MRI-Based Characterization of Intestinal Motility in Children and Young Adults With Newly Diagnosed Ileal Crohn Disease Treated by Biologic Therapy: A Controlled Prospective Study

Abstract

BACKGROUND.

Changes in intestinal motility in patients with newly diagnosed Crohn disease have historically been evaluated primarily in a subjective manner.

OBJECTIVE.

The purpose of this study was to assess longitudinal changes in objective intestinal motility scores in children and young adults with newly diagnosed ileal Crohn disease treated with biologic (anti–tumor necrosis factor-α) medical therapy compared with those in control participants.

METHODS.

This prospective study included 20 children and young adults (eight female and 12 male patients; mean age, 14.6 ± 2.1 [SD] years) with newly diagnosed ileal Crohn disease who were recruited between December 2018 and October 2021 as well as 15 control participants without any known gastrointestinal conditions (eight female and seven male patients; mean age, 18.1 ± 4.4 years). All participants underwent research MRI examinations of the small bowel, including dynamic cine 2D SSFP sequences. Patients with Crohn disease underwent additional research MRI examinations performed at both 6 weeks and 6 months after initiation of biologic therapy. Two operators independently derived terminal ileal intestinal motility scores from the dynamic cine sequences by use of FDA-approved software (with higher scores indicating greater intestinal motility). Intestinal motility scores were compared between patient and control groups by use of t tests, whereas changes in intestinal motility scores after treatment were assessed using linear mixed models. Interoperator absolute agreement was assessed using the intraclass correlation coefficient (ICC).

RESULTS.

Mean terminal ileal intestinal motility scores were not significantly different between patients with newly diagnosed ileal Crohn disease and control participants (for operator 1, 180.9 ± 63.3 vs 229.7 ± 115.2, respectively [p = .12]; for operator 2, 175.0 ± 62.2 vs 236.4 ± 117.4, respectively [p = .05]). Mean intestinal motility scores changed over time compared with baseline in response to biologic therapy, for operator 1 (180.9 ± 63.3 at baseline, 248.1 ± 104.9 at 6 weeks after treatment initiation, and 249.1 ± 73.2 at 6 months after treatment initiation [p = .04]) and operator 2 (175.0 ± 62.2 at baseline, 247.8 ± 112.7 at 6 weeks after treatment initiation, and 239.6 ± 72.7 at 6 months after treatment initiation [p = .03]). Absolute agreement in intestinal motility scores was excellent between operators (ICC = 0.89).

CONCLUSION.

MRI measurements of intestinal motility are dynamic in children and adults with newly diagnosed small-bowel Crohn disease, showing early increases in response to biologic therapy.

CLINICAL IMPACT.

MRI-based intestinal motility scores may aid individualized assessment of disease activity and treatment response in patients with small-bowel Crohn disease.

MR enterography plays a critical role in diagnosing and monitoring small-bowel Crohn disease in children and adults [1]. Evaluation of intestinal inflammation typically entails subjective qualitative assessment of a variety of findings, including the degree of bowel wall thickening, intramural edema, and postcontrast hyperenhancement [2, 3]. Currently, a paucity of quantitative MRI-based tools are used in clinical practice to assess the inflammatory activity and functional behavior of bowel affected by Crohn disease. Such objective tools could be used to aid in diagnosing small-bowel Crohn disease and to assess treatment response and better direct personalized patient care.

Evaluation of intestinal motility, or peristalsis, can be performed using MRI-based dynamic cine imaging. The motility of bowel segments affected by Crohn disease can be appraised both subjectively and objectively with the use of such images. Subjective assessment of intestinal motility has been shown to aid in the detection of inflamed bowel segments, with diseased bowel segments showing decreased peristalsis (i.e., the “frozen bowel” sign) [4]. Quantitative assessment of intestinal motility performed using nonrigid registration of dynamic MR images has been shown to distinguish patients with Crohn disease from control patients and to correlate with both endoscopic and histopathologic inflammatory activity [5–7]. Intestinal motility scores measured from MR images also correlate with abdominal symptoms and laboratory measurements of inflammation in patients with Crohn disease [8, 9]. There is a paucity of data showing whether these measurements of small-bowel motility can accurately detect the response to medical therapy in small-bowel Crohn disease, although a study of 46 patients (35 of whom were identified retrospectively and 11 of whom were recruited prospectively) showed associations between increasing peristalsis and normalization of the C-reactive protein (CRP) level and decreasing MR index of activity (MaRIA) score [10].

The purpose of the present study was to assess longitudinal changes in objective intestinal motility scores in children and young adults with newly diagnosed ileal Crohn disease treated with biologic (anti–tumor necrosis factor-α [anti–TNF-α]) medical therapy compared with those in control participants. We hypothesized that MRI-based measurements of intestinal motility would increase in response to efficacious medical treatment and correlate with clinical measures of inflammation and disease activity.

Methods

Participants

This prospective study received institutional review board approval and was performed in compliance with HIPAA. Written informed consent was obtained for all participants, including informed assent for participants younger than 18 years old.

We recruited pediatric and young adult patients from the Cincinnati Children’s Hospital Medical Center inflammatory bowel disease clinic who had newly diagnosed ileal Crohn disease between December 2018 and October 2021. Patients were eligible for participation if the following four criteria were fulfilled: they had newly diagnosed small-bowel (ileal) Crohn disease confirmed by endoscopy and either CT or MRI, medical treatment with biologic (anti–TNF-α) medical therapy was planned, they were able to undergo research MRI before initiation of biologic therapy and to return for follow-up MRI examinations, and they were from 10 to 25 years old. Potentially eligible patients were also screened for the following exclusion criteria: known or suspected pregnancy and having undergone prior intestinal surgery that included removal of the terminal ileum. Concomitant treatment with nonbiologic medical therapies (e.g., immunomodulator or corticosteroid medication) was not considered an exclusion criterion.

Control participants who were 10–25 years old were recruited using hospital-approved e-mail communication. Control participants were not enrolled if they had a history of any medical condition affecting the gastrointestinal tract (including any form of inflammatory bowel disease), had undergone previous bowel surgery, had a contraindication to MRI, or if they were known to be or were suspected of being pregnant.

All patients had been included in a prior study that assessed the utility of quantitative mesenteric blood flow measurements derived from ECG-gated velocity-encoded phase contrast MRI in children and young adults with small-bowel Crohn disease [11]. That previous study did not evaluate intestinal motility.

MRI Protocol

Patients with Crohn disease underwent three research MRI examinations, which were performed at baseline (before initiation of biologic medical therapy) and at approximately 6 weeks (± 2 weeks) and 6 months (± 4 weeks) after treatment initiation. Control participants underwent a single research MRI examination. All MRI examinations used the same protocol.

All research MRI examinations were performed using a single 1.5-T system (Ingenia, Philips Healthcare). The protocol included both standardized anatomic and quantitative acquisitions as well as dynamic cine imaging. For the current study, cine images were analyzed to generate intestinal motility scores (GIQuant scores, determined using a postprocessing algorithm [GIQuant, version 1.5.4, Motilent]), which served as a surrogate marker of peristalsis. Cine imaging was performed in the coronal plane using a 2D balanced SSFP sequence. The dynamic cine dataset was acquired separately at six to eight slice locations, depending on the body habitus of the participant, and used an FOV that extended from the anterior abdominal wall through the retroperitoneum. Pulse sequence parameters of the dynamic cine sequence were as follows: TR/TE, 2.5/1.5; flip angle, 45°; FOV, 300 × 300 mm; matrix, 200 × 200; slice thickness, 10 mm; sensitivity encoding (SENSE) acceleration, 2; number of dynamic acquisitions, 120; scanning time per dynamic acquisition, 0.48 ms; and total scanning time per slice location, 59.2 seconds. Motilent did not provide input regarding the dynamic cine imaging protocol (e.g., in terms of temporal resolution, section thickness, in-plane spatial resolution, or image contrast).

Image Analysis, Including Measurement of Intestinal Motility

The coronal dynamic cine images from research MRI examinations were deidentified and analyzed independently by two operators: a nonauthor medical imaging physicist employed by Motilent (operator 1, who had 3 years of experience using study-related software) and a study investigator (operator 2 [J.R.D.], a fellowship-trained pediatric radiologist with 13 years of clinical practice after training and with no prior experience using the study-related software). The readers used cine imaging to generate an intestinal motility score for each examination reflecting the small bowel through the level of the terminal ileum; a lower score indicated relatively decreased intestinal motility, and a higher score indicated relatively increased intestinal motility. Both operators were familiar with interpretation of MR enterography examinations and were blinded to each other’s interpretations and to relevant clinical data (e.g., clinical inflammatory markers) at the time of analysis.

The image analysis method that was used for this study has been previously described [5–7]. The algorithm, GIQuant, has received 510(k) clearance from the U.S. FDA. In brief, the tool processes dynamic cine images for a given slice location and anatomic position by use of a previously validated registration algorithm designed to assess bowel motility [7]. Deformation fields generated with the registration are used to produce a surrogate motility metric and representative parametric maps, which are defined as the SD of the Jacobian (i.e., a measure of local area change) determinant of the deformation fields, on a voxel-by-voxel basis. A score of 0 corresponded to no peristalsis or motility. Terminal ileal motility was measured on these parametric maps of bowel motion. The operators placed a freehand ROI on a coronal anatomic image that encompassed the distalmost portion (approximately 5 cm, extending to the ileocecal valve) of the terminal ileum, with the ROI located along the serosal surface of the bowel and including bowel wall and lumen. This ROI was then automatically copied by the viewing platform onto the corresponding parametric motility map, with the mean pixel value reported as a measure of the overall intestinal motility score (arbitrary units [AU]) for that segment.

Operator 2 also placed a large ROI over fluid-filled loops of noninflamed (i.e., normal-appearing) proximal and mid ileum. The intestinal motility score from this large ROI was used to normalize the conventional terminal ileal intestinal motility score (hereafter referred to as the normalized intestinal motility score).

The axial and coronal SSFSE sequences of the small bowel from the research MRI examinations (section thickness, 5 mm; TE, 80 ms; number of signal averages, 2) were reviewed by an additional board-certified fellowship-trained pediatric radiologist (A.J.T., who had 14 years of postfellowship experience) to assess the severity of radiologic inflammation of the terminal ileum by use of the simplified MaRIA scores [12]. Specifically, the terminal ileum was evaluated for the presence of bowel wall thickening greater than 3 mm (1 point), intramural edema (1 point), mucosal ulcers (2 points), and perienteric inflammatory changes (1 point). A total score of 0 suggested no active inflammation, a score of 1 or greater suggested active inflammation, and a score of 2 or greater suggested severe active inflammation with endoscopic ulcers being likely.

Documentation of Clinical Data, Including Inflammatory Markers

At all three research visits, clinical inflammatory markers were measured for each patient with Crohn disease, including laboratory assessments of the erythrocyte sedimentation rate (ESR), CRP level, serum albumin level, and fecal calprotectin. The weighted Pediatric Crohn Disease Activity Index (wPCDAI), a validated tool used in both research and clinical settings to assess inflammatory disease activity, also was documented. Demographic information (e.g., age and sex) as well as height and weight were recorded.

Assessment of Clinical Remission

Clinical remission at 6 months after the initiation of biologic medical therapy was evaluated using a composite endpoint, as described in the Supplemental Methods in the online supplement.

Statistical Analysis

Continuous variables were summarized as means and SD, whereas categoric variables were summarized as counts and percentages. A t test was used for pairwise comparisons of continuous variables, whereas the Fisher exact test was used to compare categoric variables. ROC curves were used to assess the diagnostic accuracy of intestinal motility scores for distinguishing patients with newly diagnosed ileal Crohn disease from control participants and for predicting clinical remission at 6 months after initiation of medical therapy based on the composite endpoint. Sensitivity and specificity were obtained from any ROC analysis that showed a statistically significant AUC, with a cutoff value selected using the Youden index.

Linear mixed models were used to assess whether intestinal motility scores changed over time in response to medical therapy, with visit (occurring at baseline, at 6 weeks, and at 6 months) used as a fixed effect and individual participants considered a random effect. This analysis allowed inclusion of participants for whom data were missing (e.g., participants who underwent surgical removal of the terminal ileum before the 6-month research encounter.) A Tukey multiple comparisons test was used for post hoc comparison of intestinal motility scores between pairs of study visits (e.g., baseline vs 6 weeks), with p values adjusted for multiple comparisons. The number of participants who had categoric increases and decreases in intestinal motility scores, respectively, between baseline and 6 weeks after treatment initiation and between 6 weeks and 6 months after treatment initiation also were recorded.

Pearson correlation coefficients were used to evaluate associations between intestinal motility scores and clinical inflammatory markers, whereas intraclass correlation coefficients (ICCs) were used to assess interobserver absolute agreement.

These analyses were performed independently for the terminal ileal intestinal motility scores obtained by the two operators. The analyses were also performed for the terminal ileal normalized intestinal motility score obtained by operator 2.

A p < .05 was considered statistically significant for all inference testing. The 95% CIs were calculated as appropriate. Statistical analyses were performed using GraphPad Prism software (version 9.3.1 for Windows, GraphPad Software) and MedCalc statistical software (version 20.027, MedCalc Software). A formal sample size calculation was not performed given that this was an ancillary study to an earlier primary research investigation.

Results

Participants

During the study period, a total of 42 patients with newly diagnosed ileal Crohn disease for whom medical treatment with biologic (anti–TNF-α) therapy was planned were recruited for potential participation. Twenty-two patients declined participation, and no patients were excluded due to the additional screening criteria, resulting in a final sample of 20 patients with newly diagnosed Crohn disease. A total of 15 control participants were recruited. The mean age of patients with Crohn disease was 14.6 ± 2.1 years, and that of control participants was 18.1 ± 4.4 years (p = .004). The Crohn disease cohort included eight female patients and 12 male patients, whereas the control cohort included eight female participants and seven male participants (p = .51). Baseline clinical and radiologic measures of inflammatory activity are presented in Table 1.

TABLE 1:

Baseline Characteristics, Including Clinical and Radiologic Inflammatory Measures, in Children and Young Adults With Newly Diagnosed Ileal Crohn Disease

Characteristic	Mean ± SD	Range
Patients with Crohn disease (n = 20)a
Age (y)	14.6 ± 2.1	12–18
Erythrocyte sedimentation rate (mm/min)	24.2 ± 22.6	4–103
C-reactive protein level (mg/dL)	1.7 ± 2.2	0.4–9.5
Serum albumin level (g/dL)	3.6 ± 0.5	2.8–4.5
Fecal calprotectin (μg/g of feces)	469.9 ± 432.2	64.2–1405.6
wPCDAI (score of 0–125)	35.6 ± 25.0	0–93
Simplified MaRIA (score of 0–5)	3.3 ± 1.5	0–5
Control participants (n = 15)b
Age (y)	18.1 ± 4.4	11–25

Note—wPCDAI = weighted pediatric Crohn disease activity index, MaRIA = MR index of activity.

^aEight female and 12 male participants.

^bEight female and seven male participants.

All 20 patients with Crohn disease completed the baseline research MRI examination, which showed nonstricturing, nonpenetrating disease in 19 patients and penetrating disease in one patient. Nineteen of 20 patients underwent research MRI at 6 weeks after treatment initiation; a single patient withdrew from the study after undergoing baseline imaging and was lost to follow-up due to the COVID-19 pandemic. Seventeen of 20 patients underwent research MRI at 6 months after treatment initiation; one patient did not undergo MRI again due to the COVID-19 pandemic, and two additional patients underwent ileal surgical resection between the study visit at 6 weeks after treatment initiation and the scheduled 6-month follow-up encounter. The three participants who did not undergo all follow-up research MRI examinations were included in the analyses based on available relevant data. A flow diagram of study participant selection is presented in Figure 1.

Fig. 1—

Flow diagram shows selection of study participants. Fifteen control participants also were enrolled and underwent single research MRI examination. anti–TNF-α = anti–tumor necrosis factor-α.

Intestinal Motility Scores in Patients With Crohn Disease and Control Participants

For operator 1, the mean intestinal motility score was not significantly different between patients with newly diagnosed ileal Crohn disease and control participants (180.9 ± 63.3 vs 229.7 ± 115.2; p = .12) (Figs. 2 and S1, the latter of which is found in the online supplement). The intestinal motility score had an AUC of 0.62 (95% CI, 0.44–0.78; p = .26) for differentiating patients with newly diagnosed small-bowel Crohn disease from control participants.

Fig. 2—

Distributions of terminal ileal intestinal motility scores. Horizontal lines within boxes indicate medians, ends of boxes denote IQRs, and whiskers represent minimums and maximums exclusive of outliers. AU = arbitrary units, CD = Crohn disease.

A, Tukey box plot shows distributions of terminal ileal intestinal motility scores in patients with newly diagnosed ileal CD and control participants for operator 1 (p = .12) and operator 2 (p = .05).

B, Tukey box plot shows distributions of terminal ileal intestinal motility scores normalized to intestinal motility scores from fluid-filled loops of proximal and mid ileum for operator 2 (p < .001).

For operator 2, the mean intestinal motility score also was not significantly different between patients with newly diagnosed ileal Crohn disease and control participants (175.0 ± 62.2 vs 236.4 ± 117.4; p = .05). The intestinal motility score had an AUC of 0.66 (95% CI, 0.48–0.81; p = .11) for differentiating patients with newly diagnosed small-bowel Crohn disease from control participants.

The mean normalized intestinal motility score was significantly lower in patients with newly diagnosed ileal Crohn disease compared with control participants (0.38 ± 0.12 vs 0.60 ± 0.15; p < .001) (Fig. 2). The normalized intestinal motility score had an AUC of 0.88 (95% CI, 0.72–0.96; p < .001) for differentiating patients with newly diagnosed small-bowel Crohn disease from control participants, with a sensitivity of 80.0% (16/20; 95% CI, 56.3–94.3%) and specificity of 93.3% (14/15; 95% CI, 68.1–99.8%) at a cutoff of 0.46 or less.

Change in Peristalsis Score in Response to Medical Therapy

For operator 1, the mean intestinal motility score increased significantly over time compared with baseline in response to biologic medical therapy among patients with Crohn disease, measuring 180.9 ± 63.3 at baseline, 248.1 ± 104.9 at approximately 6 weeks after initiation of biologic therapy, and 249.1 ± 73.2 at approximately 6 months after initiation of biologic therapy (p = .04) (Figs. 3 and 4). Operator 1 was unable to quantify peristalsis for a single patient at 6 months because the terminal ileum was thought to be anatomically normal on SSFSE sequences and was inadequately imaged on cine imaging (residing between two adjacent slice locations). At post hoc assessment, the mean intestinal motility score was greater between baseline and 6 months (p = .007) but not between either baseline and 6 weeks (p =.07) or 6 weeks and 6 months (p > .99).

Fig. 3—

12-year-old patient with newly diagnosed ileal Crohn disease. R = right, L = left, Min = minimum, Max = maximum.

A–C, Side-by-side MRI anatomic (left) and intestinal motility (right) parametric maps were obtained at baseline (A) (simplified MR index of activity [MaRIA] score, 5) and at 6 weeks (B) (simplified MaRIA score, 3) and 6 months (C) after initiation of biologic therapy (simplified MaRIA score, 2). Intestinal motility score (denoted by mean value in green box on parametric maps) increased between baseline (160.3) and 6 weeks (395.5) before decreasing between 6 weeks and 6 months (211.8). Bowel wall thickness did not change between 6 weeks and 6 months.

Fig. 4—

Distribution of terminal ileal intestinal motility scores in patients with newly diagnosed ileal Crohn disease (CD). AU = arbitrary units.

A and B, Tukey box plot (A) and dot plot (B) for operator 1 show distribution of terminal ileal intestinal motility scores at baseline and at 6 weeks and 6 months after initiation of biologic therapy. Scores significantly increased over time in response to treatment (p = .04). In A, horizontal lines within boxes indicate medians, ends of boxes indicate IQRs, whiskers represent minimums and maximums exclusive of outliers, and circle beyond whisker denotes outlier. In B, circles denote scores at baseline, squares represent scores at 6 weeks, and triangles indicate scores at 6 months.

C and D, Tukey box plot (C) and dot plot (D) for operator 2 show distribution of terminal ileal intestinal motility scores at baseline and at 6 weeks and 6 months after initiation of biologic therapy. Scores significantly increased over time in response to treatment (p = .03). In C, horizontal lines within boxes indicate medians, ends of boxes indicate IQRs, whiskers represent minimums and maximums exclusive of outliers, and circle beyond whisker denotes outlier. In D, circles denote scores.

E and F, Tukey box plot (E) and dot plot (F) for operator 2 show distribution of terminal ileal normalized intestinal motility scores at baseline and at 6 weeks and 6 months after initiation of biologic therapy. Scores significantly increased over time in response to treatment (p = .004). In E, horizontal lines within boxes indicate medians, ends of boxes indicate IQRs, whiskers represent minimums and maximums. In F, circles denote scores.

For operator 2, the mean intestinal motility score changed significantly over time compared with baseline in response to biologic medical therapy in patients with Crohn disease, measuring 175.0 ± 62.2 at baseline, 247.8 ± 112.7 at approximately 6 weeks after initiation of biologic therapy, and 239.6 ± 72.7 at approximately 6 months after initiation of biologic therapy (p = .03) (Fig. 4). Similar to operator 1, at post hoc assessment, operator 2 found that the mean intestinal motility score was greater between baseline and 6 months (p = .008) but not between either baseline and 6 weeks (p = .06) or 6 weeks and 6 months (p = .97).

The mean normalized intestinal motility score also changed significantly over time compared with baseline in response to treatment, measuring 0.38 ± 0.12 at baseline, 0.57 ± 0.20 at approximately 6 weeks after initiation of biologic therapy, and 0.57 ± 0.21 at approximately 6 months after initiation of biologic therapy (p = .004) (Fig. 4). At post hoc assessment, the mean intestinal motility score was greater between baseline and 6 weeks (p = .01) as well as between baseline and 6 months (p = .003), but it was not greater between 6 weeks and 6 months (p > .99).

Categoric Assessment of Individual Intestinal Motility Scores Over Time

For operator 1, 14 of 19 patients (73.7%) showed an increase in the intestinal motility score between baseline and the 6-week follow-up in response to treatment, whereas five of 19 patients (26.3%) showed a decrease in the intestinal motility score. A total of eight of 17 patients (47.1%) had an increase in the intestinal motility score between the 6-week and 6-month follow-ups in response to treatment, whereas nine of 17 patients (52.9%) had a decrease.

For operator 2, 13 of 19 patients (68.4%) had an increase in intestinal motility score between baseline and 6-week follow-up in response to treatment, whereas six of 19 patients (31.6%) had a decrease in the score. A total of eight of 17 patients (47.1%) had an increase in the intestinal motility score between the 6-week and 6-month follow-ups in response to treatment, whereas nine of 17 patients (52.9%) had a decrease.

A total of 14 of 19 patients (73.7%) had an increase in the normalized intestinal motility score between baseline and 6-week follow-up in response to treatment, whereas five of 19 patients (26.3%) had a decrease in the score. A total of six of 17 patients (35.3%) showed an increase in the normalized intestinal motility score between the 6-week and 6-month follow-ups in response to treatment, whereas 11 of 17 patients (64.7%) had a decrease.

Correlation With Clinical and Radiologic Measures of Inflammation

For operator 1, the intestinal motility score showed a significant negative correlation with clinical measures of inflammation and disease activity, including the ESR (n = 55, r = −0.35; p = .009), CRP level (n = 55, r = −0.33; p = .02), and wPCDAI (n = 55, r = −0.43; p = .001) (Fig. 5). The intestinal motility score was not significantly correlated with the serum albumin level (n = 54, r = 0.27; p = .05), fecal calprotectin (n = 55, r = −0.05; p = .70), radiologic simplified MaRIA score (n = 55, r = −0.24; p = .08), or age at baseline imaging (n = 20, r = −0.24; p = .31).

Fig. 5—

Scatterplot from operators 1 (circles) and 2 (squares) shows relationship between terminal ileal intestinal motility scores and weighted pediatric Crohn disease activity index (wPCDAI) (operator 1: r = −0.43, p = .001; operator 2: r = −0.36, p = .006). Solid lines represent least squares lines of best fit, and dotted lines denote 95% confidence limits of fit lines. AU = arbitrary units.

For operator 2, the intestinal motility score showed significant negative correlations with wPCDAI (n = 56, r = −0.36; p = .006) and the radiologic simplified MaRIA score (n = 56, r = −0.30; p = .03). The intestinal motility score was not significantly correlated with the ESR (n = 56, r = −0.24; p = .08), CRP level (n = 56, r = −0.25; p = .06), fecal calprotectin (n = 56, r = −0.10; p = .45), serum albumin level (n = 55, r = 0.22; p = .11), or age at baseline imaging (n = 20, r = −0.24; p = .30).

The normalized intestinal motility scores showed significant negative correlations with the CRP level (n = 56, r = −0.30; p = .02), wPCDAI (n = 56, r = −0.29; p = .03), and radiologic simplified MaRIA score (n = 56, r = −0.27; p = .04). The normalized intestinal motility score was not significantly correlated with the ESR (n = 56, r = −0.26; p = .06), serum albumin level (n = 55, r = 0.26; p = .05), fecal calprotectin (n = 56, r = −0.13; p = .33), or age at baseline imaging (n = 20, r = −0.25; p = .29).

Prediction of Clinical Remission

Eleven of 19 patients (57.9%) achieved clinical remission at 6 months based on the composite endpoint. Neither the baseline intestinal motility score (AUC = 0.71; p = .13) nor the change in the intestinal motility score between baseline and 6 weeks after treatment initiation (AUC = 0.50; p > .99) significantly predicted treatment response at 6 months for operator 1. Similarly, for operator 2, neither the baseline intestinal motility score (AUC = 0.55; p = .77) nor the change in the intestinal motility score between baseline and 6 weeks after treatment initiation (AUC = 0.50; p > .99) significantly predicted treatment response at 6 months.

Neither the baseline normalized intestinal motility score (AUC = 0.57; p = .63) nor the change in the normalized intestinal motility score between baseline and 6 weeks after treatment initiation (AUC = 0.55; p = .77) significantly predicted treatment response at 6 months.

Interobserver Agreement

Interobserver absolute agreement between the two operators for the intestinal motility score obtained, pooled across all MRI examinations performed at the three time points in patients with Crohn disease and at the single time point in control participants (n = 70 examinations), was excellent (ICC = 0.89; 95% CI, 0.83–0.93) (Fig. 6). The mean intestinal motility score was not significantly different between operator 1 (225.2 ± 93.2) and operator 2 (224.3 ± 96.0) (p = .87).

Fig. 6—

Scatterplot shows relationship between terminal ileal intestinal motility scores from operators 1 and 2 (intraclass correlation coefficient [ICC] = 0.89). Inner set of dotted lines denotes 95% confidence limits, outer set of dotted lines indicates 95% prediction limits, thick black line represents least squares line of best fit, and thin black line denotes line of identity. AU = arbitrary units.

Discussion

We have shown in a prospective cohort that intestinal motility is altered in children and young adults with newly diagnosed Crohn disease and that it changes dynamically over time in response to biologic medical therapy. These results suggest that MRI-based scores of intestinal motility (a surrogate for peristalsis) provide a potential marker of small-bowel inflammation that does not require IV contrast material or spasmolytic medication and, perhaps more importantly, may allow objective assessment of response to medical therapy. In particular, we have shown increases in intestinal motility between baseline and 6 weeks after initiation of biologic treatment in individual patients and, on average, between baseline and 6 months after treatment initiation for the entire study cohort. Patients with Crohn disease showed no significant change in intestinal motility scores between 6 weeks and 6 months after treatment initiation, which suggests that population-level treatment effects that impact peristalsis may be most consistently observed early after initiation of biologic therapy. These patterns were documented for two independent observers, thereby supporting the robustness of the measurements as a noninvasive marker of early treatment response. We speculate that patients who do not show an increase in peristalsis between the initiation of biologic therapy and the end of induction (i.e., 6 weeks after the initiation of therapy) may benefit from more aggressive medical therapy, such as increased dosing or changing medications; however, further study is needed.

Although mean intestinal motility scores did not change for either observer between 6 weeks and 6 months after initiation of medical therapy, a categoric review of the data for individual patients provides additional insights. Some patients with Crohn disease had a continued increase in motility scores during this period, which is suggestive of improvement in disease (presumably due to decreasing inflammation, although histologic confirmation was not available), whereas other patients experienced an interval decrease in motility scores, which is suggestive of worsening disease (presumably due to increasing inflammation). Potentially, the patients who had a decrease in intestinal motility between 6 weeks and 6 months after treatment initiation may have had an initial response to biologic medical therapy during induction, followed by a relative loss of response during the maintenance phase of treatment that may have been initially subclinical (a known phenomenon with an annual risk of 20.9%) [13]. These findings also suggest that intestinal motility scores have the potential to be the first objective (and U.S. FDA–cleared) MRI-based marker of disease severity in small-bowel Crohn disease that can show treatment response at the individual patient level and, perhaps, direct more personalized medical care. Although additional studies are needed, it is also possible that this MRI-based metric could become a qualified biomarker for use in future clinical trials evaluating candidate antiinflammatory medications in the setting of Crohn disease.

The present study also shows that intestinal motility scores are correlated with a variety of clinical markers of inflammatory burden, including the ESR, CRP level, and wPCDAI. These significant associations suggest that increasing peristalsis is associated with decreasing inflammation. However, the strengths of these correlations were moderate at best, suggesting a complementary role of MRI-based measurements of peristalsis in assessing disease activity and treatment response. Peristalsis did not show significant correlation with fecal calprotectin for either observer or with radiologic simplified MaRIA score for one of the two observers. The lack of a statistically significant association with fecal calprotectin is in contrast to findings from a study by Menys et al. [9] that showed a significant (albeit weak) association between the motility score and fecal calprotectin (ρ = −0.33; p = .015). However, a recent study of 12 children showed results similar to those of our study, with no significant relationship identified between quantified intestinal motility and fecal calprotectin (r = −0.27, p = .40) [14]. The lack of a significant association between radiologic simplified MaRIA scores and intestinal motility for one of two observers suggests that the morphologic changes that affect the bowel in Crohn disease may be decoupled from peristalsis and that these two assessments may be complementary. In addition, the intestinal motility score was not significantly associated with age for either operator, although the range of ages included in the study was narrow given the inclusion of only children and young adults.

Unlike prior studies, the present study failed to show a significant difference in conventional intestinal motility scores between patients with newly diagnosed Crohn disease and control participants when conventional terminal ileal intestinal motility scores were used. For example, Menys et al. [6] showed significantly greater intestinal motility in noninflamed terminal ileum compared with inflamed terminal ileum (p = .002). The lack of difference between groups in the present study may be because patients with newly diagnosed Crohn disease had a range of disease inflammatory activity, as shown by the ranges of simplified MaRIA scores and various clinical inflammatory markers measured at baseline. Another possible cause is a lack of statistical power, as the present study was exploratory in design and not specifically powered to identify such a particular difference between cohorts.

To our knowledge, the present study is the first to assess a novel normalized terminal ileal intestinal motility score that normalizes the conventional motility score of the terminal ileum to that of upstream noninflamed fluid-filled proximal and mid ileum. Unlike conventional terminal ileal intestinal motility scores, the normalized score showed a highly significant difference between patients with Crohn disease and control participants (p < .001). The novel score allowed differentiation of participants in the two cohorts (patients vs control participants), with an AUC of 0.88. In addition, normalized intestinal motility scores showed a more robust change over time in response to biologic therapy when compared with conventional scores, with post hoc analysis revealing a significant increase between baseline and 6 weeks after treatment initiation in patients with Crohn disease. These apparently stronger associations shown by the normalized intestinal motility score may be due to a couple of factors. Normalizing intestinal motility scores to proximal noninflamed bowel may help account for individual differences in basal peristalsis rates, which vary between individuals due to a variety of factors (e.g., time of day and gastrointestinal-related hormone levels). Furthermore, in patients with Crohn disease, more upstream small-bowel loops may have been mildly hyperperistaltic, even in the absence of stricturing disease, yielding a decrease in the terminal ileal normalized intestinal motility score.

Finally, measurement of intestinal motility scores was robust, with excellent interreader agreement (ICC = 0.89) and no significant mean difference between operators. In comparison, a recent study of four MRI-based scoring systems of intestinal inflammation, all of which were based on combinations of anatomic, diffusion-weighted, or postcontrast imaging findings, showed some-what lower ICCs (range, 0.70–0.78), including 0.70 for standard MaRIA scoring [15].

The present study has limitations. First, the MRI dynamic cine imaging protocol was developed without the input of Motilent, which may have impacted algorithm performance. Second, the number of patients with Crohn disease was small. Third, intestinal motility was evaluated in the terminal ileum only. The generalizability of the results to other bowel segments, including the colon, remains uncertain. Fourth, all imaging was performed using a 1.5-T MRI system from a single manufacturer. Finally, intestinal motility scores were not directly correlated with endoscopic or histopathologic assessments.

In conclusion, intestinal motility, a surrogate for peristalsis, can be quantified in children and young adults with Crohn disease. The present study suggests that intestinal motility is dynamic, changing longitudinally over time in response to medical treatment, potentially as early as 6 weeks after initiation of biologic therapy. Differing trajectories in intestinal motility scores over time, and in particular between 6 weeks and 6 months after treatment initiation, suggest that measurements of intestinal motility may provide a useful noninvasive objective tool for assessing treatment response, possibly facilitating personalized medical decision making. Measurements of intestinal motility have excellent interobserver agreement. Finally, measurements of intestinal motility may benefit from normalization to motility scores from proximal small bowel. This tool could become the first qualified biomarker for objectively assessing treatment response in Crohn disease, thus serving as a useful clinical trial endpoint and a treatment target in the clinic.

HIGHLIGHTS.

Key Finding

• MRI-based terminal ileal intestinal motility scores showed excellent agreement between two operators (ICC = 0.89) and exhibited early changes in response to biologic medical therapy (mean at baseline vs 6 weeks after therapy initiation: operator 1, 180.9 ± 63.3 [SD] vs 248.1 ± 104.9; operator 2, 175.0 ± 62.2 vs 247.8 ± 112.7).

Importance

• MRI-based intestinal motility scores are a dynamic measure of intestinal inflammation in Crohn disease and may allow individualized treatment response assessment to biologic medical therapy.