Crohn’s Disease Stricture Response to Treatment Assessed with Magnetic Resonance Imaging and Intestinal Ultrasound: STRIDENT Randomized Trial

Abstract

Background

Transmural healing is emerging as a key treatment target in Crohn’s disease. This study aimed to determine the role of magnetic resonance imaging (MRI) and intestinal ultrasound (IUS) in the assessment of the radiologic response of Crohn’s disease strictures to treatment.

Methods

The STRIDENT (Stricture Definition and Treatment) study was a randomized controlled trial of (2:1) intensive high-dose adalimumab combined with a thiopurine vs standard dose monotherapy adalimumab in patients with stricturing Crohn’s disease. Clinical response was defined as a reduction in the Obstructive Symptom Score at 12 months. Intestinal ultrasound was performed at baseline, 4, 8, and 12 months and MRI at baseline and 12 months. This study examines secondary outcomes of stricture resolution and changes in stricture morphology with treatment.

Results

Of 77 patients, 52 were in the intensive treatment group and 25 in the standard therapy group. Clinical response was achieved in 56 of 77 patients (73%). Complete stricture resolution occurred in 17 patients on IUS (29%) and 16 patients on MRI (22%). Stricture improvement occurred in 23 of 59 patients on IUS (39%) and 24 of 72 patients on MRI (33%). Bowel wall thickness improved at 12 months on both IUS (P < .0001) and MRI (P < .001) and was significantly lower in clinical responders (IUS P = .003) and those with fecal calprotectin < 100 µg/g (IUS P < .001; MRI P = .001).

Conclusions

Radiologic severity of Crohn’s disease strictures can improve with drug treatment, with complete stricture resolution observed in some. Intestinal ultrasound and MRI are effective modalities for monitoring the treatment response in patients with stricturing Crohn’s disease (STRIDENT Drug Therapy Study: NCT03220841).

Introduction

Stricturing is a common complication of Crohn’s disease.^1,2 It has traditionally been managed with endoscopic or surgical intervention.^3,4 We have recently demonstrated that the majority of Crohn’s disease strictures contain both inflammatory and fibrotic components, and are responsive to anti-inflammatory drug therapy.^4–8

Transmural healing is emerging as an important therapeutic target in Crohn’s disease; however, assessment of transmural disease and healing is challenging and relies on imaging.⁹ Intestinal ultrasound (IUS) is a radiation-free, noninvasive bedside assessment of transmural inflammation which is increasingly being used to monitor Crohn’s disease activity and response to therapy.^10–12 Computed tomography (CT) and magnetic resonance imaging (MRI) are commonly used to assess transmural inflammation, with MRI the preferred modality due to the lack of radiation exposure.¹³ However, MRI is costly, and timely access is challenging.^11,14 While several studies have evaluated the prognostic value of MRI, there is little data examining the role of IUS in monitoring stricturing Crohn’s disease.^15–24

Key parameters for assessing transmural and mucosal disease activity include bowel wall thickening, stratification, color Doppler signal (hyperemia), and the development of complications such as strictures or fistulae.²⁵ Given the ease of application and patient preference for IUS over other imaging modalities, it shows promise as a surveillance tool for early treatment response.^10,26,27 However there has been limited evaluation of IUS as a modality to assess stricturing Crohn’s disease, and its response to therapy, particularly in prospective studies.^15–17

The randomized STRIDENT (Stricture Definition and Treatment) Study aimed to determine whether intensive drug therapy (adalimumab and thiopurine) is more effective in improving Crohn’s disease stricture symptoms and stricture morphology compared to standard therapy (standard dose adalimumab monotherapy). The primary endpoint was improvement in the symptom response defined by as a ≥1 point reduction in the Crohn’s Disease Obstructive Symptom Score (OSS) at 12 months compared to baseline. A majority of patients in both treatment arms had symptomatic improvement after 12 months of therapy.⁴ Patients in the intensive treatment arm experienced a lower rate of treatment failure.

The purpose of this study is to assess the value of IUS and MRI in monitoring the response to drug therapy in patients with stricturing Crohn’s disease.

Methods

The STRIDENT Study was an open-label, single-center, randomized controlled trial assessing the efficacy of drug therapy for the treatment of patients with symptomatic Crohn’s disease-related inflammatory intestinal strictures. The STRIDENT study design and the primary outcome have previously been published.⁴ The study was approved by the St Vincent’s Hospital Melbourne Human Research Ethics Committee (HREC 055/17). Active inflammation was defined by the presence of one or more criteria (Table 1) and confirmed by MRI or endoscopy at enrollment. Three MRI criteria were required for radiological diagnosis of a stricture: a bowel segment with ≥ 80% lumen diameter reduction compared to adjacent proximal intestine, bowel wall thickness (BWT) > 3 mm, and narrowing present on all scan acquisitions. Magnetic resonance imaging criteria for stricture diagnosis were informed by the Society of Abdominal Radiology and American Gastroenterology Association joint guidelines.¹³ The IUS parameters for identification of stricture included BWT > 3 mm, luminal narrowing +/- prestenotic dilatation, although IUS was not used to diagnose strictures for entry to the study.²⁹

Patients were excluded if there was evidence of active perianal disease, acute bowel obstruction requiring urgent surgical intervention, internal fistulizing disease in association with strictures, rectal or anal strictures, or dysplastic/malignancy-associated strictures.

Patients were block randomized in a 2:1 ratio (block size 3) to receive intensive treatment (high-dose adalimumab plus thiopurine) or standard treatment (standard adalimumab monotherapy). Data were collected using REDCap data management platform.^30,31 The intensive treatment group received 160 mg adalimumab weekly for 4 weeks and then 40 mg fortnightly in addition to either azathioprine 2.5 mg/kg or mercaptopurine 1.5 mg/kg daily. The standard treatment group received 160 mg adalimumab at week 0, 80 mg at week 2, and then 40 mg every fortnight. The intensive treatment group received dose escalation to 40 mg weekly or 80 mg weekly (if already receiving 40 mg weekly) at 4 and 8 months if there was evidence of persistent inflammation defined as C-reactive protein (CRP) ≥ 5 mg/L, fecal calprotectin > 100 μg/g, or BWT > 3 mm on IUS with increased mucosal blood flow as measured by color Doppler signal (Limberg score ≥ 2).

Fecal calprotectin and serum adalimumab levels were measured every 4 months.

Patients underwent ileocolonoscopy (or gastroscopy for upper gastrointestinal strictures) and MRI at baseline and 12 months. Magnetic resonance imaging examinations were performed using a 1.5T Siemens scanner (Avanto and Symphony, Siemens AG; Table S1). The radiologist reporting MRI findings was blinded to randomization and treatment. Intestinal ultrasound was performed on all patients at baseline and every 4 months by one of 2 experienced ultrasound examiners (E.W. or J.S.) who were not blinded to treatment. Ultrasonography was performed without fasting, using a 3.5-5 MHz convex and/or a 7 MHz linear transducer (Supersonic Imagine Premium Aixplorer SIH514). Bowel wall thickness was the average of one measurement each in longitudinal and axial views. Intestinal ultrasound scan quality was assessed as poor, satisfactory, or excellent by the performing clinician. Magnetic resonance imaging, IUS, and endoscopy were also performed at the time of withdrawal or treatment failure when possible.

Outcomes

The primary outcome of the STRIDENT study was symptom clinical response defined by as a ≥ 1 point reduction in the Crohn’s Disease OSS at 12 months compared to baseline. The OSS was developed for the STRIDENT studies and is a composite measure of the total number of days of obstructive abdominal pain and the severity over a 14-day period.⁴

This analysis of the imaging assessments addressed the secondary outcome of stricture improvement on IUS which required a reduction in stricture BWT (≥25%), resolution of hyperemia (Limberg score ≤ 1), and normalization of prestenotic dilatation (<2.5 cm). Stricture resolution on IUS required a normalized BWT (<3 mm), resolution of hyperemia, and normalization of prestenotic dilatation. Other outcomes of interest included bowel wall vascularity on IUS, assessed using the Limberg score (with grades 0 and 1 considered normal and grades 2-4 considered abnormal), BWT, stricture length, mesenteric hyperechogenicity, bowel wall stratification, and prestenotic dilatation.³² Stricture resolution on MRI was defined as normal BWT (≤3 mm) with normal luminal diameter and resolution of prestenotic dilatation (<3 cm). Stricture improvement on MRI included normalization of prestenotic dilatation, improvement in stricture-specific Magnetic Resonance Index of Activity (MaRIA) score²⁸ (≥25%), and reduction in stricture simplified MaRIA score (≥1 point).³³ Magnetic resonance imaging was assessed for BWT, prestenotic dilatation, stricture length, MaRIA score (Table S2), edema, diffusion-weighted imaging (DWI) hyperintensity, and enhancement.²⁸ The MaRIA score is a measure of disease activity (BWT, edema, fat stranding, and ulceration), with the global score a score of all colonic segments and the terminal ileum, as opposed to the simplified MaRIA score where normal segments do not contribute to the final score.³³ A stricture MRI risk score was calculated based on a combination of pretreatment factors: prestenotic dilatation ≥30 mm, stricture length > 50 mm, and BWT ≥ 10 mm (each factor being assigned a score of 1).²³

Statistical Analysis

To detect a difference in symptom improvement between the treatment groups with 80% power at 5% significance, allowing for up to 20% attrition, a total of 78 patients were required (52 in the intensive treatment group and 26 in the standard treatment group).⁴

Baseline characteristics are presented as frequency (%) and median (interquartile range [IQR]). Treatment failure was defined as symptoms requiring a change in treatment or endoscopic dilation, bowel obstruction requiring surgery, failure to cease corticosteroids by week 4, more than 2 4-week courses of corticosteroids, or loss to follow-up within the first 4 months. Missing data due to treatment failure or protocol violation were replaced by the value at the patient’s study withdrawal where available. In the event of stricture resolution on IUS, BWT was not always recorded for the normalized segment, in these instances a value of 3 mm was imputed to allow for continuous data analysis. Only patients with good quality baseline and 12-month IUS, where strictures were visible at baseline, were included for stricture resolution and improvement analysis. Analyses are based on individual patients rather than the number of strictures.

Wilcoxon signed-rank test was used to compare matched continuous variables, and Mann-Whitney U test was used for continuous variables. The chi-square test was performed for dichotomous variables. Spearman’s rank correlation was used for continuous variables, while the chi-square test of association, with reporting of Cramer’s V/Phi coefficient, was performed for correlation of dichotomous variables. Bland-Altman analysis with mixed effects modeling was performed to assess the agreement between MRI and IUS at baseline and 12 months for measuring dominant stricture length. Bowel wall thickness variability at baseline and 12 months was compared using intraclass correlation coefficient (ICC) analysis with a mixed effects model. A cutoff adalimumab drug level for BWT was determined by a receiver operating characteristic (ROC) analysis using the Youden index. A P-value of < .05 was considered significant. Data were analyzed using STATA Version 17.0 (StataCorp LLC). All authors had access to the study data and reviewed and approved the final manuscript.

Results

Between September 2017 and September 2019, 77 patients underwent randomization with 52 patients (68%) randomized to intensive treatment and 25 (33%) to standard treatment. Baseline characteristics are presented in Table 2. All 77 patients had baseline IUS performed (12 deemed poor quality and 8 were unable to identify the stricture) with 60 patients having IUS performed at 12 months (5 poor quality and 24 unable to identify the stricture). A further 13 patients failed treatment and underwent IUS at the time of study exit. Fifty-nine patients had good quality IUS scans at baseline and 12 months, with visible strictures at baseline. All 77 patients had a baseline MRI and 64 at 12 months. The 13 treatment failure patients had an MRI at the time of study exit. Seventy-two patients had visible strictures at baseline on MRI. Details of baseline and serial IUS and MRI scans are presented in Table S3. Those with poor quality IUS scans had a significantly higher median body mass index than those with good quality scans (30 vs 25 kg/m²; P = .003; Table S4). According to the primary outcome of the clinical study, 56/77 (73%) of patients achieved clinical response at 12 months.

Stricture Resolution

Stricture resolution on IUS occurred in 17/59 (29%) patients, and stricture improvement was seen in a further 23/59 (39%) patients at 12 months. On MRI, resolution was seen in 16/72 (22%) and improvement occurred in 24/72 (33%) at 12 months. There was a strong association between stricture resolution on IUS and MRI (χ²₁(n = 59) = 41.19, Cramer’s V ϕ=0.84; P < .001).

There were no significant differences in rates of stricture resolution on MRI according to treatment group or clinical response; however, on IUS, clinical responders achieved stricture resolution more often than non-responders (37% vs 6%, P = .02). Stricture resolution occurred more frequently when fecal calprotectin was < 100 µg/g at 12 months (IUS: P < .001; MRI: P = .004). Baseline IUS factors associated with resolution at 12 months include BWT < 5 mm with an odds ratio (OR) of 5.3 (95% CI, 1.47-19.30, P = .01) and stricture length < 5 cm with an OR of 3.9 (95% CI, 1.2-13.1, P = .03). The presence of multiple strictures at baseline was a negative predictor of stricture resolution on MRI at 12 months. Each additional stricture identified was associated with an OR of 0.39 (95% CI, 0.16–0.99, P = .046), suggesting that a higher number of strictures at baseline reduces the likelihood of overall resolution. A baseline MRI risk score of zero predicted resolution at 12 months with an OR of 4.56 (95% CI, 1.32-15.72, P = .02). There was a relatively strong association between stricture resolution on MRI and passable strictures at endoscopy (χ²₁(n = 64) = 10.54, Cramer’s V ϕ=0.41; P = .002).

Stricture Morphology

Bowel wall thickness

Median IUS BWT across the entire cohort (regardless of treatment assignment) reduced from 6.4 mm (IQR 5.2-7.2 mm) at baseline to 4.9 mm (IQR 3.0-6.4 mm) at 12 months (P < .0001) irrespective of treatment intensity. Median MRI BWT across the entire cohort reduced from 9 mm (IQR 8-10 mm) at baseline to 7 mm (IQR 5-9 mm) at 12 months (P < .0001) irrespective of treatment intensity or clinical response (Table 3, Figure 1). Twenty-five patients (42%) achieved a 25% improvement in BWT on both modalities at 12 months compared to baseline. In those with clinical response at 12 months or those with 4-month adalimumab drug levels > 9.7 µg/L, IUS BWT was lower at 4 months (Figure 1, Tables 3 and 4) compared to those without clinical response or with lower drug level. MRI BWT at 12 months was lower in patients with an adalimumab drug level > 9.7 µg/L compared to ≤9.7 µg/L at 4 months (P = .003; Table 4). Fecal calprotectin ≥ 100 µg/g was associated with greater BWT than those with normal fecal calprotectin at 12 months on IUS and MRI (Table 5, Figure 1).

Stricture length

The median stricture length on both IUS and MRI significantly reduced in our cohort between baseline and 12 months irrespective of treatment assignment, adalimumab drug level, or fecal calprotectin. Clinical responders had shorter stricture lengths compared to non-responders at 12 months on both imaging modalities (Table 3).

Prestenotic dilatation

Prestenotic dilatation did not significantly change from baseline to 12 months (IUS: 15% vs 19%, P = .77; MRI: 36% vs 31%, P > .99, Figure 1). No differences between treatment groups, clinical response, adalimumab drug level, or fecal calprotectin were observed with respect to the resolution of dilatation at 12 months. Stricture resolution at 12 months occurred more often in patients without prestenotic dilatation at baseline (IUS 16/17 patients [94%]; MRI 12/16 patients [75%]).

Other IUS parameters

Intestinal ultrasound bowel wall hyperemia (Limberg ≥ 2) was present in 51% of patients at baseline, reduced significantly by 12 months, and was not influenced by treatment intensity, clinical response, or adalimumab level (Figure 1). Hyperemia at 12 months occurred more commonly in those with fecal calprotectin ≥ 100 µg/g (P = .009).

Intestinal ultrasound mesenteric hyperechogenicity was seen in 61% of patients at baseline and reduced to 25% of patients at 12 months (P < .001; Figure 1). Treatment intensity did not impact this outcome. Mesenteric hyperechogenicity at 12 months was more common in patients with lower adalimumab drug level at 4 months (P = .03) and in patients with fecal calprotectin ≥ 100 µg/g at 12 months (P = .002).

Preserved bowel wall stratification on IUS was seen in 59% of patients at baseline, improving to 75% of patients at 12 months (P < .001; Figure 1). There was no significant difference between treatment groups, primary outcome, or according to adalimumab level. Bowel wall stratification at 12 months was preserved more frequently in those with a fecal calprotectin < 100 µg/g at 12 months (P = .009).

Other MRI parameters

The median stricture MaRIA score reduced from 21.1 (IQR 15.0-28.6) at baseline to 13.2 (IQR 6.7-18.5) at 12 months (P < .0001). The 12-month MaRIA score was lower in those with an adalimumab drug level > 9.7 µg/L at 4 months (P = .02) compared to those with ≤9.7 µg/L and in those with a fecal calprotectin < 100 µg/g at 4, 8, or 12 months (P = .009; P = .001; P = .005, respectively) compared to those with calprotectin ≥ 100 µg/g at the same timepoints (Table S5). There was a 25% reduction in MaRIA score at 12 months in 56% of patients, with more achieving this in the intensive treatment group (P = .009). The median simplified MaRIA score reduced from 3 (IQR 2-5) at baseline to 1 (IQR 1-3) at 12 months (P < .0001) with the 12 month simplified MaRIA score lower in clinical responders (P = .03) and in those with a fecal calprotectin < 100 µg/g at 8 or 12 months (P = .02, P = .02) compared to those with ≥ 100 µg/g at 8 or 12 months. There was a ≥ 1 point reduction in simplified MaRIA at 12 months in 66% of patients, with no difference between clinical response, adalimumab level, or fecal calprotectin.

Stricture-associated edema was identified in 30/73 (41%) patients at baseline, improving to 4/69 (6%) patients at 12 months (P < .001, Figure 1). There was no significant difference in rates of stricture-associated edema at 12 months according to treatment group, clinical response, adalimumab drug level or fecal calprotectin.

Diffusion-weighted imaging (DWI) hyperintensity was absent in 8/73 (11%) at baseline, increasing to 30/68 (44%) at 12 months (P < .001). There was no significant difference in DWI hyperintensity at 12 months according to treatment group, primary outcome, or adalimumab level.

Almost all patients’ strictures showed delayed enhancement at baseline 71/77 (92%), compared to 41/67 (61%) at 12 months (P < .001). At 12 months, rates of delayed enhancement were lower in clinical responders compared to non-responders (54% vs 92%, P = .01).

Correlation Between MRI, IUS, and Endoscopy

At baseline endoscopy, strictures were not passable in 64% of patients, this improved to 50% at 12 months although this change was not statistically significant (P = .09). There was a relatively strong association between stricture resolution on MRI and passable strictures at endoscopy (χ²₁(n = 64) = 10.54, Cramer’s V ϕ=0.41; P = .002). Stricture resolution on both IUS and MRI (13/59 patients) was strongly associated (χ²₁(n = 59) = 41.19, Cramer’s V ϕ=0.84; P < .001).

There was moderate agreement between IUS and MRI BWT (ICC 0.23, 95% CI, 0.13-0.33). The agreement between MRI and IUS worsens for larger BWT (with every millimeter increase in BWT, bias increases for around 8% (95% CI, 5.7-10.8), P < .001). For smaller BWT, agreement is better (for 4 mm BWT, regression-based bias 6.1%, limits of agreement −62%, 74%) and worsens with larger BWT (9 mm BWT, regression-based bias 48%, limits of agreement −5%, 100%; Figure S2). There was a relatively strong association between ≥ 25% improvement in BWT on IUS and MRI at 12 months (χ²₁(n = 57) = 12.45, Cramer’s V ϕ=0.47; P = .001). Prestenotic dilatation on IUS and MRI at 12 months were relatively strongly associated (χ²₁(n = 59) = 10.73, ϕ=0.43; P = .003).

Magnetic resonance imaging stricture length differed significantly from endoscopic stricture length at baseline and 12 months (baseline: n = 39; MRI 6 cm [IQR 4-12 cm] vs endoscopic 1.8 cm [IQR 1-3 cm], P < .0001; 12 months: n = 42: MRI 4 cm [IQR 0.5-8 cm] vs endoscopic 1 cm [IQR 0-2 cm], P = .0001). Intestinal ultrasound stricture length also differed significantly from endoscopic stricture length at baseline and 12 months (baseline: IUS 4 cm [IQR 3-8 cm] vs endoscopic 1.8 cm [IQR 1-3 cm], P < .0001; 12 months: IUS 4 cm [IQR 2-6 cm] vs endoscopic 1 cm [IQR 0-2 cm], P < .0001). Magnetic resonance imaging stricture length strongly correlated overall with endoscopic length (n = 81, ρ=0.67; P < .0001), while IUS moderately correlated with endoscopic length (n = 57, ρ=0.51; P = .0001). There was moderate agreement between IUS and MRI stricture length (ICC 0.35, 95% CI, 0.26-0.43). The agreement between MRI and IUS slightly worsens for longer strictures (with every cm increase in stricture length, bias increases 2.3% (95% CI, 0.5-4.0), P = .01). While agreement is good for shorter strictures (mean bias −0.4% for 2 cm stricture), limits of agreement are beyond clinically relevant margins (−121%, 120%). Similarly wide limits of agreement are observed for longer strictures (for 20 cm long stricture: mean bias 60%, limits of agreement −48%, 170%), suggesting MRI tends to overestimate the stricture length compared to IUS (Figure S2).

Adalimumab Drug Level

Receiver operating characteristic analysis of the 4-month adalimumab level defined 11.0 µg/L as the threshold discriminating patients with and without stricture resolution at 4 months (sensitivity 100%, specificity 59%, Area Under The Curve [AUC] 0.72 [AUC at cut point 0.8], Youden index 0.59, OR 1.0002 [95% CI, 1.0000-1.0003]; P = .045; Figure S1). In those with 4-month adalimumab drug levels > 9.7 µg/L, BWT was lower at 4 months on IUS and at 12 months on MRI (Table 4). The MaRIA score at 12 months was also lower in those with an adalimumab drug level > 9.7 µg/L at 4 months (P = .02).

Discussion

We investigated the role of IUS and MRI in a prospective cohort of symptomatic stricturing Crohn’s disease patients undergoing drug therapy. Patients with strictures were included in the study if stricture inflammation was evident biochemically or on MRI. Both MRI and IUS were sensitive to the diagnosis of endoscopically confirmed strictures and to the anatomical changes in those strictures that occurred with drug treatment. As far as we are aware, this is the first study to prospectively assess MRI, IUS, and endoscopic outcomes in a stricturing Crohn’s disease cohort undergoing drug therapy.

Stricture resolution (observed on imaging as well as endoscopy) occurred in about a quarter of patients after anti-tumor necrosis factor (TNF) therapy. Stricture improvement was seen in approximately a third of all patients, and with longer follow-up we anticipate that a proportion of patients with stricture improvement will eventually achieve stricture resolution. Further studies with larger sample sizes and extended follow-up periods are required to test this hypothesis and to assess outcomes of long-term anti-TNF therapy (and other advanced therapies) on stricture morphology. Improvements in stricture morphology were strongly associated with improvement in symptom scores and lower fecal calprotectin levels, irrespective of anti-TNF treatment intensity. A higher adalimumab drug level was associated with improvement in some radiological outcomes. Stricture resolution occurred more often in patients without prestenotic dilatation at baseline, while improvement in symptoms at 12 months was more frequent in patients with shorter strictures. Perhaps these stricture features reflect a lower burden of inflammation and fibrosis and are more treatment-responsive and in turn, patients with longer strictures and prestenotic dilatation may represent a more severe or advanced phenotype.

There was good concordance between IUS and MRI with respect to stricture resolution, and although assessment of stricture length on imaging compared to endoscopy was not always comparable, there was a modest correlation at 12 months. Magnetic resonance imaging overestimated stricture length when compared to endoscopy and IUS. This may have implications for clinical practice where longer strictures, deemed unsuitable for endoscopic balloon dilatation (EBD) due to length on MRI, could be reassessed with IUS for consideration of endoscopic therapy. Further studies are required to confirm this hypothesis. The assessment of strictures without the need for patient fasting or ingestion of contrast agent is an advantage of IUS when compared to MRI. Magnetic resonance imaging however, appeared better for the identification of upper gastrointestinal strictures (3% of the cohort), whereas shorter strictures were more often seen on IUS.

Few studies have directly compared the diagnostic accuracy of IUS and MRI; however, the METRIC trial found high sensitivity for the detection of active Crohn’s disease, which required increased BWT on both modalities, among other parameters of active disease (including fecal calprotectin > 250 µg/g).³⁴ They did note higher sensitivity of MRI for small bowel disease presence (97% [95% CI, 91-99] vs 92% [95% CI, 84-96], P = .03) and a non-significant higher specificity (96% [95% CI, 86-99] vs 84% [95% CI, 65-94], P = .05).³⁴ Similar to our findings, Martínez et al. found a significant difference in the magnitude of BWT between modalities, with higher measurements on MRI (6.6 mm ± 2.1 mm) compared to IUS (5.8 mm ± 1.5 mm, P = .01).³⁵ They did find excellent agreement between modalities for sites of disease activity (κ = 0.91).³⁵ We found moderate agreement between MRI and IUS for BWT in our study. Our findings support the utility of IUS in the assessment and monitoring of stricturing Crohn’s disease, when strictures are visible on IUS.

Fecal calprotectin correlates well with endoscopic disease activity, mucosal healing, and clinical remission.^9,14 However, it is unable to discriminate between locations of intestinal inflammation.^9,14 Few studies have examined the relationship between fecal calprotectin and transmural healing. In this study, fecal calprotectin < 100 µg/g at baseline and 4 and 8 months was associated with a lower median BWT and a ≥25% reduction in BWT at 12 months on IUS and MRI. Improvement in IUS bowel wall hyperemia and mesenteric hyperechogenicity was seen more frequently at 12 months in those with fecal calprotectin < 100 µg/g. This suggests that improvement in strictures may be observed more frequently in patients with a lower inflammatory burden. Our study found a moderate to strong association between low fecal calprotectin and stricture resolution.

Multiple studies have examined adalimumab trough levels and optimal cutoffs for achieving clinical remission, with suggested target levels of > 5 µg/L during the maintenance phase.^36,37 However, there is little literature examining adalimumab drug levels and radiological outcomes for stricturing Crohn’s disease. A retrospective observational study by Ungar et al. of 44 patients on adalimumab maintenance therapy found those with an adalimumab drug level of > 6.7 µg/L had significantly lower ileal or colonic BWT, when drug level and IUS were performed after a median of 10.8 months of adalimumab therapy.³⁸ In our study, ROC analysis identified an adalimumab trough level of 11 µg/L or more was best associated with stricture resolution at 4 months (sensitivity 100%, specificity 59%, AUC 0.8, P = .045). Adalimumab drug level at 4 months was associated with improvements in IUS and MRI BWT, IUS bowel wall hyperemia, and disease activity score (MaRIA) score at 12 months. In patients with Crohn’s disease strictures being treated with anti-TNF, proactive therapeutic drug monitoring may improve transmural healing rates although further prospective studies are required to confirm this association.

Limitations to this study include the lack of blinding of IUS clinicians to treatment arm and clinical progress which potentially may impact assessment. However, the MRI radiologist was blinded to treatment. The quality of IUS scans was also assessed at the discretion of the clinician, although there is currently no standardized assessment of IUS scan quality. Three-quarters of patients in this study had paired good quality scans that allowed for assessment of strictures. Patients with poor quality scans were found to have a higher median body mass index (Table S4); however, further study into factors affecting the quality of IUS scans is required. While proximal small bowel strictures were uncommon in this study, IUS has lower sensitivity for the detection of disease in the proximal small bowel.³⁹ Studies with larger sample sizes would be useful to confirm the results found in this study.

In this study, prestenotic dilatation was not a requirement for stricture diagnosis; however, 67% of strictures were not passable endoscopically and current radiologic guidelines support making a stricture diagnosis based on luminal narrowing and increased BWT with or without prestenotic dilatation.^14,40

Normal parameters for intestinal stricture criteria vary between IUS and MRI, hence slightly different definitions were used for IUS and MRI in this study.⁴⁰ Recently the Stenosis Therapy and Anti-Fibrotic Research (STAR) consortium defined small bowel strictures on IUS as requiring increased BWT (>3 mm), luminal narrowing +/− prestenotic dilatation (>2.5 cm).⁴¹ In this study, response to treatment was defined as a reduction in stricture length, BWT, luminal narrowing, and prestenotic dilation. There is no universally accepted definition of stricture or transmural healing in Crohn’s disease on IUS, although normalization of BWT is the most important parameter.⁴² While the STAR consortium has emphasized the importance of stricture-specific transmural healing, this is yet to be explicitly defined. It is important to note that stricture healing does not equate to transmural healing of all Crohn’s disease segments. Universal definitions to standardize IUS-based stricture assessment and transmural healing are needed to improve intra and interobserver reproducibility.

Cross-sectional imaging of MRI and IUS provides different but complementary information to endoscopy, the latter observing a passable lumen diameter, but being unable to assess features of the intestinal wall such as thickness and transmural inflammation. In non-passable strictures, cross-sectional imaging is mandatory to assess stricture length. We would support the routine use of IUS and MRI in the diagnosis and monitoring of Crohn’s disease strictures. Changes observed on IUS appear to inform clinical and endoscopic outcomes and may assist with treatment optimization.

In summary, bowel damage associated with inflammatory stricturing Crohn’s disease is often reversible, with radiologic stricture improvement and resolution observed on both IUS and MRI with anti-TNF therapy. Improvement in BWT was seen more commonly in clinical responders to anti-TNF therapy irrespective of treatment intensity. Higher drug levels (greater than 11 µg/L) are associated with improved outcomes. Intestinal ultrasound and MRI performed similarly and can be used to monitor the treatment response of Crohn’s disease strictures to drug therapy. Intestinal ultrasound response in patients with Crohn’s strictures 4 months following the commencement of drug therapy predicts durable clinical response and should be considered for incorporation into routine clinical care.

Author Contributions

G.L.: data analysis, writing—original draft, writing—review and editing. J.S.: conceptualization, methodology, data curation, writing—review and editing, supervision. A.H.: conceptualization, methodology, data curation, writing—review and editing, supervision. T.S.: data curation. AR: data curation. E.W.: conceptualization, methodology, data curation, writing—review and editing, supervision. M.K.: conceptualization, methodology, data curation, writing—review and editing, supervision.

All authors approved the final version of the article and authorship list.

Funding

This work was supported by the National Health and Medical Research Council; Gastroenterological Society of Australia IBD Clinician Establishment Award; Australasian Gastrointestinal Research Foundation; AbbVie; Spotlight Foundation; The University of Melbourne.

Conflicts of Interest

T.R.S. reports personal fees from Siemens and Bayer. M.A.K. reports grants and personal fees from AbbVie and Janssen; and personal fees from Takeda, Pfizer, and Ferring. The remaining authors declared no competing interests.

Data Availability

The deidentified data underlying this article will be shared on reasonable request to the corresponding author. Review by both the requesters and the authors’ institutional Human Research Ethics Committee may be required.

The STRIDENT (Stricture Definition and Treatment) study (US NCT 03220841)

https://clinicaltrials.gov/study/NCT03218202?term=STRIDENT&rank=1

Ethics Statement

The study was approved by the St Vincent’s Hospital Melbourne Human Research Ethics Committee (HREC 055/17).

Guarantor of the article

Professor Michael Kamm.

REFERENCES