- Email: [email protected]
Concurrent resections of pancreatic islet cell cancers with synchronous hepatic metastases: Outcomes of an aggressive approach
Concurrent resections of pancreatic islet cell cancers with synchronous hepatic metastases: Outcomes of an aggressive approach Juan M. Sarmiento, MD, Florencia G. Que, MD, Clive S. Grant, MD, Geoffrey B. Thompson, MD, Michael B. Farnell, MD, and David M. Nagorney, MD, Rochester, Minn
Background. Pancreatic islet cell cancers are often characterized by the presence of endocrinopathies, an indolent clinical course, and a propensity for hepatic metastases. Hepatic metastases are associated with a negative impact on survival. The role of concurrent resections of pancreatic islet cell cancers and the hepatic metastases has not been defined. Methods. The records of all consecutive patients undergoing concurrent resections of pancreatic islet cell cancers and their hepatic metastases between 1980 and 1998 were reviewed. Outcomes regarding overall progression-free and symptom-free survival and perioperative morbidity and mortality were assessed. Results. All 23 patients underwent distal pancreatectomy and splenectomy. Six major (≥ 3 segments) and 17 minor (c3 segments) partial hepatectomies were performed. Complete gross resection of cancer (R0/R1) were performed in 9 patients and debulking resections (R2) (<10% residual tumor volume) in 14 patients. There were no perioperative deaths. Major complications occurred in 4 patients (18%). Overall, progression-free, and symptom-free survival was 71% (median: 76 months), 5% (median: 21 months), and 24% (median: 26 months), respectively, at 5 years. Conclusions. These data support aggressive concurrent resection of the pancreatic islet cell cancers and synchronic hepatic metastases when technically feasible. Because disease progression is frequent and the major cause of death, investigations of adjuvant and adjunctive therapies are warranted. (Surgery 2002;132:976-83.) From the Division of Gastroenterologic and General Surgery, Mayo Clinic, Rochester, Minn
RESECTION OF HEPATIC METASTASES from neuroendocrine malignancies can provide significant symptomatic benefit in patients with functioning tumors and may improve overall survival regardless of clinical syndromes.1,2 Neuroendocrine tumors are usually associated with several features that lend themselves to aggressive resection: cancer growth rate is indolent, primary tumor growth is characterized predominantly by gross tumor expansion rather than invasion, and growth characteristics of hepatic metastases are similar. Accumulating experiences with these uncommon malignancies continue to support surgical intervention if both the primary cancer and the metastases are resectable. Unfortunately, in practice, the incidence of pancreatic islet cell cancers (PIC) restricts the implePresented at the 23rd Annual Meeting of the American Association of Endocrine Surgeons, Banff, Alberta, Canada, April 7-9, 2002. Reprint requests: David M. Nagorney, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905. © 2002, Mosby, Inc. All rights reserved. 0039-6060/2002/$35.00 + 0 doi:10.1067/msy.2002.128615
976 SURGERY
11/6/128615
mentation and power of randomized clinical trials comparing an aggressive surgical approach in patients with PIC and hepatic metastases with nonsurgical or medical alternatives that would determine the treatment of choice in such patients. Although resection of hepatic metastases is wellestablished for colorectal cancer,3 data supporting a similar approach for metastatic neuroendocrine cancers is still evolving. As with metastatic colorectal cancer, control of the primary cancer before resection or ablation of the hepatic metastases is imperative. Whether concurrent or staged resections of the primary cancer and the hepatic metastases is preferable in patients presenting with overt metastatic disease remains debatable. If concurrent resections of the primary cancer and the hepatic metastases can be performed as safely as staged resections, potential benefits in terms of overall survival, reduced duration of hospitalization, and, as a corollary, reduced hospitalization costs and earlier return to gainful employment would further justify this approach.4 Similar to patients with colorectal cancer, resection of hepatic metastases from PIC has been generally undertaken after
Surgery Volume 132, Number 6
resection of the primary pancreatic cancer because of the postulate that perioperative risk is increased with concurrent pancreatic and hepatic resections. Interestingly, current data regarding the perioperative risks for staged hepatic resection of either colorectal (mortality 2-4%)5 or neuroendocrine metastases (morbidity 24%, mortality 3%)2 are similar, and overall perioperative risks for pancreatic (morbidity 31%, mortality 1%)6 and colorectal resections (mortality 3%)7 are not dissimilar. Moreover, prohibitive morbidity and mortality risk for concurrent resections of colorectal cancers and hepatic metastases has not been observed by us4 or others.8 We wondered whether our experience with concurrent resections of PIC and hepatic metastases was similar. To further define the role of hepatic resection for metastatic PIC, we reviewed our experience. Our aims were to determine the actual perioperative risks of concurrent resections of PIC and hepatic metastases and to assess overall, symptom-free, and disease-free survival. METHODS The records of all consecutive patients undergoing concurrent resections of PIC and hepatic metastases at the Mayo Clinic in Rochester, Minn, between 1980 and 1998 were studied. This study period was chosen to provide a minimum following-up of 3 years, given the prognosis and indolent course of PIC from our past study.2 Patients with hepatic metastases who had resection of the PIC and biopsy of hepatic metastases or patients who had staged resections of the primary cancer and hepatic metastases despite recognition of both at initial presentation were excluded. PIC cancers were classified clinically by assay of the predominant serum hormone or immunohistochemic staining the resected PIC cancer. Patients without clinical manifestations of hormone hypersecretion or predominant hormone markers were classified as nonfunctioning islet cell cancers. Patient demographics, clinical findings, and perioperative morbidity and mortality were recorded. All pancreatic resections were limited to distal pancreatectomy, splenectomy, and regional lymphadenectomy. Hepatic resections were classified as major if ≥3 segments were removed and minor if ≤2 segments were removed. Most hepatic resections were accompanied by subsegmental nonanatomic excisions of smaller metastases (≤1 cm). Concurrent resections were classified as complete gross resection (R0/R1) if all visible tumor was removed and debulking (R2) if gross tumor remained. Residual tumor (R2) was located within the liver and was ≤10% of the hepatic metastases by volume.
Sarmiento et al 977
Hepatic resections were performed if preoperative imaging and intraoperative findings suggested that at least 90% (by volume) of the metastases were resectable. Computed tomography of the abdomen was the primary imaging modality to assess resectability of both the primary cancer and metastases. Magnetic resonance imaging was used as a complimentary imaging modality in the latter years of the study. From a technical standpoint, the distal pancreatectomy, splenectomy, and lymphadenectomy were performed initially in every patient. Hepatic resection was performed with minimal inflow vascular occlusion to reduce venous congestion of the pancreatic stump. Soft suction drains were left adjacent to the pancreatic stump and transected hepatic surface if major or extended hepatic resection was performed. Patients were followed-up for clinical recurrence of endocrinopathies (if present preoperatively), cancer recurrence or progression, and survival. Follow-up was on the basis of outpatient evaluations and updated by correspondence. Survival was calculated from the date of histologic diagnosis. The method of Kaplan and Meier was used to estimate overall progression-free and symptom-free survival. The Wilcoxon test was used for comparison of survival estimates between groups. RESULTS Twenty-eight patients presented with PIC and hepatic metastases. Five patients underwent staged resections at the discretion of the primary surgeon. The remaining 23 patients are the subject of this report. Gender distribution was even with 12 men and 11 women. Mean (± SD) age was 53 ± 12 years. The frequency of cancers by hormonal markers was: nonfunctioning, 7; glucagonoma, 6; insulinoma, 2; corticotropin-producing tumors, 2; polyfunctional tumors, 2; gastrinoma, 1; VIPoma, 1; carcinoid, 1; and 1 patient with hypercalcemia without confirmation of parathormone-like hormone production. Although 16 patients had functioning cancers, 1 patient with a glucagonoma presented with upper gastrointestinal (GI) bleeding but no clinical endocrinopathy. Therefore, endocrine symptoms were seen in 15 patients. The other 8 patients experienced abdominal pain (n = 5) sometimes associated to nausea and vomiting, presence of abdominal mass (n = 1), weight loss (n = 1), and upper GI bleeding (n = 1). Each patient in this study had a PIC located in the pancreatic body or tail. All pancreatic resections consisted of distal pancreatectomy, splenectomy, and regional lymphadenectomy and were RO/R1. Ten patients (44%) had sinistral portal
978 Sarmiento et al
Surgery December 2002
Fig 1. Overall survival of 23 patients with simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 71%; median, 76 months).
hypertension from splenic vein obstruction or invasion. Two patients required a venotomy of the portal vein to extract tumor thrombus extending from the primary tumor. Hepatic resections included: right hepatectomy, 4; left hepatectomy, 1; extended left hepatectomy, 1; bisegmentectomy, 2; segmentectomy, 7; and multiple subsegmentectomies, 8. Major hepatic resections (≥ 3 segments) were performed in 26% of the patients. Complete macroscopic resection (RO/R1) was achieved in 9 patients (39%). Fourteen patients had a debulking procedure with macroscopic residual disease in the liver (R2) but each had an estimated 90% (by volume) of the hepatic metastases resected. Hormonal markers in the 9 patients who had R0/R1 resections normalized in 6 (the other 3 had nonfunctioning tumors). Regional lymph nodes were involved in 15 patients. Regional peripancreatic lymph nodes were positive in 14 patients and celiac nodes were positive in 1 patient. Lymphatic involvement was absent in 6 patients. No lymph nodes were identified pathologically in 2 patients. There was no perioperative mortality (≤ 30 days from operation). Complications occurred in 7 patients (30%). Complications included: bile leaks, 2; pancreatic leaks, 2; wound infection, 1; delirium tremens, 1; and urinary tract infection, 1. The major complication rate related to either hepatic or pancreatic resection was 18%. Each bile or pancreatic fistula resolved with continuous suction drainage from operatively placed drains and required no further intervention. Prophylactic perioperative somatostatin analogues were not used routinely. Median duration of operation was 275 minutes (range: 120-390 minutes). Median
transfusion requirement was 1 U of packed red blood cells (range: 0-20). Blood products were necessary in 12 patients (52%). Median length of stay was 8 days (range: 5-22). On follow-up, 4 patients (17%) had new signs of pancreatic insufficiency develop that were endocrine, 2; exocrine, 1; or combined, 1. No patients required late reoperation for intra-abdominal complications from local or regional recurrence of the PIC. Overall survival was 71% at 5 years. Median survival was 76 months (Fig 1). Of the 18 patients who had resections 5 or more years before the date of analysis, 11 of these patients were actual 5-year survivors (61%). Recurrence-free or progression-free survival was only 5% at 5 years. Median recurrencefree or progression-free survival was 21 months (Fig 2). All patients who have had recurrence of the PIC had liver involvement (demonstrated by computed tomography in all but one). One patient was censored from recurrence or progression analysis because the actual date of documented recurrence was unknown but that patient died from recurrent cancer. One patient with a recurrent VIPoma underwent orthotopic liver transplantation and remains alive and disease-free at 36 months after transplantation and is the only patient without cancer recurrence currently. Four patients have undergone repeated partial hepatic resections for progression. Among the 15 patients with clinical endocrinopathies, symptom-free survival was 24% at 5 years and median symptom-free survival was 26 months (Fig 3). Each patient with an endocrinopathy had a complete (or almost complete) response after resection. There was no difference in survival between patients who had R0/R1 versus R2 resections. Median survival was 65 and 76 months for
Surgery Volume 132, Number 6
Sarmiento et al 979
Fig 2. Recurrence-free (or progression-free) survival of 23 patients with simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 5%; median, 21 months). All patients had recurrence.
RO/R1 and R2 resections, respectively (P = .90) (Fig 4). DISCUSSION We believe that the treatment for hepatic metastases from neuroendocrine cancers is resection if the primary cancer has been resected or is grossly resectable and more than 90% (by volume) of the hepatic metastases are resectable.2 Whether this approach is preferable for patients with PIC presenting with overt hepatic metastases is undefined. Concurrent resection of both the primary cancer and hepatic metastases has been criticized because of the presumed perioperative risks associated with concurrent pancreatic and hepatic resections and the poor prognosis associated with advanced-stage cancers whether functioning or nonfunctioning.9,10 In fact, data addressing the subset of patients with metastatic PIC undergoing resection is sparse and we are unaware of any series contraindicating this approach. The major findings of our study suggest that combined pancreatic and hepatic resections can be performed safely with acceptable morbidity and mortality. Moreover, this approach provides effective palliation of endocrine symptoms when present and an overall 5-year survival of 71%. Because cancer progression remains the major cause of death, both adjuvant and adjunctive treatment strategies to prevent or delay progression are needed. Our experience with concurrent resection of both PIC and hepatic metastases demonstrates that this approach can be performed safely and with acceptable morbidity. Admittedly, selection of patients was limited to PIC arising from the distal
pancreas and the extent of hepatic metastases required major resections in only 26% of patients. However, these initial data do provide a basis for further evaluation of this approach. The frequency of pancreatic and hepatic complications was similar herein. In fact, the frequency of complications for either major component of the concurrent approach herein does not differ from that reported for isolated pancreatectomies (30-40%)6 or partial hepatectomies (30%)11 for malignancy. We do not believe that this incurred morbidity precludes further pursuit of the concurrent approach. However, careful selection of patients is necessary. Clinical performance status should exceed a Karnofsky score of 50% (Eastern Cooperative Oncology Group performance status <2) and other comorbidities should be limited. Complete resection of the PIC should precede the partial hepatectomy. If intraoperative problems during the pancreatic resection arise, partial hepatectomy should be deferred. We have not performed concurrent pancreaticoduodenectomy and major hepatic resection and currently favor staged resections in these patients. We would not be reluctant to combine pancreaticoduodenectomy with minor hepatic resections. The natural history of unresected PIC with hepatic metastases is poor. Despite a more prolonged survival compared with ductal pancreatic cancers of similar stage, survival only has ranged from 1.7 to 3.1 years.12,13 Cancer progression is the primary cause of death in most of these patients and intra-abdominal complications, such as intestinal hemorrhage or obstruction related to the primary cancer, has accounted for a majority of the
980 Sarmiento et al
Surgery December 2002
Fig 3. Symptom-free survival of 15 patients with clinical endocrinopathy submitted to simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 24%; median, 26 months).
Fig 4. Comparative survival of 23 with simultaneous resection of PIC tumors and synchronic hepatic metastases; 9 patients received a complete gross (R0/R1) resection (5-year survival, 76%; median, 65 months) versus 14 patients with a debulking (R2) resection (5-year survival, 67%; median, 76 months); P = .90 by the Wilcoxon test. Solid line, R0/R1; broken line, R2.
remainder. Quality of life in patients with metastatic PIC is impaired not only from complications of the primary cancer and the usual constitutional symptoms of cancer but also clinical endocrinopathies. Despite improvements in clinical control of endocrine symptoms with somatostatin analogues,14 and objective responses to multimodality chemotherapy,15 overall survival has improved min-
imally. The current outcomes for nonsurgical approaches do not warrant categoric exclusion of surgical approaches. Indeed, accumulating data have suggested that resection of PIC and their metastases in selected patients can prolong survival compared with patients of similar stage historically.16 Consequently, we have continued to explore the role of resection in the multimodality treat-
Surgery Volume 132, Number 6
ment of patients with advanced stage PIC. We believe that resection, though not often curative in this clinical setting, may delay other treatments and does not prevent their eventual implementation. The indications for resection of PIC in patients presenting with overt hepatic metastases have been primarily palliative in intent. Resection of the PIC in the presence of hepatic metastases has been advocated specifically to relieve intractable pain or to palliate GI obstruction or hemorrhage from advanced local tumor growth for improvement in quality of life.13 Although PIC of the head of the pancreas more frequently are associated with these clinical problems, patients with PIC in the body or tail of the pancreas can also have life-threatening complications develop. Indeed, 44% of the patients in this series had sinistral portal hypertension from obstruction or thrombosis of the splenic vein. GI hemorrhage from gastric varices can be life threatening and warrant distal pancreatectomy and splenectomy in selected patients.17,18 However, only a minority of patients with metastatic PIC will require palliative resections of the primary cancer for complications. We believe that resection of PIC is contraindicated in the presence of undebulkable hepatic or distant metastases if patients are asymptomatic. Most of these patients with metastatic PIC should be treated with palliative chemotherapy, somatostatin analogues, or both depending on the presence of functional endocrinopathies. However, patients in whom the primary cancer is resectable and in whom near total hepatic debulking is possible should be assessed for resection. These patients are the focus of this report further exploring the role of concurrent debulking of both disease sites. Cytoreductive or debulking resections for functioning neuroendocrine metastases after resection of the primary cancer has evolved over the last decade. The role of debulking of metastatic PIC was examined in small series in the early 1990s.1,2 Clinical endocrinopathies were the indications for hepatic resections. Relief from endocrine symptoms and prolongation of survival was evident. In one report, patients who were unresected had an overall survival of 28% at 5 years, but a 79% survival at 5 years debulking.16 Although 14 patients underwent partial hepatectomy and 11 patients had a partial pancreatectomy, these data did not specify how many patients had concurrent resections. Resections were curative in intent in 80% of these patients. There were no deaths and the complication rate was 30%. Despite disease recurrence in all patients, all experienced reduction of antisecretory medications and their symptoms. Similar findings have been reported previously by ourselves1,2 and
Sarmiento et al 981
others.13,19 Although some of these patients had concurrent resections, our data herein comprise the largest series of patients undergoing concurrent resections of the primary PIC and a hepatic metastases to our knowledge. The outcomes in terms of symptomatic response and survival were similar to other reports that predominantly represent outcomes from staged resections. We are unaware of any data comparing staged versus concurrent resections for such patients. Whether concurrent resections are warranted in patients with nonfunctioning PIC is unknown. In contrast to patients with functional endocrinopathies in whom the degree of achievable palliation justifies the perioperative risks, significant improvement in survival is required to justify the perioperative risks with nonfunctioning PIC. Although we continue to explore the role of resection of nonfunctioning PIC with hepatic metastases in selected patients to determine whether survival will improve, we remain concerned regarding this aggressive approach in such patients. Clearly, overall survival for such patients is limited if no tumor is resected.12,20 However, despite selection, survival was greater in patients undergoing resection of either the primary nonfunctioning islet cell cancer, alone, or in patients after resection of both the primary and the hepatic metastases.13 Interestingly, in a prior study, we were unable to demonstrate a difference in survival between functioning and nonfunctioning PIC for similar extent of debulking.20 These findings suggest that resection may improve survival in patients with metastatic PIC and may be unique to GI neuroendocrine cancers compared with other solid abdominal malignancies. Finally, regardless of functional status or timing of hepatic resection, disease progression is typical and the most common cause of death.13 We agree that both adjunctive (present of gross residual disease) and adjuvant (absence of gross residual disease) treatment strategies should be investigated. Currently, we are evaluating the role of ongoing percutaneous radiofrequency ablation for hepatic metastases after hepatic regeneration and orthotopic liver transplantation for hepatic metastases that do not meet criteria for debulking. Because the liver is the major site of progression, perhaps ablation of hepatic metastases as they arise may affect survival and control of endocrine symptoms if present. Systemic chemotherapy or regional chemoembolization should compliment ablative and resective approaches. Nonetheless we believe that resections of PIC with hepatic metastases in selected patients meeting the previously stated criteria should be undertaken initially in patients who
982 Sarmiento et al
are fit and, if technically feasible, concurrently. Surgical resection does not preclude implementation of other strategies.
REFERENCES 1. McEntee GP, Nagorney DM, Kvols LK, Moertel CG, Grant CS. Cytoreductive hepatic surgery for neuroendocrine tumors. Surgery 1990;108:1091-6. 2. Que FG, Nagorney DM, Batts KP, Linz LJ, Kvols LK. Hepatic resection for metastatic neuroendocrine tumors. Am J Surg 1995;169:36-42. 3. Treatment of hepatic metastases from colorectal cancer: the SSAT, AGA, ASLD, ASGE, AHPBA consensus panel. J Gastrointest Surg 1997;1:396-7. 4. Chua HK, Wolff BG, Sordenna K, Tsiotos GG, Larson DR, Nagorney DM. Is combined colectomy and hepatectomy for synchronous metastatic colorectal cancer more efficient than staged? Gastroenterology 2000;118:1027. 5. Jamison RL, Donohue JH, Nagorney DM, Rosen CB, Harmsen WS, Ilstrup DM. Hepatic resection for metastatic colorectal cancer results in cure for some patients. Arch Surg 1997;132:505-10. 6. Lillemoe KD, Kaushal S, Cameron JL, Sohn TA, Pitt HA, Yeo CJ. Distal pancreatectomy: indications and outcomes in 235 patients. Ann Surg 1999;229:693-8. 7. Harmon JW, Tan DG, Gordon TA, Bowman HM, Choti MA, Kaufman HS, et al. Hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg 1999;230:404-13. 8. Rosen SA, Buell JF, Yoshida A, Kazsuba S, Hurst R, Michelassi F, et al. Initial presentation with stage IV colorectal cancer: how aggressive should we be? Arch Surg 2000;135:530-5. 9. Chen H, Hardacre JM, Uzar A, Cameron JL, Choti MA. Isolated liver metastases from neuroendocrine tumors: does resection prolong survival? J Am Coll Surg 1998;187:88-93. 10. Chamberlain RS, Canes D, Brown KT, Saltz L, Jarnagin W, Fong Y, et al. Hepatic neuroendocrine metastases: does intervention alter outcomes? J Am Coll Surg 2000;190: 432-45. 11. Blumgart LH, Fong Y. Surgical options in the treatment of hepatic metastases from colorectal cancer. Curr Probl Surg 1995;32:333-421. 12. Thompson GB, van Heerden JA, Grant CS, Carney JA, Ilstrup DM. Islet cell carcinomas of the pancreas; a twentyyear experience. Surgery 1988;104:1011-7. 13. Solorzano CC, Lee JE, Pisters PW, Vauthey JN, Ayers JD, Jean ME, et al. Nonfunctioning islet cell carcinoma of the pancreas: survival results in a contemporary series of 163 patients. Surgery 2001;130:1078-85. 14. Ericksson B, Oberg K. Summing up 15 years of somatostatin analog therapy in neuroendocrine tumors: future outlook. Ann Oncol 1999;10:531-8. 15. Moertel CG, Lefkopoulo M, Lipsitz S, Hahn RG, Klaassen D. Streptozocin-doxorubicin, streptozocin-fluorouracil or chlorozotocin in the treatment of advanced islet-cell carcinoma. New Engl J Med 1992;326:519-23. 16. Carty SE, Jensen RT, Norton JA. Prospective study of aggressive resection of metastatic pancreatic endocrine tumors. Surgery 1992;112:1024-32. 17. Metz DC, Benjamin SB. Islet cell carcinoma of the pancreas presenting as bleeding from isolated gastric varices. Dig Dis Sci 1991;36:241-4.
Surgery December 2002
18. Loftus JP, Nagorney DM, Ilstrup D, Kunselman AR. Sinistral portal hypertension: splenectomy or expectant management. Ann Surg 1993;217:35-40. 19. Norton JA, Doherty GM, Fraker DL, Alexander HR, Doppman JL, Venzon DJ, et al. Surgical treatment of localized gastrinoma within the liver: a prospective study. Surgery 1998;124:1145-52. 20. Sarmiento JM, Heywood G, Ilstrup DM, Nagorney DM, Que FG. Surgical treatment of neuroendocrine metastases to the liver. Hepatology 2001;34:236A.
DISCUSSION Dr Janice L. Pasieka (Calgary, Alberta, Canada). I see you were very fortunate that all of your primary tumors were in the body and tail of the pancreas. Could you please tell us what you would do in a situation where the primary is in the head of the pancreas with liver metastases? Would you proceed with a Whipple resection and resection of the liver metastases? Dr Sarmiento. This is not a matter of luck; this is a matter of selection. Obviously, we are not doing a Whipple procedure plus a right hepatectomy. But we think we can still do a Whipple procedure and a minor resection of the liver if necessary. Dr Pasieka. How small is small? Dr Sarmiento. (Essen, Germany) One or 2 segments of the liver. Everything depends on the intraoperative evaluation of the patient. Dr Andrea Frilling (Essen, Germany). May I address 2 questions? You said that all of your patients had recurrent tumor. In the group of patients in whom liver resection was performed, there were 6 patients with R0 resection. Where did these patients get their recurrence? Was it within the liver or extrahepatically? The second question would be, are there patients in your series in whom you performed metachronous liver resection? That means first pancreatic resection and, a few weeks later, liver resection? Dr Sarmiento. All the recurrences that we saw were in the liver. That is true. We cannot say that we are curing these patients even when we achieve complete microscopic clearance of the tumor. The issue is that, if in a group of 23 patients with a 100% recurrence, speaking of R0 versus R1 resections is just semantics. Again, all the patients had recurrence of the liver. We had 28 patients presenting with synchronous metastases. Twenty-three were resected concurrently and 5 were resected as staged procedures. Therefore, with only 5 patients in the control group, it is very risky to compare the complication rate and the survival. That is why we did not perform that comparison. Dr Frilling. Just 1 additional question. Do you perform intraoperative ultrasound before you resect the patient? Dr Sarmiento. Definitely. Dr Sally E. Carty (Pittsburgh, Pennsylvania). It is very admirable that you were able to look at a large enough series of patients in this disease. I note that you had fewer gastrinomas than many comparable series and also a higher rate of palliative operation compared with complete resection. That is not your fault, that is the extent
Sarmiento et al 983
Surgery Volume 132, Number 6 of the patients’ disease. But I would just like to inquire about your last conclusion, improvement in survival, I may have missed something, but your data showed no difference in survival among patients with palliative compared with complete resection. And it is good to be careful in making statements that compare results with historic controls. What we need is a randomized, prospective study. The problem is, who is going to sign up for the nonoperated arm?
O
Dr Sarmiento In your article in 1992,16 you also say that it was very difficult to get enough patients for a randomized trial. And all the data that I saw from your article are from that study without control group. So, basically, we cannot produce data in a randomized fashion. What we say, here, is that performing resection, complete or incomplete resection, is better than not doing resection at all. That is what I mean when I refer to improvement in survival.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 800-654-2452 Outside the US, call 407-345-4000
Background. Pancreatic islet cell cancers are often characterized by the presence of endocrinopathies, an indolent clinical course, and a propensity for hepatic metastases. Hepatic metastases are associated with a negative impact on survival. The role of concurrent resections of pancreatic islet cell cancers and the hepatic metastases has not been defined. Methods. The records of all consecutive patients undergoing concurrent resections of pancreatic islet cell cancers and their hepatic metastases between 1980 and 1998 were reviewed. Outcomes regarding overall progression-free and symptom-free survival and perioperative morbidity and mortality were assessed. Results. All 23 patients underwent distal pancreatectomy and splenectomy. Six major (≥ 3 segments) and 17 minor (c3 segments) partial hepatectomies were performed. Complete gross resection of cancer (R0/R1) were performed in 9 patients and debulking resections (R2) (<10% residual tumor volume) in 14 patients. There were no perioperative deaths. Major complications occurred in 4 patients (18%). Overall, progression-free, and symptom-free survival was 71% (median: 76 months), 5% (median: 21 months), and 24% (median: 26 months), respectively, at 5 years. Conclusions. These data support aggressive concurrent resection of the pancreatic islet cell cancers and synchronic hepatic metastases when technically feasible. Because disease progression is frequent and the major cause of death, investigations of adjuvant and adjunctive therapies are warranted. (Surgery 2002;132:976-83.) From the Division of Gastroenterologic and General Surgery, Mayo Clinic, Rochester, Minn
RESECTION OF HEPATIC METASTASES from neuroendocrine malignancies can provide significant symptomatic benefit in patients with functioning tumors and may improve overall survival regardless of clinical syndromes.1,2 Neuroendocrine tumors are usually associated with several features that lend themselves to aggressive resection: cancer growth rate is indolent, primary tumor growth is characterized predominantly by gross tumor expansion rather than invasion, and growth characteristics of hepatic metastases are similar. Accumulating experiences with these uncommon malignancies continue to support surgical intervention if both the primary cancer and the metastases are resectable. Unfortunately, in practice, the incidence of pancreatic islet cell cancers (PIC) restricts the implePresented at the 23rd Annual Meeting of the American Association of Endocrine Surgeons, Banff, Alberta, Canada, April 7-9, 2002. Reprint requests: David M. Nagorney, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905. © 2002, Mosby, Inc. All rights reserved. 0039-6060/2002/$35.00 + 0 doi:10.1067/msy.2002.128615
976 SURGERY
11/6/128615
mentation and power of randomized clinical trials comparing an aggressive surgical approach in patients with PIC and hepatic metastases with nonsurgical or medical alternatives that would determine the treatment of choice in such patients. Although resection of hepatic metastases is wellestablished for colorectal cancer,3 data supporting a similar approach for metastatic neuroendocrine cancers is still evolving. As with metastatic colorectal cancer, control of the primary cancer before resection or ablation of the hepatic metastases is imperative. Whether concurrent or staged resections of the primary cancer and the hepatic metastases is preferable in patients presenting with overt metastatic disease remains debatable. If concurrent resections of the primary cancer and the hepatic metastases can be performed as safely as staged resections, potential benefits in terms of overall survival, reduced duration of hospitalization, and, as a corollary, reduced hospitalization costs and earlier return to gainful employment would further justify this approach.4 Similar to patients with colorectal cancer, resection of hepatic metastases from PIC has been generally undertaken after
Surgery Volume 132, Number 6
resection of the primary pancreatic cancer because of the postulate that perioperative risk is increased with concurrent pancreatic and hepatic resections. Interestingly, current data regarding the perioperative risks for staged hepatic resection of either colorectal (mortality 2-4%)5 or neuroendocrine metastases (morbidity 24%, mortality 3%)2 are similar, and overall perioperative risks for pancreatic (morbidity 31%, mortality 1%)6 and colorectal resections (mortality 3%)7 are not dissimilar. Moreover, prohibitive morbidity and mortality risk for concurrent resections of colorectal cancers and hepatic metastases has not been observed by us4 or others.8 We wondered whether our experience with concurrent resections of PIC and hepatic metastases was similar. To further define the role of hepatic resection for metastatic PIC, we reviewed our experience. Our aims were to determine the actual perioperative risks of concurrent resections of PIC and hepatic metastases and to assess overall, symptom-free, and disease-free survival. METHODS The records of all consecutive patients undergoing concurrent resections of PIC and hepatic metastases at the Mayo Clinic in Rochester, Minn, between 1980 and 1998 were studied. This study period was chosen to provide a minimum following-up of 3 years, given the prognosis and indolent course of PIC from our past study.2 Patients with hepatic metastases who had resection of the PIC and biopsy of hepatic metastases or patients who had staged resections of the primary cancer and hepatic metastases despite recognition of both at initial presentation were excluded. PIC cancers were classified clinically by assay of the predominant serum hormone or immunohistochemic staining the resected PIC cancer. Patients without clinical manifestations of hormone hypersecretion or predominant hormone markers were classified as nonfunctioning islet cell cancers. Patient demographics, clinical findings, and perioperative morbidity and mortality were recorded. All pancreatic resections were limited to distal pancreatectomy, splenectomy, and regional lymphadenectomy. Hepatic resections were classified as major if ≥3 segments were removed and minor if ≤2 segments were removed. Most hepatic resections were accompanied by subsegmental nonanatomic excisions of smaller metastases (≤1 cm). Concurrent resections were classified as complete gross resection (R0/R1) if all visible tumor was removed and debulking (R2) if gross tumor remained. Residual tumor (R2) was located within the liver and was ≤10% of the hepatic metastases by volume.
Sarmiento et al 977
Hepatic resections were performed if preoperative imaging and intraoperative findings suggested that at least 90% (by volume) of the metastases were resectable. Computed tomography of the abdomen was the primary imaging modality to assess resectability of both the primary cancer and metastases. Magnetic resonance imaging was used as a complimentary imaging modality in the latter years of the study. From a technical standpoint, the distal pancreatectomy, splenectomy, and lymphadenectomy were performed initially in every patient. Hepatic resection was performed with minimal inflow vascular occlusion to reduce venous congestion of the pancreatic stump. Soft suction drains were left adjacent to the pancreatic stump and transected hepatic surface if major or extended hepatic resection was performed. Patients were followed-up for clinical recurrence of endocrinopathies (if present preoperatively), cancer recurrence or progression, and survival. Follow-up was on the basis of outpatient evaluations and updated by correspondence. Survival was calculated from the date of histologic diagnosis. The method of Kaplan and Meier was used to estimate overall progression-free and symptom-free survival. The Wilcoxon test was used for comparison of survival estimates between groups. RESULTS Twenty-eight patients presented with PIC and hepatic metastases. Five patients underwent staged resections at the discretion of the primary surgeon. The remaining 23 patients are the subject of this report. Gender distribution was even with 12 men and 11 women. Mean (± SD) age was 53 ± 12 years. The frequency of cancers by hormonal markers was: nonfunctioning, 7; glucagonoma, 6; insulinoma, 2; corticotropin-producing tumors, 2; polyfunctional tumors, 2; gastrinoma, 1; VIPoma, 1; carcinoid, 1; and 1 patient with hypercalcemia without confirmation of parathormone-like hormone production. Although 16 patients had functioning cancers, 1 patient with a glucagonoma presented with upper gastrointestinal (GI) bleeding but no clinical endocrinopathy. Therefore, endocrine symptoms were seen in 15 patients. The other 8 patients experienced abdominal pain (n = 5) sometimes associated to nausea and vomiting, presence of abdominal mass (n = 1), weight loss (n = 1), and upper GI bleeding (n = 1). Each patient in this study had a PIC located in the pancreatic body or tail. All pancreatic resections consisted of distal pancreatectomy, splenectomy, and regional lymphadenectomy and were RO/R1. Ten patients (44%) had sinistral portal
978 Sarmiento et al
Surgery December 2002
Fig 1. Overall survival of 23 patients with simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 71%; median, 76 months).
hypertension from splenic vein obstruction or invasion. Two patients required a venotomy of the portal vein to extract tumor thrombus extending from the primary tumor. Hepatic resections included: right hepatectomy, 4; left hepatectomy, 1; extended left hepatectomy, 1; bisegmentectomy, 2; segmentectomy, 7; and multiple subsegmentectomies, 8. Major hepatic resections (≥ 3 segments) were performed in 26% of the patients. Complete macroscopic resection (RO/R1) was achieved in 9 patients (39%). Fourteen patients had a debulking procedure with macroscopic residual disease in the liver (R2) but each had an estimated 90% (by volume) of the hepatic metastases resected. Hormonal markers in the 9 patients who had R0/R1 resections normalized in 6 (the other 3 had nonfunctioning tumors). Regional lymph nodes were involved in 15 patients. Regional peripancreatic lymph nodes were positive in 14 patients and celiac nodes were positive in 1 patient. Lymphatic involvement was absent in 6 patients. No lymph nodes were identified pathologically in 2 patients. There was no perioperative mortality (≤ 30 days from operation). Complications occurred in 7 patients (30%). Complications included: bile leaks, 2; pancreatic leaks, 2; wound infection, 1; delirium tremens, 1; and urinary tract infection, 1. The major complication rate related to either hepatic or pancreatic resection was 18%. Each bile or pancreatic fistula resolved with continuous suction drainage from operatively placed drains and required no further intervention. Prophylactic perioperative somatostatin analogues were not used routinely. Median duration of operation was 275 minutes (range: 120-390 minutes). Median
transfusion requirement was 1 U of packed red blood cells (range: 0-20). Blood products were necessary in 12 patients (52%). Median length of stay was 8 days (range: 5-22). On follow-up, 4 patients (17%) had new signs of pancreatic insufficiency develop that were endocrine, 2; exocrine, 1; or combined, 1. No patients required late reoperation for intra-abdominal complications from local or regional recurrence of the PIC. Overall survival was 71% at 5 years. Median survival was 76 months (Fig 1). Of the 18 patients who had resections 5 or more years before the date of analysis, 11 of these patients were actual 5-year survivors (61%). Recurrence-free or progression-free survival was only 5% at 5 years. Median recurrencefree or progression-free survival was 21 months (Fig 2). All patients who have had recurrence of the PIC had liver involvement (demonstrated by computed tomography in all but one). One patient was censored from recurrence or progression analysis because the actual date of documented recurrence was unknown but that patient died from recurrent cancer. One patient with a recurrent VIPoma underwent orthotopic liver transplantation and remains alive and disease-free at 36 months after transplantation and is the only patient without cancer recurrence currently. Four patients have undergone repeated partial hepatic resections for progression. Among the 15 patients with clinical endocrinopathies, symptom-free survival was 24% at 5 years and median symptom-free survival was 26 months (Fig 3). Each patient with an endocrinopathy had a complete (or almost complete) response after resection. There was no difference in survival between patients who had R0/R1 versus R2 resections. Median survival was 65 and 76 months for
Surgery Volume 132, Number 6
Sarmiento et al 979
Fig 2. Recurrence-free (or progression-free) survival of 23 patients with simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 5%; median, 21 months). All patients had recurrence.
RO/R1 and R2 resections, respectively (P = .90) (Fig 4). DISCUSSION We believe that the treatment for hepatic metastases from neuroendocrine cancers is resection if the primary cancer has been resected or is grossly resectable and more than 90% (by volume) of the hepatic metastases are resectable.2 Whether this approach is preferable for patients with PIC presenting with overt hepatic metastases is undefined. Concurrent resection of both the primary cancer and hepatic metastases has been criticized because of the presumed perioperative risks associated with concurrent pancreatic and hepatic resections and the poor prognosis associated with advanced-stage cancers whether functioning or nonfunctioning.9,10 In fact, data addressing the subset of patients with metastatic PIC undergoing resection is sparse and we are unaware of any series contraindicating this approach. The major findings of our study suggest that combined pancreatic and hepatic resections can be performed safely with acceptable morbidity and mortality. Moreover, this approach provides effective palliation of endocrine symptoms when present and an overall 5-year survival of 71%. Because cancer progression remains the major cause of death, both adjuvant and adjunctive treatment strategies to prevent or delay progression are needed. Our experience with concurrent resection of both PIC and hepatic metastases demonstrates that this approach can be performed safely and with acceptable morbidity. Admittedly, selection of patients was limited to PIC arising from the distal
pancreas and the extent of hepatic metastases required major resections in only 26% of patients. However, these initial data do provide a basis for further evaluation of this approach. The frequency of pancreatic and hepatic complications was similar herein. In fact, the frequency of complications for either major component of the concurrent approach herein does not differ from that reported for isolated pancreatectomies (30-40%)6 or partial hepatectomies (30%)11 for malignancy. We do not believe that this incurred morbidity precludes further pursuit of the concurrent approach. However, careful selection of patients is necessary. Clinical performance status should exceed a Karnofsky score of 50% (Eastern Cooperative Oncology Group performance status <2) and other comorbidities should be limited. Complete resection of the PIC should precede the partial hepatectomy. If intraoperative problems during the pancreatic resection arise, partial hepatectomy should be deferred. We have not performed concurrent pancreaticoduodenectomy and major hepatic resection and currently favor staged resections in these patients. We would not be reluctant to combine pancreaticoduodenectomy with minor hepatic resections. The natural history of unresected PIC with hepatic metastases is poor. Despite a more prolonged survival compared with ductal pancreatic cancers of similar stage, survival only has ranged from 1.7 to 3.1 years.12,13 Cancer progression is the primary cause of death in most of these patients and intra-abdominal complications, such as intestinal hemorrhage or obstruction related to the primary cancer, has accounted for a majority of the
980 Sarmiento et al
Surgery December 2002
Fig 3. Symptom-free survival of 15 patients with clinical endocrinopathy submitted to simultaneous resection of PIC tumors and synchronic hepatic metastases (5-year survival, 24%; median, 26 months).
Fig 4. Comparative survival of 23 with simultaneous resection of PIC tumors and synchronic hepatic metastases; 9 patients received a complete gross (R0/R1) resection (5-year survival, 76%; median, 65 months) versus 14 patients with a debulking (R2) resection (5-year survival, 67%; median, 76 months); P = .90 by the Wilcoxon test. Solid line, R0/R1; broken line, R2.
remainder. Quality of life in patients with metastatic PIC is impaired not only from complications of the primary cancer and the usual constitutional symptoms of cancer but also clinical endocrinopathies. Despite improvements in clinical control of endocrine symptoms with somatostatin analogues,14 and objective responses to multimodality chemotherapy,15 overall survival has improved min-
imally. The current outcomes for nonsurgical approaches do not warrant categoric exclusion of surgical approaches. Indeed, accumulating data have suggested that resection of PIC and their metastases in selected patients can prolong survival compared with patients of similar stage historically.16 Consequently, we have continued to explore the role of resection in the multimodality treat-
Surgery Volume 132, Number 6
ment of patients with advanced stage PIC. We believe that resection, though not often curative in this clinical setting, may delay other treatments and does not prevent their eventual implementation. The indications for resection of PIC in patients presenting with overt hepatic metastases have been primarily palliative in intent. Resection of the PIC in the presence of hepatic metastases has been advocated specifically to relieve intractable pain or to palliate GI obstruction or hemorrhage from advanced local tumor growth for improvement in quality of life.13 Although PIC of the head of the pancreas more frequently are associated with these clinical problems, patients with PIC in the body or tail of the pancreas can also have life-threatening complications develop. Indeed, 44% of the patients in this series had sinistral portal hypertension from obstruction or thrombosis of the splenic vein. GI hemorrhage from gastric varices can be life threatening and warrant distal pancreatectomy and splenectomy in selected patients.17,18 However, only a minority of patients with metastatic PIC will require palliative resections of the primary cancer for complications. We believe that resection of PIC is contraindicated in the presence of undebulkable hepatic or distant metastases if patients are asymptomatic. Most of these patients with metastatic PIC should be treated with palliative chemotherapy, somatostatin analogues, or both depending on the presence of functional endocrinopathies. However, patients in whom the primary cancer is resectable and in whom near total hepatic debulking is possible should be assessed for resection. These patients are the focus of this report further exploring the role of concurrent debulking of both disease sites. Cytoreductive or debulking resections for functioning neuroendocrine metastases after resection of the primary cancer has evolved over the last decade. The role of debulking of metastatic PIC was examined in small series in the early 1990s.1,2 Clinical endocrinopathies were the indications for hepatic resections. Relief from endocrine symptoms and prolongation of survival was evident. In one report, patients who were unresected had an overall survival of 28% at 5 years, but a 79% survival at 5 years debulking.16 Although 14 patients underwent partial hepatectomy and 11 patients had a partial pancreatectomy, these data did not specify how many patients had concurrent resections. Resections were curative in intent in 80% of these patients. There were no deaths and the complication rate was 30%. Despite disease recurrence in all patients, all experienced reduction of antisecretory medications and their symptoms. Similar findings have been reported previously by ourselves1,2 and
Sarmiento et al 981
others.13,19 Although some of these patients had concurrent resections, our data herein comprise the largest series of patients undergoing concurrent resections of the primary PIC and a hepatic metastases to our knowledge. The outcomes in terms of symptomatic response and survival were similar to other reports that predominantly represent outcomes from staged resections. We are unaware of any data comparing staged versus concurrent resections for such patients. Whether concurrent resections are warranted in patients with nonfunctioning PIC is unknown. In contrast to patients with functional endocrinopathies in whom the degree of achievable palliation justifies the perioperative risks, significant improvement in survival is required to justify the perioperative risks with nonfunctioning PIC. Although we continue to explore the role of resection of nonfunctioning PIC with hepatic metastases in selected patients to determine whether survival will improve, we remain concerned regarding this aggressive approach in such patients. Clearly, overall survival for such patients is limited if no tumor is resected.12,20 However, despite selection, survival was greater in patients undergoing resection of either the primary nonfunctioning islet cell cancer, alone, or in patients after resection of both the primary and the hepatic metastases.13 Interestingly, in a prior study, we were unable to demonstrate a difference in survival between functioning and nonfunctioning PIC for similar extent of debulking.20 These findings suggest that resection may improve survival in patients with metastatic PIC and may be unique to GI neuroendocrine cancers compared with other solid abdominal malignancies. Finally, regardless of functional status or timing of hepatic resection, disease progression is typical and the most common cause of death.13 We agree that both adjunctive (present of gross residual disease) and adjuvant (absence of gross residual disease) treatment strategies should be investigated. Currently, we are evaluating the role of ongoing percutaneous radiofrequency ablation for hepatic metastases after hepatic regeneration and orthotopic liver transplantation for hepatic metastases that do not meet criteria for debulking. Because the liver is the major site of progression, perhaps ablation of hepatic metastases as they arise may affect survival and control of endocrine symptoms if present. Systemic chemotherapy or regional chemoembolization should compliment ablative and resective approaches. Nonetheless we believe that resections of PIC with hepatic metastases in selected patients meeting the previously stated criteria should be undertaken initially in patients who
982 Sarmiento et al
are fit and, if technically feasible, concurrently. Surgical resection does not preclude implementation of other strategies.
REFERENCES 1. McEntee GP, Nagorney DM, Kvols LK, Moertel CG, Grant CS. Cytoreductive hepatic surgery for neuroendocrine tumors. Surgery 1990;108:1091-6. 2. Que FG, Nagorney DM, Batts KP, Linz LJ, Kvols LK. Hepatic resection for metastatic neuroendocrine tumors. Am J Surg 1995;169:36-42. 3. Treatment of hepatic metastases from colorectal cancer: the SSAT, AGA, ASLD, ASGE, AHPBA consensus panel. J Gastrointest Surg 1997;1:396-7. 4. Chua HK, Wolff BG, Sordenna K, Tsiotos GG, Larson DR, Nagorney DM. Is combined colectomy and hepatectomy for synchronous metastatic colorectal cancer more efficient than staged? Gastroenterology 2000;118:1027. 5. Jamison RL, Donohue JH, Nagorney DM, Rosen CB, Harmsen WS, Ilstrup DM. Hepatic resection for metastatic colorectal cancer results in cure for some patients. Arch Surg 1997;132:505-10. 6. Lillemoe KD, Kaushal S, Cameron JL, Sohn TA, Pitt HA, Yeo CJ. Distal pancreatectomy: indications and outcomes in 235 patients. Ann Surg 1999;229:693-8. 7. Harmon JW, Tan DG, Gordon TA, Bowman HM, Choti MA, Kaufman HS, et al. Hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg 1999;230:404-13. 8. Rosen SA, Buell JF, Yoshida A, Kazsuba S, Hurst R, Michelassi F, et al. Initial presentation with stage IV colorectal cancer: how aggressive should we be? Arch Surg 2000;135:530-5. 9. Chen H, Hardacre JM, Uzar A, Cameron JL, Choti MA. Isolated liver metastases from neuroendocrine tumors: does resection prolong survival? J Am Coll Surg 1998;187:88-93. 10. Chamberlain RS, Canes D, Brown KT, Saltz L, Jarnagin W, Fong Y, et al. Hepatic neuroendocrine metastases: does intervention alter outcomes? J Am Coll Surg 2000;190: 432-45. 11. Blumgart LH, Fong Y. Surgical options in the treatment of hepatic metastases from colorectal cancer. Curr Probl Surg 1995;32:333-421. 12. Thompson GB, van Heerden JA, Grant CS, Carney JA, Ilstrup DM. Islet cell carcinomas of the pancreas; a twentyyear experience. Surgery 1988;104:1011-7. 13. Solorzano CC, Lee JE, Pisters PW, Vauthey JN, Ayers JD, Jean ME, et al. Nonfunctioning islet cell carcinoma of the pancreas: survival results in a contemporary series of 163 patients. Surgery 2001;130:1078-85. 14. Ericksson B, Oberg K. Summing up 15 years of somatostatin analog therapy in neuroendocrine tumors: future outlook. Ann Oncol 1999;10:531-8. 15. Moertel CG, Lefkopoulo M, Lipsitz S, Hahn RG, Klaassen D. Streptozocin-doxorubicin, streptozocin-fluorouracil or chlorozotocin in the treatment of advanced islet-cell carcinoma. New Engl J Med 1992;326:519-23. 16. Carty SE, Jensen RT, Norton JA. Prospective study of aggressive resection of metastatic pancreatic endocrine tumors. Surgery 1992;112:1024-32. 17. Metz DC, Benjamin SB. Islet cell carcinoma of the pancreas presenting as bleeding from isolated gastric varices. Dig Dis Sci 1991;36:241-4.
Surgery December 2002
18. Loftus JP, Nagorney DM, Ilstrup D, Kunselman AR. Sinistral portal hypertension: splenectomy or expectant management. Ann Surg 1993;217:35-40. 19. Norton JA, Doherty GM, Fraker DL, Alexander HR, Doppman JL, Venzon DJ, et al. Surgical treatment of localized gastrinoma within the liver: a prospective study. Surgery 1998;124:1145-52. 20. Sarmiento JM, Heywood G, Ilstrup DM, Nagorney DM, Que FG. Surgical treatment of neuroendocrine metastases to the liver. Hepatology 2001;34:236A.
DISCUSSION Dr Janice L. Pasieka (Calgary, Alberta, Canada). I see you were very fortunate that all of your primary tumors were in the body and tail of the pancreas. Could you please tell us what you would do in a situation where the primary is in the head of the pancreas with liver metastases? Would you proceed with a Whipple resection and resection of the liver metastases? Dr Sarmiento. This is not a matter of luck; this is a matter of selection. Obviously, we are not doing a Whipple procedure plus a right hepatectomy. But we think we can still do a Whipple procedure and a minor resection of the liver if necessary. Dr Pasieka. How small is small? Dr Sarmiento. (Essen, Germany) One or 2 segments of the liver. Everything depends on the intraoperative evaluation of the patient. Dr Andrea Frilling (Essen, Germany). May I address 2 questions? You said that all of your patients had recurrent tumor. In the group of patients in whom liver resection was performed, there were 6 patients with R0 resection. Where did these patients get their recurrence? Was it within the liver or extrahepatically? The second question would be, are there patients in your series in whom you performed metachronous liver resection? That means first pancreatic resection and, a few weeks later, liver resection? Dr Sarmiento. All the recurrences that we saw were in the liver. That is true. We cannot say that we are curing these patients even when we achieve complete microscopic clearance of the tumor. The issue is that, if in a group of 23 patients with a 100% recurrence, speaking of R0 versus R1 resections is just semantics. Again, all the patients had recurrence of the liver. We had 28 patients presenting with synchronous metastases. Twenty-three were resected concurrently and 5 were resected as staged procedures. Therefore, with only 5 patients in the control group, it is very risky to compare the complication rate and the survival. That is why we did not perform that comparison. Dr Frilling. Just 1 additional question. Do you perform intraoperative ultrasound before you resect the patient? Dr Sarmiento. Definitely. Dr Sally E. Carty (Pittsburgh, Pennsylvania). It is very admirable that you were able to look at a large enough series of patients in this disease. I note that you had fewer gastrinomas than many comparable series and also a higher rate of palliative operation compared with complete resection. That is not your fault, that is the extent
Sarmiento et al 983
Surgery Volume 132, Number 6 of the patients’ disease. But I would just like to inquire about your last conclusion, improvement in survival, I may have missed something, but your data showed no difference in survival among patients with palliative compared with complete resection. And it is good to be careful in making statements that compare results with historic controls. What we need is a randomized, prospective study. The problem is, who is going to sign up for the nonoperated arm?
O
Dr Sarmiento In your article in 1992,16 you also say that it was very difficult to get enough patients for a randomized trial. And all the data that I saw from your article are from that study without control group. So, basically, we cannot produce data in a randomized fashion. What we say, here, is that performing resection, complete or incomplete resection, is better than not doing resection at all. That is what I mean when I refer to improvement in survival.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 800-654-2452 Outside the US, call 407-345-4000