Abdom Imaging 21:195–201 (1996)
Abdominal Imaging q Springer-Verlag New York Inc. 1996
Spread of gallbladder carcinoma: CT evaluation with pathologic correlation T. Ohtani, Y. Shirai, K. Tsukada, T. Muto, K. Hatakeyama Department of Surgery, Niigata University School of Medicine, 1-757 Asahimachi-dori, Niigata 951, Japan Received: 5 July 1995/Accepted: 8 August 1995
Abstract Background: To assess the accuracy of computed tomographic (CT) imaging in the detection of spread and staging of gallbladder carcinoma. Methods: CT findings of spread of gallbladder carcinoma in 59 Japanese patients who underwent radical surgery were correlated retrospectively with pathologic findings. Results: The incidence of histologically proven nodal involvement was 54% (32 patients) and the most common spread of gallbladder carcinoma. The sensitivities in CT detection of N1 and N2 nodal involvement were 36% and 47%, respectively; positive predictive values were 94% and 92%, respectively. Direct extension to the liver, extrahepatic bile duct, and gastrointestinal tract or pancreas were histologically confirmed in 24, 18, and five patients. The sensitivities in the CT detection of direct spread to the liver of less than 2 cm, more than 2 cm, the extrahepatic bile duct, and the gastrointestinal tract or pancreas were 65%, 100%, 50%, and 57%, respectively; positive predictive values were 77%, 100%, 90%, and 100%, respectively. The incidence of liver metastases and involvement of interaortocaval nodes were 7% and 16%, respectively. The sensitivities in CT detection of liver metastases and involvement of interaortocaval nodes were 75% and 21%, respectively; positive predictive values were 100% and 86%, respectively. CT could not detect direct spread to omentum and peritoneal seedings. Conclusion: For detecting the spread of gallbladder carcinoma, CT imaging has low to moderate sensitivity; however, CT imaging can help in determining resectability and in planning the treatment, especially in advanced-stage gallbladder carcinoma, because of a high positive predictive value. Correspondence to: T. Ohtani
Key words: Gallbladder, carcinoma—Spread, gallbladder carcinoma—Gallbladder carcinoma, CT studies— Computed tomography, gallbladder neoplasms.
Primary carcinoma of the gallbladder is uncommon; however, it is the most common malignancy of the biliary tract [1, 2]. Prognosis is poor, with unsatisfactory results reported after surgery [3–5]. In recent years, radical surgery has been advocated for the treatment of advanced-stage gallbladder carcinoma [6–8]. As radical procedure with en bloc dissection is the only hope for a potential cure, an accurate preoperative assessment of the extent of disease is essential. Some investigators have reported that computed tomographic (CT) imaging is useful in demonstrating the primary lesion and local spread of gallbladder carcinoma [9–12]. However, these studies lacked histologic vertification and/or comparison with the nonresected patients. In this study, we compared findings obtained by CT imaging with results of pathologic examination in patients with gallbladder carcinoma who underwent radical surgery. Our aims were to assess the incidence of the spread of gallbladder carcinoma and to evaluate preoperatively the extent of tumor spread and a staging accuracy of CT imaging.
Materials and Methods We reviewed retrospectively the records of 59 Japanese patients who were operated on for histologically proven gallbladder carcinoma between 1982 and 1993. There were 19 men and 40 women, with a mean age of 65.3 years (range Å 37–85 years). All patients underwent radical surgery with wedge resection of the liver and/or lobectomy of the liver, dissection of regional lymph nodes located in the hepatoduodenal ligament and on the posterior superior aspect of the head of the pancreas, and/or resection of the extrahepatic bile duct, and/or pancreatoduodenectomy.
sv 4v03 0006 Mp 195 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
196
T. Ohtani et al.: Spread of gallbladder carcinoma
Fig. 1. CT features of gallbladder carcinoma. A A 74-year-old woman with pT1a tumor. CT shows an intraluminal mass protruding into the gallbladder (arrow). B A 68-year-old woman with pT4 tumor. Irregularity of the gallbladder wall is demonstrated (arrow). Tumor spreads directly to the colon (arrowhead). C A 71-year-old woman with pT3 tumor. A mass that has replaced the gallbladder is demonstrated (arrows).
The TNM tumor stage classification was used in this study, and the extent of tumor spread was defined according to the TNM system [13]. CT findings were analyzed for detection of the primary tumor and the extent of tumor spread to the adjacent organs: the liver, extrahepatic bile duct, pancreas, and other parts of the gastrointestinal tract. Regional nodes were evaluated for involvement by gallbladder carcinoma. We adopted the names previously proposed for regional lymph nodes: cystic, pericholedochal, posterior superior pancreatoduodenal, retroportal, right celiac, hepatic (i.e., nodes along the proper hepatic artery and along the common hepatic artery), superior mesenteric, and interaortocaval [14, 15]. The N1 nodal station was comprised of the cystic and pericholedochal nodes. The N2 nodal station was comprised of the posterior superior pancreatoduodenal, retroportal, right celiac, hepatic (nodes along the proper hepatic artery and along the common hepatic artery), and superior mesenteric nodes. Spread to the interaortocaval nodes was classified as distant metastasis (M1) by the TNM staging system [13]. The CT criteria for nodal metastasis were a soft tissue mass with an anteroposterior dimension of 10 mm or larger and ringlike or heterogeneous contrast enhancement [15]. Distant metastasis was analyzed for the presence or absence of liver metastasis and peritoneal seedings. CT findings were reviewed by two of the authors (T.O. and K.T.). Following a retrospective review of all CT scans, results of preoperative CT findings and pathologic findings of resected specimens were correlated. The scanners used were the Hitachi CT-W1000HR (Hitachi, Tokyo, Japan), the General Electric CT/T 8800 (GE Medical Systems, Milwaukee, WI), or the Siemens Somatom DR (Siemens, Erlangen, Germany). CT imaging was performed by using contiguous axial sections that measured 4, 5, 8, or 10 mm. All contrast-enhanced scans were performed by using a 50-ml bolus followed by a 50-ml drip infusion of intravenous contrast material.
The diagnostic accuracy was calculated for each parameter (sensitivity, specificity, and predictive values) and evaluated by CT imaging. Probabilities were determined by chi-square tests.
Results Detection of Primary Tumor Tumor stage was pathologic T1a (pT1a) in six patients, pT1b in one patient, pT2 in 19 patients, pT3 in 19 patients, and pT4 in 14 patients. The most common CT findings in 40 patients (68%) with gallbladder carcinoma was an intraluminal mass that protruded into the gallbladder (Fig. 1A). Ten patients (17%) showed an asymmetric thickening and/or irregularity of the gallbladder wall (Fig. 1B). Four patients (7%) presented with a mass that filled or replaced the gallbladder (Fig. 1C). The relation between CT findings of the primary tumor and pathologic observations is shown in Table 1. CT failed to detect a gallbladder abnormality in five of 59 patients (8%). Two of these patients had gallstones; the diagnosis of cancer was made intraoperatively in one patient and was made postoperatively in the other. Three patients, pT3 in two and pT4 in one,
sv 4v03 0006 Mp 196 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
T. Ohtani et al.: Spread of gallbladder carcinoma
197
Table 1. Correlation between CT features and T categories T categories* No. of patients
pT1
pT2
pT3
pT4
An intralumiral mass protruding into the gallbladder
40
5
13
13
9
Asymmetric thickening and/or irregularity of the gallbladder wall
CT features
10
2
4
2
2
Mass filling replacing the gallbladder
4
0
0
2
2
Tumor not visualized
5
0
2
2
1
* Refer to the TNM staging system [13]
showed a soft tissue mass in the extrahepatic bile duct due to intraductal spread of the primary tumor. Accuracy of Detecting Direct Spread Direct spread to the hepatic parenchyma was histologically proven in 24 patients. Of these, 20 patients had a tumor extension to the liver of less than 2 cm, and four showed an extension of more than 2 cm. CT imaging identified a direct spread to the liver in 17 of 24 patients (71%) on CT imaging (Fig. 2). CT imaging demonstrated false-positive findings in four patients and failed to demonstrate direct spread to the liver in seven. For detecting direct spread to the liver, CT showed a sensitivity of 71%, a specificity of 89%, a positive predictive value of 81%, and a negative predictive value of 82%. Tumor extension into the hepatic parenchyma of less than 2 cm was correctly predicted in 13 of 20 patients (65%). Tumor extension of the hepatic parenchyma of more than 2 cm was correctly predicted in all four patients (100%). To detect direct spread to the hepatic parenchyma, no statistically significant difference between groups with liver infiltration of less than 2 cm and more than 2 cm in either sensitivity and specificity was noted. Dilated biliary ducts were demonstrated on CT imaging in 16 patients. Extension of the gallbladder carcinoma to the extrahepatic bile duct was histologically proven in 13 patients, and extrahepatic bile duct carcinoma concomitant with gallbladder carcinoma was proven in three. An additional five patients without dilated biliary ducts on CT imaging showed tumor spread to the extrahepatic bile duct that was not detected by CT imaging. Thus, nine of 18 patients (50%) with histologically proven extension to the extrahepatic bile duct were correctly diagnosed on CT imaging (Fig. 3). A false-positive finding on CT imaging was determined in another patient. For detecting spread to the extrahepatic bile duct, a statistically significant difference in
sensitivity was noted between groups with presence or absence of dilated biliary ducts (p õ 0.01). Direct extension to the gastrointestinal tract or to the pancreas was proven histologically in seven lesions of five patients: three lesions were a direct extension to the duodenum, two to the transverse colon, one to the stomach, and one to the pancreas. A direct extension to the gastrointestinal tract or to the pancreas was correctly demonstrated on CT imaging in four (57%) of seven lesions (Figs. 1B, 2). Five patients had a histologically proven direct spread to the omentum, none of which were detected on CT imaging. Table 2 compares CT findings and pathologic observations of spread of the primary tumor. Accuracy of Detecting Nodal Spread Thirty-two of 59 patients (54%) had a total of 170 histologically proven positive nodes. Twenty-seven patients had involvement of the N1 nodal station, and 14 (52%) had involvement of the N2 nodal station. Eighteen patients had involvement of the N2 nodal station, and 14 (78%) had involvement of N1 and N2 nodal stations. Of the patients who were positive for one of the N2 nodal stations, 16 of 18 patients (89%) showed involvement of two or more sites at the N2 nodal station. Table 3 summarizes the CT detection of positive nodes correlated with pathologic findings. Of the 170 histologically proven positive nodes, 67 (40%) were correctly categorized as positive on CT imaging according to our criteria (Fig. 4). Seventeen of the 47 histologically proven positive nodes in the N1 nodal station were correctly categorized as positive, but false-positive findings were determined at CT for one node and false-negative findings for 30 nodes. On CT imaging, 48 nodes in the N2 nodal station were categorized as positive according to our criteria, 44 of which were correctly diagnosed as positive and four of which were incorrectly categorized as negative. According to our criteria, the sensitivity of CT imaging for detecting positive nodes was 36% for the N1 nodal station and 47% for the N2 nodal station, with a specificity of 99% and 99% respectively. For detecting nodal involvement, no statistically significant difference was noted between N1 and N2 nodal stations in either sensitivity and specificity. Accuracy of Detecting Distant Metastasis Thirty-seven patients underwent dissection of the interaortocaval nodes in addition to the usual dissection of nodes in the hepatoduodenal ligament and on the posterosuperior aspect of the head of the pancreas. Six of these patients (16%) had involvement of interaortocaval nodes. Five of the six patients also had involvement of
sv 4v03 0006 Mp 197 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
198
T. Ohtani et al.: Spread of gallbladder carcinoma
Fig. 2. A 79-year-old woman with histologically proven direct spread to the hepatic parenchyma and duodenum. CT shows an intraluminal mass of the gallbladder extending into the hepatic parenchyma (arrows) and demonstrates the contrast-enhanced lateral wall of the duodenum (arrowhead). Fig. 3. A 68-year-old woman with histologically proven direct spread to the extrahepatic bile duct. CT demonstrates contrast-enhanced ex-
both N1 and N2 nodal stations; one patient had no other nodal involvement. These six patients had 29 histologically proven positive nodes. Six of the 29 histologically proven positive nodes were correctly categorized as positive. A false-positive finding was determined at CT for one node, and metastasis was not depicted in the other 23 nodes. Liver metastases were predicted by CT in three of four patients (Fig. 5). CT failed to detect a small liver metastasis in one patient. One patient had peritonitis carcinomatosis; the tumor was smaller than 5 mm on the surface of liver, which was not demonstrated on CT imaging. Discussion The prognosis of gallbladder carcinoma is still poor, and surgical resection represents the only potentially curative
trahepatic bile duct that has been infiltrated with a low-density mass (arrow). Fig. 4. A 70-year-old woman with nodal involvement. CT depicts the right celiac node and interaortocaval nodes showing heterogenous contrast enhancement (arrows). Fig. 5. A 62-year-old woman with liver metastasis. CT shows a lowdensity mass near the gallbladder bed (arrow).
treatment [6, 16]. Despite advances in diagnostic modalities, nearly all patients diagnosed with gallbladder carcinoma have advanced disease [3–5, 16]. Gallbladder carcinoma commonly spreads to the liver, the extrahepatic bile duct, or lymphatic vessels by the time surgical exploration is done. Surgical management varies according to the extent of the primary tumor. Thus, it is important to stage this disease preoperatively. There has been relatively little attention to the staging of gallbladder carcinoma; more attention has been given to diagnosis. We conducted the present retrospective study to assess the incidence of the spread of gallbladder carcinoma and evaluate the accuracy of CT imaging in detecting preoperative spread and staging in patients with gallbladder carcinoma who had undergone radical surgery. The diagnosis of gallbladder carcinoma is still difficult, regardless of the recent development of imaging modalities [17, 18]. Shirai et al. reported a preoperative
sv 4v03 0006 Mp 198 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
T. Ohtani et al.: Spread of gallbladder carcinoma
199
Table 2. CT detection of spread of gallbladder carcinoma Predictive value Modes of extension Direct spread Liver 0 õ distance õ 20 mm Distance ú 20 mm Bile duct Gastrointestinal tract or pancreas Nodal involvement N1 nodal station N2 nodal station Distant metastasis Liver Interaortocaval nodes
% Sensitivity
% Specificity
% Positive
% Negative
% Accuracy
65 100 50 57
90 100 98 100
77 100 90 100
83 100 82 95
81 100 78 95
36 47
99 99
94 92
82 92
83 92
75 21
100 99
100 86
98 92
98 92
Table 3. CT detection of histologically proven positive nodes CT findings Pathologic findingsa
True positive
False positive
False negative
Cystic (18) Pericholedochal (29) Posteriorsuperior (18) Pancreatoduodenal Retroportal (27) Right celiac (10) Hepatic nodes Along proper hepatic artery (19) Along common artery (13) Superior mesenteric (7) Interaortocaval nodes (29)
5 12 10
1 0 1
12 17 8
12 9
1 1
15 1
7 4 2 6
1 0 0 1
12 9 5 23
a
Numbers in parentheses are numbers of positive nodes
success rate of diagnosis of 42% in 241 patients with gallbladder carcinoma in their multiinstitutional study [18]. In this retrospective study, CT failed to detect gallbladder abnormality in five of 59 patients (8%). The five patients not diagnosed had pT2 (two patients), pT3 (two patients), and pT4 (one patient). Thus, failure of CT to detect gallbladder carcinoma does not necessarily mean the presence of ‘‘early carcinoma.’’ The pT1 tumors were observed as an intraluminal mass in five patients and as gallbladder wall thickening in two on CT imaging. It is difficult to identify an early carcinoma of the gallbladder according to CT features of the primary tumor. The most common spread of gallbladder carcinoma has been reported as direct invasion of the liver [17, 19, 20]. The prevalence of a direct spread to the liver was seen in 24 of 59 patients (41%) and was the second most common in this study. Tumor extension of more than 2 cm into the hepatic parenchyma was correctly predicted in all four of our cases and direct extension of less than 2 cm in 13 of 20 patients (65%). False-positive results
were found in four patients when the criterion for direct spread to the liver was as an unclear demarcation between the gallbladder wall and the hepatic parenchyma. To demonstrate minimal infiltration into hepatic parenchyma, one should use a narrow collimation of section thickness because a thin slice section may avoid the partial volume phenomenon. As this study is retrospective, lack of uniformity in imaging may hinder the ability to detect a spread into the liver. The use of a single type of CT scanner and the acquisition of images of uniform section thickness may demonstrate spread to the liver more effectively than was observed in this study. The incidence of a direct spread to the extrahepatic bile duct was 31% (18 of 59 patients), the third most common spread of gallbladder carcinoma in this study. Either a direct invasion of the extrahepatic bile duct or lymphatic spread of suprapancreatic portion can lead to obstruction of the extrahepatic bile duct [21, 22]. In addition, because extrahepatic bile duct carcinoma is sometimes accompanied by gallbladder carcinoma, a specific diagnosis of gallbladder carcinoma that in-
sv 4v03 0006 Mp 199 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
200
T. Ohtani et al.: Spread of gallbladder carcinoma
volves the extrahepatic bile duct or of gallbladder carcinoma concomitant with extrahepatic bile duct carcinoma should be made [23]. In this study, three of 16 patients (19%) who demonstrated a dilated biliary duct on CT imaging had an extrahepatic bile duct carcinoma concomitant with gallbladder carcinoma. Careful examination of the gallbladder wall is necessary in jaundiced patients who demonstrate masses in the extrahepatic bile duct or around the suprapancreatic part of the bile duct. In this study, five patients without dilated biliary ducts on CT imaging had disease spread to the extrahepatic bile duct that was not detected by CT. Modalities such as cholangiography may help to detect a minimal invasion of the extrahepatic bile duct in patients with gallbladder carcinoma. Advanced-stage gallbladder carcinoma eventually spreads to the gastrointestinal tract and pancreas. It is important to establish the presence of direct spread to these organs because it may help in determining the treatment, including the combined resection of these organs. It has been reported that CT is not sensitive for detecting direct spread to these organs [12]. In our series, a direct extension to the gastrointestinal tract or the pancreas was correctly demonstrated in four of seven lesions (57%) by CT imaging. However, a direct spread to the omentum could not be demonstrated on CT imaging. The incidence of nodal involvement was 54% (32 of 59 patients) and was the most common spread of gallbladder carcinoma in this study. We previously reported that patients with nodal metastases who underwent radical surgery had a 5-year survival rate of 45%, indicating that an acceptable rate of long-term survival can be achieved, even with nodal metastases [6]. Therefore, the accurate assessment of regional nodal spread is essential before the operation. To identify positive nodes on CT imaging, one should be familiar with the pattern of nodal spread in gallbladder carcinoma and the criteria for positive nodes on CT imaging. Fahim et al. suggested that mural lymphatics of the gallbladder drain directly or via the cystic nodes or pericholedochal nodes to the superior pancreatoduodenal nodes [20]. We documented in a dye-injection study that mural lymphatics of the gallbladder descend around the bile duct, flow into the lymph nodes posterior to the pancreas, the portal vein, and the common hepatic artery, and finally draw into the interaortocaval nodes adjacent to the left renal vein [14]. It is generally accepted that lymph nodes 10 mm or more in diameter on CT imaging are abnormal. In this study, the CT criteria for nodal metastasis was used previously as a soft tissue mass with an anteroposterior dimension of 10 mm or more and with a ringlike or heterogeneous contrast enhancement [15]. Based on this criteria, the sensitivity for detecting positive nodes by CT imaging was 36% for the N1 nodal station and 47% for the N2 nodal station; the specificity
was 99% and 99%, respectively. The diagnostic accuracy for the N1 nodal station was lower than that for the N2 nodal station. The difficulty in detecting the N1 nodal station is that it might be confused with the primary lesion. However, structures around the N2 nodal station including the portal vein, the hepatic artery, the gastroduodenal artery, and the posterior–superior pancreatoduodenal artery were useful landmarks for detecting the N2 nodal station. When one of the sites at N2 nodal station is positive for metastatic disease, clinicians should be aware that 89% of patients show involvement of two or more sites at the N2 nodal station, and 78% have involvement of the N1 nodal station. Interaortocaval nodes are classified as distant metastases (M1 disease) in the classification system of the American Joint Committee on Cancer Staging [13]. In 37 patients who underwent dissection of these nodes, we found six (16%) with involvement of the interaortocaval nodes. We suggest that interaortocaval nodes should be classified as the final regional nodes of the gallbladder from results of our dye-injection study [14]. CT imaging is widely accepted as a useful for detecting liver tumors. The prognosis of gallbladder carcinoma with hepatic metastases is grave. Nevertheless, hepatic resection can be performed for selected patients with an adequate cancer-free surgical margin [24]. We examined the pattern of hepatic metastases and concluded that a hepatectomy with cancer-free margin with a radical lymphadenectomy can benefit selected patients with hepatic metastases [24]. Thus, it is critical to diagnose the extent of hepatic metastases to assess the prognosis and to plan surgical treatment whether palliative, potentially curative, or unresectable. Peritoneal seeding of the primary tumor is classified as a distant metastasis (M1 disease) according to the TNM staging system [13]. Although Engels et al. correctly predicted the presence of peritoneal carcinomatosis by CT imaging in seven of nine patients with gallbladder carcinoma, it is generally accepted that CT imaging is insensitive for detecting minimal peritoneal seeding [25]. In our evaluation of 59 patients who had undergone radical surgery, peritoneal seeding was encountered in only one patient and was not demonstrated on CT imaging. Follow-up CT is important in identifying peritoneal seeding in patients with gallbladder carcinoma who have undergone radical surgery [25]. In conclusion, nodal involvement is the most common spread of gallbladder carcinoma and detecting positive nodes on CT imaging is essential to assess the staging of this disease. Advanced-stage gallbladder carcinoma commonly spreads to the liver and the extrahepatic bile duct and sometimes to the gastrointestinal tract or pancreas. It is important to establish the presence of direct spread to these organs because it may help in determining the treatment. For detecting spread of gallbladder carcinoma, CT imaging has low to moderate
sv 4v03 0006 Mp 200 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
T. Ohtani et al.: Spread of gallbladder carcinoma
sensitivity. However, especially in advanced-stage gallbladder carcinoma, CT imaging can help in assessing the disease stage preoperatively and in planning the treatment, be it a palliative or potentially curative surgical resection or to identify unresectable disease, because of its highly positive predictive value.
References 1. Piehler JM, Crichlow RW. Primary carcinoma of the gallbladder. Surg Gynecol Obstet 1987;147:929–942 2. Klein JB, Finck FM. Primary carcinoma of the gallbladder: review of 28 cases. Arch Surg 1972;104:769–777 3. Moosa AR, Anagnost M, Hall AW, et al. The continuing challenge of gallbladder cancer: survey of thirty years’ experience at the university of Chicago. Am J Surg 1975;130:57–62 4. Wanebo JH, Castle WN, Fechner RE. Is carcinoma of the gallbladder a curable lesion? Ann Surg 1982;195:624–631 5. Donohue JH, Nagoney DM, Grant CS, et al. Carcinoma of the gallbladder: does radical resection improve outcome? Arch Surg 1990;125:237–241 6. Shirai Y, Yoshida K, Tsukada K, et al. Radical surgery for gallbladder carcinoma. Ann Surg 1992;216:565–568 7. Nakamura S, Sakaguchi S, Suzuki S, et al. Aggressive surgery for carcinoma of the gallbladder. Surgery 1989;106:467–473 8. Nimura Y, Hayakawa N, Kamiya J, et al. Hepatopancreatoduodenectomy for advanced carcinoma of the biliary tract. Hepatogastroenterology 1991;38:170–175 9. Itai Y, Araki T, Yoshikawa K, et al. Computed tomography of gallbladder carcinoma. Radiology 1980;137:713–718 10. Weiner SN, Koenigsberg M, Morehouse H, et al. Sonography and computed tomography in the diagnosis of carcinoma of the gallbladder. AJR 1984;142:735–739
201 11. Franquet T, Montes M, Azua R, et al. Primary gallbladder carcinoma: imaging findings in 50 patients with pathologic correlation. Gastrointest Radiol 1991;16:143–148 12. Kumar A, Aggarwal S. Carcinoma of the gallbladder: CT findings in 50 cases. Abdom Imaging 1994;19:304–308 13. American Joint Committee on Cancer. Manual for staging of cancer, 4th ed. Philadelphia: JB Lippincott, 1992 14. Shirai Y, Yoshida K, Tsukada K, et al. Identification of the regional lymphatic system of the gallbladder by vital staining. Br J Surg 1992;79:659–662 15. Ohtani T, Shirai Y, Tsukada K, et al. Carcinoma of the gallbladder: CT evaluation of lymphatic spread. Radiology 1993;189: 875–880 16. Ogura Y, Mizumoto R, Isaji S, et al. Radical operations for carcinoma of the gallbladder: present status in Japan. World J Surg 1991;15:337–343 17. Fultz PJ, Skucas J, Weiss SL. Comparative imaging of gallbladder cancer. J Clin Gastroenterol 1988;10:683–692 18. Shirai Y, Yoshida K, Tsukada K, et al. Inapparent carcinoma of the gallbladder: an appraisal of a radical second operation after simple cholecystectomy. Ann Surg 1992;215:326–331 19. Thorsen MK, Quioz F, Lawson TL, et al. Primary biliary carcinoma: CT evaluation. Radiology 1984;152:479–483 20. Fahim RB, McDonald JR, Richards JC, et al. Carcinoma of the gallbladder: a study of its modes of spread. Ann Surg 1962; 156:114–124 21. Zeman RK, Schiebler M, Clark LR, et al. The clinical and imaging spectrum of pancreaticoduodenal lymph node enlargement. AJR 1985;144:1223–1227 22. Reinman TH, Balfe DM, Weyman PJ. Suprapancreatic biliary obstruction: CT evaluation. Radiology 1987;163:49–56 23. Ohtani T, Yoshio S, Tsukada K, et al. The association between extrahepatic biliary carcinoma and the junction of the cystic duct and the biliary tree. Eur J Surg 1994;160:37–40 24. Shirai Y, Tsukada K, Ohtani T, et al. Hepatic metastases from carcinoma of the gallbladder. Cancer 1995;75:2063–2068 25. Engels JT, Balfe DM, Lee JK. Biliary carcinoma. CT evaluation of extrahepatic spread. Radiology 1989;172:35–40
sv 4v03 0006 Mp 201 Monday Apr 01 02:38 PM SV-Ab Image (v. 21, #3) 0006(1277)
