Monitoring of gallbladder and gastric coordination by EPI


  • Presented in part as a poster to the 12th Annual Meeting of ISMRM, Kyoto, Japan, 2004



To assess for the first time the potential of echo-planar magnetic resonance imaging (EPI) for measuring simultaneously both gallbladder and gastric emptying.

Materials and Methods

Eight healthy subjects ingested 500 mL of an acid-stable liquid test meal containing 15% olive oil and flavoring. Every 20 minutes for three hours thereafter, a rapid EPI multislice set was acquired across the whole abdomen, using a dedicated whole-body 0.5-T EPI scanner.


The bile in the gallbladder and the test meal in the stomach appeared bright in the EPI images, aiding localization and region of interest analysis. We measured the gallbladder emptying curve and fitted the data to a simple analytical model. The mean fasted gallbladder volume was 25 ± 4 mL, comparable to previously published MRI and ultrasound values. Gastric emptying data fitted well to a linear model linear (R2 = 0.99), and we observed an exponential (R2 = 0.98) relationship between gallbladder and gastric volumes for the first 90 minutes.


This study shows the potential of EPI to monitor simultaneously and noninvasively the emptying of the gallbladder and of the gastric lumen. No contrast enhancing agents are needed. This method could overcome the limitations of previous gamma scintigraphy and ultrasound techniques. J. Magn. Reson. Imaging 2005;21:82–85. © 2004 Wiley-Liss, Inc.

SINCE THE EARLY 1980s (1–3), up until today (4–10), there has been a continual interest in assessing gallbladder and gastric emptying simultaneously, to provide an insight into the physiology of the gallbladder contraction and its coordination with gastric emptying and the delivery of nutrients to the duodenum (1–3). Such data could also aid studies of gut hormone signaling (8, 10) and the action of pharmacologic interventions (5, 7). Importantly, such data could also help to clarify objectively the uncertain association between symptoms and gastrointestinal involvement in diseases such as diabetes with suspect visceral neuropathy (4) and hepatitis (4, 9).

Such investigations have previously been conducted using gamma scintigraphy techniques (1, 2, 4, 8), ultrasound (9), a combination of both ultrasound and scintigraphy (6), or ultrasound and the breath hydrogen test (3, 5). However, such techniques suffer from known limitations such as poor spatial resolution, lack of three-dimensional capability, or use of radioactive materials ingested orally for the stomach and intravenously for the gallbladder.

Recently, magnetic resonance imaging (MRI) has been shown to be able to measure the emptying of the gallbladder accurately, noninvasively, and with high spatial resolution (11, 12). It has also been shown in the past that MRI can measure gastric emptying (13, 14). In this study we aimed to assess for the first time the potential of echo-planar imaging (EPI) for the simultaneous assessment of gallbladder and gastric emptying, with a view of developing a simple method that could improve upon some of the ultrasound and scintigraphic techniques mentioned above for assessing gallbladder and gastric coordination.


Eight healthy subjects ingested 500 mL of a liquid test meal containing 15% olive oil (stabilized with Tween 60 surfactant and designed to be stable at the low pH found in the gastric environment) and flavoring after an overnight fast. Subjects were asked to ingest the meal within 10 minutes. A dedicated whole-body 0.5-T EPI scanner, built in-house and equipped with actively shielded gradients and a 50 cm diameter bird-cage coil was used. Transverse, 10-mm thick slices were acquired with in-plane image resolution of 3.5 mm × 2.5 mm (with a 128 × 128 pixels matrix, effective echo time TE = 40 msec and repetition time 180 msec). The EPI module used a switched gradient that was sinusoidally oscillated at 0.5 kHz with a maximum amplitude of 14.4 mT/m. In all experiments a rapid modulus blipped echo-planar single pulse technique (MBEST) EPI multislice set of 10 slices, repeated with the bed moved between blocks to encompass both the gallbladder and the gastric lumen, was acquired in a period of less than 30 seconds before ingestion, immediately after ingestion and every 20 minutes for three hours thereafter. The subjects were kept sitting upright between scanning periods and were repositioned in the magnet at each imaging time. This protocol was approved by the local Ethics Committee, and subjects gave informed written consent.

Data are presented as mean ± standard error of the mean (SEM). Gallbladder and gastric volumes were measured by manually drawing regions of interest on Analyze (Mayo Foundation, Rochester, MN) and then summing across the slices.

The gallbladder contraction with time comprehends both an emptying and a delayed refilling component. We used the model shown in Eq. [1] to fit the data:

equation image(1)

where V1 and V2 are the maximum volume reached by each component, and D1 and D2 are their decay constants, respectively. τ represents the time delay of the second component's peak.


Due to the intrinsic T2* weighting of echo-planar images, the bile in the gallbladder and the test meal in the stomach appeared bright in the images, aiding localization and region-of-interest analysis (Fig. 1).

Figure 1.

Multislice EPI set across the abdomen of a volunteer. The gallbladder and the test meal in the stomach appear very bright in the images aiding localization and analysis. This contrast could be improved by simply increasing the T2-weighting (TE = 40 msec only here).

The mean gallbladder emptying curve is shown in Fig. 2. The solid line is the line of best fit to a simple model, which is described in the Methods section. The mean fasted gallbladder volume was 25 ± 4 mL, comparable to previously published MRI and ultrasound values (12, and references therein).

Figure 2.

Gallbladder volumes plotted vs. time of acquisition. t = 0 is before ingestion of the test meal, t = 10 minutes is immediately after the test meal. The solid line shows the best fit of the data to the simple model described in the Methods section.

The average gastric emptying curves, measured simultaneously, fitted well to a linear model (Fig. 3, R2 = 0.99), as expected for this test meal, which was rich in fat.

Figure 3.

Gastric volumes plotted vs. time of acquisition. The solid line shows the linear fit to the data (R2 = 0.99).

An exponential relationship was observed between gallbladder volume and gastric volume for the first 90 minutes (i.e., before the gallbladder started refilling) as shown in Fig. 4 (R2 = 0.98).

Figure 4.

Gallbladder volume values plotted against gastric volume values at corresponding time points for the first 90 minutes (i.e., before the gallbladder started refilling). Coordination of gallbladder and gastric emptying in this case is exponential (R2 = 0.98).


This study shows the potential of EPI to monitor simultaneously and noninvasively the emptying of the gallbladder and of the gastric lumen. No contrast enhancing agents are needed. When drawing the regions of interest on Analyze, the operator defined the intensity levels for the gallbladder and the stomach separately, and defined their edge manually. The stomach is a clearly defined region usually well separated from the gallbladder, whose characteristic globular shape clearly distinguishes it from the cystic duct. The duodenum, being a very contractile region, is usually empty and again easily distinguished from the relatively well-defined gallbladder and stomach regions. The maximum and minimum gallbladder volumes allow easy estimates of gallbladder contraction (11). Importantly, by knowing meal energy density from measuring meal dilution by MRI (15, 16), it would also be possible to estimate calories delivered to the duodenum and also correlate this to gallbladder coordination, in a fashion similar to Fig. 4. We have carried out an assessment of the intra-observer and inter-observer errors in measuring gallbladder volumes. The same operator measured the volume of a single gallbladder data set manually, five times, with a resulting standard deviation error of 10%. Afterwards, five different operators measured manually the same gallbladder set once each with a standard deviation error of 11%. In the past we have carried out validation studies (17) where the error in measuring gastric volumes with EPI was found to be of the same order (10%). An EPI multislice scan across the abdomen takes only a matter of seconds to perform. On a commercial clinical scanner one would allow about five minutes for patient setup, about 50 seconds for autoshim/tune/flip angle determination and a quick scout sequence to locate the stomach, plus about one minute to take the patient out. Hence, three subjects could be interleaved within the same three-hour experimental session with a data point every 25 minutes per patient, saving costs and machine time and allowing serial studies with blood sampling and comparison against gut signaling peptides and/or action of drugs. Semiautomation of the volume measurements could be aided by increasing the T2* weighting of the EPI module. Although EPI has been used in this study, another ultrafast T2-weighted sequence (such as half Fourier single shot turbo spin echo (HASTE)) could be used. The data has potentially high temporal and spatial resolution, allowing a sensitive measurement of the gallbladder emptying curve. This has allowed us to develop a simple model of gallbladder emptying, which yielded a good fit, as shown in Fig. 2. This will be refined as more data become available for different meals. In particular it would be interesting to investigate the possible presence of two gallbladder emptying components (fast and slow) that have previously been reported (3). The fat emulsion test meal that we used in this study was intentionally artificial. Test meals that are easier to prepare and more palatable could be used instead, such as commercially available vegetable soups, providing they contain enough fat to stimulate a robust small intestinal hormone peptide response. Gamma scintigraphic studies themselves use a variety of test meals like pancakes and milkshakes with half emptying times depending principally on the calorie content of the components.