Inflammatory Bowel Disease

Fig. 12.1

Fourteen-year-old girl treated for UC. Sagittal US section visualizes a mild thickening of the left colon wall with preserved bowel signature

Fig. 12.2

Twelve-year-old boy known for CD. Sagittal US section reveals a severe thickening of the bowel wall and loss of the layered pattern

An abnormal bowel loop often appears fixed with increased Doppler flow within the wall, and is frequently separated from other loops by an increased volume of mesenteric fat (fat wrapping).

In a recent meta-analysis, including both adult and pediatric patients, Calabrese et al. [16] reported that US has a 79.7% sensitivity and a 96.7% specificity for the diagnosis of suspected CD, and 89% sensitivity and 94.3% specificity for initial assessment in established patients with CD. Ileal CD was identified with 92.7% sensitivity, 88.2% specificity, and colonic CD with 81.8% sensitivity and 95.3% specificity, with lower accuracy for detecting proximal lesions. The absence of bowel wall thickening had a high negative predictive value for rapidly excluding CD [17, 18]. Concordance between US and MRE for bowel thickening and vascularization has been reported to be excellent and fairly good, respectively [19], but reliability of US is lower in the mid and proximal portions of the small bowel [20]. Furthermore, in recent study, the utility US for assessing response to medical therapy in CD or detecting and following disease-related complications has been questioned [21].

Other common findings on US in patients with CD include isolated hyperemia of the appendix (73%) and appendiceal involvement in 21% [22]. Enlarged and numerous mesenteric lymph nodes are frequent associated findings, although they are nonspecific. As noted above, some data suggests administration of oral contrast agent (hydrosonography) may improve sensitivity and specificity in identifying CD, and in assessing location and disease extent [16].

12.3.1.2 Complications

Complications are less frequent in children than in the adult population.

Stenosis (Fig. 12.3) presents as a rigid and thickened bowel associated with a dilated upstream normal loop, often with preserved peristalsis. The diagnosis of stenosis is easier in the small bowel than in the colon. Associated hyperemia within the area of the stenosis is helpful in differentiating potentially reversible inflammation from a fixed fibrotic lesion which will usually require surgery. Low vascularization with retraction of the adjacent loops towards the stenosis supports a fibrotic phenotype [23]. However, fibrosis and inflammation frequently coexist within the same segment of stenosed bowel and distinction between these pathological entities using US remains controversial [10–12]. In general, increased bowel wall thickness, loss of mural stratification, and increased fibrofatty proliferation may predict patients who might require surgical treatment [24]. Across adult and pediatric populations, US has sensitivity of 79.7% and specificity of 94.7% in diagnosing stenosis [16].
A fistula (Fig. 12.4) expresses as a hypoechoic tract, for instance, between a bowel loop and another bowel loop, between a bowel loop and the bladder or skin. Detection of mobile hyperechoic gas bubbles within this tract facilitates the diagnosis, and video clips are a useful for their detection. US performs as well as MR imaging for the diagnosis of enteric fistula, with a sensitivity approaching 80%, in adults [25]. A sinus is a blind tract and manifests in a similar way, although does not communicate with another organ.
An abscess is typically a thick walled ill-defined collection, often with increased Doppler signal in the wall. The echogenicity of the central core is variable and depends on the content (gas, debris, blood, pus) (Fig. 12.5). In a recent meta-analysis [16], US has comparable sensitivity (79.7%) and specificity (94.7%) with CT or MR imaging for the detection of CD abscesses in adults and children.

Fig. 12.3

Sixteen-year-old boy known for CD presenting an ileal stenosis. (a) Sagittal US section visualizes thickening of the wall and loss of bowel signature. The bowel lumen is not visible anymore and there is a distension of the upstream loop. The aspect of both segments did not modify during the entire examination. (b) Transverse Doppler US at the stenosis site demonstrates associated hyperemia of the thickened bowel wall

Fig. 12.4

Eleven-year-old boy explored by US; follow-up of a known ileal CD. (a) Markedly heterogeneous ill-defined hypoechoic area between the transverse colon (yellow star) and the diseased ileal loop (yellow triangle). (b) A flow of hyperechoic air bubbles (yellow arrows) was seen during the examination coming from the ileum towards the transverse colon establishing the presence of a fistulous tract between these two loops

Fig. 12.5

Eleven-year-old boy known for active ileal CD. US demonstrates an ill-defined hypoechoic area surrounded by a hyperechoic fat infiltration adjacent to a thickened ileal loop. The presence of air bubbles (yellow arrow) within this hypoechoic area ascertains the diagnosis of abscess

Disease activity: US has been advocated as an efficient tool to monitor disease progression and response to medical therapy [16], although others have questioned its utility [21], for example due to its limited reproducibility and restricted evaluation of some bowel segments such as the jejunum and rectum.

Increased vessel density and hyperemia (Fig. 12.6) on Color Doppler [26] is a validated marker of disease activity. However, the correlation is weak against the Crohn Disease Activity Index (CDAI) [16] and is not in itself able to precisely map the severity of disease [13, 27].
Contrast-US may provide additional diagnostic information [28]. Rosembaum et al. [24] have reported that the US findings most often demonstrate in children operated for CD include: bowel wall thickness above 4.3 mm (mean, 6.1 mm), more frequent loss of mural stratification, and fibrofatty proliferation.

Fig. 12.6

Twelve-year-old boy followed for CD. US Doppler demonstrates thickening and hyperemia of a jejunal loop associated with hyperechoic fat thickening (sclerolipomatosis)

There is a high concordance between US and endoscopy (90%) for the extension evaluation of UC. Multiple regression analysis shows that US measurement of increased bowel wall thickness, increased vascularity, loss of haustra, and loss of stratification of the bowel wall were all independent predictive value of severity at endoscopy. A US score based on these parameters strongly correlates with clinical and endoscopic activity of disease [29].

12.3.2 Magnetic Resonance Enterography (MRE)

Thanks to its high contrast resolution, multiplanar capability, absence of radiation exposure and ability to evaluate both the bowel and extra-bowel diseases with high accuracy, MRI is increasingly used to evaluate IBD. Furthermore, the images are in general more easily understood than US by referring clinicians.

12.3.2.1 Technique

A consensus statement on an optimized MRE protocol has been recently published by the ESGAR and the ESPR Societies [9], and is summarized below.

Preparation:

Depending on their age, children should avoid solid oral intake from 2 to 6 h prior to the examination. Fluid restriction is not recommended, although carbonated drinks are best avoided.

Many hyperosmolar contrast agents are acceptable in order to achieve bowel distension including polyethylene glycol, methylcellulose, sorbitol, mannitol, lactose, psyllium fiber, 0.1% low-density barium sulfate suspension, with none proven superior to the others. The ingestion should commence 45–60 min prior to MRE. The recommended volume in children is 20 mL/kg, up to a maximum up to 25 mL/kg. A full explanation of the drinking requirements to both the child and parents is important prior to MRE, and the use of a flavoring to the oral contrast agent will facilitate the child’s compliance.

The use of spasmolytic agents is considered optional in children. Whilst image quality may be improved by reducing peristaltic artifact and improving distention, to date there is no strong evidence this translates into improved diagnostic accuracy [30, 31]. Furthermore, use of spasmolytics may prolong the examination time and such agents attract side effects such as nausea and vomiting [30]. If a spasmolytic is employed, the recommended first-line spasmolytic agent is i.v. hyoscine butylbromide (0.5 mg/kg i.v.). The recommended second-line agent is i.v. glucagon 0.5 mg (<24.9 kg) and 1 mg (>24.9 kg), given as a slow infusion with i.v. saline at an infusion rate at 1 mL/s).

A rectal contrast enema is not required.

MRE can be performed at either at 1.5 or 3 T, although the use of higher magnetic field strength can increase chemical shift and susceptibility artifacts.

Scanning in the prone position has been demonstrated to improve small bowel distension, compared to supine although there is no evidence that this translate into improved lesion detection [31, 32]. Large multi-elements coils are needed to cover from the perineum up to the left colonic flexure, in high resolution.

Sequences:

Both steady state free precession gradient echo and 2D—T2-weighted images should be performed in the axial and coronal planes. Fat saturation in one of these planes is recommended. A maximal slice thickness of 5 mm is also recommended.

Currently the use of pre- and post-gadolinium contrast 3D T1-Fat saturation weighted sequences is recommended with a slice thickness does not exceed 3 mm. The post contrast sequence should be acquired at the portal phase of injection.

Additional sequences are also widely used but are currently considered optional whilst supporting evidence accumulates [9]:

Diffusion-weighted imaging. Recommended values range from 0 or 50 up to 600–900. The slice thickness should be 5 mm. Motion artifacts are problematic due to the length of acquisition but can be reduced with appropriate gating, fast imaging techniques (i.e., echoplanar imaging or parallel imaging), and the use of intravenous antiperistaltic agents before image acquisition. The axial plane may be less prone to artifact than the coronal plane.
Dynamic cine steady-state motility images [33, 34] are useful, particularly in suspected luminal stenosis. A better appreciation of the upstream distension and real time opening of the lumen of the abnormal bowel loop are possible using cine imaging.
Magnetization transfer sequences may have utility in quantifying fibrosis, but currently remain a research tool.

The total scan duration should equal to or be less than 45 min.

Given the success of limited sequences rapid MR protocols for suspected appendicitis [35], a fast MR IBD protocol has been proposed as an alternative to emergency US for investigating the acute pediatric abdomen [36]. This fast MR protocol is limited to a morphologic T2 sequence in two planes associated with a DWI sequence to allow identification of IBD and its complications, including diagnosis of bowel stenosis, abscess, and fistula. In the absence of bowel obstruction (which provides intrinsic bowel distension), one of the limiting factors for the use of MRI in the acute setting could be the need for an oral contrast agent in a potential surgical patient. However, it has been established in the adult literature that neither oral or rectal preparation is mandatory in order to exclude UC [37] or CD [38]. Another potential limiting factor for acute MRI is the time taken for the examination. However, rapid MRI protocols can be very short (less than 10 min), do not require spasmolytics or gadolinium chelates and can be performed with easy patient positioning such as dorsal decubitus [32].

The utility of enteric MRI performed without spasmolytics or gadolinium contrast is debated [9, 36, 39–47] but skipping medication and contrast injection reduces examinations time and increases test acceptability in children. Furthermore, there is an increasing evidence base supporting the use of DWI as an alternative to gadolinium enhanced imaging. Shenoy et al. [39] have recently reported in a cohort of 27 pediatric patients that although DWI in isolation does not perform as well as standard MRE for detection of active Crohn disease, a combination of DWI and MRE increases imaging accuracy for disease activity compared with either technique alone. In 44 young adults, Seo et al. [40] found that DWI MRE was at least equal to contrast-enhanced MRE for the evaluation of inflammation in CD, and in a large study of 130 adults and children with CD, DWI was also found to be sufficiently accurate to replace gadolinium administration [14]. Some authors have reported that DWI may indeed be even superior to contrast enhanced imaging. Sirin et al. [41] reported that in a cohort of 37 children, DWI revealed lesions not detected with MRE performed with gadolinium injection. Similarly, both Dubron et al. [42] and Neubauer et al. [36] reported superior performance of DWI in comparison to gadolinium enhanced imaging in cohorts of 48 children, and 33 children and young adults, respectively. Avoiding IV injection and spasmolytics, together with a reduction of repeated breath holds acquisitions are strong motivations for the further development and validation of rapid and “noninvasive” MRI protocols.

12.3.2.2 Disease Phenotype

MRE is a powerful tool in characterizing the underlying phenotype of IBD. Classically, multifocal bowel affecting the colon and the terminal ileum are typical of CD [48] whereas contiguous abnormal bowel wall from the rectum to the cecum is typical of UC. However, as noted above, atypical presentations of IBD exist in pediatric practice and so caution is required when making a final diagnosis.

12.3.2.3 Positive Findings

The ability to evaluate the bowel wall is significantly enhanced by luminal distension. Collapsed bowel can both hide disease and mimic pathology, generating false positive diagnosis. Wall thickening with luminal restriction is typical of CD, although in the absence of significant bowel wall thickening, abnormal DWI signal, for example, should raise the possibility of early mucosal disease (Fig. 12.7).

Fig. 12.7

Eleven-year-old girl who presents a light pancolitis on endoscopy (UC). (a) Coronal balanced turbo field echo (bTFE) sequence reveals a discrete thickening of the left colon with no evidence of thickening of the cecum. (b) Coronal T2-weighted turbo spin-echo (bTSE) sequence confirms the findings in (a). (c) DWI sequence (b1300) reveals abnormal bowel hypersignal. Early mucosal damage appears better depicted on diffusion

In parallel with US, abnormal wall thickness is usually defined as greater than 3 mm. Bowel wall signal changes is usually assessed by comparing to a nearby normal bowel loop, or fixed structure such as the psoas muscle. Active IBD results in restricted diffusion and increased mural T2 signal. This increased T2 signal is secondary to mural edema and inflammatory cell infiltration but can also be seen secondary to fat deposition seen in chronic strictures. Various contrast enhancement patterns may be seen after gadolinium injection including mucosal, layered mucosal and serosal, homogenous or predominantly serosal (Fig. 12.8). Correlation between imaging and surgical specimens histologically explored suggests increased enhancement the wall is positively correlated with disease chronicity, and the increased microvessels density commonly found in CD influences the rate of enhancement [49].

Fig. 12.8

Fourteen-year-old girl treated for ileal CD. Abnormal MR findings on different sequences. (a) Axial T2-weighted single-shot turbo spin-echo (SshSE) sequence demonstrating thickening of an ileal loop. (b) Axial balanced Turbo Field Echo (bTFE) sequence confirms the findings in (a). (c) Axial T1-weighted fat saturation after gadolinium injection sequence reveals marked enhancement of the thickened loop. (d) Axial DWI sequence (b1300) demonstrates an abnormal bowel hypersignal

In addition, MR imaging is able to demonstrate additional findings in IBD including ulceration, pseudopolyps, and mural abscesses (typically bright on T2 weighted imaging and hypointense on T1 with ring enhancement following IV gadolinium administration).

Bowel affected by IBD has a variety of appearances depending on the disease chronicity and current level of activity. It may look fixed, dilated, exhibit pseudo-sacculations or become strictured.

The use of several sequences over the time course of the examination, and motility sequences are all useful in differentiating bowel wall collapsed with pseudo-thickening from true pathological bowel wall thickening. In this situation, the absence of injection of antiperistaltic agent is interesting.

MRI also provides a detailed evaluation of extramural findings including: fibrofatty proliferation, engorged mesenteric vessels supplying an inflamed bowel (the comb sign), fistula, abscess, and inflammatory lymphadenopathy.

12.3.2.4 Complications

Abscess, sinus tract, and fistula are typically found in CD and are infrequent in UC.

Stenosis: Diagnosis is based on bowel thickening with associated upstream distension greater than 3 cm which is consistent across all sequences, particularly on the cine motility (Cine-MR) (Fig. 12.9).
Fistula and Sinus tracts: These present as high T2 signal linear tracts usually with restricted diffusion and increased enhancement following IV gadolinium. The perineum can usually be included in the field of exploration, particularly in children, thereby allowing evaluation of perianal fistulas. In this regard, MRI holds a significant advantage over US (Fig. 12.10).
Abscess: An abscess presents as a well-defined encapsulated lesion, frequently with heterogeneous hyperintense T2 signal, central restricted diffusion and low T1 signal with avid rim enhancement following gadolinium.