Longterm Survival after Extended Resections in Patients with Gallbladder Cancer

Transcription

1 Longterm Survival after Extended Resections in Patients with Gallbladder Cancer Anu Behari, MS, Sadiq S Sikora, MS, FACS, Gajanan D Wagholikar, MS, MCh, Ashok Kumar, MS, MCh, Rajan Saxena, MS, Vinay K Kapoor, MS, FACS, FRCS BACKGROUND: STUDY DESIGN: RESULTS: CONCLUSIONS: Surgery is the treatment of choice for gallbladder cancer, but the extent of resection and its benefits remain unclear. Survival analysis of 42 patients who underwent extended resections for gallbladder cancer was performed. Resections were labeled R0 (curative) or R1 (noncurative) based on histopathologic evaluation. Survival curves were constructed using the Kaplan-Meier method, and survival data were analyzed by univariate and multivariate analyses to identify factors associated with longterm ( 2 years) survival. R0 status was achieved in 18 patients (43%): 100%, 100%, 45%, and 0% in stages I, II, III, and IV, respectively. Patients with R0 resections had a significantly better survival than those with R1 resections (median 25.8 months versus 17.0 months; p 0.03). R0 status was achieved in only 3 of 20 patients (15%) with node positive (N1) disease compared with 14 of 17 patients (82%) with node negative (N0) disease. Patients with N0 disease had a significantly better survival than those with N1 disease (median not reached versus 17 months; p 0.01). None of the patients with N1 disease survived 5 years; 5-year survival for N0 patients was 58%. Adjuvant therapy did not have a significant effect on survival. In patients with gallbladder cancer, R0 status could be achieved in only 43% of patients undergoing extended resections. R0 status and N0 disease were associated with better longterm survival. ( J Am Coll Surg 2003;196: by the American College of Surgeons) Gallbladder cancer (GBC) is the most common malignancy of the biliary tract. 1 It is the most frequently encountered malignant lesion of the gastrointestinal tract and the most common cause of malignant obstructive jaundice in northern India. 2,3 A large number of patients present at an advanced stage of disease that precludes curative treatment. Recent improvements in imaging techniques allow diagnosis at an earlier stage, and this might ultimately translate into better overall survival. 4 In earlier series, longterm survival was reported only among patients with incidental GBC detected after cholecystectomy for presumed benign gallbladder disease. 5,6 In patients with locally advanced GBC, extended resections encompassing the entire locoregional extent of disease might offer notable improvements in survival with No competing interests declared. Received April 5, 2002; Accepted August 13, From the Department of Surgical Gastroenterology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India. Correspondence address: Sadiq S Sikora, MS, FACS, Department of Surgical Gastroenterology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Rae Bareli Rd, Lucknow, , India. acceptable morbidity and mortality. Recent reports, mainly from Japan, reveal an encouraging trend, with reports of longterm survivors after extended resections. 7-9 The role of adjuvant therapy is evolving and addition of radiotherapy can lead to better survival in patients with microscopic residual disease after extended resections. 10 We report our experience with extended resections in patients with GBC. METHODS From January 1989 to December 1999, 357 patients with GBC (men:women, 1:2.3) were operated on in the Department of Surgical Gastroenterology at the Sanjay Gandhi Postgraduate Institute of Medical Sciences, located in Lucknow, in northern India. Surgical resection was performed in 142 patients (40%). Of these, simple cholecystectomy was performed in 100 patients, and extended resections in 42 patients. Bypass procedures were performed in 139 patients, 92 of whom underwent biliary bypass procedures. A concomitant gastric bypass was performed in 50 of these patients, and 47 patients had a gastric bypass alone by the American College of Surgeons ISSN /03/$21.00 Published by Elsevier Science Inc. 82 PII S (02)

2 Vol. 196, No. 1, January 2003 Behari et al Gallbladder Cancer 83 Records of the 42 patients who underwent extended resections were analyzed to study morbidity, mortality, and longterm outcomes after these procedures. There were 14 men and 28 women, aged 30 to 70 years (median 51 years). The most common presenting symptom was pain, which was present in 35 patients (82%). Jaundice was present in seven patients (17%). The cause of jaundice was concurrent common bile duct (CBD) stones (n 4), gallbladder neck mass with CBD obstruction (n 2), and CBD obstruction from an extrinsic compression by enlarged lymph nodes (n 1). Gallstones associated with GBC were present in 27 patients (64%). Although 24 of 28 female patients (86%) had associated gallstone disease, only 3 of 14 male patients (21%) had gallstones. Concurrent CBD stones were present in seven patients (all women). All patients had a preoperative abdominal ultrasonographic (US) examination, and in 29 patients (69%), a diagnosis of GBC was considered because of the presence of a mass lesion in the gallbladder (n 28), with or without increased wall thickness, or focal increase in wall thickness alone (n 1). In nine patients, the diagnosis of GBC was made intraoperatively; the preoperative ultrasonographic diagnosis in these patients had been porcelain gallbladder, acute cholecystitis, and chronic cholecystitis with thick-walled gallbladder in one patient each and gallstone disease in the remaining six patients. In two patients, the diagnosis of GBC was suspected when the gallbladder was opened after laparoscopic cholecystectomy and confirmed on histopathology of the specimen. In one patient, partial cholecystectomy and CBD exploration had been done at another hospital and the patient was referred to us for further management when the histopathologic report of GBC was obtained. Ultrasonographically guided fine-needle aspiration cytology from a suspicious lesion was done preoperatively in 22 patients (52%). A positive diagnosis of malignancy was made in 17 of these 22 patients (77%) on cytologic examination. Procedures Extended cholecystectomy consisting of en bloc cholecystectomy with a 2-cm wedge of liver and lymph node dissection was performed in 19 of 42 patients (45%). Lymph node clearance included the lymph nodes in the hepatoduodenal ligament (cystic, hilar, and pericholedochal) and the retroportal nodes, nodes along the hepatic artery, and behind the head of the pancreas. Celiac and superior mesenteric nodes were not removed routinely. Extended cholecystectomy with excision of the bile duct was done in another six patients (14%), and two patients (5%) underwent a right hepatectomy. Other variations of extended resections (including cholecystectomy with lymph node clearance but without a wedge resection of liver and cholecystectomy with lymph node clearance with adjacent organ resection) were performed in 15 patients (36%). Four of these patients also had a CBD excision. Among the 10 patients who underwent a CBD excision, CBD was excised because of a gallbladder neck or cystic duct mass (n 4), suspicion of CBD involvement (n 2), presence of large lymph nodes in the hepatoduodenal ligament (n 1), and inadvertent injury to the bile duct during lymph node clearance (n 1). In two patients, the CBD was dilated and full of stones. In 39 patients, the extended resection was done as the first procedure. In two patients in whom the diagnosis was first suspected on examination of the gallbladder after laparoscopic cholecystectomy, wedge resection of liver and lymph node clearance along with port site excision was done a week later, after histopathologic confirmation. In one patient, completion cholecystectomy with wedge resection of liver and lymph node clearance was done when he was referred to us after partial cholecystectomy and CBD exploration at another hospital. Twenty-six patients (12 after R0 resection and 14 after R1 resection) received adjuvant treatment in the form of postoperative radiotherapy alone (n 10), chemotherapy alone (n 2), or radiotherapy with chemotherapy (n 13); one patient received preoperative intraarterial chemoimmunotherapy as part of a pilot project. All specimens were subjected to pathologic examination, and staging was done using the American Joint Committee on Cancer (TNM) classification. 11 Resections were considered complete (R0) if surgical resection margins were negative and excised lymph nodes were either free of tumor or the extent of node dissection included a level beyond the microscopic nodal spread. If surgical margins were positive or the extent of lymph node dissection did not include uninvolved nodes one level beyond the involved lymph nodes, the resections were labeled as being incomplete (R1). Followup was obtained by hospital visits, mailed questionnaires, or both. Survival curves were constructed by the Kaplan-Meier method. Patients who survived for more than 2 years were compared with those who died within 2 years to identify factors associated

3 84 Behari et al Gallbladder Cancer J Am Coll Surg Table 1. Stage Distribution of Patients Undergoing Extended Cholecystectomy (n 42) TNM stage n Stage I T1N0M0 2 Stage II T2N0M0 2 T2NxM0 1 Stage III T3N0M0 13 T3N1M0 14 T3NxM0 2 Stage IV T3N2M0 5 T3NxM1 1 T4N1M1 1 T4NxM1 1 with longterm survival. Univariate analysis was done using the test for proportions and calculating significance with Fischer s exact z test. Factors found to significantly affect longterm survival as independent variates were then subjected to multivariate logistic regression (Forward Wald stepwise) analysis. RESULTS In the final pathologic staging, there were 2 patients with stage I, 3 patients with stage II, 29 patients with stage III, and 8 patients with stage IV disease (Table 1). R0 resection was achieved in 18 patients (stage I [n 2], stage II [n 3], and stage III [n 13]). None of the eight patients with stage IV disease had an R0 resection. R0 Figure 1. Overall stagewise survival of patients undergoing extended resections for gallbladder cancer. Table 2. T and N Status and TNM Stage Distribution of Patients with R0/R1 Resections R0 resection R1 resection Status/stage Total T status T T T T N status N N Nx TNM stage Stage I Stage II Stage III Stage IV Total resection was achieved in all of 5 patients (100%) with T1 or T2 disease, 13 of 35 patients (37%) with T3 disease, and neither of the two patients with T4 disease. In stage III patients, R0 resection was achieved in 10 of 13 patients (77%) with N0 (no nodal disease) and in 3 of 14 patients (21%) with nodal disease (N1). R0 resection was achieved in only 15% (3 of 20) of patients with lymph node metastases, compared with 82% (14 of 17) without lymph node disease (Table 2). In three patients without nodal disease, R0 status could not be achieved because of a positive cystic duct margin, a positive right duct margin, and a positive colonic resection margin in one patient each. In the three patients with nodal metastases in whom R0 status could be achieved, the lymph nodal spread was limited to the cystic lymph node alone in two patients and cystic and pericholedochal nodes in one patient. There were 18 complications in 14 patients (33%). These included bile leak (n 6), major wound infection (n 7), and intraabdominal bleeding (n 1). Two patients (5%) required reexploration, one each for intraabdominal bleeding and bile leak. Both these patients died after reoperation for these complications. Among the surviving 40 patients, followup data were available for 38 patients (95%) (17 of 17 patients with R0 resection and 21 of 23 patients with R1 resection). Median survivals in patients with stages I II, III, and IV disease were 23 months, 21.8 months, and 11.1 months, respectively. The 5-year survival in stages I II,

4 Vol. 196, No. 1, January 2003 Behari et al Gallbladder Cancer 85 Figure 2. Survival rates in patients with node negative and node positive patients. p Figure 3. Survival in patients with R0 and R1 resections. p III, and IV was 80%, 28%, and 0%, respectively (Fig. 1). Survival of patients without nodal disease (median not reached) was significantly better than those with nodal disease (median 17 months) (p 0.01). None of the patients with nodal disease survived for 5 years, compared with a 58% 5-year survival in node negative patients (Fig. 2). Median survival of patients with R0 resection was 25.8 months compared with 17.0 months for patients with R1 resection (p 0.03). Five-year survivals of patients with R0 and R1 resection were 49% and 19%, respectively (Fig. 3). Overall median and 5-year survivals of 26 patients who received adjuvant treatment (21.8 months and 20%) were not significantly different from the 16 patients who did not receive adjuvant therapy (16.7 months and 36%) (Fig. 4). Patients surviving more than 2 years (n 10) were compared with the rest of the patients (n 28) for identifying factors associated with longterm ( 2 years) survival. On univariate analysis, male gender, absence of gallstones, absence of jaundice, N0 disease, and R0 status were found to be significantly associated with longterm survival (Table 3). On multivariate analysis only the absence of gallstones was found to be significantly associated with longterm survival (p 0.01). Patients with associated gallstones were otherwise comparable with those without stones in terms of stage of disease, nodal status, and R0 status (Table 4). DISCUSSION The overall prognosis for GBC remains poor, primarily because of the advanced stage of the disease at presentation. Several factors influencing prognosis have been identified, and guidelines for surgical therapy are emerging. The predominant locoregional pattern of spread of GBC, 12 primarily to the liver and structures in the hepatoduodenal ligament, duodenum, stomach, and colon, makes it amenable to surgical resection, provided the entire extent of the lesion can be encompassed. Nevertheless, the proximity or direct invasion of the vascular and biliary structures in the hepatoduodenal ligament (especially in tumors located at the gallbladder neck) and perineural invasion not only make resections technically hazardous, but also diminish the possibility of complete (R0) resection. Based on the site and depth of the lesion Figure 4. Survival in patients receiving postoperative adjuvant therapy and those not receiving adjuvant therapy.

5 86 Behari et al Gallbladder Cancer J Am Coll Surg Table 3. Prognostic Factors for Longterm ( 2 Years) Survival Factor n >2-Year survivors p Value Age 50 y y (NS) Gender Male 12 6 Female Jaundice Present 7 0 Absent Gallstones* Present 24 3 Absent CBD resection Done 9 1 Not done (NS) Nodal status Node negative 17 7 Node positive Resection status R R Adjuvant therapy Given 25 7 Not given (NS) Stage I II 5 2 III 25 8 IV (NS) *Absence of gallstone disease was the only significant factor on multivariate analysis for longterm survival (p 0.01). CBD, common bile duct; NS, not significant. and the extent of regional spread, curative resectional procedures can range from simple cholecystectomy to extensive resections, including hepatopancreatoduodenectomy with resection of adjacent organs. It is becoming clear that, although simple cholecystectomy suffices for T1a tumors, tumors beyond this depth require a more extended procedure. 13,14 For tumors with the hepatic-hilar mode of spread and those localized to the neck of the gallbladder, more extensive hepatic resection might be required. 15,16 Lymph node involvement has been widely reported to be an indicator of poor outcomes. Bartlett and associates 17 reported no longterm survivals in patients with lymph node metastases and considered N2 (retroportal, peripancreatic, celiac lymph nodes) involvement a contraindication for any attempts at resection, given the Table 4. Comparison of Gallbladder Cancer Patients With and Without Gallstones Gallstones present (n 24) Gallstones absent (n 14) p Value Variables Age (y) Gender Male Female Stage I II III IV Nodal status N N Resection status R R ominous prognosis. Benoist and colleagues 18 also reported no 5-year survivors among patients with nodal metastases, compared with a 43% 5-year survival in patients with node negative disease. Reports from Japan 19,20 have observed that, although patients with lymph node metastases have markedly reduced survival (compared with those without nodal disease), notable improvement in survival can be achieved in patients with spread limited to the hepatoduodenal ligament after complete resection. Chijiwa and coworkers 19 reported a 50% 5-year survival in patients with stage III disease and N1 (hepatoduodenal) lymph node involvement. No patient with N2 (retropancreatic) or more distant lymph node involvement survived more than 3 years even after more aggressive operations, including hepatopancreatoduodenectomy. In node-positive patients, Shirai and colleagues 20 reported a 69% 5-year survival rate after R0 resection, and none of the patients with R1 or R2 resections survived beyond 5 years. Shimada and associates 21 reported no 5-year survival in patients with involved celiac and superior mesenteric nodes, though some patients with N2 nodes along the hepatic artery and the retropancreatoduodenal nodes survived for 5 years. Although the American Joint Committee on Cancer classification groups the celiac and superior mesenteric nodes along with the hepatic artery and retropancreatic nodes as N2, involvement of the former is invariably associated with an ominous progno-

6 Vol. 196, No. 1, January 2003 Behari et al Gallbladder Cancer 87 sis. Considering the poor prognosis associated with the involvement of celiac and superior mesenteric nodes, it has been proposed that involvement of these nodal stations be designated as M1 disease. 22 In our experience, hepatic involvement alone, in the absence of simultaneous nodal disease, never precluded an R0 resection, but lymph node involvement usually precluded a complete resection. Only 3 of 14 patients (21%) with node positive stage III disease could have an R0 resection compared with 10 of 13 patients (77%) with N0 disease. Whether more extensive lymph node dissections in the former group of patients could have resulted in more complete resections and better survival is open for debate. Only 2 of the 16 patients with nodal disease survived more than 2 years and none survived 5 years. In contrast, 7 of 17 node negative patients survived more than 2 years, and 5-year survival was 58%. Of the 10 longterm ( 2 years) survivors in our series, 7 did not have nodal disease, and 2 patients had involvement of cystic and pericholedochal lymph nodes only. This is similar to the experience of Bartlett and colleagues, 17 who observed that none of their node positive patients survived more than 18 months. The importance of achieving a complete (R0) resection in improving survival has been widely reported. Bloechle and associates 23 observed that after 2 years, 46% of patients with an R0 resection were alive, but none of the patients with residual tumor were alive. De Aretxabala and colleagues 24 and Gall and coworkers 25 also reported a markedly longer survival in patients after curative resections, compared with patients having noncurative resections. In a recent report, Todoroki and associates 10 reported a 73% 5-year survival for patients undergoing R0 resections, compared with 15% in those with microscopic residual disease. The importance of residual tumor was underscored by the finding of marked benefit in survival in patients with microscopic versus macroscopic residual disease after resections in T4 disease. Addition of radiotherapy further improved survival in patients with microscopic residual disease. 10 In our series, patients with R0 resections had a notably longer median survival and better 5-year survival than those with R1 resections (25.8 months versus 17 months p 0.03). The majority of these patients had stage III disease. In five of seven patients with predominant neck masses, extended cholecystectomy failed to achieve R0 status, and in three of these patients, the positive resection margin was on the CBD or the right hepatic duct. Analysis of longterm survivors reiterated that node negative disease and R0 resection were favorable factors for survival. Observation of a better survival in men and those without associated gallstones provides an interesting avenue for epidemiologic research. Considering that patients without stones were otherwise comparable with those with stones, the cancers in gallbladders with and without stones might be biologically different tumors with a different behavior and prognosis. Better survival in men might be related to the influence of hormonal factors on GBC. The surgical treatment of GBC might be guided by the principle of intent to achieve a complete (R0) resection whenever feasible. Apart from general assessment of tumor extent, intraoperative lymph node sampling with imprint cytology or frozen section examination might be used as a guide to determine the likelihood of complete resection. If the tumor is confined to the gallbladder or if liver infiltration is less than 2 cm in the area of fundus or body of the gallbladder, extended cholecystectomy results in a negative liver margin. If the tumor is in the neck of the gallbladder or if liver infiltration is more than 2 cm, more extensive hepatic resections are required to achieve negative surgical margins. Again, if the retropancreatic nodes are negative, extended cholecystectomy will ensure complete tumor removal with negative surgical margins. If the nodes in the retropancreatic area are involved, extended lymph node dissection (with or without pancreaticoduodenectomy) might be indicated. Positive paraaortic, celiac, and superior mesenteric nodes contraindicate any major resection. But once the disease has spread beyond the gallbladder or to the distant lymph nodes, the outlook remains poor. 26 Author Contributions Study conception and design: Sikora, Behari Acquisition of data: Behari, Wagholikar, Kapoor Analysis and interpretation of data: Sikora, Behari, Kapoor Drafting of manuscript: Sikora, Saxena, Kumar Critical revision: Sikora, Behari, Kapoor Statistical expertise: Kapoor Supervision: Sikora, Kapoor, Saxena

7 88 Behari et al Gallbladder Cancer J Am Coll Surg REFERENCES 1. Diehl AK. Epidemiology of gallbladder cancer: a synthesis of recent data. J Natl Cancer Inst 1980;65: National Cancer Registry Programme. Biennial report New Delhi: Indian Council of Medical Research; 1992: Sikora SS, Kapoor R, Pradeep R, et al. Palliative surgical treatment of malignant obstructive jaundice. Eur J Surg Oncol 1994; 20: Ouchi K, Suzuki M, Saijo S, et al. Do recent advances in diagnosis and operative management improve the management of gallbladder carcinoma? Surgery 1993;113: Peihler JM, Crichlow RW. Primary carcinoma of the gallbladder. Surg Gynecol Obstet 1978;147: Nevin JE, Moran TJ, Kay S, King R. Carcinoma of the gallbladder: staging, treatment and prognosis. Cancer 1976;37: Nakamura S, Sakaguchi S, Suzuki S, Muro H. Aggressive surgery for carcinoma of the gallbladder. Surgery 1989;106: Nimura Y, Hayakawa N, Kamiya J, et al. Hepatopancreaticoduodenectomy for advanced carcinoma of the biliary tract. Hepatogastroenterology 1991;38: Nakamura S, Nishiyama S, Yokoi Y, et al. Hepatopancreaticoduodenectomy for advanced gallbladder carcinoma. Arch Surg 1994;129: Todoroki T, Kawamato T, Takahashi H, et al. Treatment of gallbladder cancer by radical resection. Br J Surg 1999;86: Fleming ID, Cooper JS, Henson DE, et al, eds. AJCC cancer staging manual, 5th ed. Philadelphia: Lippincott-Raven; 1997: Borema EJ. Towards an oncological resection of gallbladder cancer. Eur J Surg Oncol 1994;20: Kapoor VK. Incidental gallbladder cancer. Am J Gastroenterol 2001;96: Wagholikar GD, Behari A, Krishnani N, et al. Early gallbladder cancer. J Am Coll Surg 2002;194: Ogura Y, Tabata M, Kawarada Y, Mizumoto R. Effect of hepatic invasion on the choice of hepatic resection for advanced carcinoma of the gallbladder: histologic analysis of 32 surgical cases. World J Surg 1998;22: Yamaguchi K, Chijiwa K, Shimizi S, et al. Anatomical limit of extended cholecystectomy for gallbladder carcinoma involving the neck of the gallbladder. Int Surg 1998;83: Bartlett D, Fong Y, Fortner JG, et al. Long term results after resection for gallbladder cancer. Ann Surg 1996;224: Benoist S, Panis Y, Fagniez PL, and the French University Association of Surgical Research. Long term results after curative resection for carcinoma of the gallbladder. Am J Surg 1998;175: Chijiwa K, Tanaka M. Carcinoma of the gallbladder: an appraisal of surgical resection. Surgery 1994;115: Shirai Y, Yoshida K, Tsukuda K, et al. Radical surgery for gallbladder carcinoma: long term results. Ann Surg 1992;216: Shimada H, Endo I, Togo S, et al. The role of lymph node dissection in the treatment of gallbladder carcinoma. Cancer 1997;79: Kapoor VK, Sonawane RN, Haribhakti SP, et al. Gallbladder cancer: proposal for a modification of the TNM classification. Eur J Surg Oncol 1998;24: Bloechle C, Izbicki JR, Passlick MD, et al. Is radical surgery in locally advanced gallbladder carcinoma justified? Am J Gastroenterol 1995;90: De Aretxabala X, Roa I, Burgos L, et al. Gallbladder cancer in Chile. Cancer 1992;69: Gall FP, Kockerling F, Scheele J, et al. Radical operations for carcinoma of the gallbladder: present status in Germany. World J Surg 1991;15: Kapoor VK, Benjamin IS. Resectional surgery in gallbladder cancer. Br J Surg 1998;85: