Background and aims: The postoperative course of Crohn's disease (CD) is best predicted by ileocolonoscopy. Ultrasonography (US) has been proposed as indicator for postsurgical recurrence (PSR), but further confirmation is needed. We performed a systemic review with meta-analysis to assess the pooled diagnostic accuracy of US in the evaluation of PSR.
Methods: The systematic review was performed in PubMed/MEDLINE, EMBASE, SCOPUS, and Cochrane databases to identify studies assessing the US accuracy in PSR diagnosis. A sub-analysis between bowel sonography (BS), small-intestine contrast ultrasound (SICUS), and contrast-enhanced ultrasound (CEUS) was performed. Pooling was performed using diagnostic fixed or random-effect model according with heterogeneity.
Results: Ten studies (536 patients) met the inclusion criteria. There was no publication bias. Pooled sensitivity and specificity of US in detecting PSR were 0.94 (95% CI, 0.86-0.97) and 0.84 (95% CI, 0.62-0.94; diagnostic accuracy 90%), respectively. At sub-analysis, pooled sensitivity and specificity were 0.82 (95% CI, 0.76-0.88) and 0.88 (95% CI, 0.74-0.95) respectively for BS, with 0.99 (95% CI, 0.99-1.00) and 0.74 (95% CI, 0.73-0.74) for SICUS. Finally, an SROC curve was built to establish the best bowel wall thickness (BWT) cutoff able to predict the presence of severe PSR (Rutgeerts ≥3): a BWT ≥5.5 mm at US revealed sensitivity of 83.8% (95% CI, 73.6%-90.6%), specificity of 97.7% (95% CI, 93%-99%).
Conclusions: US shows high sensitivity and specificity for the diagnosis of PSR. SICUS appears more sensitive-but less specific-than BS, while the role of CEUS needs further investigation. A cutoff value of BWT ≥5.5 mm is strongly indicative of severe PSR. 10.1093/ibd/izy012_video1izy012.video15775249754001.