Role of contrast-enhanced ultrasound in evaluation of the bowel

CEUS is performed on a fasting patient, using a microbubble contrast agent that enhances the ultrasound signals and allows for real-time evaluation of the vascularity [20]. In Canada, our experience has been with perflutren microspheres, marketed as Definity (Lantheus Medical Imaging, Billerica, MA) [18]. Sulfur hexafluoride, marketed as Sonovue (Bracco, Italy) in Canada, Europe, and Asia and Lumason in USA, is a similar agent with equivalent performance. These contrast-enhancing agents are gas-filled microbubbles with a phospholipid shell and are administered to the patient as a bolus intravenous injection [18, 21]. Microbubbles remain within the capillaries throughout the examination and they oscillate in the bloodstream in response to the application of an ultrasound field, producing non-linear harmonic frequencies that are detected at CEUS [22]. Our standard volume of contrast agent for a bowel quantification study is generally twice the standard dose for liver mass evaluation: with perflutren lipid microspheres (Definity, Lantheus Medical Imaging) 0.4 ml, and with sulfur hexafluoride (Lumason, Bracco Diagnostics Inc) 4.8 ml, both followed by a 5 ml saline flush.

The selected bowel is imaged with a high-resolution probe, in the frequency range of 5–9 MHz, using contrast-specific imaging techniques that currently are available on most high-end ultrasound systems. We choose the frequency optimal for the bowel evaluation which on our equipment is a C9-2 (EPIQ, Philips Bothell, WA) with a lower frequency reserved for occasional very large patients with deep bowel. Endovaginal probes with contrast capability also work exceptionally well to study deep ileal loops, in particular. Linear probes and high-resolution convex probes are also highly successful for evaluation of the perineum and anal canal.

Contrast imagining technique parameters include a low mechanical index and contrast-specific subtraction, whereby background tissue echoes are removed from the image so that a microbubble-only image is created.

Standard CEUS settings are contrast gain of 50–60%, mechanical index of 0.05, dynamic range of 40 dB, and focal zone set at maximal depth. We maintain the same parameters between examinations for standardization of the procedure. Obese patients and deeper bowel segments will reduce visualization and signal intensity; therefore, grayscale optimization is an important factor when performing bowel studies.

Although most US systems do allow for CEUS of the bowel and also generation of time–intensity quantification curves (TIC), there is considerable variability in the output of the machines which hampers the progress of this exciting technique. Additionally, many workstations, for quantification of blood flow, are machine/vendor specific which is additionally limiting. In our institution, we perform all of our quantifications with Q-Station (Philips Healthcare Ultrasound, Bothell WA). However, a more versatile system is VueBox (Bracco, Geneva, Switzerland) which is able to analyze data for quantification from all US systems following entry of 2D DICOM datasets provided by the vendors to Bracco.

Patients generally lie comfortably in a supine position. Quiet respiration is encouraged. To perform CEUS, the transducer is placed at the region of interest (ROI) selected on the initial grayscale assessment. The abnormal bowel segment is evaluated in a sagittal orientation to minimize out-of-plane effects of respiratory motion and to facilitate generation of quality TIC allowing for quantification analysis. After contrast is injected through the patient’s IV, a continuous acquisition is initiated, even before the arrival of the first bubble in the field of view, and lasting for 2 min with no motion of the transducer. Most ultrasound equipment will allow for initial evaluation of the quantification curves at the bedside to assess technical quality of the time–intensity curve and to obtain preliminary results. If CEUS fails, or is technically inadequate, a second contrast injection may be performed, after a several minute delay.

Subjective evaluation of CEUS includes assessment of the degree and pattern of mural and mesenteric enhancement [23, 24]. Generally, inflamed bowel shows transmural enhancement with rapid enhancement of the bowel wall to a higher level than the adjacent tissue and in proportion to the level of inflammation, initially leading to a peak intensity before declining due to the excretion and distribution of the contrast agent [21, 22] (Fig. 1 and Supplementary Video 1). Generally, enhancement is transmural although less often, only the submucosal layer may be involved.

With experience, observation of the wash-in and decline of contrast agent in the bowel wall may be interpreted as reflective of mild disease with low peak and rapid decline and of more severe disease with a higher peak intensity and longer duration of enhancement.

Additionally, the vascularization of the mesentery can be evaluated subjectively by demonstration of a comb sign (Fig. 2E and Supplementary Video 2), named after its more familiar appearance on CT scan and representing the filling of prominent straight intestinal arterial branches in the mesenteric arcade. This is seen invariably with active disease.

In cases where the abnormal loop of bowel is deep within the pelvis, endovaginal B Mode and CEUS can be performed (Fig. 2). This excellent technique allows for close evaluation of the bowel with high-resolution US probes with great definition of wall abnormalities, in addition to superb capability to show blood flow and complications (Supplementary Videos 2 and 3). CEUS can also be performed using the endovaginal approach with creation of quantification curves (Fig. 2F).

Further, perianal disease and evaluation of perianal complications are well performed with US and CEUS from a perianal approach (Fig. 3 Supplementary Video 4). The use of CEUS with this technique markedly improves differentiation of phlegmon from perianal abscess.

An advantage of CEUS is the generation of time–intensity curves as an objective parameter of bowel enhancement. The recorded TICs can be analyzed using built in software on the ultrasound machine or using remote workstations. The raw native data from the 2-min Cine-clip stored in the US system are log compressed and exported in DICOM format to a PACS workstation or other analysis system, where the data are re-linearized. From these data, the linear TICs with appropriate curve fitting are generated allowing for measurement of blood flow parameters within the bowel wall.

The TIC is created by placing a ROI within the enhanced bowel wall (Supplementary Video 5). In our institution, we draw free hand at least three ROIs (~ 1 cm²) centered in the most enhanced segment, and only including the bowel wall. The placement of the ROI depends on technical factors. Most often it is easiest to place the ROI in the anterior wall if luminal gas is present but our own unpublished reproducibility data show that placement within a visibly uniformly enhancing portion of the bowel wall, either anterior or posterior, is suitable, (Fig. 1F) This is supported by Greis C who shows a direct correlation of Peak Enhancement and concentration of microbubbles [25]. The ROI should not overlap and its size can be also standardized by software. From this, an enhancement curve is created based on the wash-in and wash-out characteristics. This curve shows the wash-in slope, time to peak, peak intensity, area under the curve (AUC), mean transit time, time from peak to ½, and rise time, all curved fitted and presented in log-normal (Fig. 4A, 4B) [26]. We draw three regions of interest to confirm reliability of the curve and its measurements. As there is a lot of information associated with a 2- or 3-min acquisition of data for TIC generation, acquiring information from a single ROI might introduce error from peristalsis or other movement, for example. The use of three non-overlapping ROIs will detect if one or more datasets are unreliable and ensures higher standard of results. From these three ROIs, we select the linear curve allowing for the best curve fitting for the final result.

Most analytic software programs offer two types of graphic display for viewing a TIC, log and linear representations. Although the linear graphs show the most realistic presentation of the raw data, allowing for the most accurate and precise measurement of inflammatory parameters (Fig. 4D), its large range of numerical values makes it difficult for presentation of data that would be of a reasonable clinical use. For example, a value of 100 in the linear display corresponds to 20 dB logarithmic display and 1000 linear would be 30 dB. Therefore, we view a log display for measurement of peak enhancement in dB only, as it makes a range of values which are more manageable and compatible with classification ranges of disease activity based on the bowel wall thickness for clinical use (Fig. 4C) [20]. Our own published data show that the clinical values for peak enhancement range between 10 dB and 30 dB and encompass patients with quiescent disease at the lower spectrum, less than 15 dB, and those with severe active inflammation in the order of 23 dB or greater, with more moderate inflammation showing values in between (Fig. 5) [18]. Additionally, following treatment, a change of the peak enhancement reflects a real change, whereas using linear data, huge differences in the values may reflect virtually no significant change at all.

Our software analysis system generates a printout for every patient including all of the parameters listed above, with the peak enhancement in dB and all other parameters generated from the curve of the log-normal linear data. To this date, we predominantly utilize the peak enhancement in dB, the AUC, measured in arbitrary intensity units, and the time to peak, in seconds for establishing an initial baseline activity assessment and to show changes over time with treatment or with natural disease progression (Fig. 4C, D). Subjective and objective observations of the TICs show that with severe disease, the PE is high and the decline may be delayed and at a higher enhancement level (yellow curve, Fig. 5B). On the other hand, with mild disease the PE enhancement is low and the decline is rapid and more closely reaches the baseline (red curve, Fig. 5B). These same assessments are more difficult to recognize when looking at the linear curves (Fig. 5A).

In centers in Europe, CEUS quantification analysis is often expressed as a percentage change of peak enhancement and of the AUC before and after treatment [21, 27‐29]. In the study by Quaia [30], two ROIs are drawn, one in the thickest part of the anterior wall or covering the entire bowel wall in cases when the lumen was not visible and the second as an internal reference at the adjacent mesentery. TICs are also fitted with a non-linear least-squares regression method to a proprietary log-normal model with assessment of multiple curve parameters [30, 31]. These studies only compare exams between the same patient over time.

We prefer, instead, that this robust technique allows for integration of CEUS parameters with other established subjective and objective measures of disease activity, especially wall thickness and signal on color Doppler imaging, [32] to allow for categorization of disease between large numbers of patients and for more easily applied determinations of disease activity [18]. In addition, it is invaluable for follow-up of patients over time.

Today, there is considerable variation in the output of US equipment for quantification techniques. This inconsistency in display makes acceptance of this technique for routine clinical applications a challenge and it is hoped that standardization of methods, displays, and reporting will occur in the near future as it has for other imaging techniques over time.

Determination of disease activity is most important when evaluating bowel inflammation. This is essential at the time of new diagnosis, during surveillance, to monitor treatment response and for the total assessment of symptomatic patients.

Evaluation of disease activity is performed in our institution using an ultrasound global assessment based on objective measurement of the bowel wall thickness and semiquantitative assessment of the inflammatory fat and signals on CDI [11, 18]. Wall thickness is considered as the strongest predictor of disease activity followed by hyperemia on CDI, as shown by multiple prior studies [11, 32‐37] and in a recent publication by Novak, et al. [32]. CEUS has a good direct correlation between the level of PE and degree of bowel wall thickness for assessment of disease activity [18]. Observations are classified in four categories, according to the severity of the findings shown: normal/remission, mild, moderate, or severe disease. If the bowel wall is more than 4 mm in thickness, CEUS may be performed to confirm grayscale findings and to record baseline parameters for use if there is a treatment change or for future reference. CEUS parameters are now integrated within our global assessment as an additional objective parameter of disease activity (Table 1). This allows for a more systematic approach to better follow therapeutic change. CEUS is now widely accepted and recommended in inflammatory bowel disease in the European Federation of societies for Ultrasound in Medicine and Biology (EFSUMB) guidelines [38].

Both wall thickness and hyperemia of the bowel wall when present are good predictors of disease activity [3, 18]. However, inactive disease can be present with normal bowel wall thickness or with persistently abnormal bowel wall thickening. Differentiation of active from inactive disease then relies on detection of increased vascularity within the bowel wall. This is initially performed with color Doppler, where inactivity will show little or no blood flow. However, in some cases CDI is not reliable due to patient body habitus or technical factors. In these instances, failure to show color Doppler signal within a thickened and abnormal appearing loop of bowel poses a dilemma between inactive disease or merely a technical failure resulting in poor signal detection. Therefore, activity is indeterminate. In this circumstance, CEUS is an invaluable tool often showing transmural enhancement and high CEUS parameters in cases of active disease where no color signal can be detected at all due to technical failure (Fig. 6, Supplementary Video 6) [39]. On the other hand, in patients with very long standing disease where fat infiltration of the submucosal layer often creates thickened bowel wall with no color Doppler, chronic/quiescent disease needs to be differentiated from acute on chronic active disease.

Ultrasound is long recognized as a very reliable technique for the detection of strictures and associated incomplete mechanical bowel obstruction. This dynamic real-time technique shows dilated prestenotic segments and also evidence of dysfunctional peristalsis. Differentiation of strictures poses a frequent dilemma for imagers. This can be easily resolved by the addition of subjective and objective evaluation of the bowel blood flow with CEUS by showing the contribution of inflammation to the stricture and may be aided further by stiffness determination with elastography. Strictures in IBD may be inflammatory or chronic in nature. Differentiation of these two broad categories is recognized as a major role for imaging techniques as management differs. Several investigators have addressed the role of CEUS [40‐42] and also elastography to characterize these strictures with ultrasound [42, 43]. Our own study shows that, in patients with chronic stenotic disease, smooth muscle hyperplasia/hypertrophy may contribute more to stricture than fibrosis [44] similar that to prior study by Baumgart [45]. Additionally, we show an inverse relationship between velocity measurements on shear wave elastography and peak enhancement on CEUS. [30]. Quaiai et al. show differentiation of inflammatory vs fibrotic strictures in which determination of active inflammation would favor the role of medical therapy [40].

Medical therapy for patients with inflammatory bowel disease has evolved from the treatment of patient’s symptoms and complications to active efforts to modify the natural course of the disease by achieving mucosal healing as shown at endoscopy. Today, the use of long-term biologic therapies has changed the imaging management of this patient population with recognition that interval surveillance is a requisite component of treatment management. Additionally, medical therapy with steroids, immune modulators, and biologics all necessitate imaging surveillance to assess for treatment response and detection of complications in these immunosuppressed patients. Various publications have shown that the use of CEUS with TIC analysis can differentiate pharmacologic therapy response and identify non-responders [29, 30]. This distinction is very important for patient outcome and prognosis. Early recognition of inadequate treatment response would alleviate a long, expensive, and ineffective treatment [46].

Therapy response can be categorized as showing complete positive response, no response, or various grades of partial response. We do this grading by evaluating all parameters shown on grayscale and also on CEUS. While a complete positive response may lead to virtual resolution of all observations, more often a partial response may only show reduction of the CEUS parameters with the grayscale and CDI parameters continuing to suggest active disease. CEUS with grayscale evaluation may also show complete lack of response suggesting the necessity of dose escalation or change of therapeutic regime. Surveillance scans are performed of all patients on biologic therapy as inflammatory parameters are recognized as unreliable determinants of disease.

In IBD and other bowel conditions, there is a potential for development of inflammatory masses representing fluid containing abscesses or phlegmons, where there may be severe inflammation but no actual drainable pus. Differentiation of these masses on grayscale US is often challenging as there is a strong tendency for overlap in their US imaging appearance with all frequently showing as a hypoechoic mass. This distinction is of extreme clinical significance, as management, prognosis, and therapy differ [47].

Phlegmonous masses may show diffuse hyperenhancement reflecting acute inflammatory changes, demonstrated on color Doppler or CEUS (Fig. 7A, B, Supplementary Video 7). A phlegmon is generally treated conservatively with antibiotic therapy. An abscess, by comparison, will show regions of avascularity corresponding to pockets of pus with peripheral areas of enhancement, reflective of reactive inflammation, and the abscess wall [38] (Fig. 7C, D, Supplementary Video 8). Drainage of large abscesses is the general rule although small pockets of pus may also be treated conservatively. US/CEUS may be additionally helpful for guidance of percutaneous placement of drainage catheters and for monitoring response to therapy over time.

The study of inflammatory masses with US may be influenced by probe selection and placement. It includes the advantage of endovaginal scan in women, where the frequently involved pouch of Douglas is ideally evaluated and also the transperineal placement, in men and women, for confirmation of perianal abscess (Fig. 3). Inflammatory polyps may be suspected and confirmed as intraluminal vascular masses (Fig. 8A and Supplementary Video 9).

Acute inflammatory conditions of the bowel include especially appendicitis and diverticulitis. At presentation, patients may have bowel perforation and/or an inflammatory mass, similar to those already shown with IBD. As previously described, these inflammatory masses can be further characterized with CEUS to differentiate phlegmon and abscess, as their clinical management would be different (Fig. 7).

Furthermore, CEUS may be a helpful adjuvant in early detection of wall ischemia, as a sign of gangrenous appendicitis that can be a precursor of perforation [37, 48]. In a study including 50 patients with suspected appendicitis, CEUS was 100% accurate in the identification of suppurative and gangrenous appendicitis when compare to surgical and histological findings [49].

Chronic diverticular disease may occasionally show colonic wall thickening and pseudomasses of uncertain etiology. This persistent hypoechoic wall thickening, in the absence of acute inflammatory changes, is most often related to familiar muscular hypertrophy.

This rare self-limiting inflammatory/ischemic process of the appendix epiploica of the colon may present as a hypo- and hyperechoic mass-like area next to the colonic wall at the site of patient’s pain that will enhance on CEUS as an inflammatory mass with a central area of non-enhancement that represents the central thrombosed pedicle, as occasionally seen on CT as the hyperdense central dot sign [50].

CEUS is only helpful for identification and grading of inflammation. However, it is not useful for differentiation of infectious colitis. CEUS has being described to be used only in a few instances where detection of wall necrosis or abscess may be a feature of condition, such as in cases of neutropenic enterocolitis [48, 51].

A few studies with a small number of patients also hypothesized that CEUS could detect microcirculatory changes of the bowel wall during evaluation of GVHD and to monitor therapy in this condition [48, 52].

Bowel ischemia can be present due to bowel strangulation, thrombosis of the SMA, or non-occlusive mesenteric ischemia in patients with bowel obstruction. Patients with ischemia may have thin bowel walls as a result of loss of bowel wall tissue, vascularity, and muscular tone [39].

Studies have shown that lack of or decreased CEUS enhancement of the bowel wall is a hallmark for bowel ischemia with sensitivity of 85–94% and specificity of 100% [53]. In the case of bowel obstruction, this is seen most commonly in the least peristaltic segment and/or in the most dilated bowel segment [54].

In addition, CEUS has been used in other centers for adequate visualization of the visceral arteries. A European study combined CEUS with color and spectral Doppler allowing for an unequivocal diagnosis of visceral artery stenosis in patients with abdominal angina [55].

A number of GI tumors can be incidentally encountered during bowel assessment including neuroendocrine tumors, gastrointestinal stromal tumors, adenocarcinoma, and lymphoma. Bowel wall neoplasms, regardless of their histology, generally show as hypoechoic masses on grayscale imaging which may grow with a polypoid intraluminal, mural, or serosal (Fig. 8). Their US study should always include careful assessment of the regional fatty mesentery and the lymph nodes.

As in other contrast imaging modalities, CEUS can delineate the tumor extent and vascularity very well. Some studies suggest that dynamic contrast-enhanced ultrasound with vascular recognition imaging software enables detection of micro-vessels and quantification of tumor perfusion. This has been used for assessment of treatment response and was correlated with survival [56].

CEUS can easily detect viable enhancing tumor versus areas of necrosis to guide target biopsy sampling and to minimize inadequate or insufficient sampling errors. This is particularly important in gastrointestinal stromal tumors (GIST) which can present as small or large masses within the peritoneal cavity, with or without an obvious attachment to a loop of bowel. They frequently show large areas of central necrosis or cystic degeneration, which is frequently suggestive of their diagnosis (Fig. 8F and Supplementary Video 14). On CEUS, GIST tumors show as a hypervascular mass in the arterial phase followed by slow wash-out during the portal venous phase [48]. In addition, in a small study of 24 patients with GIST, CEUS allowed for early prediction of tumor response to Imatinib treatment [57].

As compared with inflammation of the bowel which most often will involve a long segment, generally showing a stratified wall layering, most tumors show instead as a focal black mass, generally with complete wall layer destruction. Focal round masses of small and medium size identified incidentally during scans performed for IBD surveillance are often carcinoid (Fig. 8B, Supplementary Video 10 and 8E, Supplementary Video 13) or neuroendocrine tumors. They are highly vascular on CEUS and their wash-out correlates with their malignant potential [58].

Lymphoma may show additionally an aneurysmal pattern of luminal expansion and lymph node involvement. (Figure 8C, Supplementary Video 11). Perienteric fat infiltration may be associated with all tumors.

Adenocarcinoma is seen most often in the large bowel although small bowel tumors are more prevalent in the population with IBD as compared with the general population (Fig. 8D, Supplementary Video 12).

The exquisite resolution of endovaginal scans may improve evaluation of rectal cancer in women (Fig. 9, Supplementary Video 15) [59, 60].