P243 Multi-modality intestinal ultrasound is accurate and reproducible to distinguish inflammatory from chronic fibrotic strictures in Crohn’s disease – results from the STRICTURE study

Abstract

Abstract Background In 30%-50% of patients with Crohn’s disease (CD) stenotic complications occur1. There is currently no imaging modality to identify stricture composition, which would allow early targeted (anti-inflammatory vs surgical) treatment. Intestinal ultrasound (IUS) and advanced modalities (contrast-enhanced ultrasound (CEUS) and shear-wave elastography (SWE)) have the potential for transmural disease evaluation. The STRICTURE study investigated whether (advanced) IUS techniques could distinguish between inflammatory- (IP) and non-inflammatory phenotypes (non-IP) of stricturing CD. Methods In this prospective, cross-sectional study consecutive patients with small bowel CD undergoing surgery were included. Patients were eligible if they had a non-passable stricture during endoscopy in the small bowel and/or a stricture at cross-sectional imaging (IUS or MRE)2. Prior to surgery, IUS, CEUS and SWE were performed. Two blinded sonographers scored the cine-loops for IUS and CEUS. After surgery, histological slides were retrieved and location matched with IUS. Two blinded pathologists scored for inflammation, adipocytes and fibrosis. Subsequently the predominant phenotype was determined as: [1] inflammatory (IP; Nancy score 4 with no marked fibrosis/adipocytes), [2] fibrotic (FP; structural changes due to marked fibrosis/adipocytes3) or [3] mixed phenotype (MP; inflammatory and fibrotic aspects but no predominant phenotype). FP and MP were both classified as non-IP. Results A total of 36 patients (age: 42±18 years) with a mean BWT of 6.7±1.7 mm were included. Median time between IUS and surgery was 14 [3-50] days. A total of 7 patients had an IP, 18 a FP and 11 a MP. For the conventional IUS parameters, loss of wall layer stratification (WLS) was more frequently found in IP strictures(OR: 7.87 [1.24-50.00], p=0.029). For CEUS, most parameters were significantly higher in IP versus non-IP (Table 1) and at multi-variable logistic regression Wash-in area under the curve remained the only accurate parameter to distinguish IP from non-IP (OR:1.55 [1.03-2.34], p=0.035) [Figure 1]. SWE inversely correlated with CEUS (Table 1, Figure 1) but did not differentiate between IP and non-IP (33.30 kPa vs 43.49 kPa, p=0.48). The agreement for BWT and loss of WLS was good (ICC: 0.77, p<0.001) and moderate (κ: 0.56, p<0.001), respectively. For CEUS, the most accurate parameters had good to excellent agreement [Table 1]. Conclusion Loss of WLS and CEUS are accurate to distinguish an IP from a non-IP stricture in CD and CEUS inversely correlates with SWE. In addition to accuracy, reproducibility was high and multi-modality IUS could be of additional value in this specific population to select patients most suitable for anti-inflammatory treatment. Table 1 Figure 1