MRI assessment of the postprandial gastrointestinal motility and peptide response in healthy humans

Feeding triggers inter‐related gastrointestinal (GI) motor, peptide and appetite responses. These are rarely studied together due to methodological limitations. Recent MRI advances allow pan‐intestinal, non‐invasive assessment of motility in the undisturbed gut.


| INTRODUCTION
The clinical relevance of measuring intestinal motility has been recently shown through its correlation with intestinal inflammation in Crohn's disease. 1,2 MRI methodology has recently come to the fore offering repeatable means to measure global and segmental motility. 3  Glucagon-like peptide 1 (GLP-1). 7 CCK delays gastric emptying and stimulates gall bladder contraction, 8 PYY slows meal transit through a delay in gastric emptying and increasing the GI transit time (known as ileal brake), 9 while GLP-1 delays gastric emptying 10 and decreases small bowel motility separately. 11,12 It thus comes as rather a major diagnostic limitation that most investigative MR paradigms are undertaken in the fasting patient and not in the physiological state with an abnormally distended bowel with contrast agent. Advances in the availability of MRI scanners, increases in the speed of acquisition and growing acceptance of this method for the investigation gastrointestinal disease has driven its role in the GI physiology. A number of techniques to assess fasting motility in ta fasted and prepared during an MR enterography (MRE) are now available 13 enabling rapid, reproducible, and sensitive assessment of global and segmental motility to complement morphological changes seen through structural imaging. 14,15 In this study, we aim to develop a methodology to assess panintestinal motility assessment in a single session using MRI in a fed and physiological state and to compare imaging findings to postprandial GI peptide responses and key patient symptoms in a healthy volunteer's cohort. This work will serve as a platform to enhance our understanding of GI physiology and diagnostic capabilities in relevant disease groups.

| Study design
This study was open label. MRI marker capsules were used to measure whole gut transit time. 19 Participants were given five MRI marker capsules (20 mm×7 mm) filled with 0.4 mL −1 15 μmol L −1 Gadoteric acid, an MRI contrast agent, to take home. They were instructed to swallow these 24 hour before attending the unit for their study day and undergoing the baseline MRI scans as previously described. 19 They were asked to fill in a questionnaire to ensure adherence to the study day restrictions.
The subjects were asked to fast from 2000 hour the previous evening and to avoid alcohol, caffeine, strenuous exercise and any medication that could affect gut function for 18 hour before the experiment ( Figure S1).

Key Points
• The clinical relevance of measuring intestinal motility using MRI has been recently shown. However, most investigative MRI paradigms are undertaken in the fasting state using bowel preparation. Here, we develop pan-intestinal motility assessment in a single session using MRI in a fed and physiological state and compare imaging findings to postprandial GI peptide responses and symptoms in healthy volunteers.
• The test meal challenge was effective in inducing a change in multiple physiological quantified end-points and in monitoring markers of GI motility in a single MRI study session which was acceptable to the subjects.
• Mapping out postprandial physiological changes in disease groups will allow us to better understand patient symptoms and perhaps identify GI peptides as possible biomarkers of dysmotility and patient symptoms. This work will serve as a platform to enhance our diagnostic capabilities in relevant disease groups such as Crohn's.
On the day of the scan, Participants attended the 1.5 T Philips Achieva MRI scanner (Philips Healthcare, Best, the Netherlands) at

| Magnetic resonance imaging
MRI scanning was carried out supine. Participants were scanned using a range of sequences. At each time point scans were acquired to assess gastric volume, 20 gall bladder volume, 21 small bowel water content 22 and small bowel motility. 15 In addition at baseline the position of the MRI marker capsules were determined to measure whole gut transit time. 19 Gastric emptying was assessed using a balanced gradient echo sequence (bTFE) acquiring 50 contiguous axial slices with recon- The content of apparent freely mobile water in the small bowel was assessed as previously described 22 using a single-shot fast spin echo sequence acquiring 24 contiguous coronal slices with reconstructed in-plane resolution 0.78×0.78 mm 2 , slice thickness 7 mm and slice gap 0mm, TE=320 ms, TR=8000 ms, SPIR fat saturation within one breath hold of 24 seconds. This sequence yields high-intensity signals from areas with freely mobile fluid and dark signals from poorly mobile or bound water and all other body tissues.
Small bowel motility was assessed using a single slice cine-MRI acquisition set at six contiguous parallel coronal planes through the small bowel. Data were acquired using a balanced gradient echo sequence (bTFE) with reconstructed in-plane resolution 1. Gall bladder volume was measured pre-and postprandial, at every acquisition time point up to 60 minutes postprandially. 21 This was carried out using the same images as for the gastric volumes as previously shown.

| Plasma collection and peptides assays analysis
On the morning of the test, 0.325 mL −1 of aprotinin was added to vacutainer tubes (BD-361017, BD Diagnostics, Oxford) for collection for each time point aiming for a final volume of 6.5 mL −1 .
Fasting 10 mL −1 blood sample was drawn and collected in the tubes. After the test meal, data were acquired every 15 minutes for the first 60 minutes and every 30 minutes thereafter to 270 minutes. Twelve samples were taken totalling 120 mL −1 . Samples were centrifuged at 3000 rpm for 10 minutes and stored on ice. 17 Plasma peptides (total GLP-1, total PYY) were analysed through enzymelinked immunosorbent assay (ELISA) techniques (Millipore, UK) as previously shown. 17 The concentrations of serum CCK were Two independent observers (AK, CH) drew ROIs manually over all the loops of the small bowel in all the slices for each imaging datasets ( Figure 1). From these ROIs the mean total power across all small bowel pixels was calculated. A larger total power motility index represents higher small bowel motility.

| Small bowel water content (SBWC)
SBWC was measured as previously validated, 22  Briefly, this method assumed that any pixel with signal intensity above the calculated threshold in the heavily T2-weighted coronal images is filled with free water. Structures such as blood vessels, bladder, and gall bladder are manually excluded.

| Gall bladder and gastric volumes
Gall bladder and gastric volumes were quantified using in house software written in IDL (Research Systems Inc. Boulder, Colorado, USA).
This method uses a semiautomatic previously validated thresholdingregion growing technique to define the content of the stomach and gas within the stomach on each image slice. 20 For total gastric volume was calculated as the sum of the stomach contents and any gas.
Postprandial gastric content volumes were fitted to a 5 parameter equation 27 to model the emptying process and allow the calculation of the gastric half-emptying time (T 1/2 ).

| Whole gut transit (WGT)
Whole gut transit was assessed as previously described. 19 From the two sets of MRI images a transit score was calculated by subdividing the bowel into eight sections and each capsule was scored according to its position in the colon at 24 hours. A weighting factor was calculated for each capsule depending on the difference of the capsule score from the median capsule score.

| Visual analogue scale (VAS)
Symptoms regarding appetite, satiety and abdominal pain were scored at each time point using previously validated questionnaire. 17

| Statistical analyses
Due to the pilot nature of this study, it was not possible to power it but similar studies done by our group have used similar-sized cohorts. 17,28,29 The data are expressed as mean±standard error of the mean (SEM). Normality of the data was assessed using Shapiro-Wilk's test. One-way analysis of variance (ANOVA) was used to assess the significance of differences. When the analysis of variance was F I G U R E 1 Motility assessment. Reference image (A) and a motility map (B) for a healthy volunteer. ROIs were placed on the reference image to include small bowel loops only significant, post hoc test assessments of the individual time points were performed using the Dunnett's for parametric data or Dunn's test for non-parametric data, to account for multiple comparisons.
All statistical analyses were performed using GraphPad Prism 7.01 (La Jolla, USA). A P-value less than .05 was considered statistically significant.

| RESULTS
All fifteen healthy volunteers (9 female, 6 male, age 29.3±2.7 years and BMI 20.1±1.2 kg m −2 ) completed the study and tolerated the experimental procedures well without any adverse event.

| Gallbladder volumes
The changes in gallbladder volumes with time are shown in Figure S2.

| Gastric volumes
The baseline gastric volumes showed small amount of resting gastric juices of 26±7 mL −1 . Gastric content volumes rose significantly upon feeding to 418±17 mL −1 at t=0 (P≤.0001) after which the volume of the stomach declined and went back to baseline (37±7 mL −1 ) at 150 minutes with an average time to empty half of the stomach contents (T 1/2 ) of 46±5 minute (Figure 2).

| Small bowel motility
From Figure 3 it can be seen that the motility index increased significantly from fasting 39±3 arbitrary units (a.u.) to a maximum of 87±7 a.u. immediately postprandial (t=0 minute) (P≤.001) after feeding and then gradually decreased back to around baseline (44±4 a.u.) in 90 minutes. Motility index rose slightly again at 120 minutes to 55±4 a.u. and decreased back again to 38±4 a.u. at 240 minutes.

| Small bowel water content
The data in Figure 4 shows a small amount of fasting (t=−20 minutes) small bowel water content of 39±2 mL −1 . The test meal induced a significant change in small bowel water content. Immediately, after the all soup meal was ingested, this increases to a maximum of 51±2 mL −1 (P≤.05) at 15 minutes. The volume decreased toward baseline (38±2 mL −1 ) at 60 minutes after which a second peak at 180 minutes is clearly seen with a volume of 65±3 mL −1 .

| Whole gut transit
The median average weighted position score (WAPS) of the MRI capsules was 1.0 (0-3.8). As described previously, 19 the WAPS at 24 hours was converted to WGT in hours, giving a median of 33 hours.

| Symptom VAS data
The meal induced a significant increase in fullness. VAS increased from 9±5 to 44±5 at 0 minute and 41±6 at 30 minutes postprandial (P<.001). The feeling returned to baseline thereafter. Bloating, distention, pain, and nausea did not change significantly. (Table 1).

| DISCUSSION
The test meal challenge was effective in inducing a change in multiple physiological quantified end-points and in monitoring markers of GI motility in a single MRI study session which was acceptable to the subjects. To our knowledge this is the first study measuring gut motility in an undisturbed bowel after a nutrient meal without artificially distending the bowel.
The physiological parameters measured in this study were in the expected range for a normal healthy cohort as we have shown repeatedly in our previous studies. [28][29][30] The test meal was ingested within a maximum of 20 minutes by the 15 subjects and hence acted as a food bolus in the stomach. The meal itself was richer in carbohydrate content (4.5%) and fat (2.9%) rather than protein (1.5%) and hence would be a good stimulus for GLP-1, PYY and CCK secretion by the enteroendocrine L cells and I cells respectively.
The meal used for this experimental work was a homogenous soup rather than a solid meal. T 1/2 for the gastric emptying was 57±5 minute starting off with a maximal volume of 458±20 mL −1 . The gastric emptying was approximately linear from 0 minute to 90 minutes postprandial. This finding is similar to our previous observations whereby following a solid meal with a drink, there is an initial rapid gastric The plasma levels of GLP-1 increased rapidly from 15±3 ug mL −1 to 22±4 ug mL −1 in response to the carbohydrate load within the meal.
It is known that one of the principal factors governing gastric emptying is the total carbohydrate load, 32  tions. Such a small bowel distention could make small bowel motility more apparent and easier to measure or otherwise be a stimulus itself for smooth muscle contraction. 35 To our knowledge, this is the first time that such detailed and non-invasive assessment of small bowel physiology has ever been described in the postprandial state with MRI.
The whole gut transit time reported in this study was a median of 33 hours. This is comparable to our previous reports of 31 hours using the same methodology. 19 There are limitations in the study.