Appraisal of surgical resection of gallbladder cancer with special reference to lymph node dissection

Langenbeck s Arch Surg (2000) 385:509 514 DOI 10.1007/s004230000163 CURRENT CONCEPTS IN CLINICAL SURGERY Hiroshi Shimada Itaru Endo Yoshiro Fujii Noriyuki Kamiya Hideki Masunari Osamu Kunihiro Kuniya Tanaka Kouichiro Misuta Shinji Togo Appraisal of surgical resection of gallbladder cancer with special reference to lymph node dissection Received: 8 June 2000 Accepted: 14 June 2000 Published online: 1 November 2000 Springer-Verlag 2000 H. Shimada ( ) I. Endo Y. Fujii N. Kamiya H. Masunari O. Kunihiro K. Tanaka K. Misuta S. Togo Department of Surgery II, Yokohama City University School of Medicine, 3-9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan e-mail: hs440312@med.yokohama-cu.ac.jp Tel.: +81-45-7872650 Fax: +81-45-7829161 Abstract Background: Radical lymph node dissection in surgery for advanced gallbladder cancer is controversial. The purpose of this study is to evaluate the role of lymph node dissection based on the clinico-pathologic results. Patients: Seventy-three patients who underwent radical surgery including systematic dissection of the N1+N2 region lymph node plus some of the para-aortic nodes were reviewed. Results: pt1 patients had no lymph node metastasis, but pt2 and pt3/pt4 patients had lymph node metastasis at a rate of 50.0% (13/26) and 83.3% (25/30), respectively. As infiltration of the hepatoduodenal ligament (Binf) became severe, the rate and extent of lymph node metastasis increased. There were four 5-year survivors with lymph node involvement. The 5-year survival rates are 77.0% in pn0 cases and 27.3% in pn1 cases (P<0.01). There was no difference in survival between pn1 and pn2 patients. However, significant differences in survival were observed between pn0/1 and pn2/3 patients when these patients were limited to Binf0/1. Examination of the recurrence pattern showed that most patients with pn0/1/2 had no regional lymph node recurrence, but there was para-aortic lymph node recurrence in patients with pn3 outside the dissected region. Significant prognostic factors influencing survival after surgery by multivariate analysis were pn2/3, pt, and residual tumor. Conclusion: Systematic lymph node dissection of N1, N2, and part of the para-aortic region improves survival in advanced gallbladder cancer patients, especially in those without either para-aortic lymph node metastases or Binf2/3. Keywords Systematic lymph node dissection Radical surgery Recurrence pattern Clinical outcome Indication Introduction Gallbladder cancer is a late symptomatic disease, and most patients are treated only at an advanced stage. Prognosis is therefore poor. The surgical management of gallbladder cancer is controversial, especially regarding the indications for radical resection in advanced cases. Recent Japanese reports suggest that extended operations combining hepatic resection with extensive lymph node dissection can improve long-term survival [1, 2, 3,4]. On the other hand, a French multicenter study found that the 5-year survival rate of patients with lymph node metastasis was 0%. This result led those authors to propose that radical resection should only be considered in the absence of regional lymph node metastasis [5]. The present authors [3] have already discussed the role of extended lymph node dissection in the treatment of advanced gallbladder cancer and have proposed indications for patients without moderate or severe infiltration of the hepatoduodenal ligament. However, it is still unclear which subset of patients can benefit from radical surgery.

510 The purpose of the present study is to analyze the results of curative resection for gallbladder cancer and to determine the most appropriate indications for extended lymph node dissection in advanced gallbladder cancer. Patients and methods During the 15 years between 1983 and 1991 at Fukui Medical School, and between 1992 and 1999 at Yokohama City University, 73 patients with gallbladder cancer underwent radical surgery. There were 18 men and 55 women, with an average age of 65.7 (range 33 84) years. According to the TNM classification of the Union internationale contre le cancer (UICC), patients were classified as follows: pt1 (n=17); pt2 (n=26); pt3/4 (n=30); pn0 (n=35); pn1 (n=11); pn2 (n=13); 14 patients had positive paraaortic lymph nodes (pn3), and all patients were M0 (Table 1). According to the Japanese Rules for Surgical and Pathological Studies [7], infiltration of the hepatoduodenal ligament (Binf) is Table 1 Profile of patients with resected gallbladder carcinoma a According to the Japanese Rules for Surgical and Pathological Studies [7] b According to the Union internationale contre le cancer (UICC) Age (years) 65.7 (33 84) Gender Male 18 Female 55 Histology b pap 24 tub1 21 tub2 12 por 11 ad-sq 2 muc 2 undiff 1 pt a 1a 10 1b 7 2 26 3 11 4 19 pn a 0 35 1 11 2 13 3 (para-aortic) 14 Stage a 1 17 2 13 3 10 4A 6 4B 27 Binf b 0 1 53 2 3 20 Hinf b 0 1 56 2 3 17 Residual tumor a R0 31 R1 28 R2 14 Table 2 Surgical procedures Procedure Cholecystectomy with or without bile duct resection 19 Liver bed resection with or without bile duct resection 23 S4a + S5 hepatectomy 19 Anterior segmentectomy 2 Posterior segmentectomy + S5 + S4a 1 Right trisegmentectomy with caudate lobectomy 2 Extended right hepatectomy with caudate lobectomy 6 Central bisegmentectomy 1 categorized as negative (Binf0), mild (Binf1), moderate (Binf2), or severe (Binf3); and hepatic infiltration through the gallbladder bed (Hinf) is categorized as negative (Hinf0), suspicious (Hinf1), mild and confined to the surrounding tissues (Hinf2), or mass forming (Hinf3). The surgical procedures included extended hepatectomy and pancreatoduodenectomy (PD), and are listed in Table 2. In the authors department, the standard radical surgical procedure for gallbladder cancer patients with pt2 and some with pt3 and pt4 ranges from cholecystectomy with hepatic wedge resection of the gallbladder bed to extended right hepatectomy, according to the extent of hepatic invasion, but includes resection of the suprapancreatic extrahepatic bile duct, sometimes with PD, and en bloc dissection of the regional lymph nodes with the para-aortic lymph nodes, as previously reported [3]. PD was indicated for patients with marked lymph node metastasis around the head of the pancreas or direct invasion of the duodenum. The names of the regional lymph nodes of the gallbladder have not been standardized internationally. Therefore, the authors [3] have used names and stages such as N1, N2, and para-aortic, as proposed by Shirai et al. [8], as a modification of the nomenclature of Fahin et al. [9]. The specimens were cut into 5-mm-thick tissue sections after fixation in formaldehyde for 7 10 days. The fixed lymph nodes were sectioned in the region of their maximum diameters. Histologic findings were also evaluated according to the Japanese Rules for Surgical and Pathological Study [7]. The postoperative survival of patients was calculated by the product-limit mode of the Kaplan-Meier method, and differences between the curves were measured using the generalized Wilcoxon test. Statistical comparisons were made by chi-squared analysis. Univariate and multivariate analyses were carried out using the Cox proportional hazards model for significant prognostic factors influencing survival. Results Lymph node involvement relative to depth of cancer invasion Number of cases Patients with pt1 disease had no lymph node metastasis. However, 50% (13/26) and 83.3% (25/30) of patients with pt2 and pt3/pt4 disease, respectively, had lymphatic metastasis (Table 3). As the depth of cancer invasion became greater, the frequency and extent of lymph node metastasis also increased.

511 Lymph node involvement relative to invasion of the hepatoduodenal ligament The proportion of lymph node involvement in patients with Binf0/1 was 35.8%. However, 95.0% of patients with Binf2/3 had lymph node metastasis, metastasis to the paraaortic lymph nodes being especially frequent (Table 4). Survival rate There were four 5-year survivors with lymph node metastasis (one pn1, two pn2, and one pn3 patients). The Table 3 Lymph node metastasis in relation to depth of invasion. Definitions according to the Union Internationale Contre le Cancer (UICC) pn0 pn1 pn2 pn3 (para-aortic) pt1a 10 0 0 0 pt1b 7 0 0 0 * pt2 13 5 4 4 pt3/pt4 5 6 9 10 pn0 vs pn1+2+3: *P<0.01; **P<0.05; ***P<0.01 ** *** Table 4 Lymph node metastasis in relation to infiltration of the hepatoduodenal ligament pn0 a pn1 pn2 pn3 (para-aortic) Binf0 1 b 34 9 6 4 Binf 2 3 1 2 7 10 pn0 vs pn1+2+3: *P<0.01 a According to the Union internationale contre le cancer (UICC) b According to the Japanese Rules for Surgical and Pathological Studies [7] * overall 3- and 5-year survival rates were 49.3% and 44.0%, respectively, based on the depth of cancer invasion; the 5-year survival rate was 85.7% in pt1 patients, 48.5% in pt2 patients (P<0.05), and 18.4% in pt3/pt4 patients (pt2 vs pt3/4: P<0.01). The 5-year survival rates were 77.0% in pn0 patients, 27.3% in pn1 (P<0.01), 10.0% in pn2, and 8.3% in pn3 patients and were 58.3% in Binf0/1, 0% in Binf2/3 patients (P<0.01), and 50.7% in Hinf0/1, and 0% in Hinf2/3 patients (P<0.01; Table 5). The survival curve of pn0/1 in the patients with Binf0/1 was significantly higher than that of pn2/3 regardless of the extent of Hinf, but this was not the case in the patients with Binf2 or 3 (Fig. 1). Pattern of recurrence after radical operation Of 15 patients without lymph node involvement, 3 died as a result of recurrence. One of these three cases (pt4 and Hinf3) showed para-aortic lymph node recurrence, while the other two patients (both pt2 and Hinf0) died of carcinomatous peritonitis and liver metastasis, respectively. Of the ten pn1 cases, seven patients died because of recurrence. Two of the seven had hepatic metastasis, one had liver and lung metastasis, two had carcinomatous peritonitis, while one had pt4, Hinf2, and Binf3; and the last one, with pt2, fell victim to para-aortic lymph node recurrence. Out of nine pn2 cases, recurrence in the para-aortic lymph nodes was seen in three. In four of the six cases of para-aortic lymph node metastasis, para-aortic lymph node metastasis was found outside the dissected region, but control of the other two cases was achieved by extensive lymph node dissection (Table 6). Table 5 Cumulative survival rates (excluding hospital deaths and death unrelated to gallbladder cancer) a Subject to the Union internationale contre le cancer (UICC) b Subject to Japanese Rules for Surgical and Pathological Studies [7] n One Three Five year years years pt a 1 15 100 86.7 86.7 pt2 26 88.0 52.9 48.5 pt3/4 28 48.9 27.6 18.4 Binf0+1 b 51 88.2 65.4 58.3 Binf2+3 18 32.7 0 0 Hinf0+1 b 54 81.1 57.4 50.7 Hinf2+3 15 50.9 21.8 0 pn0 a 33 100 77.0 77.0 pn1 11 54.5 27.3 27.3 pn2 13 59.8 39.9 10.0 pn3 13 33.3 8.3 8.3 Stage a 1 15 100 86.7 86.7 Stage 2 13 100 73.3 73.3 Stage 3 10 70.0 40.0 40.0 Stage 4A 6 66.7 22.2 0 Stage 4B 25 48.5 24.8 9.9 P<0.05 P<0.01 P<0.01 ns ns pn0+1 vs pn2+3: P<0.01

512 Fig. 1 Cumulative survival of patients with pt*2/3/4 according to lymph node metastasis (excluding hospital deaths and R*0+1 patients). *According to the Union internationale contre le cancer (UICC) ; According to the Japanese Rules for Surgical and Pathological Studies [7] Table 6 Recurrence patterns in patients with pt2/3/4 (excluding R*0+1) Local Liver Lymph Peritoneum Distant node organ Hinf0 1 a pn0 b (2/12) 1 1 pn1 (4/7) 2 1 2 pn2+3 (7/10) 1 2 5 2 1 Hinf2/3 pn0 (1/3) 1 pn1 (3/3) 2 1 1 1 pn2+3 (5/6) 2 4 2 1 a According to the Union internationale contre le cancer (UICC) b According to the Japanese Rules for Surgical and Pathological Studies [7] Table 7 Multivariate analysis of prognostic factors using Cox s proportional hazard regression model (Bm bile duct margins, Em excisional margins) Variables Grading Regression Standard Chi- P value coefficient error square pn a 1: pn0+1 0.7535 0.3814 3.9037 0.0482 2: pn2+3 Bm b 1: Bm0 0.7592 0.3086 6.0520 0.0139 2: Bm1 3: Bm2 Em b 1: Em0 0.7846 0.2409 10.6131 0.0011 2: Em1 3: Em2 pt a 1: pt2 0.5232 0.22346 5.4761 0.0193 2: pt3 3: pt4 a According to the Union Internationale Contre le Cancer (UICC) b SAccording to the Japanese Rules for Surgical and Pathological Studies [7] Prognostic factors influencing survival Univariate analysis revealed that survival was significantly related to seven factors: pt, pn (pn0/1 vs pn2/3), TNM stage, residual tumor, Binf (Binf0/1 vs Binf2/3), Hinf (Hinf0/1 vs Hinf2/3), and hepatectomy with vascular resection. Multivariate analysis revealed that of these seven factors, pn (pn0/1 vs pn2/3), residual tumors and pt residual tumor and pt were independent prognostic factors influencing survival (Table 7). Discussion In the past, most surgeons had a pessimistic view of surgical treatment for patients with advanced gallbladder cancer, believing that surgery contributes to cure only in the early stages of the disease [10,11]. However, most Japanese surgeons have become aware of the various spread patterns of advanced gallbladder cancer, which include lymphatic invasion, hepatic invasion via the gallbladder bed, bile duct invasion, and invasion to adjacent organs. Of these, lymphatic invasion and hepatic invasion via the gallbladder bed are classified as spread patterns curable by radical surgery [12, 13, 14,15]. Lymph node involvement is an important prognostic factor in gallbladder cancer. In 1962, Fahin et al. [9] advocated lymph node dissection for gallbladder cancer. Shirai et al. [16] and the present authors [3] reported several cases of long-term survival of patients with advanced gallbladder cancer and lymph node involvement of N1 and part of the N2 region in which treatment by systematic lymph node dissection of N1+N2+part of the para-aortic region was performed. By contrast, a recent French multicenter study [5] found that the 5-year survival rate for patients with lymph node metastasis was 0%. No information was provided regarding the rigorousness of the approach to the indications, the surgical procedures employed for lymph

513 node dissection, or the histologic examination of the resected lymph node specimens. However, it is well within the bounds of possibility that strictly adhering to the protocol in a multicenter study presented many difficulties, and the reason for the absence of any long-term survivor with lymph node metastasis may have been that most patients with lymph node metastasis also had Binf2/3. It is necessary to clarify whether lymph node dissection improves the survival of patients with advanced gallbladder cancer and at what stage systematic lymph node dissection is advisable or valuable. The authors [3] have already reported on the use of systematic lymph node dissection with or without PD for advanced gallbladder cancer and on the correlation of the frequency and extent of lymph node involvement with the depth of cancer invasion and with the extent of hepatoduodenal ligament invasion. This result was almost the same as that stated in the Japanese literature [13,15]. In the current series, the patient with pt1 disease had no lymph node involvement, and those with pt2 disease with invasion of the subserosal layer had a high frequency of lymph node involvement, as was the case in the series reported by Yoshikawa et al. [17]. This result is also supported by the fact that in the subserosal layer of the gallbladder wall, there are rich networks of blood and lymph vessels through which invading cancer cells can spread, but it is controversial as to whether or not gallbladder cancer that has invaded as deep as the muscular layer is actually accompanied by lymph node metastasis. Regarding infiltration of the hepatoduodenal ligament, half of the patients with moderate-to-severe infiltration of the hepatoduodenal ligament had undergone para-aortic lymph node metastasis. This result is also supported by the fact there is a rich abundance of lymph and blood vessels and of autonomic nerve fibers in the hepatoduodenal ligament. In the present study, four patients with positive lymph nodes, comprising one pn1, two pn2, and one pn3 patients, survived for more than 5 years. The 5-year survival were 77.0% for pn0 patients, 27.3% for pn1, and 10% for pn2/3 patients. These results, which were almost the same as those of the other Japanese reports, were superior to those reported in the French survey [5]. Systematic lymph node dissection is performed in the N1+N2 regions and in part of the para-aortic region, and a significant difference in survival between pn1 and pn2 patients or between pn1/2 and pn3 patients might be expected. However, there was no such significant difference in this study. When the patients were limited to those with Binf0/1, there was a significant difference in survival between pn0/1 and pn2/3 patients. The present authors [3] also proposed that lymph node dissection of N1 and the posterior pancreaticoduodenal (N2) lymph nodes and partial para-aortic lymph node is important to improve the prognosis of pt2 disease. Noie et al. [18] reported that based on the relationship between recurrence patterns and the extent of lymph node dissection in the patients undergoing radical surgery, two out of eight patients with recurrence might have been cured by extensive regional lymph node dissection, including dissection of the para-aortic lymph nodes and the nodes around the head of the pancreas. In comparison with Noie et al. s report [18], our series showed less recurrence, involving fewer lymph node sites, with no regional bias because of the wider scope of the lymph node dissection. Shirai et al. [19] has also reported that the 5-year survival of stages II IV was 29% and that N2 tumors should be considered for curative resection because prolonged survival is possible. To determine the usefulness and limitations of our extended lymph node dissection, we analyzed the most likely predictive factors for survival by univariate and multivariate analysis. N2 (not N1), para-aortic node metastasis, pt, and residual tumors influenced 5-year survival. These results led the authors to conclude that patients with regional lymph node involvement must be considered for curative resection, because long-term survival is indeed possible. However, for the pn2, pn3 patient with Binf2,3, adjuvant chemotherapy to prevent recurrence is necessary to achieve long-term survival. References 1. Tsukada K, Hatakeyama K, Kurosaki I, Uchida K, Shirai Y, Muto T, Yoshida K (1996) Outcome of radical surgery for carcinoma of the gallbladder according to TNM stage. Surgery 120:816 822 2. Bartlett DL, Fong Y, Fortner JG, Brennan MF, Blumgart LH (1996) Long-term results after resection for gallbladder cancer: implications for staging and management. Ann Surg 5:639 646 3. Shimada H, Endo I, Togo S, Nakano A, Izumi T, Nakagawara G (1997) The role of lymph node dissection in the treatment of gallbladder carcinoma. Cancer 79:892 899 4. Muralore A, Polastri R, Bouzari H, Vergara V, Capussotti L (2000) Radical surgery for gallbladder cancer: a worthwhile operation? Eur J Surg Oncol 26:160 163 5. Benoist S, Panis Y, Fagniez P-L, the French University Association for Surgical Research (1998) Long-term results after curative resection for carcinoma of the gallbladder. Am J Surg 175:118 122 6. Union internationale contre le cancer (UICC) (1987) TNM classification of malignant tumors, 4th edn. Springer, Berlin Heidelberg New York 7. Japanese Society of Biliary Surgery (1986) General rules for surgical and pathological studies on cancer of the biliary tract, 2nd edn. Kanehara, Tokyo

514 8. Shirai Y, Yoshida K, Tsukada K, Ohtani T, Muto T (1992) Identification of the regional lymphatic system of the gallbladder by vital staining. Br J Surg 79:659 662 9. Fahin RB, McDonald JR, Richard JC, Ferris DO (1962) Carcinoma of the gallbladder: a study of its mode of spread. Ann Surg 156:114 124 10. Gonzalez EM (1990) Gallbladder carcinoma: radical surgery? HPB Surg 2:295 297 11. Wanebo HJ, Castle WN, Fechner RE (1982) Is carcinoma of the gallbladder a curable lesion? Ann Surg 195:624 631 12. Nakamura S, Sakaguchi S, Suzuki S, Muro H (1989) Aggressive surgery for carcinoma of the gallbladder. Surgery 106:467 473 13. Ogura Y, Mizumoto R, Isaji S, Kusuda T, Matsuda S (1991) Radical operations for carcinoma of the gallbladder: present status in Japan. World J Surg 15:337 343 14. Jones RS (1990) Carcinoma of the gallbladder. Surg Clin North Am 70:1419 1428 15. Matsumoto Y, Fujii H, Yamamoto M (1992) Surgical treatment of primary carcinoma of the gallbladder based on the histologic analysis of 48 surgical specimens. Am J Surg 163:239 245 16. Shirai Y, Yoshida K, Tsukada K, Muto T, Watanabe H (1992) Radical surgery for gallbladder carcinoma, long-term results. Ann Surg 156:114 124 17. Yoshikawa T, Hanyu F, Nakamura M (1989) The role of pancreatoduodenectomy in the lymph node dissection for advanced gallbladder cancer [in Japanese with English abstract]. Rinsho Geka (J Clin Surg) 44:1751 1757 18. Noie T, Kubota K, Abe H, Kimura W, Harihara Y, Takayama T, Masatoshi M (1999) Proposal on the extent of lymph node dissection of gallbladder carcinoma. Hepatogastroenterology 46:2122 2127 19. Shirai Y, Ohtani T, Tsukada K, Hatakeyama K (1997) Combined pancreaticoduodenectomy and hepatectomy for patients with locally advanced gallbladder carcinoma. Long term results. Cancer 80:1904 1909