of the minimal post-operative recurrence may be relevant in view of the development of more effective treatment that can be used to prevent the progressive development from mild to severe lesions that affects more than 70% of patients .
Since the US methods are operator-dependent the present study should have also assessed the inter-and intra-observer variability of SICUS findings. To limit intra-observer variability three measurements of ICA wall thickness and of the increased wall thickness extension in the neo-terminal ileum were performed and their mean calculated. Inter-observer variability should be evaluated in future studies.