- Email: [email protected]
Locally advanced gallbladder cancer: Which patients benefit from resection?
Available online at www.sciencedirect.com
ScienceDirect EJSO xx (2013) 1e8
www.ejso.com
Locally advanced gallbladder cancer: Which patients benefit from resection? D.J. Birnbaum a, L. Vigan o*,a, A. Ferrero, S. Langella, N. Russolillo, L. Capussotti Dept. of HPB and Digestive Surgery, Ospedale Mauriziano Umberto I, Largo Turati 62, 10128 Torino, Italy Accepted 19 October 2013 Available online - - -
Abstract Objectives: Patients with T3e4 gallbladder cancers (GBCs) often require extended surgical procedures, and up to 30% of patients have N2 metastases. This study investigated which patients with T3e4 GBC benefit from resection. Methods: Consecutive patients (n ¼ 78) with T3e4 GBC who underwent resection between 1990 and 2011 were analysed (38 before 2003, 40 in 2003e2011). Forty patients required common bile duct (CBD) resection, 10 pancreatoduodenectomy, 4 right colectomy and 2 gastric resection. Fifty-two (67%) patients had LN metastases, including 22 with N2 metastases. Results: The in-hospital mortality rate was 8%, 11% before 2003 vs. 5% in 2003e2011. The morbidity rate (47%) remained stable during the study. Undergoing liver and pancreatic resection did not increase severe morbidity (0%) or mortality (10%). Sixty-seven (86%) patients had R0 resection. The 5-year survival rate was 17% (median follow-up, 65 months). Survival improved after 2002 (26% vs. 9%, p ¼ 0.04). R1 patients had poor 3-year survival (0% vs. 32%, p ¼ 0.001). Nþ patients also had low survival (5-year survival, 10% vs. 32% in N0, p ¼ 0.019), but N1 and N2 patients had similar outcomes. CBD resection and major hepatectomy did not worsen prognosis. Patients requiring pancreatoduodenectomy, gastric or colonic resection had 0% 3-year survival ( p ¼ 0.036 in multivariate analysis). Conclusions: Resection of T3e4 GBC is worthwhile only if R0 surgery is achievable. Outcomes improved in most recent years. N2 metastases should not preclude surgery. Good results are possible even with CBD resection or major hepatectomy, while benefits from surgery are doubtful if pancreatoduodenectomy or other organ resection is needed. Ó 2013 Elsevier Ltd. All rights reserved. Keywords: Gallbladder cancer resection; Pancreatoduodenectomy; Common bile duct resection; Survival; Liver surgery; Prognostic factors
Introduction Gallbladder cancer (GBC) is the most common biliary tract malignancy worldwide.1 Complete surgery is the only curative therapy, whereas chemotherapy offers a limited contribution to survival.1e4 Because it causes few symptoms, GBC is often diagnosed at an advanced stage and is resectable in only 15%e40% of patients.1e3 Surgical outcome depends strongly on tumour stage: while excellent survival rates of up to 100% at 5 years have been reported for early stage disease (T1e2), overall survival rates range from 0% to 30% in locally advanced tumours (T3e4).5e8 To be completely resected, locally advanced GBCs often require extended surgical procedures, such as major hepatectomy, common bile duct (CBD) resection or pancreatoduodenectomy. In addition, lymph node (LN) metastases
in the celiac and retropancreatic area are detected in up to 30% of patients.8,9 The surgical indications in these patients are still being debated. A few series involving a limited number of patients, mainly in Asian centres, have reported conflicting results.9e15 During the last 20 years, we have adopted an aggressive surgical policy in patients affected by locally advanced GBC. Extended surgical procedures have been performed as long as complete resection was achievable. The present analysis reviewed the outcomes of this aggressive surgical approach, with particular attention to patients who required resection of other organs, e.g. CBD resection or pancreatoduodenectomy. The aim of the study was to identify which patients with locally advanced GBC benefit from resection. Materials and methods
* Corresponding author. Tel.: þ39 011 5082590; fax: þ39 011 5082592. E-mail address: [email protected] (L. Vigano). a Authors contributed equally.
A total of 126 consecutive patients underwent resection for GBC at the Mauriziano Umberto I Hospital of Torino
0748-7983/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejso.2013.10.014 Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
2
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
between January 1990 and December 2011. This study analysed the 78 (62%) patients with T3 or T4 GBC. Patient management Preoperative staging routinely included carcinoembryonic antigen (CEA) and CA 19-9 value determination and thoraco-abdominal computed tomography (CT). Since 1998, magnetic resonance cholangio-pancreatography (MRCP) has been performed in all jaundiced patients or whenever CBD infiltration was suspected at CT. In recent years, contrast-enhanced magnetic resonance imaging and positron emission tomography CT were performed in case of uncertain diagnosis or suspected extra-hepatic disease, respectively. Since 1998, CT-volumetry has been performed regularly in patients scheduled for major hepatectomy to estimate the future liver remnant (FLR) volume. The FLR was considered adequate if >25% of the total liver volume. In jaundiced patients, a higher cut-off of >30% was adopted. If the FLR was inadequate, preoperative portal vein embolization (PVE) was performed, and CT-volumetry was repeated 4 weeks later. Surgery was scheduled only if adequate FLR hypertrophy was observed. In jaundiced patients, preoperative biliary drainage was not performed on a routine basis. Its indications were cholangitis, malnutrition or the need for PVE (drainage of the FLR only). Indications Surgery was performed whenever complete resection was achievable. Distant metastases, extensive infiltration of the hepatoduodenal ligament and interaorto-caval LN metastases were contraindications to surgery. The surgical indication was evaluated on a case-by-case basis when celiac or retropancreatic LN metastases were present. The indications for CBD resection were as follows: macroscopic CBD infiltration; LN metastases of the pedicle infiltrating the CBD; tumoral deposits along the hepatic pedicle; and neoplastic infiltration of the cystic duct stump in frozen section analysis. Pancreatoduodenectomy was performed in case of infiltration of the duodeno-pancreatic area by the GBC or by large retropancreatic LN metastases. Colonic or gastric resection was performed in case of macroscopic infiltration. Surgical procedure The standard scheduled procedure for T3e4 GBC was segment 4b-5 resection (anatomic bisegmentectomy or wide wedge resection), en-bloc cholecystectomy and LN dissection. LN dissection was limited to the hepatic pedicle (D1) in the first part of the study (before 1998) and was extended to the retropancreatic and celiac area (D2) since 1998. Frozen section analysis of the cystic duct stump was performed regularly. In patients with a postoperative
diagnosis of GBC after laparoscopic cholecystectomy, the trocar sites were systematically removed. The surgical techniques used for liver resection, CBD resection and pancreatoduodenectomy have been described previously.16,17 Since 2000, staging laparoscopy with laparoscopic ultrasonography always preceded laparotomy to exclude peritoneal carcinomatosis or liver metastases. Adjuvant treatment and follow-up Postoperative chemotherapy was delivered according to patient performance status and pathological findings. All patients were followed-up every 3 months with a physical examination, CEA and CA 19-9 determinations and abdominal ultrasonography or thoraco-abdominal CT. No patient was lost during the follow-up period. This analysis includes follow-up to August 31, 2012. Definitions Major hepatectomy was defined as resection of 3 Couinaud segments. Operative mortality was defined as death within 90 days after surgery or before discharge from the hospital. Morbidity included all postoperative complications and was classified according to the ClavieneDindo classification.18 Tumour TNM staging was reviewed according to the 7th edition of the AJCC manual.19 Statistical analysis Data were collected prospectively and analysed retrospectively. The series was divided into two periods in order to analyse outcome evolution during the study period. Continuous variables were compared between groups using the unpaired t test or the ManneWhitney U test, as appropriate; categorical variables were compared using the chisquare test or Fisher’s exact test, as appropriate. The KaplaneMeier method was used to estimate overall survival (OS) and recurrence-free survival (RFS) probabilities, which were compared using the log-rank test. Patients with operative mortality or incomplete surgery (R2) were excluded from survival analysis. Multivariate analysis was performed using a Cox proportional hazard model to identify independent prognostic factors of OS and RFS after liver resection. Multivariate analysis was completed for factors with a p value 0.10 in the univariate analysis. A pvalue <0.05 was considered significant for all tests. Results During the study period (1990e2011), 78 patients underwent surgery for T3e4 GBC, 38 before 2003 and 40 in 2003e2011. Of these, 34 (44%) patients were male; the median age was 67 years (range 39e83). One third of the patients (n ¼ 25) was jaundiced at diagnosis [median bilirubin value 195.8 mmol/L (59.9e335.2)]; of these, 14
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
(56%) underwent preoperative biliary drainage, with an endoscopic approach used in 64% of cases. In 17 (22%) patients, the diagnosis of GBC was postoperative, and 16 of the 17 patients were referred to the author’s department after undergoing cholecystectomy at another hospital. One of these patients underwent an emergency cholecystectomy for acute cholecystitis in the author’s department. All 17 patients underwent a second operation after cholecystectomy after a median delay of 63 days (13e102). The patient characteristics are summarized in Table 1. Surgical procedures Liver resection was performed in all but 1 patient (99%). That patient was affected by T3 GBC of the infundibulum that mimicked a mid-CBD cancer and underwent isolated CBD resection. Of the remaining 77 patients, 22 (29%) underwent major hepatectomy (8 underwent preoperative PVE), including 19 extended resections (>4 liver segments). LN dissection was D2 in 60 (77%) patients. Fifty-four patients (69%) required resection of contiguous organs: pancreatoduodenectomy in 10, isolated CBD resection in 40, right colectomy in 4 (associated with pancreatoduodenectomy in 2) and distal gastrectomy in 2. Pancreatoduodenectomy was performed because of macroscopic neoplastic infiltration in 7 patients and because of large retropancreatic LN metastases in 3. CBD resection was performed because of macroscopic neoplastic infiltration in 34 patients, non-contiguous neoplastic tissue along the hepatic pedicle in 3, large hepatic pedicle LN metastases in 2 and cystic duct stump infiltration in 1.
Table 1 Patient characteristics and pathological data. Patients (n ¼ 78) n (%) Patient characteristics and surgical procedures Age, years, median (range) Gender, Female/Male Gallblader lithiasis CEA ng/mL, median (range) Ca 19-9 U/mL, median (range) Preoperative jaundice Gallbladder cancer diagnosis Surgical procedures
Associated resections
Lymph-node dissection Pathology data TNM stagingb
Postoperative outcomes The 90-day mortality rate was 8% (n ¼ 6). The causes of death included the following: hepatic failure in 2 patients, sepsis in 1, hemoperitoneum in 1, acute myocardial infarction in 1 and acute respiratory distress syndrome in 1. The overall morbidity rate was 47% (n ¼ 37), including 16 (21%) grade III/IV complications. The complications are listed in Table 2. Patients undergoing standard surgical procedures (bisegmentectomy Sg4b-5 and LN dissection) had no mortality and no severe complications (grade III/IV). Patients requiring resection of contiguous organs had higher mortality (11% vs. 0%, p ¼ 0.090), overall morbidity (63% vs. 13%, p ¼ 0.009) and grade IIIeIV morbidity rates (30% vs. 0%, p ¼ 0.002). Of the patients requiring pancreatoduodenectomy, there was 1 operative mortality (10%) and no grade IIIeIV morbidities. The mortality rate decreased across the study period (11% before 2003 vs. 5% in 2003e2011, p ¼ n.s.) Overall and grade IIIeIV morbidity rates remained stable during the study period. Adjuvant chemotherapy was administered to 30 (38%) patients.
3
Completeness of surgery
Preoperative Intraoperative Postoperative Cholecystectomy Sg4b-5 resection Sg 5,6,7,8 [Sg1] Sg2,3,4,5 [Sg1] Sg4,5,6,7,8 [Sg1] Associated Sg1 resection Isolated CBD resection Pancreatoduodenectomy Right colectomy Distal gastrectomy D1 D2 T3 T4 N0 N1 N2 Stage 3A Stage 3B Stage 4A Stage 4B R0 R1 Liver resection margin Biliary margin [proximal/distal] Radial margin R2
Perineural infiltration Biliary duct infiltration/Suspected macroscopic infiltration Pancreatic infiltration/Suspected macroscopic infiltration Colonic & gastric infiltration/Suspected macroscopic infiltration
67 (39e83) 44 (56)/34 (44) 55 (71) 2.8 (0.6e657) 89.6 (0.8e4778) 25 (32) 49 (63) 12 (15) 17 (22) 1 (1) 55 (71) 3 (4) 2 (3) 17 (22) 8 (10) 40 (51) 10 (13)a 4 (5)a 2 (3) 18 (23) 60 (77) 62 (79) 16 (21) 26 (33) 30 (38) 22 (28) 23 (29) 25 (32) 8 (10) 22 (28) 67 (86) 10 (13) 2 5 [3/2] 5 1 (1)d 63 (81) 27/34c 7/7 6/6
CEA: Carcinoembryonic antigen; Sg: segment; CBD: common bile duct. a Two patients had combined pancreatoduodenectomy and right colectomy. b According to the 7th edition of the AJCC manual.22 c Patients with isolated CBD resection were considered. d One patient had misdiagnosed liver metastasis (R2 patient).
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
4
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
Pathological data The pathological data are summarized in Table 1. The tumour staging was T3 in 62 (79%) patients and T4 in 16 (21%). The histologic tumour type was adenocarcinoma in 72 (92%) patients and squamous carcinoma in 6 (8%). The median number of retrieved LNs was 7 (1e35), 4 after D1 dissection and 9 after D2 ( p ¼ 0.005). The number of retrieved LNs was similar between patients with and without CBD resection. Fifty-two (67%) patients had LN metastases; the metastases extended to the retropancreatic or the celiac area (N2) in 22 (28%). Pathological examination showed that 33 (42%) patients had CBD infiltration (27 during isolated CBD resections and 6 during pancreatoduodenectomy). Suspected macroscopic infiltration of the CBD was confirmed in 79% of patients, while suspected infiltration of the pancreas, the stomach and the right colon was confirmed in all patients. The resection was complete (R0) in 67 (86%) patients, while 10 (13%) patients had R1 resection (positive biliary margin in 5). One patient had R2 resection: intraoperative ultrasonography detected a suspected liver metastasis, and the biopsy was negative in frozen section analysis but positive for adenocarcinoma at final pathology. Patients requiring resection of contiguous organs had a higher R1 resection rate, but the difference was not significant (16% vs. 8%, p ¼ n.s.) All patients that underwent pancreatoduodenectomy had R0 resection. Overall survival data After a median follow-up of 65 months (6e114), the 1-, 3- and 5-year OS rates were 66%, 27% and 17%, Table 2 Short-term outcomes. Short-term outcome (n ¼ 78) n (%) 90-day mortality Overall morbidity Liver failure Bile leak Hemoperitoneum Abdominal collection Ascites Sepsis Pancreatic leak Duodenal leak Pulmonary morbidity Pulmonary embolism Renal failure Acute myocardial infarction Small bowel disorders Dindo & Clavien classification18 Grade IeII Grade III [a/b] Grade IV [a/b] Blood transfusion Hospital stay days, median (range)
6 37 7 15 4 3 6 8 3 1 12 1 3 1 1
(8) (47) (9) (19) (5) (4) (8) (10) (4) (1) (15) (1) (4) (1) (1)
15 (19) 11 [7/4] (14) 5 [4/1] (6) 12 (15) 13 (0e83)
respectively. The median survival was 16 months. Nine (12%) patients survived more than 5 years after resection: all 9 had T3 GBC and R0 resection; 2 were N1 and 1 was N2; 3 had CBD resection (1 had CBD infiltration at final pathology). OS improved in recent years: 5-year survival was 9% in patients operated on before 2003 vs. 26% in patients treated in 2003e2011 ( p ¼ 0.037). Recurrence analysis and recurrence-free survival data Fifty-nine patients (76%) had recurrence: hepatic in 26 (33%), extra-hepatic in 29 (37%) and both hepatic and extra-hepatic in 4 (5%). Of the extra-hepatic recurrence sites, the most common ones were the peritoneum (n ¼ 11), the hepatoduodenal ligament (n ¼ 9) and the lung (n ¼ 6). Of the 9 patients with hepatoduodenal ligament recurrence, 7 had CBD resection because of infiltration and 5 had R1 resection. All 9 of the patients had perineural infiltration. The risk of local recurrence was increased in patients with CBD infiltration (24% vs. 5%, p ¼ 0.027) and R1 status (50% vs. 7%, p ¼ 0.0002). N status and the extension of LN dissection did not correlate with local recurrence risk. The 1-, 3- and 5-year RFS rates were 44%, 22% and 16%, respectively (median 10 months). Four patients had recurrence resection (1 liver, 1 lung, 1 carcinosis, 1 adrenal gland); the median OS after re-resection was 49 months (15e114). RFS improved in patients treated in 2003e2011 (21% vs. 7% in patients treated before 2003), but the difference was not significant ( p ¼ 0.056). Prognostic factors of OS and RFS Table 3 shows the results of univariate and multivariate analysis of OS and RFS prognostic factors. A survival benefit was evident only if complete surgery was achieved: 3-year OS was 32% in R0 patients vs. 0% in R1 patients ( p ¼ 0.002) (Fig. 1(a)); 3-year RFS was 26% vs. 0%, respectively ( p ¼ 0.0004). Similar unfavourable results were observed in patients with T4 tumours (0% 3-year OS vs. 33% in T3; 0% 3-year RFS vs. 26% in T3), but the difference was not significant (Fig. 1(b)). LN status correlated with prognosis (5-year OS 32% if N0 vs. 10% if Nþ, p ¼ 0.019; RFS 28% vs. 10%, respectively, p ¼ 0.013), but OS and RFS were similar in N1 and N2 patients (11% vs. 8% and 8% vs. 7%, respectively, p ¼ n.s.) (Fig. 1(c)). According to the TNM staging, 5-year OS was 35% in patients with stage 3A tumours, 13% in 3B, 0% in 4A and 8% in 4B ( p ¼ n.s.; stage 3A vs. others p ¼ 0.019). Similarly, RFS was better in patients with stage 3A tumours (31% vs. 9%, p ¼ 0.010), while no differences were observed among patients with stage 3B, 4A and 4B tumours. Patients with a diagnosis of GBC after a simple cholecystectomy who underwent a second operation had OS and
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
5
Table 3 Univariate and multivariate analysis of prognostic factors of overall and recurrence-free survival. Parameter
Age (years) Sex Jaundice at diagnosis Major hepatectomy CBD resection CBD infiltration DP/Gastric/Colonic res. T stage N stage TNM staging
Complete surgery Perineural infiltration Adjuvant CTx
70 >70 Male Female Y N Y N Y N Y N Y N T3 T4 N0 N+ 3A 3B 4A 4B R0 R1 Y N Y N
Median OS (months)
5-yr OS (%)
OS
15 23 24 15 24 16 13 16 15 17 15 16 16 16 15 16 24 14 24 11 25 15 17 10 15 95 15 16
16 20 26 11 15 19 14 19 11 29 5 24 0a 21 21 0a 32 10 35 13 0a 8 20 0a 6 58 24 14
n.s. 0.048
RFS
Univariate
Multivariate analysis
Univariate
Multivariate analysis
p
p
HR (CI95%)
p
p
e
n.s.
e
n.s.
0.017
n.s.
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
0.081
n.s.
2.962 (1.064e8.246) 1 e
n.s.
3.877 (1.369e10.981) 1 e
0.019
n.s.
0.013
n.s.
n.s.
e
n.s.
e
1 3.0.39 (1.335e6.917) 4.208 (1.426e12.414) 1 e
0.0004
0.022
0.006
0.014
0.085
0.038
0.002
0.008
0.002
0.009
n.s.
n.s.
0.011
HR (CI95%)
1 2.929 (1.169e7.338) 3.491 (1.295e9.415) 1 e
OS: overall survival; RFS: recurrence-free survival; 5-yr: 5-year; HR: hazard ratio; CBD: common bile duct; DP: pancreatoduodenectomy; res.: resection; CTx: chemotherapy. a No patient alive at three years after surgery.
RFS rates similar to patients with pre-/intra-operative diagnosis of GBC. Patients requiring major hepatectomy had OS and RFS rates similar to those undergoing minor resection. Patients with and without CBD resection had similar OS (Fig. 2(a)) and RFS. CBD infiltration confirmed at pathology examination was associated with lower OS and RFS (at 5 years: OS of 5% vs. 24%; RFS of 6% vs. 22%), but the difference was not significant (Fig. 2(b)). Patients requiring pancreatoduodenectomy had 0% OS and RFS 2 years after surgery, and patients requiring gastric or colonic had 0% OS and RFS 3 years after surgery (vs. 21% OS and 20% RFS at 5 years in the remaining patients, p ¼ 0.085 for OS and p ¼ 0.081 for RFS) (Fig. 2(c)). The multivariate analysis showed that the need for resection of contiguous organs other than the CBD (i.e. pancreatoduodenectomy, gastric and colonic resection) was an independent negative prognostic factor of OS (hazard ratio [HR] ¼ 2.962, p ¼ 0.038). Two additional negative prognostic factors were identified: R1 resection (HR ¼ 3.039, p ¼ 0.008) and perineural infiltration (HR ¼ 4.208,
p ¼ 0.009). The same parameters were independent negative predictive factors of RFS: resection of contiguous organs other than the CBD (HR ¼ 3.877, p ¼ 0.011); R1 resection (HR ¼ 2.929, p ¼ 0.022) and perineural infiltration (HR ¼ 3.491, p ¼ 0.014). Discussion The present study helps refine indications in patients treated surgically for locally advanced (T3e4) GBC. Only patients with complete resection (R0) showed a survival benefit from resection. N status predicted outcome, but there were long-term survivors in the group of Nþ patients, even when the patients had celiac or retropancreatic LN metastases. Patients requiring major hepatectomy or CBD resection had good survival outcomes; thus, such patients should be scheduled for surgery whenever possible. In contrast, patients requiring pancreatoduodenectomy or other organ resection (stomach or right colon) had extremely poor outcomes and showed no clear survival benefit from surgery.
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
6
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
Figure 1. Overall survival curves after gallbladder cancer resection according to the pathological data: (a) R status (R0 vs. R1); (b) T stage (T3 vs. T4); (c) N stage (N0 vs. N1 vs. N2). Patients with operative mortality (n ¼ 6) and R2 resection (n ¼ 1) were excluded.
GBC is an aggressive cancer: most patients present with advanced disease at diagnosis, and only a minority can be scheduled for curative treatment.1e3 To achieve complete resection, extended procedures may be required, such as CBD resection and pancreatoduodenectomy. The benefits from surgery in such aggressive tumours are debatable. There are few studies of this population, and most have been conducted in Asian centres and include a limited number of patients treated over a long time period.9e15 When considering treatment strategy, the first thing to consider is operative risk. Even though short-term outcomes after liver surgery have improved in recent years,20,21 mortality and morbidity rates after GBC resection remain high, ranging from 5% to 19% and from 14% to 53%, respectively.9e15,22 Similar figures were obtained in the present series: the mortality rate was 8%, and morbidity occurred in half of the patients. The need for aggressive surgical procedures could explain these data. High operative risks have been reported for liver plus pancreatic resection,10,13,22 but this was not the case in our experience. Specifically, the mortality rate was not increased after combined liver and pancreatic resection and there was no grade IIIeIV morbidity. The short-term results could reflect the long time interval analysed in most series. In recent years, patient selection has improved, as has surgical expertise. Preoperative biliary drainage in
jaundiced patients and PVE in those scheduled for a major hepatectomy allow adequate preparation for resection. In fact, mortality rates decreased during the study period from 11% before 2003 to 5% in 2003e2011. Nevertheless, mortality rates are still high compared with those reported for other pathologies, such as colorectal metastases (1e2%),20,21 and the rate of severe complications did not decrease during the study period. Provided there are good short-term results, the question must be asked: which patients benefit from resection? Considering long-term outcomes, the consensus is that aggressive surgery is worthwhile as long as complete resection can be achieved.2,9,15,23e25 In the Nagoya series, as in the present one, R1 resection was associated with 0% survival at 3 years.9 This raises another question: should all resectable patients be resected? Jaundice and CBD involvement have been reported to be predictors of poor outcome.3,9,15,26 For a long time, these two conditions were considered a contraindication for resection, but recent results have been encouraging.9,27 Nishio et al. reported a 5-year OS of 36% in patients with CBD resection.9 In the present study, patients requiring CBD resection had OS and RFS that were similar to those without CBD resection, and there were even three 5-year survivors. As in the Japanese study, jaundice at diagnosis showed no prognostic impact. CBD infiltration confirmed at final pathology was associated with worse outcome and higher risk of local recurrence, but did not
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
7
Figure 2. Overall survival curves after gallbladder cancer resection according to the performed resection: (a) common bile duct (CBD) resection (Yes vs. No); (b) CBD infiltration at final pathology (Yes vs. No); (c) pancreatoduodenectomy vs. gastric/colonic resection vs. no associated resection. Patients with operative mortality (n ¼ 6) and R2 resection (n ¼ 1) were excluded.
preclude long-term survival and thus should not be considered a contraindication for surgery. In some patients, pancreatoduodenectomy is needed to achieve complete resection because of direct neoplastic infiltration of the duodeno-pancreatic area or because of large retropancreatic LN metastases. Only a few cases have been described.10e12,14,28 Some authors reported excellent survival of up to 40% at 5 years.11,12,14 Araida et al. even suggested that pancreatoduodenectomy allows removal of micrometastasis in small peri-pancreatic nodes and is associated with better survival compared with nonpancreatoduodenectomy patients (5-year OS of 87% vs. 17%).14 Both T2 and T3e4 GBC were included in that study. The benefits of pancreatoduodenectomy were lost when the analysis included just patients with hepatoduodenal ligament invasion (5-year OS of 0e5% according to LN status). In the present series, only patients with locally advanced GBC were included, and pancreatoduodenectomy was performed if needed to achieve R0 status. In such patients, survival rates were extremely poor (0% at 2 years) even if surgery was complete in all patients. Similar results were observed in patients requiring colonic or gastric resection because of neoplastic infiltration (0% OS at 3 years). These poor outcomes cast doubt on the benefits of surgery. LN status is another important issue. It is consistently one of the strongest predictors of survival in patients with
advanced GBC,2,8,22,25,30 and LN metastasis sites may stratify prognosis in Nþ patients.2,22,30 In the current series, patients with LN metastases had lower OS and RFS compared with N0 patients, but long survival was even observed in Nþ patients. Furthermore, N1 and N2 patients had similar outcomes. According to the present results, LN status has a prognostic value, but the presence of LN metastases, even in the celiac or retropancreatic area, is not a contraindication to resection. Finally, this analysis identified perineural infiltration as a strong prognostic factor (5-year OS of 6% if present). The prognostic role of perineural infiltration has been reported previously.26,29 In T2 GBC, perineural infiltration has even been proposed as a criterion for systematic CBD resection to improve the completeness of surgery and survival.29 Perineural infiltration probably reflects tumour spread in the hepatoduodenal ligament and may predict microscopic residual disease. In fact, we found that local recurrence occurred only in patients with perineural infiltration and was associated with R status. However, it is impossible to preoperatively assess perineural infiltration, which limits its clinical usefulness. The present study has some limitations. This is a retrospective study that analysed a limited number of patients that were treated over a 20-year period. Some prognostic factors, such as the T stage or TNM stage, could not be identified
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
8
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
because of the small cohort size. The rarity of GBC and its low resectability rate makes it impossible to study a large group of patients at a single centre. Only prospective multicentre studies could overcome these limitations and should be conducted in the future. Nevertheless, the present study clarifies some of the controversial issues related to surgical indications in locally advanced GBC. In conclusion, in patients with T3e4 GBC, surgery is worthwhile only if R0 resection is achievable. Overall, outcomes have improved in recent years. N status has prognostic value, but the presence of LN metastases, even in the celiac or retropancreatic area, should not preclude surgery. In resectable patients, the indication for surgery should be based on the required procedure: good results are possible if CBD resection or major hepatectomy are required, while survival benefits from surgery are doubtful if pancreatoduodenectomy or resection of other organs are needed. Conflict of interest statement None declared.
References 1. Misra S, Chaturvedi A, Misra NC, et al. Carcinoma of the gallbladder. Lancet Oncol 2003;4:167–76. 2. Hueman MT, Vollmer Jr CM, Pawlik TM. Evolving treatment strategies for gallbladder cancer. Ann Surg Oncol 2009;16:2101–15. 3. Fong Y, Jarnagin W, Blumgart LH. Gallbladder cancer: comparison of patients presenting initially for definitive operation with those presenting after prior noncurative intervention. Ann Surg 2000;232:557–69. 4. Valle J, Wasan H, Palmer DH, et al. ABC-02 Trial Investigators. Cisplatin plus gemcitabine versus gemcitabine for biliary tract cancer. N Engl J Med 2010;362(14):1273–81. 5. Wakai T, Shirai Y, Yokoyama N, et al. Early gallbladder carcinoma does not warrant radical resection. Br J Surg 2001;88(5):675–8. 6. Pawlik TM, Gleisner AL, Vigano L, et al. Incidence of finding residual disease for incidental gallbladder carcinoma: implications for re-resection. J Gastrointest Surg 2007;11(11):1478–86. 7. Muratore A, Amisano M, Vigano L, et al. Gallbladder cancer invading the perimuscular connective tissue: results of reresection after prior non-curative operation. J Surg Oncol 2003;83(4):212–5. 8. Araida T, Higuchi R, Hamano M, et al. Hepatic resection in 485 R0 pT2 and pT3 cases of advanced carcinoma of the gallbladder: results of a Japanese Society of Biliary Surgery survey e a multicenter study. J Hepatobiliary Pancreat Surg 2009;16:204–15. 9. Nishio H, Ebata T, Yokoyama Y, et al. Gallbladder cancer involving the extrahepatic bile duct is worthy of resection. Ann Surg 2011; 253:953–60. 10. Kondo S, Nimura Y, Hayakawa N, et al. Extensive surgery for carcinoma of the gallbladder. Br J Surg 2002;89:179–84. 11. Shirai Y, Ohtani T, Tsukada K, et al. Combined pancreaticoduodenectomy and hepatectomy for patients with locally advanced gallbladder carcinoma: long term results. Cancer 1997;80:1904–9.
12. Sasaki R, Takahashi M, Funato O, et al. Hepatopancreatoduodenectomy with wide lymph node dissection for locally advanced carcinoma of the gallbladder e long-term results. Hepatogastroenterology 2002; 49:912–5. 13. Miyazaki M, Itoh H, Ambiru S, et al. Radical surgery for advanced gallbladder carcinoma. Br J Surg 1996;83:478–81. 14. Araida T, Yoshikawa T, Azuma T, et al. Indications for pancreatoduodenectomy in patients undergoing lymphadenectomy for advanced gallbladder carcinoma. J Hepatobiliary Pancreat Surg 2004;11:45–9. 15. D’Angelica M, Dalal KM, DeMatteo RP, et al. Analysis of the extent of resection for adenocarcinoma of the gallbladder. Ann Surg Oncol 2009;16:806–16. 16. Capussotti L, Massucco P, Ribero D, et al. Extended lymphadenectomy and vein resection for pancreatic head cancer: outcomes and implications for therapy. Arch Surg 2003;138(12):1316–22. 17. Ferrero A, Russolillo N, Vigano L, et al. Safety of conservative management of bile leakage after hepatectomy with biliary reconstruction. J Gastrointest Surg 2008;12(12):2204–11. 18. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205–13. 19. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. 20. Simmonds PC, Primrose JN, Colquitt JL, et al. Surgical resection of hepatic metastases from colorectal cancer: a systematic review of published studies. Br J Cancer 2006;94(7):982–99. 21. Vigano L, Russolillo N, Ferrero A, et al. Evolution of long-term outcome of liver resection for colorectal metastases: analysis of actual 5-year survival rates over two decades. Ann Surg Oncol 2012;19(6): 2035–44. 22. Kondo S, Nimura Y, Kamiya J, et al. Factors influencing postoperative hospital mortality and long-term survival after radical resection for stage IV gallbladder carcinoma. World J Surg 2003;27:272–7. 23. Dixon E, Vollmer Jr CM, Sahajpal A, et al. An aggressive surgical approach leads to improved survival in patients with gallbladder cancer: a 12-year study at a North American Center. Ann Surg 2005;241: 385–94. 24. Kondo S, Nimura Y, Kamiya J, et al. Five-year survivors after aggressive surgery for stage IV gallbladder cancer. J Hepatobiliary Pancreat Surg 2001;8:511–7. 25. Murakami Y, Uemura K, Sudo T, et al. Prognostic factors of patients with advanced gallbladder carcinoma following aggressive surgical resection. J Gastrointest Surg 2011;15(6):1007–16. 26. Hawkins WG, DeMatteo RP, Jarnagin WR, et al. Jaundice predicts advanced disease and early mortality in patients with gallbladder cancer. Ann Surg Oncol 2004;11:310–5. 27. Regimbeau JM, Fuks D, Bachellier P, et al. Prognostic value of jaundice in patients with gallbladder cancer by the AFC-GBC-2009 study group. Eur J Surg Oncol 2011;37:505–12. 28. D’Angelica M, Martin 2nd RC, Jarnagin WR, et al. Major hepatectomy with simultaneous pancreatectomy for advanced hepatobiliary cancer. J Am Coll Surg 2004;198(4):570–6. 29. Sakamoto Y, Kosuge T, Shimada K, et al. Clinical significance of extrahepatic bile duct resection for advanced gallbladder cancer. J Surg Oncol 2006;94(4):298–306. 30. Miyakawa S, Ishihara S, Horiguchi A, et al. Biliary tract cancer treatment: 5,584 results from the Biliary Tract Cancer Statistics Registry from 1998 to 2004 in Japan. J Hepatobiliary Pancreat Surg 2009; 16(1):1–7.
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
ScienceDirect EJSO xx (2013) 1e8
www.ejso.com
Locally advanced gallbladder cancer: Which patients benefit from resection? D.J. Birnbaum a, L. Vigan o*,a, A. Ferrero, S. Langella, N. Russolillo, L. Capussotti Dept. of HPB and Digestive Surgery, Ospedale Mauriziano Umberto I, Largo Turati 62, 10128 Torino, Italy Accepted 19 October 2013 Available online - - -
Abstract Objectives: Patients with T3e4 gallbladder cancers (GBCs) often require extended surgical procedures, and up to 30% of patients have N2 metastases. This study investigated which patients with T3e4 GBC benefit from resection. Methods: Consecutive patients (n ¼ 78) with T3e4 GBC who underwent resection between 1990 and 2011 were analysed (38 before 2003, 40 in 2003e2011). Forty patients required common bile duct (CBD) resection, 10 pancreatoduodenectomy, 4 right colectomy and 2 gastric resection. Fifty-two (67%) patients had LN metastases, including 22 with N2 metastases. Results: The in-hospital mortality rate was 8%, 11% before 2003 vs. 5% in 2003e2011. The morbidity rate (47%) remained stable during the study. Undergoing liver and pancreatic resection did not increase severe morbidity (0%) or mortality (10%). Sixty-seven (86%) patients had R0 resection. The 5-year survival rate was 17% (median follow-up, 65 months). Survival improved after 2002 (26% vs. 9%, p ¼ 0.04). R1 patients had poor 3-year survival (0% vs. 32%, p ¼ 0.001). Nþ patients also had low survival (5-year survival, 10% vs. 32% in N0, p ¼ 0.019), but N1 and N2 patients had similar outcomes. CBD resection and major hepatectomy did not worsen prognosis. Patients requiring pancreatoduodenectomy, gastric or colonic resection had 0% 3-year survival ( p ¼ 0.036 in multivariate analysis). Conclusions: Resection of T3e4 GBC is worthwhile only if R0 surgery is achievable. Outcomes improved in most recent years. N2 metastases should not preclude surgery. Good results are possible even with CBD resection or major hepatectomy, while benefits from surgery are doubtful if pancreatoduodenectomy or other organ resection is needed. Ó 2013 Elsevier Ltd. All rights reserved. Keywords: Gallbladder cancer resection; Pancreatoduodenectomy; Common bile duct resection; Survival; Liver surgery; Prognostic factors
Introduction Gallbladder cancer (GBC) is the most common biliary tract malignancy worldwide.1 Complete surgery is the only curative therapy, whereas chemotherapy offers a limited contribution to survival.1e4 Because it causes few symptoms, GBC is often diagnosed at an advanced stage and is resectable in only 15%e40% of patients.1e3 Surgical outcome depends strongly on tumour stage: while excellent survival rates of up to 100% at 5 years have been reported for early stage disease (T1e2), overall survival rates range from 0% to 30% in locally advanced tumours (T3e4).5e8 To be completely resected, locally advanced GBCs often require extended surgical procedures, such as major hepatectomy, common bile duct (CBD) resection or pancreatoduodenectomy. In addition, lymph node (LN) metastases
in the celiac and retropancreatic area are detected in up to 30% of patients.8,9 The surgical indications in these patients are still being debated. A few series involving a limited number of patients, mainly in Asian centres, have reported conflicting results.9e15 During the last 20 years, we have adopted an aggressive surgical policy in patients affected by locally advanced GBC. Extended surgical procedures have been performed as long as complete resection was achievable. The present analysis reviewed the outcomes of this aggressive surgical approach, with particular attention to patients who required resection of other organs, e.g. CBD resection or pancreatoduodenectomy. The aim of the study was to identify which patients with locally advanced GBC benefit from resection. Materials and methods
* Corresponding author. Tel.: þ39 011 5082590; fax: þ39 011 5082592. E-mail address: [email protected] (L. Vigano). a Authors contributed equally.
A total of 126 consecutive patients underwent resection for GBC at the Mauriziano Umberto I Hospital of Torino
0748-7983/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejso.2013.10.014 Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
2
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
between January 1990 and December 2011. This study analysed the 78 (62%) patients with T3 or T4 GBC. Patient management Preoperative staging routinely included carcinoembryonic antigen (CEA) and CA 19-9 value determination and thoraco-abdominal computed tomography (CT). Since 1998, magnetic resonance cholangio-pancreatography (MRCP) has been performed in all jaundiced patients or whenever CBD infiltration was suspected at CT. In recent years, contrast-enhanced magnetic resonance imaging and positron emission tomography CT were performed in case of uncertain diagnosis or suspected extra-hepatic disease, respectively. Since 1998, CT-volumetry has been performed regularly in patients scheduled for major hepatectomy to estimate the future liver remnant (FLR) volume. The FLR was considered adequate if >25% of the total liver volume. In jaundiced patients, a higher cut-off of >30% was adopted. If the FLR was inadequate, preoperative portal vein embolization (PVE) was performed, and CT-volumetry was repeated 4 weeks later. Surgery was scheduled only if adequate FLR hypertrophy was observed. In jaundiced patients, preoperative biliary drainage was not performed on a routine basis. Its indications were cholangitis, malnutrition or the need for PVE (drainage of the FLR only). Indications Surgery was performed whenever complete resection was achievable. Distant metastases, extensive infiltration of the hepatoduodenal ligament and interaorto-caval LN metastases were contraindications to surgery. The surgical indication was evaluated on a case-by-case basis when celiac or retropancreatic LN metastases were present. The indications for CBD resection were as follows: macroscopic CBD infiltration; LN metastases of the pedicle infiltrating the CBD; tumoral deposits along the hepatic pedicle; and neoplastic infiltration of the cystic duct stump in frozen section analysis. Pancreatoduodenectomy was performed in case of infiltration of the duodeno-pancreatic area by the GBC or by large retropancreatic LN metastases. Colonic or gastric resection was performed in case of macroscopic infiltration. Surgical procedure The standard scheduled procedure for T3e4 GBC was segment 4b-5 resection (anatomic bisegmentectomy or wide wedge resection), en-bloc cholecystectomy and LN dissection. LN dissection was limited to the hepatic pedicle (D1) in the first part of the study (before 1998) and was extended to the retropancreatic and celiac area (D2) since 1998. Frozen section analysis of the cystic duct stump was performed regularly. In patients with a postoperative
diagnosis of GBC after laparoscopic cholecystectomy, the trocar sites were systematically removed. The surgical techniques used for liver resection, CBD resection and pancreatoduodenectomy have been described previously.16,17 Since 2000, staging laparoscopy with laparoscopic ultrasonography always preceded laparotomy to exclude peritoneal carcinomatosis or liver metastases. Adjuvant treatment and follow-up Postoperative chemotherapy was delivered according to patient performance status and pathological findings. All patients were followed-up every 3 months with a physical examination, CEA and CA 19-9 determinations and abdominal ultrasonography or thoraco-abdominal CT. No patient was lost during the follow-up period. This analysis includes follow-up to August 31, 2012. Definitions Major hepatectomy was defined as resection of 3 Couinaud segments. Operative mortality was defined as death within 90 days after surgery or before discharge from the hospital. Morbidity included all postoperative complications and was classified according to the ClavieneDindo classification.18 Tumour TNM staging was reviewed according to the 7th edition of the AJCC manual.19 Statistical analysis Data were collected prospectively and analysed retrospectively. The series was divided into two periods in order to analyse outcome evolution during the study period. Continuous variables were compared between groups using the unpaired t test or the ManneWhitney U test, as appropriate; categorical variables were compared using the chisquare test or Fisher’s exact test, as appropriate. The KaplaneMeier method was used to estimate overall survival (OS) and recurrence-free survival (RFS) probabilities, which were compared using the log-rank test. Patients with operative mortality or incomplete surgery (R2) were excluded from survival analysis. Multivariate analysis was performed using a Cox proportional hazard model to identify independent prognostic factors of OS and RFS after liver resection. Multivariate analysis was completed for factors with a p value 0.10 in the univariate analysis. A pvalue <0.05 was considered significant for all tests. Results During the study period (1990e2011), 78 patients underwent surgery for T3e4 GBC, 38 before 2003 and 40 in 2003e2011. Of these, 34 (44%) patients were male; the median age was 67 years (range 39e83). One third of the patients (n ¼ 25) was jaundiced at diagnosis [median bilirubin value 195.8 mmol/L (59.9e335.2)]; of these, 14
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
(56%) underwent preoperative biliary drainage, with an endoscopic approach used in 64% of cases. In 17 (22%) patients, the diagnosis of GBC was postoperative, and 16 of the 17 patients were referred to the author’s department after undergoing cholecystectomy at another hospital. One of these patients underwent an emergency cholecystectomy for acute cholecystitis in the author’s department. All 17 patients underwent a second operation after cholecystectomy after a median delay of 63 days (13e102). The patient characteristics are summarized in Table 1. Surgical procedures Liver resection was performed in all but 1 patient (99%). That patient was affected by T3 GBC of the infundibulum that mimicked a mid-CBD cancer and underwent isolated CBD resection. Of the remaining 77 patients, 22 (29%) underwent major hepatectomy (8 underwent preoperative PVE), including 19 extended resections (>4 liver segments). LN dissection was D2 in 60 (77%) patients. Fifty-four patients (69%) required resection of contiguous organs: pancreatoduodenectomy in 10, isolated CBD resection in 40, right colectomy in 4 (associated with pancreatoduodenectomy in 2) and distal gastrectomy in 2. Pancreatoduodenectomy was performed because of macroscopic neoplastic infiltration in 7 patients and because of large retropancreatic LN metastases in 3. CBD resection was performed because of macroscopic neoplastic infiltration in 34 patients, non-contiguous neoplastic tissue along the hepatic pedicle in 3, large hepatic pedicle LN metastases in 2 and cystic duct stump infiltration in 1.
Table 1 Patient characteristics and pathological data. Patients (n ¼ 78) n (%) Patient characteristics and surgical procedures Age, years, median (range) Gender, Female/Male Gallblader lithiasis CEA ng/mL, median (range) Ca 19-9 U/mL, median (range) Preoperative jaundice Gallbladder cancer diagnosis Surgical procedures
Associated resections
Lymph-node dissection Pathology data TNM stagingb
Postoperative outcomes The 90-day mortality rate was 8% (n ¼ 6). The causes of death included the following: hepatic failure in 2 patients, sepsis in 1, hemoperitoneum in 1, acute myocardial infarction in 1 and acute respiratory distress syndrome in 1. The overall morbidity rate was 47% (n ¼ 37), including 16 (21%) grade III/IV complications. The complications are listed in Table 2. Patients undergoing standard surgical procedures (bisegmentectomy Sg4b-5 and LN dissection) had no mortality and no severe complications (grade III/IV). Patients requiring resection of contiguous organs had higher mortality (11% vs. 0%, p ¼ 0.090), overall morbidity (63% vs. 13%, p ¼ 0.009) and grade IIIeIV morbidity rates (30% vs. 0%, p ¼ 0.002). Of the patients requiring pancreatoduodenectomy, there was 1 operative mortality (10%) and no grade IIIeIV morbidities. The mortality rate decreased across the study period (11% before 2003 vs. 5% in 2003e2011, p ¼ n.s.) Overall and grade IIIeIV morbidity rates remained stable during the study period. Adjuvant chemotherapy was administered to 30 (38%) patients.
3
Completeness of surgery
Preoperative Intraoperative Postoperative Cholecystectomy Sg4b-5 resection Sg 5,6,7,8 [Sg1] Sg2,3,4,5 [Sg1] Sg4,5,6,7,8 [Sg1] Associated Sg1 resection Isolated CBD resection Pancreatoduodenectomy Right colectomy Distal gastrectomy D1 D2 T3 T4 N0 N1 N2 Stage 3A Stage 3B Stage 4A Stage 4B R0 R1 Liver resection margin Biliary margin [proximal/distal] Radial margin R2
Perineural infiltration Biliary duct infiltration/Suspected macroscopic infiltration Pancreatic infiltration/Suspected macroscopic infiltration Colonic & gastric infiltration/Suspected macroscopic infiltration
67 (39e83) 44 (56)/34 (44) 55 (71) 2.8 (0.6e657) 89.6 (0.8e4778) 25 (32) 49 (63) 12 (15) 17 (22) 1 (1) 55 (71) 3 (4) 2 (3) 17 (22) 8 (10) 40 (51) 10 (13)a 4 (5)a 2 (3) 18 (23) 60 (77) 62 (79) 16 (21) 26 (33) 30 (38) 22 (28) 23 (29) 25 (32) 8 (10) 22 (28) 67 (86) 10 (13) 2 5 [3/2] 5 1 (1)d 63 (81) 27/34c 7/7 6/6
CEA: Carcinoembryonic antigen; Sg: segment; CBD: common bile duct. a Two patients had combined pancreatoduodenectomy and right colectomy. b According to the 7th edition of the AJCC manual.22 c Patients with isolated CBD resection were considered. d One patient had misdiagnosed liver metastasis (R2 patient).
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
4
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
Pathological data The pathological data are summarized in Table 1. The tumour staging was T3 in 62 (79%) patients and T4 in 16 (21%). The histologic tumour type was adenocarcinoma in 72 (92%) patients and squamous carcinoma in 6 (8%). The median number of retrieved LNs was 7 (1e35), 4 after D1 dissection and 9 after D2 ( p ¼ 0.005). The number of retrieved LNs was similar between patients with and without CBD resection. Fifty-two (67%) patients had LN metastases; the metastases extended to the retropancreatic or the celiac area (N2) in 22 (28%). Pathological examination showed that 33 (42%) patients had CBD infiltration (27 during isolated CBD resections and 6 during pancreatoduodenectomy). Suspected macroscopic infiltration of the CBD was confirmed in 79% of patients, while suspected infiltration of the pancreas, the stomach and the right colon was confirmed in all patients. The resection was complete (R0) in 67 (86%) patients, while 10 (13%) patients had R1 resection (positive biliary margin in 5). One patient had R2 resection: intraoperative ultrasonography detected a suspected liver metastasis, and the biopsy was negative in frozen section analysis but positive for adenocarcinoma at final pathology. Patients requiring resection of contiguous organs had a higher R1 resection rate, but the difference was not significant (16% vs. 8%, p ¼ n.s.) All patients that underwent pancreatoduodenectomy had R0 resection. Overall survival data After a median follow-up of 65 months (6e114), the 1-, 3- and 5-year OS rates were 66%, 27% and 17%, Table 2 Short-term outcomes. Short-term outcome (n ¼ 78) n (%) 90-day mortality Overall morbidity Liver failure Bile leak Hemoperitoneum Abdominal collection Ascites Sepsis Pancreatic leak Duodenal leak Pulmonary morbidity Pulmonary embolism Renal failure Acute myocardial infarction Small bowel disorders Dindo & Clavien classification18 Grade IeII Grade III [a/b] Grade IV [a/b] Blood transfusion Hospital stay days, median (range)
6 37 7 15 4 3 6 8 3 1 12 1 3 1 1
(8) (47) (9) (19) (5) (4) (8) (10) (4) (1) (15) (1) (4) (1) (1)
15 (19) 11 [7/4] (14) 5 [4/1] (6) 12 (15) 13 (0e83)
respectively. The median survival was 16 months. Nine (12%) patients survived more than 5 years after resection: all 9 had T3 GBC and R0 resection; 2 were N1 and 1 was N2; 3 had CBD resection (1 had CBD infiltration at final pathology). OS improved in recent years: 5-year survival was 9% in patients operated on before 2003 vs. 26% in patients treated in 2003e2011 ( p ¼ 0.037). Recurrence analysis and recurrence-free survival data Fifty-nine patients (76%) had recurrence: hepatic in 26 (33%), extra-hepatic in 29 (37%) and both hepatic and extra-hepatic in 4 (5%). Of the extra-hepatic recurrence sites, the most common ones were the peritoneum (n ¼ 11), the hepatoduodenal ligament (n ¼ 9) and the lung (n ¼ 6). Of the 9 patients with hepatoduodenal ligament recurrence, 7 had CBD resection because of infiltration and 5 had R1 resection. All 9 of the patients had perineural infiltration. The risk of local recurrence was increased in patients with CBD infiltration (24% vs. 5%, p ¼ 0.027) and R1 status (50% vs. 7%, p ¼ 0.0002). N status and the extension of LN dissection did not correlate with local recurrence risk. The 1-, 3- and 5-year RFS rates were 44%, 22% and 16%, respectively (median 10 months). Four patients had recurrence resection (1 liver, 1 lung, 1 carcinosis, 1 adrenal gland); the median OS after re-resection was 49 months (15e114). RFS improved in patients treated in 2003e2011 (21% vs. 7% in patients treated before 2003), but the difference was not significant ( p ¼ 0.056). Prognostic factors of OS and RFS Table 3 shows the results of univariate and multivariate analysis of OS and RFS prognostic factors. A survival benefit was evident only if complete surgery was achieved: 3-year OS was 32% in R0 patients vs. 0% in R1 patients ( p ¼ 0.002) (Fig. 1(a)); 3-year RFS was 26% vs. 0%, respectively ( p ¼ 0.0004). Similar unfavourable results were observed in patients with T4 tumours (0% 3-year OS vs. 33% in T3; 0% 3-year RFS vs. 26% in T3), but the difference was not significant (Fig. 1(b)). LN status correlated with prognosis (5-year OS 32% if N0 vs. 10% if Nþ, p ¼ 0.019; RFS 28% vs. 10%, respectively, p ¼ 0.013), but OS and RFS were similar in N1 and N2 patients (11% vs. 8% and 8% vs. 7%, respectively, p ¼ n.s.) (Fig. 1(c)). According to the TNM staging, 5-year OS was 35% in patients with stage 3A tumours, 13% in 3B, 0% in 4A and 8% in 4B ( p ¼ n.s.; stage 3A vs. others p ¼ 0.019). Similarly, RFS was better in patients with stage 3A tumours (31% vs. 9%, p ¼ 0.010), while no differences were observed among patients with stage 3B, 4A and 4B tumours. Patients with a diagnosis of GBC after a simple cholecystectomy who underwent a second operation had OS and
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
5
Table 3 Univariate and multivariate analysis of prognostic factors of overall and recurrence-free survival. Parameter
Age (years) Sex Jaundice at diagnosis Major hepatectomy CBD resection CBD infiltration DP/Gastric/Colonic res. T stage N stage TNM staging
Complete surgery Perineural infiltration Adjuvant CTx
70 >70 Male Female Y N Y N Y N Y N Y N T3 T4 N0 N+ 3A 3B 4A 4B R0 R1 Y N Y N
Median OS (months)
5-yr OS (%)
OS
15 23 24 15 24 16 13 16 15 17 15 16 16 16 15 16 24 14 24 11 25 15 17 10 15 95 15 16
16 20 26 11 15 19 14 19 11 29 5 24 0a 21 21 0a 32 10 35 13 0a 8 20 0a 6 58 24 14
n.s. 0.048
RFS
Univariate
Multivariate analysis
Univariate
Multivariate analysis
p
p
HR (CI95%)
p
p
e
n.s.
e
n.s.
0.017
n.s.
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
n.s.
e
0.081
n.s.
2.962 (1.064e8.246) 1 e
n.s.
3.877 (1.369e10.981) 1 e
0.019
n.s.
0.013
n.s.
n.s.
e
n.s.
e
1 3.0.39 (1.335e6.917) 4.208 (1.426e12.414) 1 e
0.0004
0.022
0.006
0.014
0.085
0.038
0.002
0.008
0.002
0.009
n.s.
n.s.
0.011
HR (CI95%)
1 2.929 (1.169e7.338) 3.491 (1.295e9.415) 1 e
OS: overall survival; RFS: recurrence-free survival; 5-yr: 5-year; HR: hazard ratio; CBD: common bile duct; DP: pancreatoduodenectomy; res.: resection; CTx: chemotherapy. a No patient alive at three years after surgery.
RFS rates similar to patients with pre-/intra-operative diagnosis of GBC. Patients requiring major hepatectomy had OS and RFS rates similar to those undergoing minor resection. Patients with and without CBD resection had similar OS (Fig. 2(a)) and RFS. CBD infiltration confirmed at pathology examination was associated with lower OS and RFS (at 5 years: OS of 5% vs. 24%; RFS of 6% vs. 22%), but the difference was not significant (Fig. 2(b)). Patients requiring pancreatoduodenectomy had 0% OS and RFS 2 years after surgery, and patients requiring gastric or colonic had 0% OS and RFS 3 years after surgery (vs. 21% OS and 20% RFS at 5 years in the remaining patients, p ¼ 0.085 for OS and p ¼ 0.081 for RFS) (Fig. 2(c)). The multivariate analysis showed that the need for resection of contiguous organs other than the CBD (i.e. pancreatoduodenectomy, gastric and colonic resection) was an independent negative prognostic factor of OS (hazard ratio [HR] ¼ 2.962, p ¼ 0.038). Two additional negative prognostic factors were identified: R1 resection (HR ¼ 3.039, p ¼ 0.008) and perineural infiltration (HR ¼ 4.208,
p ¼ 0.009). The same parameters were independent negative predictive factors of RFS: resection of contiguous organs other than the CBD (HR ¼ 3.877, p ¼ 0.011); R1 resection (HR ¼ 2.929, p ¼ 0.022) and perineural infiltration (HR ¼ 3.491, p ¼ 0.014). Discussion The present study helps refine indications in patients treated surgically for locally advanced (T3e4) GBC. Only patients with complete resection (R0) showed a survival benefit from resection. N status predicted outcome, but there were long-term survivors in the group of Nþ patients, even when the patients had celiac or retropancreatic LN metastases. Patients requiring major hepatectomy or CBD resection had good survival outcomes; thus, such patients should be scheduled for surgery whenever possible. In contrast, patients requiring pancreatoduodenectomy or other organ resection (stomach or right colon) had extremely poor outcomes and showed no clear survival benefit from surgery.
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
6
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
Figure 1. Overall survival curves after gallbladder cancer resection according to the pathological data: (a) R status (R0 vs. R1); (b) T stage (T3 vs. T4); (c) N stage (N0 vs. N1 vs. N2). Patients with operative mortality (n ¼ 6) and R2 resection (n ¼ 1) were excluded.
GBC is an aggressive cancer: most patients present with advanced disease at diagnosis, and only a minority can be scheduled for curative treatment.1e3 To achieve complete resection, extended procedures may be required, such as CBD resection and pancreatoduodenectomy. The benefits from surgery in such aggressive tumours are debatable. There are few studies of this population, and most have been conducted in Asian centres and include a limited number of patients treated over a long time period.9e15 When considering treatment strategy, the first thing to consider is operative risk. Even though short-term outcomes after liver surgery have improved in recent years,20,21 mortality and morbidity rates after GBC resection remain high, ranging from 5% to 19% and from 14% to 53%, respectively.9e15,22 Similar figures were obtained in the present series: the mortality rate was 8%, and morbidity occurred in half of the patients. The need for aggressive surgical procedures could explain these data. High operative risks have been reported for liver plus pancreatic resection,10,13,22 but this was not the case in our experience. Specifically, the mortality rate was not increased after combined liver and pancreatic resection and there was no grade IIIeIV morbidity. The short-term results could reflect the long time interval analysed in most series. In recent years, patient selection has improved, as has surgical expertise. Preoperative biliary drainage in
jaundiced patients and PVE in those scheduled for a major hepatectomy allow adequate preparation for resection. In fact, mortality rates decreased during the study period from 11% before 2003 to 5% in 2003e2011. Nevertheless, mortality rates are still high compared with those reported for other pathologies, such as colorectal metastases (1e2%),20,21 and the rate of severe complications did not decrease during the study period. Provided there are good short-term results, the question must be asked: which patients benefit from resection? Considering long-term outcomes, the consensus is that aggressive surgery is worthwhile as long as complete resection can be achieved.2,9,15,23e25 In the Nagoya series, as in the present one, R1 resection was associated with 0% survival at 3 years.9 This raises another question: should all resectable patients be resected? Jaundice and CBD involvement have been reported to be predictors of poor outcome.3,9,15,26 For a long time, these two conditions were considered a contraindication for resection, but recent results have been encouraging.9,27 Nishio et al. reported a 5-year OS of 36% in patients with CBD resection.9 In the present study, patients requiring CBD resection had OS and RFS that were similar to those without CBD resection, and there were even three 5-year survivors. As in the Japanese study, jaundice at diagnosis showed no prognostic impact. CBD infiltration confirmed at final pathology was associated with worse outcome and higher risk of local recurrence, but did not
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
7
Figure 2. Overall survival curves after gallbladder cancer resection according to the performed resection: (a) common bile duct (CBD) resection (Yes vs. No); (b) CBD infiltration at final pathology (Yes vs. No); (c) pancreatoduodenectomy vs. gastric/colonic resection vs. no associated resection. Patients with operative mortality (n ¼ 6) and R2 resection (n ¼ 1) were excluded.
preclude long-term survival and thus should not be considered a contraindication for surgery. In some patients, pancreatoduodenectomy is needed to achieve complete resection because of direct neoplastic infiltration of the duodeno-pancreatic area or because of large retropancreatic LN metastases. Only a few cases have been described.10e12,14,28 Some authors reported excellent survival of up to 40% at 5 years.11,12,14 Araida et al. even suggested that pancreatoduodenectomy allows removal of micrometastasis in small peri-pancreatic nodes and is associated with better survival compared with nonpancreatoduodenectomy patients (5-year OS of 87% vs. 17%).14 Both T2 and T3e4 GBC were included in that study. The benefits of pancreatoduodenectomy were lost when the analysis included just patients with hepatoduodenal ligament invasion (5-year OS of 0e5% according to LN status). In the present series, only patients with locally advanced GBC were included, and pancreatoduodenectomy was performed if needed to achieve R0 status. In such patients, survival rates were extremely poor (0% at 2 years) even if surgery was complete in all patients. Similar results were observed in patients requiring colonic or gastric resection because of neoplastic infiltration (0% OS at 3 years). These poor outcomes cast doubt on the benefits of surgery. LN status is another important issue. It is consistently one of the strongest predictors of survival in patients with
advanced GBC,2,8,22,25,30 and LN metastasis sites may stratify prognosis in Nþ patients.2,22,30 In the current series, patients with LN metastases had lower OS and RFS compared with N0 patients, but long survival was even observed in Nþ patients. Furthermore, N1 and N2 patients had similar outcomes. According to the present results, LN status has a prognostic value, but the presence of LN metastases, even in the celiac or retropancreatic area, is not a contraindication to resection. Finally, this analysis identified perineural infiltration as a strong prognostic factor (5-year OS of 6% if present). The prognostic role of perineural infiltration has been reported previously.26,29 In T2 GBC, perineural infiltration has even been proposed as a criterion for systematic CBD resection to improve the completeness of surgery and survival.29 Perineural infiltration probably reflects tumour spread in the hepatoduodenal ligament and may predict microscopic residual disease. In fact, we found that local recurrence occurred only in patients with perineural infiltration and was associated with R status. However, it is impossible to preoperatively assess perineural infiltration, which limits its clinical usefulness. The present study has some limitations. This is a retrospective study that analysed a limited number of patients that were treated over a 20-year period. Some prognostic factors, such as the T stage or TNM stage, could not be identified
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014
8
D.J. Birnbaum et al. / EJSO xx (2013) 1e8
because of the small cohort size. The rarity of GBC and its low resectability rate makes it impossible to study a large group of patients at a single centre. Only prospective multicentre studies could overcome these limitations and should be conducted in the future. Nevertheless, the present study clarifies some of the controversial issues related to surgical indications in locally advanced GBC. In conclusion, in patients with T3e4 GBC, surgery is worthwhile only if R0 resection is achievable. Overall, outcomes have improved in recent years. N status has prognostic value, but the presence of LN metastases, even in the celiac or retropancreatic area, should not preclude surgery. In resectable patients, the indication for surgery should be based on the required procedure: good results are possible if CBD resection or major hepatectomy are required, while survival benefits from surgery are doubtful if pancreatoduodenectomy or resection of other organs are needed. Conflict of interest statement None declared.
References 1. Misra S, Chaturvedi A, Misra NC, et al. Carcinoma of the gallbladder. Lancet Oncol 2003;4:167–76. 2. Hueman MT, Vollmer Jr CM, Pawlik TM. Evolving treatment strategies for gallbladder cancer. Ann Surg Oncol 2009;16:2101–15. 3. Fong Y, Jarnagin W, Blumgart LH. Gallbladder cancer: comparison of patients presenting initially for definitive operation with those presenting after prior noncurative intervention. Ann Surg 2000;232:557–69. 4. Valle J, Wasan H, Palmer DH, et al. ABC-02 Trial Investigators. Cisplatin plus gemcitabine versus gemcitabine for biliary tract cancer. N Engl J Med 2010;362(14):1273–81. 5. Wakai T, Shirai Y, Yokoyama N, et al. Early gallbladder carcinoma does not warrant radical resection. Br J Surg 2001;88(5):675–8. 6. Pawlik TM, Gleisner AL, Vigano L, et al. Incidence of finding residual disease for incidental gallbladder carcinoma: implications for re-resection. J Gastrointest Surg 2007;11(11):1478–86. 7. Muratore A, Amisano M, Vigano L, et al. Gallbladder cancer invading the perimuscular connective tissue: results of reresection after prior non-curative operation. J Surg Oncol 2003;83(4):212–5. 8. Araida T, Higuchi R, Hamano M, et al. Hepatic resection in 485 R0 pT2 and pT3 cases of advanced carcinoma of the gallbladder: results of a Japanese Society of Biliary Surgery survey e a multicenter study. J Hepatobiliary Pancreat Surg 2009;16:204–15. 9. Nishio H, Ebata T, Yokoyama Y, et al. Gallbladder cancer involving the extrahepatic bile duct is worthy of resection. Ann Surg 2011; 253:953–60. 10. Kondo S, Nimura Y, Hayakawa N, et al. Extensive surgery for carcinoma of the gallbladder. Br J Surg 2002;89:179–84. 11. Shirai Y, Ohtani T, Tsukada K, et al. Combined pancreaticoduodenectomy and hepatectomy for patients with locally advanced gallbladder carcinoma: long term results. Cancer 1997;80:1904–9.
12. Sasaki R, Takahashi M, Funato O, et al. Hepatopancreatoduodenectomy with wide lymph node dissection for locally advanced carcinoma of the gallbladder e long-term results. Hepatogastroenterology 2002; 49:912–5. 13. Miyazaki M, Itoh H, Ambiru S, et al. Radical surgery for advanced gallbladder carcinoma. Br J Surg 1996;83:478–81. 14. Araida T, Yoshikawa T, Azuma T, et al. Indications for pancreatoduodenectomy in patients undergoing lymphadenectomy for advanced gallbladder carcinoma. J Hepatobiliary Pancreat Surg 2004;11:45–9. 15. D’Angelica M, Dalal KM, DeMatteo RP, et al. Analysis of the extent of resection for adenocarcinoma of the gallbladder. Ann Surg Oncol 2009;16:806–16. 16. Capussotti L, Massucco P, Ribero D, et al. Extended lymphadenectomy and vein resection for pancreatic head cancer: outcomes and implications for therapy. Arch Surg 2003;138(12):1316–22. 17. Ferrero A, Russolillo N, Vigano L, et al. Safety of conservative management of bile leakage after hepatectomy with biliary reconstruction. J Gastrointest Surg 2008;12(12):2204–11. 18. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205–13. 19. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. 20. Simmonds PC, Primrose JN, Colquitt JL, et al. Surgical resection of hepatic metastases from colorectal cancer: a systematic review of published studies. Br J Cancer 2006;94(7):982–99. 21. Vigano L, Russolillo N, Ferrero A, et al. Evolution of long-term outcome of liver resection for colorectal metastases: analysis of actual 5-year survival rates over two decades. Ann Surg Oncol 2012;19(6): 2035–44. 22. Kondo S, Nimura Y, Kamiya J, et al. Factors influencing postoperative hospital mortality and long-term survival after radical resection for stage IV gallbladder carcinoma. World J Surg 2003;27:272–7. 23. Dixon E, Vollmer Jr CM, Sahajpal A, et al. An aggressive surgical approach leads to improved survival in patients with gallbladder cancer: a 12-year study at a North American Center. Ann Surg 2005;241: 385–94. 24. Kondo S, Nimura Y, Kamiya J, et al. Five-year survivors after aggressive surgery for stage IV gallbladder cancer. J Hepatobiliary Pancreat Surg 2001;8:511–7. 25. Murakami Y, Uemura K, Sudo T, et al. Prognostic factors of patients with advanced gallbladder carcinoma following aggressive surgical resection. J Gastrointest Surg 2011;15(6):1007–16. 26. Hawkins WG, DeMatteo RP, Jarnagin WR, et al. Jaundice predicts advanced disease and early mortality in patients with gallbladder cancer. Ann Surg Oncol 2004;11:310–5. 27. Regimbeau JM, Fuks D, Bachellier P, et al. Prognostic value of jaundice in patients with gallbladder cancer by the AFC-GBC-2009 study group. Eur J Surg Oncol 2011;37:505–12. 28. D’Angelica M, Martin 2nd RC, Jarnagin WR, et al. Major hepatectomy with simultaneous pancreatectomy for advanced hepatobiliary cancer. J Am Coll Surg 2004;198(4):570–6. 29. Sakamoto Y, Kosuge T, Shimada K, et al. Clinical significance of extrahepatic bile duct resection for advanced gallbladder cancer. J Surg Oncol 2006;94(4):298–306. 30. Miyakawa S, Ishihara S, Horiguchi A, et al. Biliary tract cancer treatment: 5,584 results from the Biliary Tract Cancer Statistics Registry from 1998 to 2004 in Japan. J Hepatobiliary Pancreat Surg 2009; 16(1):1–7.
Please cite this article in press as: Birnbaum DJ, et al., Locally advanced gallbladder cancer: Which patients benefit from resection?, Eur J Surg Oncol (2013), http://dx.doi.org/10.1016/j.ejso.2013.10.014