Visualization of Bile Movement Using MRI Spin-Labeling Technique: Preliminary Results

Medical Physics and Informatics Original Research Gonoi et al. Bile Flow Visualized with Spin-Labeling Medical Physics and Informatics Original Research Wataru Gonoi 1 Masaaki Akahane 1 Yasushi Watanabe 2 Sachiko Isono 3 Eriko Maeda 4 Kazuchika Hagiwara 1 Kuni Ohtomo 1 Gonoi W, Akahane M, Watanabe Y, et al. Keywords: bile, biliary tract, cholagogues and choleretics, cholangiography, MRI, spin-labeling technique DOI:10.2214/AJR.12.8928 Received March 17, 2012; accepted after revision July 3, 2012. 1 Department of Radiology, Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Address correspondence to W. Gonoi (watapi-tky@umin.net). 2 Department of Radiology, The University of Tokyo Hospital, Tokyo, Japan. 3 MRI Applications, Toshiba Medical Systems, Tochigi, Japan. 4 Department of Computational Diagnostic Radiology and Preventive Medicine, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan. AJR 2013; 201:133 141 0361 803X/13/2011 133 American Roentgen Ray Society Visualization of Bile Movement Using MRI Spin-Labeling Technique: Preliminary Results OBJECTIVE. The purpose of this article is to noninvasively visualize intrabiliary bile movement using an MRI spin-labeling technique and administration of water, full-fat milk, and negative contrast agent as stimuli for bile excretion. SUBJECTS AND Methods. Six healthy volunteers underwent three studies with each of three oral liquid agents (water, full-fat milk, and manganese chloride solution) for a total of 18 MRI studies. Oblique-coronal T2-weighted images of the common bile duct were acquired at an inversion time of 1500 milliseconds after pulse labeling using a spin-labeling technique with an inversion pulse, repeated at intervals of 22 seconds. Bile flow rate was measured before and for 50 minutes after administration of the oral liquid agents, and its correlation with the change in gallbladder volume was assessed. RESULTS. Both anterograde and retrograde intermittent bile movements were successfully visualized in the common bile duct. The summation of excreted bile volume calculated from spin-labeled images correlated significantly with a decrease in gallbladder volume (p = 0.011). Milk stimulated significantly prolonged bile flow; flow was momentary with manganese chloride and mild with water; however, gallbladder volume decreased only in milk studies (p = 0.003). Biliary flow early after oral intake correlated significantly with gallbladder contractility at 50 minutes after oral intake (p = 0.049). CONCLUSION. A new method for visualizing intrabiliary bile movement in semi real time (22-second time resolution) using an MRI spin-labeling technique was proposed. Bile was shown to be excreted in a to-and-fro type of movement. Administration of water and negative contrast agent may induce temporary bile excretion. S everal investigative procedures are available for examination of the bile duct: biliary scintigraphy, ERCP, sonography, CT, MRI [1], IV contrast agent (gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid) discharged into the bile on MRI [2], and so on. However, these techniques have one or more inherent limitations, including radiation exposure from radiopharmaceutical agents or x-ray based procedures, side effects associated with contrast material, postendoscopic pancreatitis (in ERCP), inability to assess biliary function (only morphologic features), and poor time resolution. The spin-labeling method is an MRI technique that enables the study of blood flow and perfusion by selectively saturating the spin of protons in the blood of vessels in a particular region [3, 4]. Development of the half-fourier fast spin-echo imaging technique has improved the temporal resolution of this method [5 7] and has enabled the ves- sels of interest to be studied more selectively by the simultaneous acquisition of tag-on and tag-off images, using a freehand spin-labeling technique in which the vessels can be marked in any orientation [7, 8]. This method has been applied to the visualization of various arteries [5, 9, 10], intraportal venous blood flow [8, 11, 12], physiologic CSF [7], and secretory flow of pancreatic juice [13, 14]. The spin-labeling method enables immediate noninvasive assessment of physiologic condition and delivers no radiation exposure or other procedure-associated side effects. To the best of our knowledge, no previous study has successfully visualized bile flow in the common bile duct (CBD) noninvasively and in semi real time. The purpose of the current study was to noninvasively visualize bile movement in the CBD in semi real time using an MRI spin-labeling technique, without using IV contrast material or radiopharmaceuticals, and to assess the feasibility AJR:201, July 2013 133

Gonoi et al. TABLE 1: Scanning Parameters Sequence Parameter T2-Weighted Imaging T1-Weighted Imaging Spin-Labeled Imaging Sequence type Single-shot fast spin-echo Gradient-echo Single-shot spin-echo Scanning direction Transaxial 3D Oblique-coronal Fat suppression No Yes Yes Parallel imaging (reduction factor) QD Torso Speeder, Toshiba (2) QD Torso Speeder, Toshiba (2) QD Torso Speeder, Toshiba (2) Time of repetition (ms) Infinity 5.5 3000 Time of echo (ms) 80 2.5 80 Echo-train length 96 1 80 Flip angle ( ) 90 10 90 Slice thickness/overlap (mm) 5/0 3/1.5 10/not available No. of excitations 1 1 1 Matrix size 256 212 256 236 512 512 Field of view (cm) 36.2 30.0 34.7 32.0 40.0 40.0 Acquisition time (s) 18 23 353 of the method by correlating biliary velocity with gallbladder volume. Subjects and Methods Our institutional research ethics committee approved this prospective study, and written informed consent was obtained from all subjects. TABLE 2: Scanning Protocols and Timing of Acquisitions Time (min) Sequence Type 12 T2-weighted imaging 8 Spin-labeled imaging 2 T1-weighted imaging 0 Administration of liquids 10 Spin-labeled imaging 18 Spin-labeled imaging 27 Spin-labeled imaging 36 Spin-labeled imaging 45 Spin-labeled imaging 50 T1-weighted imaging 51 T2-weighted imaging Subjects and MRI Spin-labeled examinations were performed in six healthy volunteers (three women and three men; age range, 26 33 years) using a 1.5-T MRI scanner (Excelart Vantage, Toshiba) with quadrature-detected phased-array coils (eight channels; QD Torso Speeder, Toshiba). All subjects were asked to fast for at least 6 hours before each MRI study. Each subject underwent three MRI studies with each of three types of oral agents, with intervals of a few days. Thus, 18 MRI studies were obtained in total. During the study, the subjects were administered the three liquid materials orally, in the following order: first, soft water (volume, 300 ml; hardness, 38 mg/l; Crystal Geyser, Otsuka Foods) as a control; second, full-fat milk (volume, 300 ml; fat, 13.2 g; Meiji Fukami Ajiwai, Meiji Dairies) as a stimulus for gallbladder contraction, chosen because a minimum of 10 g fat induces sufficient gallbladder contraction [15, 16]; and third, a solution of manganese chloride (volume, 250 ml; manganese chloride tetrahydrate, 144 mg/l [Mg, 40 mg/l]; Bothdel Oral Solution 10, Kyowa Hakko Kirin), a negative oral contrast material routinely administered in MRCP studies in Japan. All liquids were at room temperature (15 20 C) before administration. The sequences and imaging parameters are summarized in Table 1 (T2-weighted images, T1- weighted images, and spin-labeled images). Spinlabeled images were 2D T2-weighted images using a respiratory-gated spin-labeling technique (time spatial labeled inversion pulse sequence) with nonselective inversion recovery applied to nullify background signal and a label-selective pulse applied to visualize bile movement. They were acquired in the oblique coronal plane that depicted both the CBD and the main pancreatic duct (Figs. 1A 1D), with an inversion time of 1500 milliseconds, after pulse labeling with a 2-cm-thick slab, and were repeated 15 times every 22 seconds on each sequence. When this technique is applied, the bile in the CBD (with inverted spins) travels into the region of interest during the period between the inversion pulse and acquisition (1500 milliseconds). A brief summary of the scanning protocol is shown in Table 2. Total acquisition time was approximately 70 minutes (20 minutes before administration of liquids plus 50 minutes after) for each study. The time period of 50 minutes after administration of liquids was chosen because gallbladder contraction is reported to last for 30 45 minutes after contractile stimuli [17 19]. Image Analysis Two board-certified radiologists with 10 and 8 years of experience in pancreaticobiliary imaging performed the image analysis. The radiologists were given no information about the subjects or the type of agent administered. Measurements were determined by consensus for nominal results and by averaging for numeric results. Calculations were performed using image-processing software (ImageJ 1.42q, National Institutes of Health), as follows. The amount of discharged bile was calculated by measuring the base (B) and height (H) in millimeters of the triangular area on the spin-labeled images (Fig. 1E). The amount of discharged bile (X) in cubic millimeters was calculated using δ cone approximation, as follows: X = πb 2 H / 12. Bile flow in the CBD (dx/dt = mm 3 /s) for each spin-labeled image was calculated as follows: dx / dt = X mm 3 / 1500 ms (1500 ms is the inversion time used in image acquisition). The total amount of bile flow after administration of the liquids for each MRI study, estimated from the spin-labeled images, was summed by integrating the area under the time mean velocity curve ( (dx / dt)dt = dx, in cubic millimeters) for time 0 until the end time of each study. The average bile flow rate before administration of three liquids was calculated as baseline bile drainage. Gallbladder volume (in cubic millimeters) before (V before ) and approximately 50 minutes after 134 AJR:201, July 2013

Bile Flow Visualized with Spin-Labeling (V after ) oral administration of liquids was determined on transaxial images by multiplying gallbladder area (manually traced on each slice) by slice thickness. Relative gallbladder volume at approximately 50 minutes after oral administration (%V after ) was determined as follows: %V after = V after / V before 100. Mean signal intensities of the distal third of the CBD and the left longissimus thoracis muscle were measured on the same fat-suppressed T1- weighted image and T2-weighted image slices, using a circular region of interest (4 mm 2 for CBD and 100 mm 2 for muscle), and were evaluated before and at approximately 50 minutes after oral administration of liquids. Relative CBD signal intensity was calculated by dividing CBD signal intensity by that of the muscle. The presence of any flow signal from the CBD to the main pancreatic duct (or vice versa) was also recorded. Statistical Analysis The relationship between the area under the time mean velocity curve and the difference in gallbladder volume before and after contrast agent administration, and the relationship between the average bile flow rate in the first spin-labeled sequence acquired after administration and the relative gallbladder volume were plotted; the least-squares linear regression model was used to test for correlations. Differences in relative gallbladder volume among the three liquids after oral administration were assessed using the Kruskal-Wallis test and post hoc analyses. The paired Student t test was used to assess changes in relative CBD signal intensity after administration of the three liquids. The level of statistical significance was set at 0.05. All statistical computing was performed using the free software R (version 2.9, The R Foundation for Statistical Computing). Results No subject was shown to have hepatobiliary or pancreatic disease, and no malformations were detected. The confluence of the cystic duct and the CBD was cephalad to the base of the labeled pulse in all subjects. A total of 1500 spin-labeled images were acquired from 100 spin-labeled sequences, after excluding failed acquisitions due to procedural delay. The CBD was depicted clearly on 89.6% (1344/1500) of all spin-labeled images (Fig. 2). The following patterns of bile flow were detected: stationary state (56% [756/1344]), anterograde movement (38% [515/1344]), retrograde movement (5% [69/1344]), and even simultaneous anterograde and retrograde movement (0.3% [4/1344]). The remaining images failed to depict the CBD because of inconsistent respiratory or peristaltic movement (10% TABLE 3: Relative Signal Intensity in the Distal Third of the Common Bile Duct on T1- and T2-Weighted Images, Before and After Administration of Liquids Relative Signal Intensity Type of Image, Liquid Before Administration of Liquids After Administration of Liquids p a T1-weighted images Water 0.88 (0.29) 0.89 (0.37) 0.95 Full-fat milk 0.90 (0.20) 0.79 (0.25) 0.38 Manganese chloride 0.88 (0.22) 1.07 (0.26) 0.33 T2-weighted images Water 8.07 (2.04) 9.01 (3.06) 0.51 Full-fat milk 7.41 (2.53) 7.77 (2.96) 0.46 Manganese chloride 7.00 (1.93) 7.00 (2.42) 0.50 Note Data are mean (SD) relative signal intensity, which is defined as signal intensity of the lower end of the common bile duct divided by that of the left longissimus thoracis muscle. a Paired Student t test. [156/1500]). In some images, anterograde flow was observed in the main pancreatic duct in the head of the pancreas, unrelated to the flow direction in the CBD or the type of liquid administered. No abnormal flow was observed from the CBD to the main pancreatic duct, or vice versa. Qualitatively, bile movement was less before oral administration of liquid agents than after (Fig. 3). Among the three liquids, bile movement was most extensive and longest lasting for full-fat milk, followed by manganese chloride and water. Bile movement was most prominent on the early phase spin-labeled images compared with the later phases, for all three liquids. Baseline bile drainage was calculated to be 0.4 mm 3 /s (SD, 6.8 mm 3 /s). There was a good correlation between area under the time mean velocity curve and the difference in gallbladder volume before and after contrast agent administration (least-squares linear regression model, R = 0.58 and p = 0.011) (Fig. 4). The value of relative gallbladder volume was high in water studies (mean [SD], 114% [61%]) and in manganese chloride studies (130% [68%]), but was significantly low in milk studies (16% [7%]). In comparing the three liquids in terms of relative gallbladder volume, the Kruskal-Wallis test revealed a significant difference (p = 0.003), and post hoc analyses revealed that the relative gallbladder volume of patients who had ingested full-fat milk was significantly smaller than those who had ingested water (p = 0.007) and manganese chloride (p = 0.03). However, no statistically significant difference was detected between water and manganese chloride studies (p = 0.91) in terms of relative gallbladder volume (Fig. 5). A moderate correlation was found between average bile flow rate in the first spin-labeled sequence after administration and relative gallbladder volume (least-squares linear regression model, R = 0.47 and p = 0.049). No significant change was observed in signal intensity of the CBD relative to muscle between before and after oral administration at any study (Table 3). Discussion This study successfully used a spin-labeling technique to noninvasively visualize biliary flow in the CBD in semi real time, and biliary flow was proven to correlate with gallbladder volume. We observed continual bile flow in the CBD, including sporadic and irregular retrograde flow. This result indicates that bile is not excreted in a continuous flow; rather, there is continual flow in a to-and-fro type of movement, as a result of the peristaltic contraction of smooth muscle. Sporadic flow of pancreatic juice was also visualized in the current study, as reported in recent studies that used the same spin-labeling technique [13, 14]. Full-fat milk was an effective stimulus for gallbladder contraction, in accordance with the results of previous studies [15, 16, 18 20]. Full-fat milk is known to induce the secretion of cholecystokinin (which increases hepatic bile secretion, contracts the gallbladder, and relaxes the sphincter of Oddi), to have an effect similar to that after cholecystokinin administration [18]. Manganese chloride [21] and ferric ammonium citrate [22, 23] are useful agents for suppressing signal arising from bowel; however, their effects on biliary function have yet to be fully investigated. Previous studies re- AJR:201, July 2013 135

Gonoi et al. ported that these agents did not alter the diameter of the CBD or the main pancreatic duct [24], although they could cause signal loss at the confluence of the CBD and main pancreatic duct after endoscopic sphincterotomy [25]. However, the sample size was small in the current study, and we detected no change in signal intensities of the distal third of the CBD after administration of liquids. Manganese chloride solution and water induced mild and temporary bile excretion in the early phase after administration. This result may suggest that oral intake itself can stimulate bile excretion; however, the effect was larger for manganese chloride than for water, the reason for which is not yet known. Of interest, gallbladder volume mildly increased at 50 minutes after administration of these two liquids, which may probably be explained by a previously reported phenomenon that hepatic bile enters the gallbladder continuously during fasting [26]. In the current study, the change in gallbladder volume before and after administration of liquids was used as the reference standard for gallbladder contraction because it was previously reported as reproducible, as well as cholescintigraphy [27 29], and the change in gallbladder volume correlated well with biliary flow observed on spin-labeled images. Biliary flow observed on spin-labeled images is determined by two factors: bile passing directly from the liver and biliary tree and bile moving along the cystic duct. The former factor could be approximated by the rate of bile synthesis calculated from previous studies (3.3 3.5 in mm 3 /s in healthy subjects and 7.6 mm 3 /s in patients after cholecystectomy and choledochotomy) [30, 31]. In the current study, baseline bile drainage was quite small (0.4 mm 3 /s), and the two factors mentioned earlier seemed balanced before administration of liquids. These flow rates were much smaller than those of extensive bile flow soon after administration of liquids, especially milk. Therefore, we speculate that flow observed on spin-labeled images after administration of liquids was mainly due to gallbladder contraction. In the current study, the early phase of bile excretion after oral intake was shown to predict the final state of bile contraction, which is in agreement with the findings of a previous study in which gallbladder ejection fraction at 20 minutes after meal ingestion correlated well with that at 45 minutes [19]. In a future study, the present method could be applied to evaluate biliary diseases. For example, impaired biliary flow would be observed in patients with intracystic cholelithiasis and choledocholithiasis [17, 18, 32], biliary neoplasms, biliary dyskinesia [33], and sphincter of Oddi dysfunction [34], or abnormal backflow of bile would be observed in patients with anomalous arrangement of pancreaticobiliary ducts [14] or status after sphincterotomy [25]. The present method would also be useful in assessing the patency of stenosed CBD in patients with primary or secondary bile duct neoplasms or inflammation and in distinguishing mucinous lesions from patent bile duct in patients with intraductal papillary mucinous neoplasm of bile duct because mucinous lesions and bile show similar signal intensity on MRI [35]. Limitations of the current study include the small number of subjects, enrollment of only healthy subjects, and use of gallbladder volume as a reference standard instead of scintigraphy. Although cystic contraction does not occur immediately after fatty meal administration [17], we could not image bile movement at this time point because of delays inherent in setting up the spin-labeled acquisition; however, administration of IV cholecystokinin would solve this problem. Obtaining continual measurements had the effect of blurring the spin-labeled images, thus degrading the quantitative aspects of the current study. In conclusion, we successfully visualized intermittent and irregular intrabiliary bile movement in semi real time using an MRI spin-labeling technique, proven by the correlation of biliary flow with gallbladder volume. Acknowledgments We thank Mizuho Murakami, Sachiko Inano, Masamichi Takahashi, Takana Yamakawa, Kaoru Sumida, and Kouhei Kamiya for their cooperation. References 1. O Connor OJ, O Neill S, Maher MM. Imaging of biliary tract disease. AJR 2011; 197:W551 W558 2. Takao H, Akai H, Tajima T, et al. MR imaging of the biliary tract with Gd-EOB-DTPA: effect of liver function on signal intensity. Eur J Radiol 2011; 77:325 329 3. Nishimura DG, Macovski A, Pauly JM. Considerations of magnetic resonance angiography by selective inversion recovery. Magn Reson Med 1988; 7:472 484 4. Edelman RR, Mattle HP, Kleefield J, Silver MS. Quantification of blood flow with dynamic MR imaging and presaturation bolus tracking. Radiology 1989; 171:551 556 5. Miyazaki M, Lee VS. Nonenhanced MR angiography. Radiology 2008; 248:20 43 6. Ito K, Koike S, Shimizu A, et al. Portal venous system: evaluation with unenhanced MR angiography with a single-breath-hold ECG-synchronized 3D half-fourier fast spin-echo sequence. AJR 2008; 191:550 554 7. Yamada S, Miyazaki M, Kanazawa H, et al. Visualization of cerebrospinal fluid movement with spin labeling at MR imaging: preliminary results in normal and pathophysiologic conditions. Radiology 2008; 249:644 652 8. Ito K, Koike S, Jo C, et al. Intraportal venous flow distribution: evaluation with single breath-hold ECG-triggered three-dimensional half-fourier fast spin-echo MR imaging and a selective inversionrecovery tagging pulse. AJR 2002; 178:343 348 9. Satogami N, Okada T, Koyama T, Gotoh K, Kamae T, Togashi K. Visualization of external carotid artery and its branches: non-contrast-enhanced MR angiography using balanced steady-state free-precession sequence and a timespatial labeling inversion pulse. J Magn Reson Imaging 2009; 30:678 683 10. Shonai T, Takahashi T, Ikeguchi H, Miyazaki M, Amano K, Yui M. Improved arterial visibility using short-tau inversion-recovery (STIR) fat suppression in non-contrast-enhanced time-spatial labeling inversion pulse (Time-SLIP) renal MR angiography (MRA). J Magn Reson Imaging 2009; 29:1471 1477 11. Tsukuda T, Ito K, Koike S, et al. Pre- and postprandial alterations of portal venous flow: evaluation with single breath-hold three-dimensional half- Fourier fast spin-echo MR imaging and a selective inversion recovery tagging pulse. J Magn Reson Imaging 2005; 22:527 533 12. Shimada K, Isoda H, Okada T, et al. Unenhanced MR portography with a half-fourier fast spin-echo sequence and time-space labeling inversion pulses: preliminary results. AJR 2009; 193:106 112 13. Ito K, Torigoe T, Tamada T, Yoshida K, Murakami K, Yoshimura M. The secretory flow of pancreatic juice in the main pancreatic duct: visualization by means of MRCP with spatially selective inversionrecovery pulse. Radiology 2011; 261:582 586 14. Sugita R, Furuta A, Horaguchi J, et al. Visualization of pancreatic juice movement using unenhanced MR imaging with spin labeling: preliminary results in normal and pathophysiologic conditions. J Magn Reson Imaging 2012; 35:1119 1124 15. Stone BG, Ansel HJ, Peterson FJ, Gebhard RL. Gallbladder emptying stimuli in obese and normalweight subjects. Hepatology 1992; 15:795 798 16. Inoue Y, Yoshikawa K. Fatty meal MRCP. Jpn Diagn Imaging 1999; 19:731 736 17. Omata T, Saito K, Kotake F, Mizokami Y, Matsuoka T, Abe K. Dynamic MR cholangiography after fatty meal loading: cystic contractility and 136 AJR:201, July 2013

Bile Flow Visualized with Spin-Labeling dynamic evaluation of biliary stasis. Magn Reson ue of negative oral contrast media in MR cholan- netic resonance cholangiography and comparison Med Sci 2002; 1:65 71 giopancreatography (MRCP) (in German). Rofo with hepatobiliary scintigraphy. J Magn Reson Im- 18. Inoue Y, Komatsu Y, Yoshikawa K, et al. Biliary 2000; 172:55 60 aging 2002; 15:75 81 motor function in gallstone patients evaluated by 25. Sugita R, Nomiya T. Disappearance of the com- 30. Prandi D. Canalicular bile production in man. Eur fatty-meal MR cholangiography. J Magn Reson mon bile duct signal caused by oral negative con- J Clin Invest 1975; 5:1 6 Imaging 2003; 18:196 203 19. Al-Muqbel KM. Gallbladder ejection fraction measured by fatty meal cholescintigraphy: is it affected by extended gallbladder emptying data acquisition time? Ann Nucl Med 2010; 24:29 34 20. Al-Muqbel KM. Diagnostic value of gallbladder emptying variables in chronic acalculous cholecystitis as assessed by fatty meal cholescintigraphy. Nucl Med Commun 2009; 30:669 674 21. Morita S, Ueno E, Masukawa A, et al. Prospective comparative study of negative oral contrast agents for magnetic resonance cholangiopancreatography. Jpn J Radiol 2010; 28:117 122 22. Hirohashi S, Uchida H, Yoshikawa K, et al. Large scale clinical evaluation of bowel contrast agent containing ferric ammonium citrate in MRI. Magn Reson Imaging 1994; 12:837 846 23. Hirohashi S, Hirohashi R, Uchida H, et al. MR cholangiopancreatography and MR urography: improved enhancement with a negative oral contrast agent. Radiology 1997; 203:281 285 24. Petersein J, Reisinger W, Mutze S, Hamm B. Val- trast agent on MR cholangiopancreatography. J Comput Assist Tomogr 2002; 26:448 450 26. Krishnamurthy GT, Krishnamurthy S. Hepatic bile entry into and transit pattern within the gallbladder lumen: a new quantitative cholescintigraphic technique for measurement of its concentration function. J Nucl Med 2002; 43:901 908 27. Buchpiguel CA, Sapienza MT, Vezzozzo DP, Rockman R, Cerri GG, Magalhaes AE. Gallbladder emptying in normal volunteers: comparative study between cholescintigraphy and ultrasonography. Clin Nucl Med 1996; 21:208 212 28. Uchiyama K, Kuniyasu Y, Higashi ST, et al. Precision of the gallbladder ejection fraction obtained with Tc-99m-pyridoxyl-5-methyl-tryptophan ( 99m Tc-PMT) hepatobiliary scintigraphy as compared with the contraction ratio in three-dimensional computed tomography. Ann Nucl Med 1997; 11:123 128 29. Vyas PK, Vesy TL, Konez O, Ciavellara DP, Hua K, Gaisie G. Estimation of gallbladder ejection fraction utilizing cholecystokinin-stimulated mag- 31. Duane WC. Measurement of bile acid synthesis by three different methods in hypertriglyceridemic and control subjects. J Lipid Res 1997; 38:183 188 32. Krishnamurthy GT, Bobba VR, McConnell D, Turner F, Mesgarzadeh M, Kingston E. Quantitative biliary dynamics: introduction of a new noninvasive scintigraphic technique. J Nucl Med 1983; 24:217 223 33. Krishnamurthy S, Krishnamurthy GT. Biliary dyskinesia: role of the sphincter of Oddi, gallbladder and cholecystokinin. J Nucl Med 1997; 38:1824 1830 34. Corazziari E. Sphincter of Oddi dysfunction. Dig Liver Dis 2003; 35(suppl 3):26 29 35. Oki H, Hayashida Y, Namimoto T, Aoki T, Korogi Y, Yamashita Y. Usefulness of gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acidenhanced magnetic resonance cholangiography for detecting mucin retention in bile ducts: a rare intraductal papillary mucinous neoplasm of the bile duct. Jpn J Radiol 2011; 29:590 594 (Figures start on next page) AJR:201, July 2013 137

Gonoi et al. C A Fig. 1 30-year-old man. Prescan setup process before acquisition of spin-labeled images. A D, First, on transaxial T2-weighted image (A), near-sagittal scanning plane (dotted lines) is set through confluence of main pancreatic duct and common bile duct, to include terminal end of main pancreatic duct. From first image, oblique-sagittal 2D MR cholangiopancreatography image (B) is placed to include main pancreatic duct and common bile duct (arrow). On transaxial T2-weighted image (C), scanning plane is set perpendicular to that shown in panel A, to include both main pancreatic duct and common bile duct (arrow), and oblique-coronal spin-echo single-shot T2-weighted image (D) is obtained. Labeling-selective pulse is set 3 cm upstream from terminal end of common bile duct (arrowhead, D). Spin-labeled images are also set on this image. This entire process takes approximately 6 minutes. E, Example of spin-labeled image. Amount of discharged bile (X), which appears as triangular shape (thin lines), is calculated as X = πb 2 H / 12, where B = base (horizontal solid bold line) and H = height (vertical solid bold line). Dotted lines indicate extent of labeling-selective pulse. Pancreatic juice flowing in main pancreatic duct in head of pancreas (not measured in present study) would appear as intraductal hyperintensity beneath labeling-selective pulse. D B E 138 AJR:201, July 2013

Bile Flow Visualized with Spin-Labeling A D Fig. 2 30-year-old man. Bile excretion on spin-labeled images with 22-second time resolution. A F, Spin-labeled images depict static state of bile (bile flow rate, 0 mm 3 /s) and static state of pancreatic juice (A), static state of bile (bile flow rate, 0 mm 3 /s) and anterograde flow of pancreatic juice (arrow, B), anterograde flow of bile (arrows, C) (bile flow rate, 43.4 mm 3 /s), anterograde flow of bile (arrows, D) (bile flow rate, 100.1 mm 3 /s), retrograde flow of bile (arrowhead, E) (bile flow rate, 5.4 mm 3 /s), and simultaneous anterograde and retrograde flow (arrows and arrowhead, F) (bile flow rate, 7.2 and 5.2 mm 3 /s, respectively). Dotted lines indicate bottom of labeling-selective pulse. B E C F AJR:201, July 2013 139

Gonoi et al. 140 Water Full-fat milk Manganese chloride Instantaneous Bile Flow (mm 3 /s) Instantaneous Bile Flow (mm 3 /s) 100 60 20 20 100 50 0 50 5 1 14 17 26 33 37 46 56 Time After Administration of Liquids (min) 150 Water 20-Period moving average (water) Full-fat milk 20-Period moving average (high-fat milk) Manganese chloride 20-Period moving average (manganese chloride) 5 9 19 30 Time After Administration of Liquids (min) Fig. 3 Time bile flow rate. A and B, Scatter diagrams are based on spin-labeled images for one subject (A) and for all subjects (B). Water, full-fat milk, and manganese chloride were administered orally at time 0. Of three agents, full-fat milk stimulated most prolonged and extensive excretion of bile (which was irregular and continual), followed by manganese chloride and water. Sporadic and irregular retrograde bile flows were also observed throughout study. 41 A B 140 AJR:201, July 2013

Bile Flow Visualized with Spin-Labeling Summation of Excreted Bile Volume Estimated on Spin-Labeled Images (mm 3 ) Relative Volume of Gallbladder (%) 250 200 150 100 50 500.0 400.0 300.0 200.0 100.0 0 0 0.0 10,000 Water Full-fat milk Manganese chloride 5000 0 5000 10,000 15,000 20,000 Decrease in Gallbladder Volume After Administration of Liquids (mm 3 ) Fig. 4 Relationship between summation of total excreted bile volume estimated on spin-labeled images and decrease in gallbladder volume at 50 minutes after administration. First-order regression line was calculated using least-squares approach. We found significant correlation between summation of total excreted bile volume and decrease in gallbladder volume (R = 0.58; p = 0.011). Water Full-fat milk Manganese chloride * * 10 20 30 40 50 60 Time After Administration (min) Fig. 5 Relative gallbladder volume after administration of three liquids. Full-fat milk induced extreme gallbladder contraction at 50 minutes after administration, whereas water and manganese chloride solution did not (although gallbladder volume increased slightly). Two outliers in upper right of figure are not from same subject. Asterisks (*) denote statistical significance. AJR:201, July 2013 141