Fax: (410) 614–4667
Is resection of periampullary or pancreatic adenocarcinoma with synchronous hepatic metastasis justified?†
Version of Record online: 16 OCT 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 11, pages 2484–2492, 1 December 2007
How to Cite
Gleisner, A. L., Assumpcao, L., Cameron, J. L., Wolfgang, C. L., Choti, M. A., Herman, J. M., Schulick, R. D. and Pawlik, T. M. (2007), Is resection of periampullary or pancreatic adenocarcinoma with synchronous hepatic metastasis justified?. Cancer, 110: 2484–2492. doi: 10.1002/cncr.23074
Presented at the Annual Meeting of the American Hepato-Pancreato-Biliary Association, Las Vegas, Nevada, April 18–22, 2007.
- Issue online: 19 NOV 2007
- Version of Record online: 16 OCT 2007
- Manuscript Accepted: 13 JUN 2007
- Manuscript Revised: 11 JUN 2007
- Manuscript Received: 30 APR 2007
- International Union Against Cancer/American Cancer Society
- American Cancer Society
- Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior
To date, no consensus has been reached regarding which primary tumor subtypes are managed appropriately with hepatic metastectomy. Specifically, the role of hepatic resection for metastatic periampullary or pancreatic adenocarcinoma remains controversial.
Between 1995 and 2005, 1563 patients underwent surgical resection for periampullary carcinoma (n = 608 patients) or pancreatic adenocarcinoma (head, n = 905 patients; tail, n = 50 patients). Data on demographics, operative details, primary tumor status, and—when indicated—extent of hepatic metastasis were collected.
Of the 1563 patients who underwent resection of periampullary or pancreatic adenocarcinoma, 22 patients (1.4%) underwent simultaneous hepatic resection for synchronous liver metastasis. The primary tumor site was ampullary (n = 1 patient ), duodenal (n = 2 patients), distal bile duct (n = 2 patients), or pancreas (head, n = 10 patients; tail, n = 7 patients). The majority of patients (86.4%) had a solitary hepatic metastasis, and the median size of the largest lesion was 0.6 cm. Hepatic metastectomy included wedge resection (n = 20 patients), segmentectomy (n = 1 patient), and hemihepatectomy (n = 1 patient). After matching patients on primary tumor histology and location, the median survival of patients who underwent hepatic resection of synchronous metastasis was 5.9 months compared with 5.6 months for patients who underwent palliative bypass alone (P = .46) and 14.2 months for patients with no metastatic disease who underwent primary tumor resection only (P < .001). Pancreatic (median, 5.9 months) versus nonpancreatic (median, 9.9 months) primary tumor histology was not associated with a difference in survival in patients who underwent resection of synchronous liver metastasis (P = .43).
Even in well selected patients with low-volume metastatic liver disease, simultaneous resection of periampullary or pancreatic carcinoma with synchronous liver metastases did not result in long-term survival in the overwhelming majority of patients. Cancer 2007. © 2007 American Cancer Society.
Surgical resection of adenocarcinoma of the pancreas or periampullary area provides the only possibility of long-term survival. Although resection traditionally had been associated with relatively high morbidity and mortality, over the past several decades, advances in anesthesia, operative techniques, and postoperative care have led to marked improvements in perioperative outcomes.1, 2 Despite these improvements, the 5-year survival rate after surgical resection remains poor, ranging from 26% for patients with pancreatic adenocarcinoma2–5 to 70% for patients with ampullary carcinoma.6–8 Patients who have advanced locoregional or metastatic disease identified at the time of exploration have an even worse prognosis.9 Thus, there has been considerable debate regarding the best way to manage this cohort of patients. For those individuals with advanced locoregional disease, Lillemoe et al.10 reported that palliative pancreaticoduodenectomy resulted in improved long-term survival compared with traditional surgical palliation—even in patients with residual local disease. The role of surgical resection in the setting of evident low-volume metastatic disease, however, is much more controversial. Specifically, in the setting of hepatic metastasis, resection of the primary pancreatic or periampullary adenocarcinoma has been discouraged.
Although various studies have noted encouraging long-term results after liver resection of metastatic noncolorectal, nonneuroendocrine neoplasms,11–14 specific data on hepatic resection of metastatic disease from primary hepatobiliary lesions are limited. Adam et al.15 recently reported a 20% to 25% 5-year survival rate for patients who underwent hepatic resection of metastatic lesions from pancreaticobiliary primary tumors. However, the overwhelming majority of those patients had metachronous metastatic disease and underwent a staged procedure after a significant disease-free interval. Information on whether these results can be extrapolated to patients who present with synchronous disease (eg, primary tumor plus liver metastasis) is lacking. Therefore, the objective of the current study was to assess the efficacy of simultaneous resection of primary periampullary or pancreatic adenocarcinoma with synchronous hepatic metastatic disease.
MATERIALS AND METHODS
Between January 1995 and April 2005, 1563 patients underwent surgical resection of periampullary carcinoma (n = 608 patients) or pancreatic adenocarcinoma (pancreatic head, n = 905 patients; pancreatic tail, n = 50 patients) at the Johns Hopkins Hospital. All data were collected prospectively in a database that was approved by the Institutional Review Board. Of the 1563 patients who underwent resection of periampullary or pancreatic adenocarcinoma, 22 patients (1.4%) underwent simultaneous resection of their primary periampullary/pancreatic malignancy in conjunction with simultaneous hepatic resection for synchronous liver metastasis. Only patients with histologically confirmed adenocarcinoma of the periampullary region or pancreatic adenocarcinoma who were treated with curative intent (complete extirpation of all measurable disease) were included in the current study. Patients with extrahepatic metastatic disease, including those with serosal implants or peritoneal disease, were not included in the study. In addition, patients with primary cystadenocarcinomas or neuroendocrine tumors were excluded from consideration. Patients were evaluated preoperatively with a baseline history and physical examination, serum laboratory tests, pancreatic protocol computed tomography studies, and a chest radiograph.
To assess the relative prognostic impact of resecting simultaneous hepatic metastasis, a matched- controlled analysis was performed. Specifically, 2 matched groups were used for the purpose of analytic comparison: 1) patients with no evidence of hepatic metastases who underwent resection of their primary tumor (eg, “resection controls”); and 2) patients with histologically proven liver metastases who underwent standard surgical palliation (eg, “palliative controls”). Surgical palliation entailed a traditional “double-bypass” characterized by a hepaticojejunostomy as well as a gastrojejunostomy. Only patients who were considered unresectable because of the presence of the liver metastases were included in the palliative control group. Patients with extrahepatic distant metastases were not included in the palliative control group.
The following data were collected for each group: demographics; primary tumor histology, location, and size; primary tumor operative details; disease status; date and status at last follow-up; and date of death. For patients with hepatic metastases (eg, the palliative control group and the simultaneous resection group), the location, size, and number of liver lesions also were recorded. At time of laparotomy, the location of the hepatic metastases and their relation to surrounding vascular and biliary structures dictated whether or not a formal anatomic resection was performed. In general, the extent of hepatic resection also was determined by the distribution of the metastatic lesions within the liver. Hepatic resection was classified as less than a hemihepatectomy (eg, segmentectomy or subsegmentectomy), hemihepatectomy, or extended hepatectomy.16 For both the primary tumor and the hepatic metastases, tumor size and number were defined by the resection specimen and/or by intraoperative ultrasonographic measurement. In addition, data on perioperative complications and adjuvant chemotherapy and radiation therapy were collected for all patients.
Cases (eg, patients who underwent simultaneous resection of the primary tumor plus hepatic resection) were matched separately in a 1:3 fashion with each control group (eg, resection controls and palliative controls). Specifically, cases and controls were matched on primary tumor characteristics (tumor size, primary tumor histology, primary site of tumor, year of operation). Morbidity, mortality, and overall survival were compared among the cases and the 2 control groups. Summary statistics were reported using mean or median values as appropriate with the associated intraquartile ranges. Student t tests or analyses of variance were used for mean comparison of variables that were distributed normally, and the Kruskal-Wallis test was used to compare skewed continuous variables. Chi-square statistics were used to compare frequencies of categorical variables among groups. Long-term survival was estimated using the nonparametric product-limit method (Kaplan and Meier).17 Differences in survival were examined using the log-rank test. All statistical analyses were performed using SPSS software (version 11.5; SPSS Inc., Chicago, Ill).
Table 1 shows the clinicopathologic features of the 22 patients who underwent simultaneous resection of their primary tumor as well as hepatic resection of liver metastases. The mean patient age was 65 years (range, 53–82 years). In most patients, the primary tumor site was the pancreas (head, n = 10 patients; tail, n = 7 patients). Other primary sites included the duodenum (n = 2 patients), the distal bile duct (n = 2 patients), and the ampulla (n = 1 patient). The mean size of the primary tumor was 5.0 cm (range, 3–6.6 cm), and the majority of patients (n = 19 patients; 86.4%) had lymph node metastasis associated with their primary tumor. With regard to the extent of metastatic disease within the liver, the majority of patients (n = 20 patients; 90.9%) had a solitary hepatic lesion, and 2 patients had 2 lesions each. The mean size of the largest lesion was 0.6 cm (range, 0.3–1.2 cm). Surgery was performed for known synchronous disease in 1 patient (4.5%), and hepatic metastases were identified incidentally at the time of laparotomy in 21 patients (95.5%).
|Variable||No. of patients (%), n = 22|
|Mean age ± SD, y||64.7 ± 11.4|
|Sex (% men)||10 (45.5)|
|Race (% white)||19 (86.4)|
|Primary tumor site|
|Median tumor size [IQR], cm||5.0 [3–6.6]|
|Lymph node disease||19 (86.4)|
|Origin of tumor|
|Pancreas head||10 (45.5)|
|Pancreas tail||7 (31.8)|
|Distal bile duct||2 (9.1)|
|Median size of largest metastasis [IQR], cm||0.6 [0.3–1.2]|
|Solitary metastases||19 (86.4)|
|Median no. of metastases [IQR]||1 [1–1]|
Details of Surgical and Adjuvant Therapy
At the time of operation, surgical treatment of the primary adenocarcinoma involved a pancreaticoduodenectomy (n = 15 patients; 68.2%) or distal pancreatectomy (n = 7 patients; 31.8%). With regard to the management of synchronous liver metastases, all 22 patients underwent a concomitant hepatic resection. No patient was treated with tumor ablation either as an isolated or a combined modality. Most patients underwent a nonanatomic resection (n = 20 patients; 90%), whereas a minority (n = 2 patients; 9%) underwent an anatomic resection (Table 2). Specifically, the extent of hepatic resection was a wedge resection in 20 patients (90%), a segmentectomy in 1 patient (4.5%), and a hemihepatectomy in 1 patient (4.5%). The median operative time was 370 minutes (range, 338–444 minutes), and the median estimated blood loss was 625 mL (range, 400–925 mL). On final pathologic analysis of the liver specimen, no patient had a macroscopically positive margin; the margin status was microscopically positive in 7 patients (31.8%) and microscopically negative in 15 patients (68.2%).
|Type of operative procedure||No. of patients (%)|
|Distal pancreatectomy||7 (31.8)|
|Wedge resection||20 (90.1)|
Seven patients (31.8%) received some form of adjuvant chemotherapy. Six patients had a pancreatic adenocarcinoma primary lesion and received either systemic 5-flurouracil (n = 3 patients) or gemcitabine (n = 3 patient). The other patient had a duodenal adenocarcinoma primary; this patient received adjuvant combination therapy with 5-flurouracil, leucovorin, and irinotecan (FOLFIRI).
The perioperative complication rate was 45.5%. Several complications were minor and were not necessarily related to the hepatic resection (Table 3). These included urinary retention (n = 2 patients; 9.1%), refractory pain (n = 1 patient; 4.5%), and superficial wound infection (n = 1 patient; 4.5%). Intermediate morbidity complications included events such as pleural effusion (n = 1 patient; 4.5%) and a perihepatic abscess (n = 1 patient; 4.5%). Pancreas-specific complications included delayed gastric emptying (n = 2 patients; 9.1%), pancreatic fistula (n = 2 patients; 9.1%), superior mesenteric vein injury (n = 1 patient; 4.5%), and postoperative bleeding (n = 1 patient; 4.5%). One patient developed a perihepatic abscess, sepsis, and multiorgan failure and died on postoperative Day 17. In total, 2 patients died within 30-days of surgery for a perioperative mortality rate of 9.1%.
|Type of complication||No. of patients|
|Delayed gastric empty||2|
|Portal/superior mesenteric vein injury||1|
|Total complication rate||45.5%|
The median time to recurrence was 3.2 months (range, 1.4–35.7 months). The median overall survival was 5.9 months, and the 1- and 3-year actuarial overall survival rates were 13.3% and 6.7%, respectively (Fig. 1). Univariate analyses failed to reveal any factor that was associated with overall survival, including primary tumor or hepatic margin status, tumor diameter, or primary lymph node status. Pancreatic primary tumor histology (median overall survival, 5.9 months) versus nonpancreatic primary tumor histology (median overall survival, 9.9 months) also was not associated with a difference in overall survival in patients who underwent resection of synchronous liver metastasis (P = .43) (Fig. 2). It is noteworthy that the longest actual survival at the time of this writing was 39.5 months. This patient initially presented with a duodenal adenocarcinoma and underwent a pancreaticoduodenectomy with simultaneous wedge liver resection followed by adjuvant FOLFIRI. He remained disease-free for nearly 3-years until he developed 2 recurrent lesions in segments 6 and 7 of the liver. After a short course of systemic chemotherapy, the patient underwent a repeat hepatic resection (right posterior sectionectomy) and is currently disease-free. Adjuvant therapy was planned but not administered as such in several patients because of early identified recurrences in the immediate postoperative period.
To assess the relative prognostic affect of resecting simultaneous hepatic metastasis, a matched-control analysis was then performed. Table 4 shows the clinicopathologic characteristics of the patients in the case and control groups according to the matching process. Matching was successful in identifying cohorts of patients with comparable age, sex, primary tumor characteristics, and type of primary resection performed.
|No. of patients (%)|
|Variable||Cases (N = 22)||Resected controls (N = 66)||Palliative controls (N = 66)|
|Mean age ± SD, y||64.7 ± 11.4||66.6 ± 11.5||67.2 ± 11.6|
|Sex (% men)||10 (45.5)||36 (54.5)||34 (51.5)|
|Race (% white)||19 (86.4)||49 (74.2)||52 (78.8)|
|Primary tumor site|
|Median tumor size [IQR], cm||5 [3–6.6]||4 [3.35–5.35]||—|
|Lymph node disease||19 (86.4)||57 (86.4)||—|
|Origin of tumor|
|Pancreas head||10 (45.5)||30 (45.5)||30 (45.5)|
|Pancreas tail||7 (31.8)||21 (31.8)||21 (31.8)|
|Duodenum||2 (9.1)||6 (9.1)||6 (9.1)|
|Distal bile duct||2 (9.1)||6 (9.1)||6 (9.1)|
|Ampulla||1 (4.5)||3 (4.5)||3 (4.5)|
|Operation for primary tumor site|
|Pancreaticoduodenectomy||15 (68.2)||45 (68.2)||0 (0)|
|Distal pancreatectomy||7 (31.8)||21 (31.8)||0 (0)|
|Palliative bypass||0 (0)||0 (0)||66 (100)|
Although the median operative time for patients undergoing simultaneous hepatic resection for synchronous liver metastasis (370 minutes; range 338–444 minutes) was similar to that for patients undergoing resection of their primary lesion alone (357 minutes; range 309–451 minutes; P = .46), the operative time was shorter for patients who underwent palliative bypass (218 minutes; range 152–251 minutes; P < .001) (Table 5). Similarly, the median estimated blood loss did not differ between patients in the simultaneous resection group (625 mL; range 400–925 mL) and patients who had no evidence of hepatic metastases who underwent resection of their primary tumor alone (725 mL; range 500–1000 mL; P = .53). In contrast, patients who underwent palliative bypass had significantly less blood loss at the time of surgery (150 mL; range 100-275 mL; P < .001).
|Variable||Cases (N = 22)||Resected Controls (N = 66)||Palliative Controls (N = 66)|
|Operative details and hospital stay|
|Median surgical time/IQR, min||370/338–444||357/309–451||218/152–251*|
|Estimated median blood loss [range], mL||625 [400–925]||725 [500–1000]||150 [100–275]*|
|Median length of stay [range], d||8 [7–11]||9 [8–13]||6 [5–8]*|
|Morbidity: No. of patients (%)|
|Delayed gastric emptying||2 (9.1)||9 (13.6)||0 (0)|
|Pancreatic fistula||2 (9.1)||3 (4.5)||0 (0)|
|Biliary leak||0 (0)||3 (4.5)||0 (0)|
|Cardiac||0 (0)||1 (1.5)||3 (4.5)|
|Pulmonary||0 (0)||2 (3)||2 (3)|
|Any||10 (45.5)||31 (46.7)||11 (16.7)*|
|30-D||2 (9.1)||3 (4.5)||2 (3)|
|Median survival, mo||180||429*||170|
|Actuarial survival, %|
Perioperative morbidity and mortality also were increased in patients who underwent simultaneous hepatic resection for synchronous liver metastasis. Specifically, whereas the postoperative complication rate was comparable between the simultaneous resection group and the resected controls (45.5% vs 47.7%, respectively; P = 1.00), patients who underwent palliative bypass had a significantly reduced rate of complications (27.7%) (Table 5). The length of hospital stay was shorter in the palliative control group (median, 6 days; range 5–8 days) compared with either the resected control group (median, 9 days; range, 8–13 days) or the simultaneous resected group (median, 8 days; range, 7–11 days; both P < .05). Although the difference was not statistically significant, patients who underwent simultaneous hepatic resection for synchronous liver metastasis tended to have a higher 30-day postoperative mortality rate (9.1%) compared with either the resected control group (4.5%) or the palliative control group (3.0%) (Table 5).
Regarding overall survival, patients who underwent simultaneous resection of the primary adenocarcinoma as well as liver metastasis had a shorter overall survival (median, 5.9 months) compared with patients who had no liver metastasis and, thus, underwent resection of the primary tumor alone (median, 14.1 months; P < .001) (Fig. 3). In contrast, the overall survival of patients with synchronous metastasis who underwent simultaneous resection was not different from the overall survival of patients who underwent palliative bypass alone (median, 5.9 months vs 5.6 months, respectively; P = .46).
Although the role of metastectomy in patients with hepatic metastases from colorectal or neuroendocrine primary tumors is well reported and accepted, the role of hepatic resection for metastatic disease derived from other tumor histologies or anatomic sites is less defined. Several institutions have reported improved survival for patients who have undergone hepatic resection for noncolorectal and nonneuroendocrine metastases.14, 18–20 Specifically, it has been reported that hepatic resection provides a potential survival benefit for patients with such nontraditional tumors as breast cancer,21, 22 sarcoma,23 melanoma,24 and squamous cell carcinoma.25 Few reports, however, have investigated the relative benefit of resecting hepatic pancreaticobiliary metastases.15, 26–28 Largely because primary pancreaticobiliary adenocarcinomas have such a poor prognosis, resection of metastatic disease—even if it is limited, liver-only disease—has been discouraged. Thus, scant data on resection of pancreaticobiliary liver metastases exist in the literature except for a few case reports. One exception was the study by Adam et al.,15 which reported on 67 patients who underwent resection of pancreaticobiliary hepatic metastases. In that study, the authors reported a 5-year survival rate of 25% for patients who underwent resection of metastatic pancreatic adenocarcinoma to the liver. Few if any patients, however, presented with synchronous liver metastasis with their primary tumor in situ. The current study is important, because it was focused specifically on outcomes after resection of synchronous pancreaticobiliary liver metastases. More important, in the current study, we assessed whether aggressive resection was warranted in the clinical scenario that often is faced by surgeons—the presence of limited, unsuspected, synchronous hepatic metastasis at the time of primary tumor resection.
Despite an aggressive surgical approach that involved extirpation of both the primary and metastatic disease sites, patients with synchronous pancreaticobiliary tumors and hepatic metastases had a poor prognosis. Specifically, the median overall survival after simultaneous resection of the primary tumor and the liver metastases was only 5.9 months; in addition, the 1-year survival rate was only 13.3%. The overwhelming majority of patients died of recurrent disease within 12 months of surgery (Fig. 1), and only 1 patient survived to 3 years. Although survival after simultaneous resection of the primary adenocarcinoma and liver metastasis was similar regardless of whether the primary tumor histology was of pancreatic (5.9 months) versus nonpancreatic (9.9 months) origin (P = .43), the 1 actual 3-year survivor did have a primary duodenal lesion. Results from other studies26, 27, 29 that evaluated synchronous primary tumor resection combined with hepatic resection of liver metastases also have been disappointing. Takada et al.27 reported a median survival of 6 months for patients with adenocarcinoma of the pancreas who underwent liver resection. In their study, there were no actual 1-year survivors after surgery. In contrast, Shimada et al.28 reported a single patient who survived 7 years after undergoing concomitant resection of an adenocarcinoma in the tail of the pancreas with a wedge resection of a small (0.8 cm), metastatic nodule in the anterior-inferior segment of the liver. Notwithstanding this single case report, data from the current study and from the study by Takada et al.27 strongly suggest that synchronous resection of pancreaticobiliary malignancies with hepatic metastases, in particular pancreatic adenocarcinoma, is associated with an abysmal median survival in the range of only 6 months.
The current study has important implications, because we report that, compared with simple bypass, which is the traditional approach for symptomatic patients who have hepatic metastasis discovered, simultaneous resection of the primary tumor with liver resection did not result in improved survival. In fact, resection of the primary lesion with the hepatic metastasis resulted in a near identical median survival (5.9 months) compared with patients who underwent bypass alone (5.6 months) (Fig. 3). In contrast, as expected, the patients who had local-only disease and who underwent resection of their primary tumor had the best outcome. Takada et al.27 reported comparable results; in their study, the patients who underwent simultaneous hepatic resection with pancreaticoduodenectomy had the same poor overall survival as the patients who underwent palliative bypass alone. Unlike the current study, however, patients in the Takada et al.27 report had a more extensive burden of hepatic disease (hepatic metastasis size ranged from 0.5 cm to 5 cm, and some patients had as many as 7 lesions). In contrast, the overwhelming majority of patients in our series had very limited metastatic disease; the median tumor size was only 0.6 cm, and 90.9% of patients had only 1 lesion. These data indicate that, although it may be tempting to treat patients who have a single, small metastasis as patients with only “limited” metastatic disease who still may benefit from resection of the primary tumor with simultaneous hepatic resection, synchronous hepatic metastasis—regardless of the number or size—is a marker for existent, widespread, systemic disease. Other investigators3 have suggested that only approximately 5% of patients who present with stage II pancreatic cancer (surgically resected, lymph node positive) do not have occult, systemic, micrometastatic disease. Therefore, it is reasonable to conclude that virtually all patients who actually have evidence of metastatic disease at the time of surgery—even if it is limited to 1 organ—also will have additional, unrecognized, systemic, micrometastatic disease.
Although we report a median survival of only 5.9 months and a 1-year survival rate of 13.3%, Adam et al.15 noted 5-year survival rates of 20%, 25%, and 46% for patients undergoing resection of liver metastasis from duodenal, pancreatic, and ampullary adenocarcinomas, respectively. There are several important reasons why the results of the current study differ dramatically from those reported by Adam et al.15 In the study by Adam et al.,15 several of the strongest predictors of outcome were disease-free interval >24 months and response to chemotherapy. Because all but 1 patient in the current study had unrecognized synchronous disease identified at the time of surgery, we did not have the benefit of predictive variables like response to preoperative chemotherapy and disease-free interval. Both chemotherapy response and disease-free interval help provide time to identify the vast majority of those patients with the worst tumor biology who will develop progressive disease and, thus, may not benefit from resection. It is not surprising, therefore, that patients with synchronous disease fared significantly worse. Although data from Adam and colleagues have suggested that hepatic resection may be warranted in some patients with pancreaticobiliary metastasis, our data seriously call into question whether hepatic resection for synchronous disease is warranted.
Not only was overall outcome not improved, but simultaneous resection of the primary pancreaticobiliary tumor in combination with hepatic metastectomy also led to increased morbidity and mortality. Intraoperatively, patients who underwent a simultaneous procedure had increased operative time and blood loss compared with patients who underwent palliative bypass alone (Table 5). Postoperatively, patients who underwent a simultaneous procedure also had significantly more complications and a longer hospital stay. Although the difference was not statistically significant, patients who underwent simultaneous hepatic resection for synchronous liver metastasis also tended to have a higher 30-day postoperative mortality rate (9.1%) compared with either the resected control group (4.5%) or the palliative control group (3%). Because simultaneous resection was not associated with an improved survival compared with bypass alone but was associated with increased morbidity, length of stay, and mortality, routine combined resection of primary pancreaticobiliary tumors with hepatic metastectomy cannot be advocated.
Given the poor prognosis of patients with metastatic disease, some investigators have argued that laparotomy in this cohort of patients may be unnecessary. Even if palliation of biliary or gastrointestinal symptoms is required, this often can be done with endoscopic/percutaneous biliary or duodenal wall stents, etc. Therefore, some investigators30 have advocated the routine use of diagnostic laparoscopy before laparotomy to identify the subset of patients with intra-abdominal/peritoneal metastatic disease. In this way, laparoscopy may help avoid an unnecessary laparotomy to determine that advanced disease (eg, liver metastasis) is present. This is important, because patients undergoing laparotomy versus laparoscopy require lengthier hospitalization and more pain medication, and they are more susceptible to postoperative complications, such as atelectasis, pneumonia, and wound infection. These data may have especially important implications in patients with metastatic pancreatic/periampullary cancer, who typically have a very short overall survival. Routine diagnostic laparoscopy, however, remains controversial; because, although diagnostic laparoscopy may decrease the number of nontherapeutic laparotomies, it also has several limitations. In particular, use of the rigid laparoscope, even when angled, offers limited access to observe some structures, including the dome of the liver. Thus, up to 40% of patients with peritoneal metastases may be missed by conventional laparoscopy.31 Currently, our group is investigating flexible peritoneoscopy as a means to improve laparoscopic staging.
The current study had several limitations. Despite having the largest pancreaticobiliary surgical experience to our knowledge in the country, only a relatively small sample size of patients could be identified for this study. Therefore, the current study has limited statistical power; because of this constraint, statistical analyses and inferences were limited. Another possible limitation involved our combining nonpancreatic periampullary adenocarcinomas with pancreatic adenocarcinomas for the purposes of reporting general outcome parameters. Although nonpancreatic periampullary tumors frequently require a partial pancreatectomy, the biology of these lesions differs from that of pancreatic adenocarcinomas. Thus, conclusions regarding those 5 patients need to be considered carefully. Finally, the current study did not assess other important endpoints, such as quality of life after surgical treatment with simultaneous resection versus palliative bypass alone.
In conclusion, the selection of appropriate patients for hepatic resection of metastatic pancreaticobiliary adenocarcinomas must be individualized. Although others15 have reported that hepatic resection of metastatic disease may be warranted in those patients who demonstrate good tumor biology (eg, a long disease-free period and chemoresponsive disease), simultaneous resection for patients who present with synchronous lesions does not appear to be justified. Data from the current study demonstrate that more aggressive surgical approaches combining resection of the primary lesion with hepatic resection yield similar overall survival durations compared with palliative bypass alone yet have increased morbidity, hospital stay, and, perhaps, mortality. A more rationale approach to symptomatic patients who present with synchronous disease may be to perform a palliative bypass, because this appears to be associated with less morbidity and may expedite postoperative therapy with systemic chemotherapy. In highly select patients who demonstrate chemoresponsive disease without interval development of extrahepatic metastasis, surgical resection may be considered subsequently. Careful clinical judgment and prudent selection of patients as part of a multidisciplinary approach are mandatory.
- 16for the International Hepato-Pancreato-Biliary Association Terminology Committee Survey. The Brisbane 2000 terminology of liver anatomy and resections. HPB. 2000; 2: 333–339.,