An Aggressive Surgical Approach Leads to Improved Survival in Patients With Gallbladder Cancer

FEATURE

An Aggressive Surgical Approach Leads to ImprovedSurvival in Patients With Gallbladder Cancer

A 12-Year Study at a North American Center

Elijah Dixon, MD,† Charles M. Vollmer, Jr, MD,* Ajay Sahajpal, MD,* Mark Cattral, MD,*David Grant, MD,* Christopher Doig, MD,* Al Hemming, MD,‡ Bryce Taylor, MD,*

Bernard Langer, MD,* Paul Greig, MD,* and Steven Gallinger, MD*

Objective: To determine if an aggressive surgical approach, with anincrease in R0 resections, has resulted in improved survival forpatients with gallbladder cancer.Summary Background Data: Many physicians express a relativelynihilistic approach to the treatment of gallbladder cancer; consensusamong surgeons regarding the indications for a radical surgicalapproach has not been reached.Methods: A retrospective review of all patients with gallbladdercancer admitted during the past 12 years was conducted. Ninety-ninepatients were identified. Cases treated during the 12-year period1990 to 2002 were divided into 2 time-period (TP) cohorts, thosetreated in the first 6 years (TP1, N � 35) and those treated in the last6 years (TP2, N � 64).Results: Disease stratification by stage and other demographicfeatures were similar in the 2 time periods. An operation withcurative intent was performed on 38 patients. Nine (26%) R0resections were performed in TP1 and 24 (38%) in TP2. The numberof liver resections, as well as the frequency of extrahepatic biliaryresections, was greater in TP2 (P � 0.04). In both time periods, anR0 resection was associated with improved survival (P � 0.02 TP1,P � 0.0001 TP2). Overall survival of all patients in TP2 wassignificantly greater than in TP1 (P � 0.03), with a median survivalof 9 months in TP1 and 17 months in TP2. The median 5-yearsurvival in TP1 was 7%, and 35% in TP2. The surgical mortality ratefor the entire cohort was 2%, with a 49% morbidity rate.Conclusions: A margin-negative, R0 resection leads to improvedsurvival in patients with gallbladder cancer.

(Ann Surg 2005;241: 385–394)

The surgical treatment of gallbladder cancer has tradition-ally been viewed with nihilism due to poor survival

results. In 1978, a review of nearly 6000 cases1 revealed a 5%5-year survival rate with a median survival of between 5 and8 months. Recent publications during the 1990s have shownthe same poor results, with 5% and 12% 5-year survival ratesreported from France and Australia.2,3 Factors contributing tothese dismal outcomes include the anatomic proximity of thegallbladder to the porta hepatis and the aggressive biologicnature of this cancer. Furthermore, direct extension of thetumor into the liver, the structures of the hepatoduodenalligament, as well as organs in close proximity such as theduodenum, the hepatic flexure of the colon, and the pancreas,contribute to technical challenges in achieving a marginnegative, potentially curative (R0) resection.

However, improvements in surgical and anesthetictechniques over the last 10 to 15 years have made it possibleto safely perform large, extensive liver resections with de-creased morbidity and low mortality, particularly in high-volume centers.4–6 Reports of improved outcomes from Jap-anese centers that have adopted an aggressive surgicalapproach to gallbladder cancer have led to a reappraisal ofsurgical strategy for this disease in North America. Somehepatobiliary specialty units have broadened an already ag-gressive approach that includes liver resection in combinationwith resection of the gallbladder, bile duct, and regionallymphatics by including pancreaticoduodenectomy. The ad-dition of pancreaticoduodenectomy allows complete resec-tion of the extrahepatic biliary tree and its regional lymphnode drainage to obtain an R0 resection that would nototherwise be possible.7–19 In these published reports, 78pancreaticoduodenectomies (6 in North American centers,the others Japanese) were performed to obtain negative mar-gins in cases with advanced disease and to clear the drainingnodal basin. Moreover, these series report 18 patients withstage IV disease who have actually survived 5 years after

From the *Hepatobiliary and Pancreatic Surgery, University of Toronto,Toronto, Ontario, Canada; †University of Calgary, Calgary, Alberta,Canada; and ‡University of Florida, Gainsville, Florida.

Reprints: Steven Gallinger MD, MSc, FRCSC, Professor of Surgery, 600University Avenue, Room 1225, Mount Sinai Hospital, Department ofSurgery, University of Toronto, Toronto, Ontario, Canada, M5G 2C4.E-mail: [email protected].

Copyright © 2005 by Lippincott Williams & WilkinsISSN: 0003-4932/05/24103-0385DOI: 10.1097/01.sla.0000154118.07704.ef

Annals of Surgery • Volume 241, Number 3, March 2005 385

radical resection. Whether the fact that some node-positive5-year survivors have been reported in Japan, while nonehave been identified in North America, reflects differences indisease biology between the 2 populations or simply reflectsthe higher number of resections done for advanced stagedisease in Japan is unclear.

The improved results for gallbladder cancer reportedwith extended procedures coupled with an appreciation thatradical surgical resection for hilar cholangiocarcinomas20–22

improves survival (unpublished data from our institution) ledto a reappraisal of our technical approach to gallbladdercancer in the mid-1990s. Since then, we have adopted a moreaggressive strategy that consists of liver resection and com-plete excision of the extrahepatic/pancreatic biliary tree withlymphadenectomy to achieve an R0 resection. To evaluatethe effects of this shift in management, we have divided thiscohort of 99 consecutive patients into 2 groups, comparingthe first 6 years to those in the last 6 years (seventh through12th years) of the review.

MATERIALS AND METHODSA retrospective study spanning a 12-year period from

January 1990 to May 2002 was performed, and a total of 99consecutive inpatients were identified. Cases were dividedinto 2 time-period cohorts, those treated in the first 6 years(TP1, N � 35) and those treated in the last 6 years (TP2,N � 64). Approval for chart review was obtained from theinstitutional review boards at both the Toronto General andMount Sinai Hospitals of the University of Toronto. Allinpatient admissions with the diagnosis of gallbladder cancerwere included, and each patient’s inpatient hospital and clinicchart was reviewed. Surgical mortality was considered asdeath occurring within 30 days of surgery. Morbidity in-cluded minor complications requiring no intervention, whilemajor complications required intervention.

Preoperative assessment included history, physical ex-amination, and radiographic studies: magnetic resonancecholangiopancreatography (MRCP). In TP1, patients werestaged with transabdominal ultrasound and double-contrast(IV and oral) computed tomography (CT) scan of the chest/abdomen/pelvis; jaundiced patients underwent endoscopicretrograde cholangiopancreatography (ERCP). Patients inTP2 underwent a transabdominal ultrasound, CT scan of thechest/abdomen/pelvis, and usually triphasic CT scan of theliver; jaundiced patients underwent either ERCP or MRCP.Staging laparoscopy was performed infrequently in selectedpatients in TP2; 3 patients were incidentally found to havegallbladder cancer prior to referral and underwent staginglaparoscopy at our institution. In one of these patients, peri-toneal metastases were found. Major hepatic resections wereperformed using standard techniques.6 Previous laparoscopicport sites were resected if present. In the early years of thestudy, parenchymal transection was performed using a crush

clamp technique. Later, parenchymal division was accom-plished using an ultrasonic dissector and more recently (sinceMarch 2002), a precision water dissection system (Hydro-Jet,ERBE, Teubingen, Germany) has been used. Central venouspressure was maintained at 5 cm H2O or less during theparenchymal transection phase. Complete inflow occlusionwas rarely used. Excision of the extrahepatic biliary treeincluded the gallbladder, the biliary confluence down to itscephalad junction with the pancreas in continuity, with acomplete portal lymphadenectomy to the suprapyloric lymphnode23 overlying the hepatic artery–gastroduodenal arteryjunction, skeletonizing the portal vein and hepatic artery.Biliary reconstruction, to a Roux-en-Y loop of jejunum, wasperformed using a single layer of 5–0 PDS interruptedsutures.

DefinitionsTerminology with respect to liver resections is that

proposed by the International Hepato-Pancreatico-Biliary As-sociation (IHPBA), also known as the Brisbane terminology(Terminology Committee of the IHPBA, www.ihpba.org/html/guidelines/index.html). Resections may also be de-scribed by the Couinaud segments resected.24–26 Wedgeresection denotes a nonanatomic, nonsegmental resection ofthe liver surrounding the gallbladder to a depth of 1 to 2 cm.We have elected to use the staging system proposed byMemorial Sloan-Kettering Cancer Center.7 In contradistinc-tion to the TNM (AJCC) staging system, the Memorialsystem categorizes T4N0M0 lesions as stage IIIb as opposedto stage IV (TNM).

StatisticsUnivariate comparisons were made with the Fisher

exact test or �2 test for dichotomous covariates, and the t testfor continuous variables. Patients were divided into 2 groups,depending on the time in which they received treatment(TP1 � 1990–1996, TP2 � 1997–2002). Numerical data areexpressed as the mean � standard deviation unless otherwiseindicated. Differences were considered significant at P �0.05. Patient survival was calculated using the Kaplan-Meiermethod. Comparisons of patient survival curves were madeusing the log-rank test. Cox proportional hazards regressionmodeling was used to assess the effect that independentcovariates had on the dependent variable survival. Logisticregression modeling was performed to assess the impact ofthe covariates on the presence or absence of complications.Statistical analysis was performed using SAS software (SASInstitute Inc, Cary, NC).

RESULTS

DemographicsNinety-nine patients admitted with gallbladder cancer

from 1990 to 2002 were identified. Thirty-three patients were

Dixon et al Annals of Surgery • Volume 241, Number 3, March 2005

© 2005 Lippincott Williams & Wilkins386

male and 66 female, with an average age at presentation of 64years (37–88).

Presenting Symptoms/Signs(Table 1) The most common presenting symptom was

pain in 76% of patients, and this was the only symptom thatwas significantly different between time periods (83% TP1versus 62% TP2, P � 0.03). Forty-five percent presentedwith obstructive jaundice. Overall, 22% presented with apalpable mass, and weight loss (average of 18 pounds)occurred in 32% of patients. The most common provisionaldiagnosis prior to intervention was gallbladder cancer (38%);62% of patients had the wrong diagnosis prior to intervention.These included biliary colic (24%), acute cholecystitis (12%),cholangiocarcinoma (12%), cholangitis/choledocholithiasis/gallstone pancreatitis (5%), gallbladder polyp (4%), miscel-laneous (3%), and chronic cholecystitis (1%).

Interventions Prior to Definitive TherapyTwenty-five percent of patients underwent a laparo-

scopic cholecystectomy prior to referral to our center. Onefifth of these were converted to an open procedure. Anadditional 6% of patients underwent an open cholecystec-tomy prior to referral. Of the 51 patients who were exploredwith curative intent at our institution, 50% had a biliary stent

(endobiliary or transhepatic) placed prior to surgery to relievejaundice.

Disease StageFigure 1 illustrates the distribution of patients by stage,

as well as by time period. The majority of patients (73%)presented with advanced disease (stage IIIb or IV). There wasa trend towards presentation with higher stage disease inpatients during TP2 (79% Stage IIIB or IV) that did not reachstatistical significance (P � 0.26).

Definitive TherapyTable 2 shows the operative features of the entire

cohort. Of the 99 patients treated, 51 patients were exploredfor a potentially curative resection at our institution; 38underwent an R0 resection (R0 resection, microscopic andmacroscopic negative margins), and 13 had a palliative pro-cedure or diagnostic biopsy. A further 10 patients had apalliative operation (bypass procedure) for gastrointestinalobstruction. The remaining 38 patients had advanced diseaseand did not undergo any surgery. Nine patients in TP1 and 29in TP2 had R0 resections. Thirty-five patients underwent aconcomitant liver resection (6 in TP1 versus 29 in TP2, P �0.04). The majority in TP1 underwent a nonanatomic wedgeresection of the liver surrounding the gallbladder. However,in TP2, 13 patients had a limited liver resection (wedgeresection), 8 had an anatomic resection of segments 4b and 5,and an extended right hepatectomy was performed in 8 cases.Complete resection of the extrahepatic/extrapancreatic biliarytree was performed in 3 patients in TP1 and 19 in TP2. Ofpatients undergoing R0 resections involving a concomitantliver resection, 4 of 6 were stage IIIb or IV in TP1, while 23of 28 had stage III disease or higher in TP2. One patient hada concurrent en-bloc pancreaticoduodenectomy for a positivebiliary resection margin on frozen section at the cephaladborder of the pancreas. Average overall operating time was215 minutes. Estimated average blood loss for both groupswas 566 mL.

TABLE 1. Characteristics of 99 Patients With GallbladderCancer

Variable

TimePeriod 1(N � 35)

TimePeriod 2(N � 64)

Total(P Values

ComparingTime

Periods)

Sex F 57% F 72% NSM 43% M 28%

Age 65 � 13 64 � 9 64 � 10NS

Average weight loss atpresentation (lbs.)

18 18 18 � 15

Mass at presentation 23% 22% 22%NS

Gallstones 90% 91% 90%NS

Weight loss atpresentation

35% 30% 32%

NSJaundice 50% 43% 45%

NSPain at presentation 62% 83% 76%

P � 0.03

NS, not significant.FIGURE 1. Stage of disease at presentation by time period oftreatment.

Annals of Surgery • Volume 241, Number 3, March 2005 Radical Surgery for Gallbladder Cancer

© 2005 Lippincott Williams & Wilkins 387

In total, 17 patients received chemotherapy (14 pallia-tive, 3 adjuvant), and 5 patients received radiotherapy (allpalliative). The frequency of use of chemotherapy and radio-therapy did not differ significantly between time periods.

Perioperative ComplicationsOperative mortality was 2% (1/51). The death occurred

as a result of liver failure following an extended right hepa-tectomy and biliary tree resection in a patient with unsus-pected alcoholic liver disease. The morbidity rate for allpatients was 49%. Complications included cholangitis (16%),wound infection (10%), liver failure (4%), pneumonia (4%),intraabdominal abscess (4%), delirium (2%), postoperative

bowel obstruction (2%), bile leak (2%), DVT/pulmonaryembolus (2%), upper GI bleed (2%), and renal failure (2%).Two patients required reoperation; 1 was for a persistent bileleak. The second was for a bowel obstruction following apalliative biliary and gastric bypass as a result of unrecog-nized tumor involvement of the right colon. Two otherpatients developed intraabdominal abscesses requiring percu-taneous drainage. Mean length of stay was 10.5 days. Mul-tivariable analysis revealed that preoperative biliary stentplacement, intraoperative drain placement, advanced age(�70 years), and weight loss at presentation were associatedwith the development of complications (Table 3). Although

TABLE 2. Operative Procedures for 99 Patients With Gallbladder Cancer

Time Period 1N � 35

Time Period 2N � 64

Total (P Values,Comparing

Between TimePeriods)

Curative resection 26% 45% 38%NS

Laparoscopic cholecystectomy 26% 25% 25%Prior to referral NSLaparoscopic cholecystectomy 6% 5% 5%Converted to open NSOpen cholecystectomy prior to referral 12% 3% 6%

NSPreoperative biliary stent 33% 59% 50%

NSLiver resection P � 0.04 for any

type ofresection

Wedge resection ofGB fossa

4 Patients 13 Patients 17 Patients

Segments 4 and 5resection

1 Patient 8 Patients 9 Patients

Extended righthepatectomy

1 Patient 8 Patients 9 Patients

Pancreaticoduodenectomy 0 1 Patient 1 PatientBiliary tree resection 3 Patients 19 Patients 22 Patients

P � 0.04Operating time, all procedures (min) 187 228 215

NSOR blood loss, all procedures (mL) 496 600 566

NSMean length of stay (days) 11.8 9.8 10.5

NSPalliative biliary bypass 6 Patients 7 Patients 13 PatientsLaparoscopy, staging 0 Patients 3 Patients 3 PatientsClosed suction drain 45% (10/22) 76% (22/29) 64% (32/51)

NS

GB, gallbladder; NS, not significant.



multiple factors likely contributed to the development ofcomplications, it appeared that closed suction drains mayhave resulted in wound infections and bile leaks and thatpreoperative biliary stents may have increased the risks ofboth wound infections and cholangitis. Drains may increasethe risk of biliary leak if placed in close proximity to adivided biliary radicle. Similarly, if the drain exits the skin inclose proximity to the wound, this may contribute to postop-erative wound infection.

PathologyThe distribution of histologies included adenocarci-

noma (88%), mucinous variant (4%), papillary (3%), adeno-squamous (2%), and squamous (2%). Tumor differentiationwas classified as moderately differentiated (53%), poorlydifferentiated (27%), and well differentiated (20%). Of note,17% of tumors exhibited perineural invasion. Twenty caseshad involved regional lymph nodes.

Survival AnalysisSurvival analysis and comparisons between time peri-

ods are shown in Figures 2 3, and 4. As a group, TP2 patientshad improved overall survival compared with cases treated inTP1 (Fig. 2). Five-year survival in TP1 and 2 is 7%, and 35%,respectively. Median survival in TP1 was 9 months, while inTP2 it was 17 months. Patients in both time periods hadimproved survival with complete (R0) surgical removal of allmalignant disease (P � 0.02 TP1; P � 0.0001 TP2 (Fig. 3A,B). Median survival in TP1 for those undergoing an R0resection versus those not having an R0 resection was 18months and 6 months, respectively (P � 0.02). Similarly, in

TP2, median survival was 19 months versus 7 months forR0-negative cases (P � 0.0001). Analysis of survival bystage in each time period reveals that patients in TP2 hadimproved survival, even when controlling for stage. Further-more, differences in survival by stage in each time period aresignificant (P value TP1 � 0.04, TP2 � 0.0013; Fig. 4A, B).A multivariate analysis of factors found to have significanceupon univariate analysis revealed that only those patientsundergoing an R0 resection, and with an absence of weightloss at presentation, manifest a statistically significant im-provement in overall survival (Table 4).

A description of the 8 long-term survivors is shown inTable 5. Eight patients suffered a local recurrence (mean timeto recurrence, 12.3 months) and 18 relapsed with metastaticdisease (mean time, 11.1 months; Fig. 4). One patient wasfound to have microscopic disease in the port site afterresection; this patient had an extended hepatectomy and aconcomitant pancreaticoduodenectomy. This patient wasclassified as having metastatic disease (stage IV) and recurredwith local and distant metastases within a year. A total of 4patients had port site recurrences, in 2 cases after resection ofthe port site.

DISCUSSIONFive-year survival for carcinoma of the gallbladder in

“Western” series ranges between 5% and 12%.2,3 Thesedismal results are related to the aggressive biology of thiscancer and advanced disease due to late presentation. Re-cently, 5-year survival rates as high as 36% have beenreported in patients undergoing “radical” resection in Japa-nese centers,19 which suggests that with optimal surgicalmanagement, locoregional recurrence may be reduced. Fonget al19 and Fong and Malhotra27 have reported similar resultsin North America with 38% 5-year survival in resectedpatients. These results, along with an appreciation that R0resections improve outcome in patients with hilar biliarycancer,20–22 led to a reappraisal of our surgical approach to

TABLE 3. Factors Predictive of Complications FollowingSurgery for Gallbladder Cancer: Multivariate Analysis

Covariates P Value

Liver resection (yes, no) 0.97Time period (1 or 2) 0.93Biliary tree resection (yes or no) 0.77Sex (male vs. female) 0.70Estimated blood loss (�500 mL, �500

mL)0.51

Jaundice (yes or no) 0.45Operating time (�3 h, �3 h) 0.21Drain (yes or no) 0.003 (OR � 39.3,

95% CI 3.42–450)Preoperative biliary stent (yes or no) 0.002 (OR � 71.73,

95% CI 4.53–1000)

Age (�70,�70 y) 0.012 (OR � 18.63,95% CI 1.86–186)

Weight loss at presentation (yes or no) 0.035 (OR � 14.50,95% CI 1.34–155)

FIGURE 2. Overall survival by time period: Kaplan-Meier curve.



gallbladder cancer in the mid-1990s. In this report, we com-pare the operative management and survival outcomes of 2distinct surgical approaches we have applied to this diseaseover the past 12 years.

Our results confirm that radical resections including theliver and/or biliary tree can be performed safely with minimalmortality (2%). However, complications are common, withmajor complications requiring intervention occurring in 29%

FIGURE 3. A, Effect of R0 resectionon survival (time period 1). B, Effectof R0 resection on survival (time pe-riod 2).



of cases. Multivariate analysis demonstrated the most impor-tant factor determining long-term survival is the performanceof a margin negative, R0 resection (P � 0.001). Although noR1 resections were intentionally performed in our study, prior

investigators have demonstrated that R1, or debulking oper-ations, do not improve survival.16,28,29

Our current management algorithm for patients withcarcinoma of the gallbladder is described below. For patients

FIGURE 4. A, Survival bystage (time period 1). B,Survival by stage (time pe-riod 2).



who present with early-stage (T1) cancers following removalof the gallbladder elsewhere, we conduct a review of thepathology specimen, with attention to depth of invasion,margin status, lymphatic and perineural invasion, and ana-tomic location of the mass within the gallbladder. If thesefeatures confirm early-stage (T1N0M0) disease, cholecystec-tomy alone is adequate treatment of these cases, with closefollow-up to detect local recurrence. However, if there is any

concern that the lesion is T2, ie, has penetrated the fullthickness of the muscularis into the subserosal plane (peri-muscular plane on liver side of gallbladder, generally theplane of dissection used in simple cholecystectomy), then anextended cholecystectomy including liver � bile duct resec-tion is performed; in this not-infrequent scenario where ex-ploration is performed after prior gallbladder removal, it isexceedingly difficult to differentiate scar from cancer, as has

TABLE 5. Actual Long-term Survivors: Following Surgery for Gallbladder Cancer

Age/Sex Stage R0 Resection Procedure Survival/Status

68 y/M Stage II T2N0M0 Yes Cholecystectomy 8 y 9 mo NED76 y/M Stage I T1N0M0 Yes Cholecystectomy with wedge

resection liver8 y 5 mo NED

69 y/M Stage IV T4N2M0 Yes Extended right hepatectomy/biliarytree resection/cholecystectomy

4 y 9 mo DOD

62 y/M Stage IIIb T4N0M0 Yes Extended right hepatectomy/biliarytree resection/cholecystectomy

4 yr DOD

74 y/F Stage II T2N0M0 Yes Cholecystectomy 3 y 9 mo DOD53 y/M Stage III T3N0M0 Yes Cholecystectomy with wedge

resection liver4 y 2 mo DOD

67 y/M Stage II T2N0M0 Yes Cholecystectomy with wedgeresection liver

3 y 9 mo DOD

75 y/M Stage III T3N0M0 Yes Cholecystectomy/segments 4 and 5resection liver/biliary tree resection

4 y 11 mo DOD

DOD, dead of disease; NED, no evidence of disease.

TABLE 4. Factors Predictive of Long-term Outcome

CovariatesUnivariate Analysis(Log-Rank-P Value)

Multivariate AnalysisP Value (Hazard Ratio, 95% CI)

R0 resection 0.0001 0.001 (HR � 0.282, 95% CI 0.14–0.59)Weight loss at presentation 0.0027 0.008 (HR � 2.54, 95% CI 1.29–5.00)M stage 0.0001 0.07 (HR � 2.20, 95% CI 0.99–4.72)MSKCC stage 0.001 0.18 (HR � 1.62, 95% CI 0.82–3.17)Biliary tree resection 0.005 0.33 (HR � 0.82, 95% CI 0.56–1.21)Liver resection 0.0001 0.44 (HR � 0.87, 95% CI 0.61–1.23)Time period 0.03 0.51 (HR � 0.77, 95% CI 0.38–1.54)Jaundice 0.03 0.85 (HR � 0.94, 95% CI 0.47–1.87)Perineural invasion 0.165Tumor differentiation 0.43Age (�70, �70) 0.97T stage 0.73N stage 0.84Mass at presentation 0.77INR (�1, �1) 0.70Pain at presentation 0.93Histology 0.67

INR, International Normalized Ratio; MSKCC, Memorial Sloan-Kettering Cancer Center.



been previously appreciated.27 Furthermore, port sites areexcised if the patient has undergone a laparoscopic cholecys-tectomy as the initial intervention.

Similarly, lesions that are initially staged as T2 or T3are treated with a liver resection, either a segment 4b/5resection, right hemihepatectomy or an extended right hepa-tectomy. The decision regarding the type of liver resection isbased somewhat on the anatomic location of the tumor.Lesions located predominantly in the fundus can be treatedwith a segment 4b/5 resection. Those located in Hartmannpouch, the gallbladder neck, or extending into the triangle ofCalot (ie, those close to the hilar plate and biliary confluence)require an extended liver resection to ensure negative mar-gins. This requires resection of the extrahepatic/extrapancre-atic biliary tree, including the horizontal course of the lefthepatic duct over to the umbilical fissure. In these cases, anextended right hepatectomy is generally performed with re-construction to the left hepatic duct at the base of theumbilical fissure. All T4 lesions are likewise treated withextended hepatectomy.

In the circumstance where the patient presents prior toany surgical intervention with T2 or worse gallbladder can-cer, we perform a liver resection. Our approach to the biliarytree is individualized. If patients present with obstructivejaundice, the biliary tree is resected. If the lesion is locatedpredominantly in Hartmann pouch, the neck of the gallblad-der, or appears to extend down the cystic duct, then a biliarytree resection is performed. We perform a complete portallymph node dissection, with thorough skeletonization of theportal structures, down to and including the suprapyloriclymph node overlying the hepatic–gastroduodenal arteryjunction. We agree with Fong and Malhotra27 with regard totreatment of bulky disease along the hepatoduodenal liga-ment, and in these circumstances we resect the biliary tree tofacilitate a complete nodal dissection. Segment 4b/5 resec-tions are limited to those patients with T3 or lower fundallesions and in those patients in which there is clearly anegative margin towards the right pedicle/plate. Extendedhepatectomy is performed in all patients with T4 lesions orany lesion in which a negative or close margin is a concern.Generally, this entails an extended right hepatectomy; how-ever, in rare cases where the bulk of the tumor encroaches onthe umbilical fissure, or intrabiliary extension of tumor ispredominantly along the left duct, then an extended lefthepatectomy is performed. All jaundiced patients, patientswith bulky nodal disease along the hepatoduodenal ligament,and any other patient with a positive cystic duct marginundergo a biliary resection. Other contiguous organ involve-ment is resected en bloc.

The main limitation of this study is its retrospective,nonrandomized nature. By comparing outcomes in the sev-enth through 12th years to those in the first 6 years, we areusing historical controls as a reference point for the new

surgical approach. The main drawback to this type of com-parison is that it is possible that the overall management ofthese patients has changed over time, irrespective of surgicalfactors. Some of these changes include improved diagnosticimaging and staging, variability in referral patterns, betterpatient selection, and improved perioperative care. To gaugethe impact of known possible confounders, we performed amultivariate analysis. However, this does not take into ac-count unknown confounders, but it is unlikely that anyunknown confounder could account for the magnitude ofdifference in survival between time periods. By including inthe multivariate analysis the time period in which therapyoccurred, unknown confounders related to the time period areaccounted for indirectly. A randomized trial comparing ex-tended versus limited surgical resection would be the idealmanner to test this concept, but this cannot be performedbecause of the paucity of cases, as well as the lack oftherapeutic equipoise, given the current evidence in our studyand others.8–19,27,29–36 Adjuvant and neoadjuvant therapieswere not used to any great extent in either time period, makingthem unlikely to account for the differences in outcomes.

In summary, a marked improvement in outcome ofpatients with gallbladder cancer has been achieved over thepast 6 years, primarily due to a shift towards more aggressivesurgery. Multivariate examination accounting for the timeperiod in which treatment occurred reveals that the mostimportant factor determining long-term survival is whetherthe patient underwent an R0 surgical resection. However,even in the most recent time period, a maximal 5-yearsurvival rate of only 35% has been achieved; therefore, it isessential that novel adjuvant therapies be developed andevaluated for these patients.

REFERENCES1. Piehler JM, Crichlow RW. Primary carcinoma of the gallbladder. Surg

Gynecol Obstet. 1978;147:929–942.2. Cubertafond P, Gainant A, Cucchiaro G. Surgical treatment of 724

carcinomas of the gallbladder: results of the French Surgical AssociationSurvey. Ann Surg. 1994;219:275–280.

3. Wilkinson DS. Carcinoma of the gall-bladder: an experience and reviewof the literature. Aust N Z J Surg. 1995;65:724–727.

4. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrenceafter hepatic resection for metastatic colorectal cancer: analysis of 1001consecutive cases. Ann Surg. 1999;230:309–318.

5. Jarnagin WR, Gonen M, Fong Y, et al. Improvement in perioperativeoutcome after hepatic resection: analysis of 1803 consecutive cases overthe past decade. Ann Surg. 2002;236:397–406.

6. Taylor M, Forster J, Langer B, et al. A study of prognostic factors forhepatic resection for colorectal metastases. Am J Surg. 1997;173:467–471.

7. Bartlett DL, Fong Y, Fortner JG, et al. Long-term results after resectionfor gallbladder cancer: implications for staging and management. AnnSurg. 1996;224:639–646.

8. Kondo S, Nimura Y, Hayakawa N, et al. Regional and para-aorticlymphadenectomy in radical surgery for advanced gallbladder carci-noma. Br J Surg. 2000;87:418–422.

9. Kondo S, Nimura Y, Hayakawa N, et al. Extensive surgery for carci-noma of the gallbladder. Br J Surg. 2002;89:179–184.

10. Kondo S, Nimura Y, Kamiya J, et al. Mode of tumor spread and surgical



strategy in gallbladder carcinoma. Langenbecks Arch Surg. 2002;387:222–228.

11. Kondo S, Nimura Y, Kamiya J, et al. Five-year survivors after aggres-sive surgery for stage IV gallbladder cancer. J Hepatobiliary PancreatSurg. 2001;8:511–517.

12. Shirai Y, Ohtani T, Tsukada K, et al. Radical surgery is justified forlocally advanced gallbladder carcinoma if complete resection is feasible.Am J Gastroenterol. 1997;92:181–182.

13. Shirai Y, Ohtani T, Tsukada K, et al. Pancreaticoduodenectomy forgallbladder cancer with peripancreatic nodal metastases. Hepatogastro-enterology. 1997;44:376–377.

14. Shirai Y, Ohtani T, Tsukada K, et al. Combined pancreaticoduodenec-tomy and hepatectomy for patients with locally advanced gallbladdercarcinoma: long term results. Cancer. 1997;80:1904–1909.

15. Shirai Y, Yoshida K, Tsukada K, et al. Radical surgery for gallbladdercarcinoma: long-term results. Ann Surg. 1992;216:565–568.

16. Todoroki T, Kawamoto T, Takahashi H, et al. Treatment of gallbladdercancer by radical resection. Br J Surg. 1999;86:622–627.

17. Todoroki T, Takahashi H, Koike N, et al. Outcomes of aggressivetreatment of stage IV gallbladder cancer and predictors of survival.Hepatogastroenterology. 1999;46:2114–2121.

18. Doty JR, Cameron JL, Yeo CJ, et al. Cholecystectomy, liver resection,and pylorus-preserving pancreaticoduodenectomy for gallbladder can-cer: report of five cases. J Gastrointest Surg. 2002;6:776–780.

19. Fong Y, Jarnagin W, Blumgart LH. Gallbladder cancer: comparison ofpatients presenting initially for definitive operation with those presentingafter prior noncurative intervention. Ann Surg. 2000;232:557–569.

20. Chamberlain RS, Blumgart LH. Hilar cholangiocarcinoma: a review andcommentary. Ann Surg Oncol. 2000;7:55–66.

21. Langer JC, Langer B, Taylor BR, et al. Carcinoma of the extrahepaticbile ducts: results of an aggressive surgical approach. Surgery. 1985;98:752–759.

22. Launois B, Terblanche J, Lakehal M, et al. Proximal bile duct cancer: highresectability rate and 5-year survival. Ann Surg. 1999;230:266–275.

23. Fujii K, Isozaki H, Okajima K, et al. Clinical evaluation of lymph nodemetastasis in gastric cancer defined by the fifth edition of the TNM

classification in comparison with the Japanese system. Br J Surg.1999;86:685–689.

24. Couinaud C. �Surgical anatomy of the liver: several new aspects�.Chirurgie. 1986;112:337–342.

25. Couinaud C. �The anatomy of the liver�. Ann Ital Chir. 1992;63:693–697.

26. Couinaud C. �Intrahepatic anatomy: application to liver transplantation�.Ann Radiol (Paris). 1994;37:323–333.

27. Fong Y, Malhotra S. Gallbladder cancer: recent advances and currentguidelines for surgical therapy. Adv Surg. 2001;35:1–20.

28. Shirai Y, Ohtani T, Tsukada K, et al. Predictors of survival in patientswith carcinoma of the gallbladder. Cancer. 1997;79:185–186.

29. Tsukada K, Hatakeyama K, Kurosaki I, et al. Outcome of radical surgeryfor carcinoma of the gallbladder according to the TNM stage. Surgery.1996;120:816–821.

30. Fong Y, Brennan MF, Turnbull A, et al. Gallbladder cancer discoveredduring laparoscopic surgery: potential for iatrogenic tumor dissemina-tion. Arch Surg. 1993;128:1054–1056.

31. Fong Y, Heffernan N, Blumgart LH. Gallbladder carcinoma discoveredduring laparoscopic cholecystectomy: aggressive reresection is benefi-cial. Cancer. 1998;83:423–427.

32. Kondo S, Nimura Y, Hayakawa N, et al. �Rationale of paraaortic lymphnodes dissection for advanced gallbladder cancer�. Nippon Geka GakkaiZasshi. 1990;91:223–227.

33. Kondo S, Nimura Y, Hayakawa N, et al. �Value of paraaortic lymph-adenectomy for gallbladder carcinoma�. Nippon Geka Gakkai Zasshi.1998;99:728–732.

34. Shoup M, Fong Y. Surgical indications and extent of resection ingallbladder cancer. Surg Oncol Clin North Am. 2002;11:985–994.

35. Tsukada K, Yoshida K, Aono T, et al. Major hepatectomy and pancre-atoduodenectomy for advanced carcinoma of the biliary tract. Br J Surg.1994;81:108–110.

36. Wakai T, Shirai Y, Hatakeyama K. Radical second resection providessurvival benefit for patients with T2 gallbladder carcinoma first discov-ered after laparoscopic cholecystectomy. World J Surg. 2002;26:867–871.



An Aggressive Surgical Approach Leads to Improved Survival in Patients With Gallbladder Cancer

Documents

Transcript of An Aggressive Surgical Approach Leads to Improved Survival in Patients With Gallbladder Cancer