Original Research. Jung Hoon Kim, MD, Tae Kyoung Kim, MD,* Hyo Won Eun, MD, Bong Soo Kim, MD, Moon-Gyu Lee, MD, Pyo Nyun Kim, MD, and Hyun Kwon Ha, MD

Transcription

1 JOURNAL OF MAGNETIC RESONANCE IMAGING 16: (2002) Original Research Preoperative Evaluation of Gallbladder Carcinoma: Efficacy of Combined Use of MR Imaging, MR Cholangiography, and Contrast-Enhanced Dual- Phase Three-Dimensional MR Angiography Jung Hoon Kim, MD, Tae Kyoung Kim, MD,* Hyo Won Eun, MD, Bong Soo Kim, MD, Moon-Gyu Lee, MD, Pyo Nyun Kim, MD, and Hyun Kwon Ha, MD Purpose: To determine the efficacy of the combined use of magnetic resonance (MR) imaging, MR cholangiography (MRC), and MR angiography (MRA) in the preoperative evaluation of gallbladder carcinoma. Materials and Methods: During a 20-month period, 41 patients with proven gallbladder carcinomas were referred for MR examination, including MR imaging, MRC, and gadolinium-enhanced dual-phase MRA to determine the operability of their gallbladder carcinoma. Eighteen patients who underwent surgery within six days of the MR examination were included in this study. All MR images were analyzed in order to assess bile duct invasion, vascular invasion, hepatic invasion or metastasis, lymph node metastasis, and invasion into adjacent organs. Results: Surgical and histopathologic findings revealed hepatic invasion in nine patients, bile duct invasion in nine, vascular invasion in three, and lymph node metastasis in 10. The sensitivity and specificity of MR examination were, respectively, 100% and 89% for bile duct invasion, 100% and 87% for vascular invasion, 67% and 89% for hepatic invasion, and 56% and 89% for lymph node metastasis. Conclusion: The all-in-one MR protocol, including MR imaging, MRC, and MRA, can be an effective diagnostic method in the preoperative work-up for gallbladder carcinoma. Key Words: gallbladder; gallbladder neoplasm; magnetic resonance image (MRI); MR angiography; MR cholangiopancreatography J. Magn. Reson. Imaging 2002;16: Wiley-Liss, Inc. PRIMARY CARCINOMA of the gallbladder is the most common malignancy of the biliary system. Early diagnosis of gallbladder carcinoma remains difficult because of the nonspecificity of its clinical manifestations. The ideal treatment for gallbladder carcinoma is curative surgical resection. When confined to the gallbladder, the tumor can be easily resected, but the tumor usually is not detected until it reaches an advanced state that has extended beyond the gallbladder (1 4). Recent reports suggest that aggressive surgery for advanced gallbladder carcinoma can improve the prognosis. In a recent report, the overall five-year survival rate was 52% after curative resection, compared with only 5% after noncurative resection (5 10). The decision whether or not to perform resection is usually made on the basis of the radiologic and clinical evaluations of the tumor status. Computed tomography (CT) and ultrasound have been used to demonstrate the primary lesion and local spread of gallbladder carcinoma; however, these image modalities are limited for detecting lesions, as well as vascular and biliary invasion (11 16). Several recent studies describe the usefulness of magnetic resonance (MR) imaging in the evaluation of gallbladder carcinoma, but there has not been enough attention to local tumor spread into adjacent organs, major visceral vessels, and bile ducts (17 21). Recent MR techniques, such as MR cholangiography (MRC) and MR angiography (MRA), have improved the evaluation of bile ducts and vascular structures. To our knowledge, there have been few reports of the usefulness of MR imaging in the diagnosis of gallbladder carcinoma, and there are no reports evaluating the potential use of MRC or MRA for this purpose. This study was intended to determine the efficacy of the combined use of MR imaging, MRC, and MRA in the preoperative evaluation of gallbladder carcinoma. Department of Diagnostic Radiology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. *Address reprint requests: T.K.K., Department of Diagnostic Radiology, Asan Medical Center, University of Ulsan College of Medicine, Poongnap-Dong, Songpa-Ku, Seoul , Korea. tkkim@amc.seoul.kr Received March 22, 2002; Accepted August 12, DOI /jmri Published online in Wiley InterScience ( MATERIALS AND METHODS Patients A computerized search of the pathology records between June 1998 and March 2000 identified 231 patients with pathologically proven gallbladder carcinoma. Forty-one of the patients underwent MR 2002 Wiley-Liss, Inc. 676

2 Preoperative Evaluation of Gallbladder Carcinoma 677 examination for preoperative evaluation of their gallbladder carcinomas. Among these 41 patients, 18 underwent surgery and were included in this study. The 23 patients who did not have surgery were excluded from the study. The reason for not performing surgery on these patients included hepatic metastasis (seven), extensive bile duct invasion (six), vascular invasion (three), and patient refusal (seven). The study included 11 women and seven men. These patients ranged in age from 40 to 72 years (mean 57 years). The time interval between MR examination and surgery ranged from one to six days (mean 4.5 days). MR Imaging All MR examinations were performed on a 1.5-T scanner (Magnetom Vision; Siemens, Erlangen, Germany). A circular, polarized, phased-array body coil with four elements was used. No patient required sedation. No antiperistaltic agents were administered. In all patients, MR imaging, MRC, and MRA were performed using standardized protocols. The entire MR examination was performed under the guidance of a radiologist in order to determine the location of the MRC and MRA slabs to properly fit the entire length of the common bile duct, porta hepatis, intrahepatic duct, portal vein, and hepatic artery. MR imaging sequences included T1-weighted two-dimensional breath-hold fast low-angle shot (FLASH), T2- weighted breathing-averaged turbo spin-echo (SE), and gadolinium-enhanced T1-weighted two-dimensional FLASH images. The imaging parameters for the T1- weighted two-dimensional FLASH images in which one signal was acquired were repetition time (TR) 132, echo time (TE) 4 msec, flip angle 90, and acquisition time seconds. The imaging parameters for the T2-weighted turbo SE sequence in which one signal was acquired were TR , TE msec, flip angle 180, and acquisition time seconds. Two MRC techniques were applied, i.e., single-slab rapid acquisition with relaxation enhancement (RARE) and multislice half-fourier acquisition single-shot turbo spin-echo (HASTE). The slabs of a single-shot RARE sequence were obtained at various angles to allow optimal visualization of the bile ducts. Images with coronal, oblique coronal (20 35 to the coronal plane), and sagittal planes were routinely obtained. The number of thick-slab acquisitions per patient ranged from five to 10 (mean seven acquisitions). Then multislice HASTE images were obtained in the coronal plane. Each examination was performed during a single breath-hold. The imaging parameters for the singleshot RARE sequence were as follows: TR msec, effective TE 1200 msec, echo spacing 11.5 msec, echo-train length 240, flip angle 150, slab thickness mm, field of view (FOV) 300 mm, number of signals acquired one, matrix , and acquisition time 6.32 seconds. The imaging parameters for the multislice HASTE sequence were the following: TR msec, effective TE 95 msec, echo spacing 11.9 msec, echo-train length 128, flip angle 150, section thickness 4 mm with no gap, number of slices (range of coverage mm), FOV 300 mm, number of signals acquired one, matrix , and acquisition time seconds. Fat saturation was used to reduce strong fat signal during image acquisition. The total acquisition time for all imaging steps in the MR imaging sequence was less than 15 minutes. Postprocessing of the multislice HASTE images was not performed. For MRA examination, the peak arterial enhancement time was calculated using a test bolus method. The test bolus injection was performed with 1 ml of gadopentetate dimeglumine (Gd-DTPA) followed by a 30-mL saline flush. Using a test bolus injection, a single-level axial image was chosen that best demonstrated the porta hepatis. Sequential axial images were then obtained at 1-second intervals up to 30 seconds from the start of the Gd-DTPA injection. A region of interest was positioned in the aorta at this level, and a time-intensity curve was obtained to determine the peak arterial enhancement time and optimal injectionto-scanning delay time. Three-dimensional MRA was performed in the coronal plane using fast imaging with a steady precession (FISP) sequence. The parameters of the contrast-enhanced three-dimensional MRA were as follows: TR/TE 5/2 msec or 4.6/1.8 msec, section thickness 3 or 3.5 mm, slab thickness 96 or 112 mm, number of partitions 24 or 32, matrix or , FOV 350 or 400 mm, and acquisition time 19 or 23 seconds. Thirty ml (0.2 mmol/ kg) of Gd-DTPA was administrated using an automatic power injector (MRS-50; Nemoto, Tokyo, Japan or Spectris; Medrad, Pittsburgh, PA) at a rate of 4 ml/ second followed by 10 ml of saline flush. Three-dimensional FISP MRA was obtained three times with the same parameters, i.e., precontrast, arterial phase, and portal phase images. A scan delay for the arterial phase was calculated according to a method described by Earls et al (22). The arterial phase images were systemically subtracted from the precontrast images. The portal phase was set with an inter-scan delay of 20 seconds after the arterial phase in order to fit the portal venous and hepatic venous phases. Zero filling along the z-axis was performed to obtain 64 reconstructed sections overlapping by 50%. The entire examination took approximately 40 minutes. Image Interpretation MR imaging, MRC, and MRA were retrospectively analyzed by two radiologists (JHK, HWE) who were unaware of the surgical results; consensus readings were used to determine the presence or absence of various modes of tumor spread, e.g., invasion into the adjacent organs such as the liver, duodenum, bile ducts, and visceral vessels, liver metastasis, and lymph node metastasis. The images also classified the appearance of the gallbladder carcinoma into three types: mass replacing the gall bladder, fungating mass, and diffuse wall thickening. Assessment of the vascular invasion was made up to the firstorder branches of the hepatic artery and portal vein because the peripheral branches beyond that point were usually not well imaged on MRA. Focal or eccentric luminal narrowing, luminal irregularity, and abrupt cutoff of the vascular branches were considered positive for vas-

3 678 Kim et al. Figure 1. MR imaging and MRC of gallbladder carcinoma with hepatic and bile duct invasion. a: Two-dimensional FLASH T1-weighted image shows a low signal intensity lesion, suggesting direct spread of gallbladder carcinoma to the hepatic parenchyma (arrows). b: Two-dimensional FLASH T1-weighted image shows a low signal intensity mass in the gallbladder (arrows). c and d: d is 1.6 cm caudad to c. Breathing-averaged turbo SE T2-weighted image shows ill defined, slightly high intensity lesion, suggesting direct spread of gallbladder carcinoma to the hepatic parenchyma (large arrows). There is a mass in the gallbladder (arrow heads) with pericholecystic fat infiltration and loss of low signal intensity in the wall of the gallbladder (small arrow). There is also lymph node involvement extending to portal hepatis (open arrow heads). e: Oblique coronal single-shot RARE MRC shows abrupt cut-off at the common hepatic duct (arrows) with dilatation of both intrahepatic ducts. At surgery, gallbladder carcinoma with hepatic and bile duct invasion were confirmed.

4 Preoperative Evaluation of Gallbladder Carcinoma 679 Figure 1 (Continued) cular invasion. Bile duct assessment was made up to the second-order branches because the peripheral branches beyond that were usually not well imaged on MRC. A dilated bile duct, focal or eccentric luminal narrowing, luminal irregularity, or abrupt cutoff of the bile duct were considered positive for bile duct invasion. Pericholecystic fat infiltration and loss of interface between the gallbladder and an adjacent organ was considered positive for invasion into the adjacent organ on MR imaging. We compared the MR findings with the surgical and histopathologic results. The Fisher exact test was used to compare the MR findings with surgical and pathologic results.

5 680 Kim et al. Figure 1 (Continued) RESULTS Twelve patients underwent curative resection and the remaining six underwent palliative surgery due to peritoneal seeding including omental infiltration (three), extensive lymph node metastasis (two), and bilateral extensive intrahepatic spread (one). Curative surgical procedures included cholecystectomy (three); cholecystectomy plus wedge resection of the liver (five); and cholecystectomy, extended right lobectomy, and resection of the suprapancreatic segment of the extrahepatic bile duct and en bloc dissection of the regional lymph nodes (four). Palliative surgical procedures included gastrojejunostomy (four) and partial excision of the tumor (two). The most common appearance of gallbladder carcinoma on MR imaging was a fungating mass protruding into the gallbladder lumen; this was seen in 13 patients (72%). Three patients (17%) showed a mass replacing the gallbladder and two patients (11%) showed diffuse wall thickening of the gallbladder. Extension of the gallbladder carcinoma to the bile duct was demonstrated in nine patients, and extrahepatic bile duct carcinoma concomitant with gallbladder carcinoma was seen in one patient. MRC diagnosed bile duct involvement in all 10 patients (Fig. 1). However, MRC demonstrated false-positive findings (right main bile duct invasion) in one patient. Extension of the gallbladder cancer to the portal vein was seen in three patients, and MRA diagnosed portal vein invasion in all of these patients (Fig. 2). However, MRA demonstrated false-positive findings (right portal vein invasion) in one patient. Extension of the gallbladder cancer to the hepatic artery was not noted in our study group. Direct tumor spread into the hepatic parenchyma was histologically proven in nine patients. MR imaging identified direct spread to the liver in six of nine patients (67%; Fig. 3). MR imaging demonstrated falsepositive findings in one patient and failed to demonstrate direct spread to the liver in three. Direct extension to the duodenum was proven histologically in four patients and was correctly demonstrated in three patients on MR imaging. Surgical and histopathologic findings revealed lymph node metastases in 17 patients. MR imaging diagnosed lymph node metastases in 10 patients but failed to demonstrate metastatic nodes in seven. Peritoneal seeding with omental infiltration was found in three patients at surgery; however, MR imaging failed to diagnose peritoneal seeding in all three. Overall, with the combined use of MR imaging, MRC, and contrast-enhanced dual-phase three-dimensional MRA, the sensitivity and specificity of MR examination were, respectively, 100% and 89% for bile duct invasion, 100% and 87% for vascular invasion, 67% and 89% for hepatic invasion, and 56% and 89% for lymph node metastasis. DISCUSSION The preoperative assessment of the extent of gallbladder carcinoma is important in determining the management plan. Although it has been controversial whether or not radical surgery can improve the patient outcome in advanced gallbladder carcinoma, recent reports have suggested that curative resection can improve the prognosis (5 10). Because surgical tumor resection is re-

6 Preoperative Evaluation of Gallbladder Carcinoma 681 Figure 2. Contrast-enhanced dual-phase three-dimensional MRA of gallbladder carcinoma with portal vein invasion. a: The normal-appearing hepatic artery is seen on the arterial phase. b: Venous phase shows focal or concentric luminal narrowing in the proximal portion of the portal vein (arrows). At surgery, cancer was seen to invade the proximal portion of the portal vein. garded as the only curative treatment method, it is important to determine the degree of tumor resectability by accurate staging of the tumor extent in the bile ducts, vascular invasion, and metastasis. Several recent studies describe the usefulness of MR imaging in the evaluation of gallbladder carcinoma. Yoshimitsu et al (17) and Demachi et al (18) reported that MR imaging is a sensitive modality for detection and

7 682 Kim et al. Figure 3. MR imaging of gallbladder carcinoma with hepatic invasion. a and b: b is 2.4 cm caudad to a. Gd-enhanced breath-hold FLASH T1-weighted image shows an enhanced mass in the gallbladder (arrow heads) with direct spread to the hepatic parenchyma (arrows). c: Breathing-averaged turbo SE T2-weighted image show a mass in the gallbladder (arrow heads) with pericholecystic fat infiltration and loss of low signal intensity in the wall of the gallbladder (arrows). An incidental gall bladder stone is noted (asterisk).

8 Preoperative Evaluation of Gallbladder Carcinoma 683 Figure 3 (Continued) evaluation of the primary lesion. Sagoh et al (19) reported the usefulness of MR imaging for determining the extension of gallbladder carcinoma, but not enough consideration has been given to the local tumor spread into adjacent organs, major visceral vessels, and the bile ducts. Recent progress in MR imaging techniques, including MRC and MRA, have improved the ability of MR imaging to evaluate the bile ducts and vascular structures. To use MR examination to evaluate preoperative patients with gallbladder carcinoma, MR imaging, MRA, and MRC are all necessary in order to assess the tumor extent in the bile ducts, vascular invasion, and metastasis. At our hospital, we use an all-in-one MR protocol including MR imaging, MRC, and MRA for the preoperative staging of gallbladder carcinoma. The results of our study show that the combined use of MR imaging, MRC, and MRA is capable of evaluating gallbladder carcinoma preoperatively in order to assess adjacent organ invasion, bile duct invasion, and vascular invasion with an acceptable sensitivity, specificity, and diagnostic accuracy. These same factors in the combined MR examination for the diagnosis of hepatic invasion were 67%, 89%, and 77%, respectively. The sensitivity, specificity, and diagnostic accuracy for bile duct invasion were 100%, 89%, and 94%, respectively. These factors for hepatic vascular invasion were 100%, 87%, and 89%, respectively. Tumor extension into the bile duct is one of the important factors determining resectability. The extent of the tumor in the bile duct has been evaluated with direct cholangiography. MRC is considered to be limited for detailed evaluation of the tumor extent in the bile duct because of the limitation of spatial resolution (23). However, recent MRC techniques with HASTE or RARE sequences can produce excellent cholangiographic imaging comparable to direct cholangiography. In addition, MRC has advantages over direct cholangiography including noninvasiveness and visualization of the isolated bile ducts. In our study, MRC accurately detected bile duct invasion in all study patients. However, MRC demonstrated a false-positive result in one patient with mild luminal narrowing of the right hepatic duct on MRC. Before attempting curative resection, preoperative detection of the vascular involvement is crucial. Conventional angiography is still performed in many centers; however, being an invasive procedure, it has a relatively low sensitivity (66% 70%) and only a 54% accuracy in predicting resectability (24). Recently, contrast-enhanced dual-phase three-dimensional MRA has been used for the preoperative evaluation of hepatic arterial and venous anatomies. Lee et al (25) reported that the overall accuracy for assessing the vascular involvement of biliary or pancreatic malignancy was 98.0% for arteries and 93.9% for veins. In our study, the diagnostic accuracy of hepatic vascular invasion of contrast-enhanced dual-phase three-dimensional MRA was 89%. MRA can also provide valuable information regarding the relationship between gallbladder carcinoma and adjacent vessels, rather than only the intraluminal information provided by conventional angiography. Because of the locally invasive nature of gallbladder cancer, it is important to visualize the extraluminal tumor mass. MR imaging has limitations for the evaluation of lymph node metastasis and peritoneal seeding. In our

9 684 Kim et al. study, the sensitivity, specificity, and diagnostic accuracy of lymph node metastasis were 56%, 89%, and 50%, respectively. None of the six patients with peritoneal seeding (three), extensive lymph nodes metastasis (two), and bilateral extensive intrahepatic spread (one) was correctly diagnosed on MR imaging. One of the limitations of this study is that we set surgical findings as a gold standard in determining bile duct invasion, vascular invasion, hepatic invasion or metastasis, lymph node metastasis, and invasion into adjacent organs. Therefore, there might be a selection bias because the results of MR examination influenced the decision to perform surgery. Next, our method of image analysis performed by consensus of two radiologists has a limitation for evaluation of such subtle findings. We believe that the one-stop MR examination including MR imaging, MRC, and MRA can provide the comprehensive information required for the diagnosis and assessment of resectability in gallbladder carcinoma as an all-in-one examination, which otherwise can be obtained only by performing three different exams. In the future, using a single procedure to obtain comprehensive information about the primary lesion, vessels, and biliary tree may reduce expenses and the length of hospital stay, with the additional attractive feature of being completely noninvasive. In conclusion, the all-in-one MR protocol, including MR imaging, MRC, and MRA, can become an effective diagnostic technique in the preoperative work-up of gallbladder carcinoma for both identifying the primary lesion and the extent of disease. REFERENCES 1. Piehler JM, Crichlow RW. Primary carcinoma of the gallbladder. Surg Gynecol Obstet 1978;147: Perpetuo MD, Valdivieso M, Heilbrun LK, et al. Natural history study of gallbladder cancer: a review of 36 years experience at M. D. Anderson Hospital and Tumor Institute. Cancer 1978;42: Nevin JE, Moran TJ, Kay S, et al. Carcinoma of the gallbladder: staging, treatment, and prognosis. Cancer 1976;37: Roberts JW, Daugherty SF. Primary carcinoma of the gallbladder. Surg Clin North Am 1986;66: Nakamura S, Sakaguchi S, Suzuki S, et al. Aggressive surgery for carcinoma of the gallbladder. Surgery 1989;106: Miyazaki M, Itoh H, Ambiru S, et al. Radical surgery for advanced gallbladder carcinoma. Br J Surg 1996;83: Todoroki T, Kawamoto T, Takahashi H, et al. Treatment of gallbladder cancer by radical resection. Br J Surg 1999;86: Donahue JH, Nagorney DM, Grant CS, et al. Carcinoma of the gallbladder. Arch Surg 1990;125: Chijiiwa K, Tanaka M. Carcinoma of the gallbladder: an appraisal of surgical resection. Surgery 1994;115: Nimura Y, Hayakawa N, Kamiya J, et al. Combined portal vein and liver resection for carcinoma of the biliary tract. Br J Surg 1991; 78: Franquet T, Montes M, Azua RD, et al. Primary gallbladder carcinoma: imaging findings in 50 patients with pathologic correlation. Gastrointest Radiol 1991;16: Ohtani T, Shirai Y, Tsukada K, et al. Spread of gallbladder carcinoma: CT evaluation with pathologic correlation. Abdom Imaging 1996;21: Collier NA, Carr D, Hemingway A, et al. Preoperative diagnosis and its effect of the treatment of carcinoma of the gallbladder. Surg Gynecol Obstet 1984;159: Yeh HC. Ultrasonography and computed tomography of carcinoma of the gallbladder. Radiology 1979;133: Itai Y, Araki T, Yoshikawa K, et al. Computed tomography of gallbladder carcinoma. Radiology 1980;137: Ito K, Awaya H, Mitchell DG, et al. Gallbladder disease: appearance of associated transient increased attenuation in the liver at biphasic, contrast-enhanced dynamic CT. Radiology 1997;204: Yoshimitsu K, Honda H, Kaneko K, et al. Dynamic MR imaging of the gallbladder lesions: differentiation of benign from malignant. J Magn Reson Imaging 1997;7: Demachi H, Matsui O, Hoshiba K, et al. Dynamic MR imaging using a surface coil in chronic cholecystitis and gallbladder carcinoma: radiologic and histopathologic correlation. J Comput Assist Tomogr 1997;21: Sagoh T, Itoh K, Togashi K, et al. Gallbladder carcinoma: evaluation with MR imaging. Radiology 1990;174: Roobolamini SA, Tebrani NS, Razavi MK, et al. Imaging of gallbladder carcinoma. Radiographics 1994;14: Rossmann MD, Friedman AC, Radecki PD, et al. MR imaging of gallbladder carcinoma. AJR Am J Roentgenol 1987;148: Earls JP, Rofsky NM, Decorato DR, et al. Breath-hold single-dose gadolinium-enhanced three-dimensional MR aortography: usefulness of a timing examination and MR power injector. Radiology 1996;201: Choi BI, Kim TK, Han JK. MR imaging of clonorchiasis and cholangiocarcinoma. J Magn Reson Imaging 1998;8: Stain SC, Baer HU, Dennison AR, et al. Current management of hilar cholangiocarcinoma. Surg Gynecol Obstet 1992;175: Lee MG, Jeong YK, Sung KB, et al. Breath-hold contrast-enhanced 3D MR angiography in 19 seconds: value in the assessment of vascular invasion in pancreaticobiliary diseases. In: Proceedings of the 5th Annual Meeting of ISMRM, Vancouver, Canada, 1997.