1. Findings and Accuracy of IUS in Crohn’s Disease (CD) Activity Assessment
1.1. Accuracy in Small Bowel CD Activity Assessment
Expert centres generally find intestinal ultrasound to be highly sensitive and specific for the assessment of disease activity in Crohn’s disease. In a large multicentre study, the sensitivity and specificity of IUS in detecting small bowel CD were 92% (95% confidence interval (CI), 84–96%) and 84% (95% CI, 65–94%) compared to a panel-derived reference standard
[22][1]. These results are similar to a previously performed systematic review in which IUS demonstrated a sensitivity and specificity of 80% (95% CI, 72–88%) and 97% (95% CI, 95–98%) for the diagnosis of suspected CD compared to a reference standard including ileocolonoscopy, radiology, and histopathology
[23][2]. Sensitivity improved to 89% (95% CI, 84–94%) for the initial assessment in established patients with CD and improved further for ileal-specific disease to 93% (95% CI, 87–99%)
[23][2].
In relation to determining the extent of small bowel CD, a recent study demonstrated a sensitivity and specificity of 70% (95% CI, 62–78%) and 81% (95% CI 64–91%) with IUS, respectively, which was 10% (95% CI, 1–18%;
p = 0·027) lower than the sensitivity of MRE for assessing small bowel extent in CD
[22][1]. Whilst IUS performed well in detecting small bowel CD, MRE may be a better modality for determining the length of small bowel involvement in patients with CD.
Disease activity is assessed through the measurement of sonographic parameters by the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB).
1.2. Bowel Wall Thickness
Bowel wall thickness (BWT) is the most commonly utilised parameter to assess disease activity with IUS
[24][3]. It is measured as the distance from the interface between the serosa and muscle proper to the interface between the mucosa and lumen layer
[16][4]. A BWT value > 3 mm has shown a sensitivity and specificity of 89% and 96%, respectively in detecting active luminal CD when referenced against ileocolonoscopy or histopathology
[25][5]. Higher BWT measurements ≥ 4 mm have demonstrated a similar accuracy to >3 mm with a sensitivity of 87% and a specificity of 98% so either of these cut-offs are considered to represent active disease
[25][5]. In addition, BWT has been shown to have good intraobserver concordance (
κ = 0.81 [0.69, 0.93])
[26][6]. Disruption of bowel wall stratification often occurs in conjunction with increased BWT. Disruption of bowel wall stratification is often associated with bowel inflammation and has been associated with an increased need for future surgical resection
[27,28][7][8].
1.3. Doppler Vascularity
Colour doppler signal (CDS) of the bowel wall is estimated and graded as part of routine intestinal ultrasound (IUS) assessment. In a normal bowel wall, doppler signal is absent, however in the presence of inflammation, CDS has been shown to increase. A good correlation between CDS and mucosal has been previously determined. Whilst multiple CD specific IUS indices incorporating CDS have been developed (see later section), the Limberg score is still the most widely performed CDS in IUS.
1.4. Inflammatory Mesenteric Fat and Lymphadenopathy
Mesenteric fat and associated regional lymphadenopathy (≥1 cm) can be seen around segments of inflamed bowel with IUS as increased hyperechogenicity. The presence of both inflammatory mesenteric fat and regional lymphadenopathy is often associated with other abnormal parameters in IUS and can be associated with a peri-intestinal inflammatory process and be representative of active CD
[24,29][3][9]. Novel IUS indices of activity, including the recent IBUS-SAS have recognised the importance of inflammatory mesenteric fat changes and incorporated their presence into activity-assessment scoring
[30][10].
1.5. Peristalsis
Peristalsis of small bowel loops is a normal finding with the absence of peristalsis being a key differentiating feature between the small and large bowel. The absence or a reduction of small bowel peristalsis is often an adjunct finding to small bowel thickening and inflammation
[39][11].
1.6. Strictures
In a systematic review of imaging assessment of small intestine strictures, IUS showed sensitivity and specificity ranging between 80–100% and 63–75%, respectively, in detecting strictures compared to surgical histopathological resection specimens
[40][12]. SICUS had an improved performance with sensitivity and specificity ranges 88–98% and 88–100%
[40][12]. MRE performed better than IUS at small bowel stricture detection with sensitivity and specificity ranges of 75–100% and 91–96%
[40][12]. In a recent head-to-head study, SICUS demonstrated a near perfect agreement (κ = 0.85) with MRE in detecting stricture number and location using a surgical resection specimen as the reference standard
[41][13]. This highlights the potential capacity of enhanced techniques in bridging the gap between IUS and conventional cross-sectional small-bowel imaging in detecting tructuring CD.
Differentiating between fibrotic and inflammatory strictures is difficult with IUS. A systematic review and meta-analysis demonstrated that IUS including CEUS is inadequate at differentiating between fibrotic and inflammatory strictures
[42][14].
1.7. Enteric Fistula and Abscess Detection
IUS can be used to identify both enteric fistula and abscesses. In a systematic review where IUS was performed prior to surgery, the pooled sensitivity and specificity of IUS in detection of enteric fistula was 74% (95% CI, 67–79%) and 95% (95% CI, 91–97%), respectively
[43][15]. Abscess detection with IUS had a sensitivity of 84% (95% CI, 79–88%) and a specificity of 93% (95% CI, 89–95%) compared to surgery. Both fistula and abscess detection sensitivity with IUS was similar to that of MRE
[43][15].
Contrast enhancement during IUS may further improve the identification of internal fistulae and abscesses in patients with CD. In a study of 67 patients with CD, the sensitivity of SICUS for detecting fistulae was 88% and 100% for identifying abscesses compared to surgical resection
[41][13]. The concordance between SICUS and MRE was substantial for fistula detection (
κ = 0.65)
[41][13]. CEUS is particularly useful in differentiating between an abscess and an inflammatory mass. When performing CEUS, inflammatory masses show intralesional enhancement, while abscess enhancement is limited to the wall
[21][16]. In a retrospective study of 71 patients with an intraabdominal mass, the differentiation between phlegmon and abscess with CEUS was comparable to cross-sectional imaging or surgical and percutaneous drainage (
κ = 0.972)
[18][17]. The overall sensitivity, specificity, and accuracy of CEUS for the diagnosis of abscess vs. phlegmon were 97%, 100%, and 98%
[18][17]. In clinical scenarios whereby cross-sectional imaging is contraindicated, CEUS represents an attractive alternate to the detection and differentiation of intraabdominal masses related to CD.
2. IUS Indices
Multiple IUS indices of CD activity assessment have been developed primarily for standardisation of assessments when used for clinical trials
[44,45][18][19]. The original IUS CD activity index developed by Limberg et al. used a composite of BWT and CDS. In a retrospective study of 32 patients, the Limberg score was correlated against a reference standard of ileal histology and was determined to have a significant (
p < 0.05) association with active disease (
κ = 0.66; sensitivity 95%; specificity 69%)
[36][20]. In a subsequent retrospective analysis of 108 patients a good correlation between the simplified endoscopic score (SES)-CD and the IUS-based Limberg score was observed (
Þ = 0.709,
p < 0.001)
[35][21]. Furthermore, in 22 patients a significant (
p = 0.005) difference between active and nonactive surgical histopathological resection specimens and the Limberg score was observed
[35][21].
Subsequent IUS indices have been developed primarily utilising BWT and CDS as the assessed parameters with few including other sonographic parameters such as bowel wall stratification, inflammatory fat, and the presence of lymph nodes
[31,34,38][22][23][24]. More recently, an expert consensus group from the International Bowel Ultrasound Group (IBUS) developed the IBUS-segmental activity score (IBUS-SAS)
[30][10]. This novel IUS activity index utilises a composite BWT, inflammatory fat, CDS, and bowel wall stratification with a multicentre international validation study in progress.
A key limitation of the current scores is the lack of external validation as well as significant concerns relating to the external validity of the parameters used. Expert consensus statements pertaining to the use IUS scoring in luminal CD have been published
[46][25]. Of the currently developed IUS indices assessed, none were found to be appropriate for use in luminal CD. BWT increased bowel wall vascularity, loss of bowel wall stratification, and mesenteric inflammatory fat were identified as ideal components of future IUS CD activity indices
[46][25]. In the future, IUS activity indices could represent a logical way in standardising reporting. Further studies are still needed to validate current IUS activity assessment scores in CD and determine its accuracy outside of trial settings.
3. Response to Therapy and Monitoring
Whilst endoscopic assessment for mucosal healing (MH) remains the gold standard in disease-activity assessment for Crohn’s disease, IUS has also demonstrated efficacy in assessing response to therapy
[47][26]. A key limitation in studies grading IUS response parameters is the heterogeneity of defining sonographic responses to treatment
[48][27]. More recently, an expert consensus statement defined a sonographic response in CD as a reduction of BWT >25% of baseline or >2 mm or >1.0 mm and one CDS reduction
[48][27].
The correlation between IUS disease-activity response and other measures of disease activity was evaluated in a cohort of 234 patients with active CD receiving therapies, including steroids, immunomodulators, and biological agents
[49][28]. At 3 and 12 months, all assessed IUS parameters including BWT, fibrofatty proliferation and CDS showed significant reductions following the initiation of therapy (
p < 0.01 for all parameters at months 3 and 12). The improvement in BWT correlated with the normalisation of the C-reactive protein after 3 months (
p < 0.001)
[49][28], and a reduction in the Harvey Bradshaw index (HBI) also correlated with the reduction in BWT
[49][28]. Sonographic response may also assist in the early detection of response to therapy and therefore allow changes in treatment earlier in the course of therapy. In a study of 52 patients with CD initiating antitumour necrosis factor (TNF) therapy, early sonographic response after 12 weeks of therapy was more likely to result in sonographic improvement at 52 weeks compared to individuals who did not achieve a week 12 response (85% vs. 28%;
p < 0.0001). The lack of sonographic improvement at 52 weeks was more likely to result in change or dose escalation in therapy or surgical requirement (65% vs. 11%)
[50][29].
Another potential advantage of IUS over ileocolonoscopy is the ability to assess for transmural healing (TH) in CD assessment. In a study of 133 patients with CD, TH assessed with IUS and MH assessed through endoscopy (SES-CD) showed a good correlation (
κ = 0.63;
p < 0.001)
[51][30]. More patients achieved MH compared with TH, though this was not statistically significant (38% vs. 25%)
[51][30]. The achievement of TH may be particularly relevant as a treatment target as it has been shown to independently predict steroid-free clinical remission (odds ratio (OR), 52.6;
p < 0.001), drug escalation (OR, 0.1;
p = 0.002), and hospitalisation (OR, 0.05;
p = 0.005) in patients with CD
[52][31].
CEUS and SICUS are useful modalities to assess CD response
[53,54][32][33] to therapy. In a study of 30 patients with active CD on therapy, CEUS had a good correlation with endoscopy in demonstrating disease remission (
κ = 0.73,
p < 0.001)
[54][33]. In a recent study of patients with CD receiving anti-TNF therapy, the CEUS parameters response was higher in those with clinical and endoscopic responses compared to nonresponders
[55][34]. Moreover, the response to therapy detected with SICUS has been associated with better long-term outcomes, including reduced need for surgery and corticosteroid use
[53][32]. With STRIDE-2 guidelines now recommending more frequent CD activity assessment, IUS presents a feasible alternative to alternative modalities such as MRE and ileocolonoscopy. Furthermore, there is emerging evidence that TH through IUS assessment may be a better treatment target than MH though long-term follow-up studies in this setting are needed
[56][35]. Both SICUS and CEUS have a significant capacity in monitoring response to therapy in CD though further studies are needed.
4. Novel Areas of Utility
4.1. Sonoelastography
Sonoelastography is a diagnostic ultrasound technique that measures tissue elasticity and stiffness and may have utility in discriminating fibrosis from inflammation. Strain and shear-wave elastography (SWE) have been studied in fibrosis detection for IUS
[65][36]. Strain elastography requires repeated probe pressure on the abdominal wall to determine a strain ratio. SWE mitigates the need for repeated probe pressure and measures tissue elasticity in kilopascals (kPa).
In a pilot study of 23 patients with CD, strain elastography, measured as a strain ratio had an excellent discriminatory ability for severe bowel fibrosis (AuROC: 0.917; 95% CI interval, 0.79–1.00) compared to a surgical resection as the reference standard with higher strain ratios seen in more severe grades of fibrosis
[66][37]. SWE performance was determined in a cohort of 35 patients undergoing surgical resection, with higher SWE values seen in severe grades of fibrosis
[67][38]. Using 22.55 KPa as a cut-off level, SWE had an excellent capacity in discriminating between mild or moderate and severe fibrosis (AuROC: 0.81
p = 0.002, sensitivity 70%, specificity 92%) compared to a surgical resection sample as the reference standard
[67][38].
IHerein
this study, SWE was unable to differentiate between grades of inflammatory strictures, potentially limiting its use to fibrotic disease only
[67][38].
In a recent meta-analysis inclusive of six studies, SWE and strain elastography were demonstrated to be potentially useful markers in CD-related fibrosis detection
[65][36]. The pooled standardised mean strain ratio was significantly higher in bowel segments with fibrotic strictures than in those without fibrotic strictures with a standardised mean difference of 0.85 (95% CI, 0–1.71;
p = 0.05)
[65][36]. The pooled standardised mean strain value was higher in bowel segments with fibrotic strictures than in those without fibrotic strictures, but did not reach statistical significance with a mean difference of 1.0 (95% CI, −0.11–2.10;
p = 0.08)
[65][36]. The meta-analysis had significant weaknesses due to the limited patient numbers and high heterogeneity between the performed studies. Sonoelastography is clearly an emerging imaging sonographic tool in determining bowel fibrosis, but further studies are needed to determine its true capacity in this setting.
4.2. Pregnancy
The sonographic evaluation of CD in pregnancy has received growing attention
[68,69,70][39][40][41]. Compared to alternative assessment modalities such as MRE and endoscopy, IUS has advantages such as the lack of intravenous contrast or procedural risks associated with endoscopy. IUS has demonstrated a moderate to strong correlation with clinical activity (r = 0.60,
p < 0.0001) in pregnant patients with IBD though feasibility appears to be reduced from the third trimester onwards (first vs. third trimester: 91% vs. 22%,
p < 0.0001)
[68][39]. Adequate ileal assessment has been seen in as low as 59% of pregnant patients with IBD beyond 20 weeks’ gestation compared to 91% prior to week 20
[69][40]. In the largest study of IUS in pregnant patients with IBD, Flanagan et al. demonstrated a weak positive correlation between bowel wall thickness and calprotectin (r = 0.26,
p = 0.03)
[69][40]. The overall accuracy of IUS in pregnant women with IBD showed a specificity of 83%, sensitivity of 74%, and a negative predictive value of 90% compared to faecal calprotectin
[69][40]. Significant challenges of IUS validation in pregnant CD patients will continue to remain due to the inability to correlate sonographic findings with the gold standard of ileocolonoscopy in the pregnant IBD cohort.