Curative Treatment for Pancreatic Neoplasms: Radical Resection

0039-6109/95 $0.00 + .20

PANCREATIC NEOPLASMS

CURATIVE TREATMENT FOR PANCREATIC NEOPLASMS Radical Resection Howard A. Reber, MD, Stanley W. Ashley, MD, and David McFadden, MD

BACKGROUND AND EPIDEMIOLOGY

In the United States, cancer of the exocrine pancreas is the fourth leading cause of cancer death in men and the fifth in women. In 1995, 27,000 new cases will be diagnosed and almost as many patients will die of advanced disease. Ductal adenocarcinoma accounts for more than 90% of exocrine pancreatic tumors, and it is a characteristically aggressive lesion. At the time of diagnosis, the tumor is confined to the pancreas in fewer than 10% of patients, 40% have locally advanced disease, and more than 50% have distant spread. More than 95% of patients eventually die of their disease. Even after resection for cure, the median length of survival is only 18 to 20 months, and the 5-year survival rate is about 10%. Unfortunately, these statistics have changed little in the last 25 years. Some improvements have been made, however. The diagnosis is being made up to 6 months earlier than it was a decade ago, probably because of improved diagnostic tests such as computed tomography and endoscopic retrograde cholangiopancreatography. There has also been an increase in the resectability rate from 15% up to about 25% in the United States, probably due mostly to a more critical selection of those patients who come to surgery. Some patients with obviously unresectable disease are managed nonoperatively today, such

From the Department of Surgery, University of California at Los Angeles Medical Center, Los Angeles, California

SURGICAL CLINICS OF NORTH AMERICA VOLUME 75 • NUMBER 5 • OCTOBER 1995

905

906

REBER ct al

as with stents to relieve biliary obstruction. Finally, in the hands of experienced surgeons, the mortality rate for a Whipple pancreaticoduodenectomy is now less than 5%. RATIONALE FOR RADICAL RESECTION

Cubilla et a1,1 in a classic study published in 1978, determined the lymph node groups usually involved in pancreatic cancer by analyzing surgical specimens from 22 patients with the disease. They divided the peripancreatic lymph nodes into five main groups, with specific sub: groups in each. They were designated (1) superior: superior head, superior body, gastric; (2) inferior: inferior head, inferior body; (3) anterior: anterior pancreaticoduodenal, pyloric, mesenteric; (4) posterior: posterior pancreaticoduodenal, common bile duct; and (5) splenic: splenic hilus and tail of pancreas. Pancreatic ductal adenocarcinoma tended to metastasize to multiple lymph nodes of the superior head, superior body, and posterior pancreaticoduodenal groups (88% of patients). Other nodes were involved less commonly, but no patients had metastases in the gastric or splenic nodes. They pointed out that one third of the patients had nodal metastases in areas not normally resected with the standard Whipple resection. In an effort to remove these areas of tumor extension, Fortner developed a more radical pancreatic resection, which resected adjacent soft tissue, lymph nodes, and a segment of the superior mesenteric-portal vein and superior mesenteric artery in many patients. 2 The operative mortality rate was about 15%, and the 5-year survival (20%) did not appear to be any better; thus, few surgeons adopted the radical approach. The National Cancer Institute's more recent evaluation of 20 patients treated with this so-called regional pancreatectomy confirmed higher operative morbidity (55%) and mortality (20%) rates, with no improvement in survival over historical controls treated with standard pancreaticoduodenectomy.19 However, additional reports from Japan suggested that more extensive resections might be useful. Nagai et aP4 reported their analysis of autopsy specimens from eight patients with pancreatic cancer. They found that half of the patients had metastatic tumor in the lymph nodes located between the superior and the inferior mesenteric arteries. This tissue was not routinely included in the field of the standard pancreaticoduodenectomy. Dissections of pancreaticoduodenectomy specimens by others revealed retroperitoneal invasion in 89% of patients and involvement of the extrapancreatic neural plexuses in 62%.10, 15 This occurred even with small tumors. Tsuchiya et aF2 reported on the relationship between tumor size and a variety of other prognostic factors, including lymph node involvement. The regional lymph nodes were studied in 108 cases of carcinoma of the head of the pancreas collected from 10 institutions throughout Japan. These nodes were classified according to the following scheme: (1) The primary group of nodes (N 1 ) were located closest to the tumor;

RADICAL RESECTION

907

(2) regional nodes more distant from the tumor were designated the secondary group of nodes (N2); and (3) the tertiary nodes (N3) were farthest away. Even in the smallest tumors (Tl' <2 cm in diameter), 10 of 22 specimens had nodal metastases, including 2 in the N3 group. As tumor size increased, the frequency of nodal involvement was greater. The cumulative survival rate was also influenced by nodal involvement. At 1 year, the survival was 83% when no nodes were involved, 45% when tumor had spread to the Nl group, and 17% when the N2 group was involved. There were no survivors when the cancer had spread to the N3 group. Several conclusions came from this study: (1) Nodal involvement by tumor was viewed as an important prognostic factor, and (2) small tumors «2 cm) were not necessarily to be considered early tumors because nodal involvement was present in almost 50% of cases. A short time later, Tsuchiya et aFl analyzed the Japanese experience with specific regard to small pancreatic cancers. They found that patients with negative nodes after surgical resection had an 89% likelihood of being alive 1 year later, whereas those with positive nodes had only a 59% chance. They also found that 29% of 45 patients with the most favorable tumors (i.e., <2 cm diameter, negative nodes, no invasion of surrounding structures) eventually died of tumor recurrence. Based in part on these data, they recommended that the regional nodes and adjacent soft tissue be removed. These and other similar studies convinced many Japanese surgeons to perform an extended resection that included the removal of lymph nodes and connective and neural tissue and resection of segments of the major vessels. In many ways, this extended resection was modeled after Fortner's original regional pancreatectomy.

TECHNIQUE OF EXTENDED PANCREATICODUODENECTOMY

The operation consists of a pancreaticoduodenectomy or sometimes a total pancreatectomy, accompanied by extensive retroperitoneal lymph node and soft tissue resection. Often the superior mesenteric-portal vein and, less commonly, the superior mesenteric artery are resected as well. 8, 9, 12, 13, 20, 23 Vascular reconstruction is generally accomplished by end-toend anastomosis, although saphenous vein or prosthetic grafts have been used. 8, 12, 13,20 The lymphadenectomy requires a wide retroperitoneal dissection of the lymph nodes from the level of the origin of the celiac axis superiorly to the iliac bifurcation inferiorly. The lateral extent of the dissection is limited by the kidneys. The removed nodes include the pyloric, superior and inferior pancreatic head, common bile duct, anterior pancreaticoduodenal, and superior mesenteric lymph node groupS.9 Para-aortic and celiac lymphadenectomy and neurectomy are also performed en bloc. The superior mesenteric-portal and splenic veins, hepatic and mesenteric arteries, and celiac axis are skeletonized. For lesions

908

REBER et al

in the head of the pancreas, the pancreas is transected to the left of the aorta. The additional dissection, including the vascular reconstruction, adds 60 to 90 minutes to the operation. COMPLICATIONS

Complications of the extended resection are similar to those that occur with the standard pancreaticoduodenectomy.8, 16, 18 The morbidity rate varies between 25% and 40%. One group found a significantly higher rate of pancreatic fistula in patients who had undergone extended pancreaticoduodenectomy, however.25 Operative mortality rates range from 4% to 14%, and the average (6.4%) is similar to that for the standard Whipple operation.5, 8, 9,12,13,16,18,20,21,25 Most reports describe some deaths due to postoperative hepatic failure. 8, 18,23,25 Intractable diarrhea may also be a problem for some patients who have undergone an extensive neurectomy, and most Japanese surgeons no longer include this as part of the operation. RESULTS

The influence of the extent of the nodal and soft tissue resection was studied by Ishikawa et a1. 9 The retrospective review revealed several interesting points. A total of 53 patients undergoing resection between 1971 and 1983 were evaluated. Thirty-two patients underwent standard pancreaticoduodenectomy, and 21 underwent the extended operation. The 3-year cumulative survival rates for the two groups were 13% and 38%, respectively. For patients with positive nodes, there were no 3-year survivors in the first group, but four (27%) patients with positive nodes in the second group survived. These data suggested that the more extensive resection improves survival in patients subsequently shown to have involved lymph nodes. Similar results from other Japanese surgeons have appeared in the literature. 5, 12, 13, 16, 18, 20, 23 Tashiro et apo from Kumamoto University reported on 14 patients with pancreatic cancer (most with stage III or IV disease) who underwent the extended pancreaticoduodenectomy and also received 30 Gy of intraoperative radiation. The 5-year survival rate was 33%. OzakP7 reported on 16 patients, of whom 4 had cancers of the body or tail of the gland. Survival at 1 year was 88% and at 3 years was 53%. Manabe et al13 performed 42 standard Whipple operations with a 9.5% operative mortality rate and no 3-year survivors. Thirty-two patients underwent extended pancreaticoduodenectomy, with a 6.2% operative mortality rate and a 5-year survival of 33.5%. Manabe's patients received neither chemotherapy nor radiation therapy. Nagakawa et aP6 reviewed their experience with 134 patients with pancreatic cancer treated since 1973. Sixty-one patients underwent the extended pancreaticoduodenectomy, a 45.5% resectability rate. Forty-

RADICAL RESECTION

909

nine of these patients had a "curative resection," based on macroscopic evidence at surgery and later microscopic examination of the specimen, with an operative mortality rate of 14% at 60 days. In the subgroup of 13 patients without lymph node metastases, the 5-year survival rate was 66%. Lymph nodes adjacent to the pancreas were involved in 21 patients, of whom only 2, or 9%, survived for 5 years. Lymph node metastases are found in approximately 75% of patients25 and are related to the number of lymph nodes captured within the specimen. Only 47% of lymph nodes are involved with tumor if less than 20 nodes are removed; 80% contain tumor if 30 or more nodes are resected. 2s Nagakawa et aP6 analyzed other pathologic factors to determine their prognostic significance. The 5-year survival rate of patients with pancreatic cancers less than 2 cm in diameter was 67%, compared with 37% in the 14 patients with tumors between 4 and 6 cm in diameter. The presence of invasion of the retroperitoneal tissue was also important; its absence conferred a 73% 5-year survival rate. When the tissue was involved, the 5-year survival rate was 14%. Satake et aP8 recently reported on 185 patients with pancreatic tumors less than 2 cm in size. The resectability rate for these small lesions was 98.3%. Only patients with Stage II disease (40% of the entire group) had better 5-year survival rates with the more radical operation. Otherwise, there were no differences in survival in the groups that had extended versus standard pancreaticoduodenectomy. Selected statistics for 651 Japanese pancreatic cancer patients who underwent the extended resection are shown in Table 1. Survival rates 5 years after operation range from 0% to 30%, but the mean 5-year survival rate of evaluable patients was 13%. Although the data are not truly comparable, this is similar to the survival rate of 9% in almost 1700 patients treated by standard pancreaticoduodenectomy in the United StatesY Kawarada from Mie University in Japan pointed out

Table 1. SELECTED REPORTS OF JAPANESE EXPERIENCE WITH THE EXTENDED PANCREATICODUODENECTOMY FOR PANCREATIC CANCER Reference Number

No. Patients

Operative Mortality Rate

15 16 9 5 21 20 13 8 12 18 4

66 61 59 167 106 17 21 35 1 57 17

4.5% 6.1 % 10% 5.3% 4% 5% 14% 6% 0% 7% 0%

607

6.2% Mean

Totals

Survival

3% 5-yr 26% 5-yr 9% 5-yr 8% 5-yr 30% 5-yr 6% 5-yr 0% 5-yr 23% 3-yr 12 mos 28% 5-yr 15% Mean

Other Treatment

Chemotherapy None None None None None None None Radiation None Radiation

910

REBER et al

recently that some of the apparent improvement in 5-year survival rates may be due to statistical treatment of the data (personal communication, November, 1994). Thus, Kaplan-Meier projections of survival may not reflect the actual cumulative survival rates, especially because the actual number of survivors in many series is quite low. PROBLEMS WITH EVALUATION OF RESULTS

In Japan, as elsewhere, adjuvant therapy has frequently been combined with the pancreatic resection. This has confused efforts to assess the value of the extended operations. For example, intraoperative radiation therapy supplemented the surgery in several clinics4 , 17 and one achieved an impressive 33% 5-year cumulative survival rate in 14 patients. 7 Intraoperative and postoperative chemotherapy (hepatic arterial infusion with mitomycin C) along with the extended resection produced 1- and 3-year cumulative survival rates for 16 patients of 88% and 55%.17 A number of proposals for staging of pancreatic carcinoma have been described to enable comparisons of the results from different centers. The most recent stage classifications are those of the Union Internationale Contre Ie Cancer (UICC) and the Japanese Pancreas Society, both published in 1987.6,24 The UICC stage classification, which is simple to apply, tends to classify patients in a less advanced stage than does the Japanese system. This is because the UICC system considers lymph node metastasis the most important prognostic factor. The Japanese system takes lymph node involvement into account but also considers the influence of retroperitoneal and vascular invasion, both of which also have important prognostic implications. Although the Japanese system is probably more accurate, it is more complicated to use. Although the fourth edition of the UICC TNM classification of tumors has been accepted by all national TNM committees, most of the Japanese reports have used the Japanese staging system to describe the tumors that have been treated. Many of the American and European reports either do not accurately report tumor stage or use the UICC system. Thus, it is impossible to know whether the different results are due to the different treatments or differences in the stages of the disease. The apparent differences in survival among countries have also led some to question whether the biology of the tumors varies in the different populations under study. In a preliminary analysis of the pathologic characteristics of 272 ductal adenocarcinomas resected in Europe, Japan, and the United States, the Japanese patients appeared to have a higher proportion of well-differentiated tumors.3 Although this implies a better prognosis with any form of treatment, the tumors in the Japanese patients were also more advanced at the time of diagnosis. The analysis suggested that Japanese patients did indeed survive longer than the combined European-American groups. It remained unclear whether the longer survival was due to biologic differences in tumor behavior or the more radical surgery performed in most of the Japanese cases.

RADICAL RESECTION

911

SUMMARY

The available data suggest that lymph node involvement is an important prognostic factor in patients with carcinoma of the head of the pancreas. Lymph node metastases occur in as many as 50% of the cases of even the smallest pancreatic cancers now being diagnosed and resected (i.e., those <2 cm in diameter). There is some evidence, especially from clinical experience in Japan, that wider lymphatic dissections (i.e., wider than those commonly done with the standard Whipple resection) may prolong survival. Unfortunately, many of the available data around the world are retrospective and are not randomized between the standard and the radical operation. Moreover, the pathologic material has not been staged uniformly according to accepted criteria. Thus the various series are not comparable. Comparisons between series require standardization with respect to stage of disease, pathologic classification, and treatment protocols. Before any modification of the standard pancreaticoduodenectomy is adopted, an appropriately designed study should be performed to test its efficacy. This study would also require a more comprehensive analysis of the pathologic material than is commonly performed today in the United States and Europe.

References 1. Cubilla AL, Fortner J, Fitzgerald PJ: Lymph node involvement in carcinoma of the pancreas area. Cancer 41:880-887, 1978 2. Fortner JG: Regional pancreatectomy for cancer of the pancreas, ampulla, and other related sites. Ann Surg 199:418-425, 1984 3. Glaser S, Longnecker D, Kato Y, et al: Comparison of histologic type and stage of exocrine pancreatic neoplasms from surgical series in Europe, Japan, and the United States [abstract]. Lab Invest 66:97A, 1992 4. Gotoh M, Monden M, Sakon M, et al: Intraoperative irradiation in resected carcinoma of the pancreas and portal vein. Arch Surg 127:1213-1215, 1992 5. Hanyu F, Suzuki M, Imaizumi T: The Whipple operation for pancreatic carcinoma: Japanese experience. Presented at Cancer of the Pancreas; Molecular Biology and Progress in Diagnosis and Treatment, Meersburg, Germany, June 1993 6. Hermanek P: Staging of exocrine pancreatic carcinoma. Eur J Surg Oncol 17:167-172, 1991 7. Hiraoka T, Uchino R, Kanemitsu K, et al: Combination of intraoperative radiation with resection of cancer of the pancreas. Int J Pancreatol 7:201-207, 1990 8. Ishikawa 0, Ohigashi H, Imaoka S, et al: Preoperative indications for extended pancreatectomy for locally advanced pancreas cancer involving the portal vein. Ann Surg 215:231-236, 1992 9. Ishikawa 0, Ohigashi H, Sasaki Y, et al: Practical usefulness of lymphatic and connective tissue clearance for carcinoma of the pancreas head. Ann Surg 208:215-220, 1988 10. Kayahara M, Nagakawa T, Konishi I, et al: Clinicopathological study of pancreatic carcinoma with particular reference to the invasion of the extrapancreatic neural plexus. Int J Pancreatol 10:105-111, 1991 11. Livingston EH, Welton ML, Reber HA: Surgical treatment of pancreatic cancer: The United States' experience. Int J Pancreatol 9:153-157, 1991

912

REBER et al

12. Manabe T, Baba N, Setoyama H, et al: Venous bypass grafting for celiac occlusion in radical pancreaticoduodenectomy. Pancreas 6:368--371, 1991 13. Manabe T, Suzuki T, Tobe T: Evaluation of en bloc radical pancreatectomy for carcinoma of the head of the pancreas involving the adjacent vessels. Dig Surg 2:27-30,1985 14. Nagai H, Kuroda A, Morioka Y: Lymphatic and local spread of T1 and T2 pancreatic cancer. Ann Surg 204:65-71, 1986 15. Nagakawa T, Kayahara M, Ohta T, et al: Patterns of neural and plexus invasion of human pancreatic cancer and experimental cancer. Int J Pancreatol 10:112-119, 1991 16. Nagakawa T, Konishi I, Ueno K, et al: Surgical treatment of pancreatic cancer: The Japanese experience. Int J Pancreatol 9:135-143, 1991 17. Ozaki H: Improvement of pancreatic cancer treatment. Int J Pancreatol 12:5-9, 1992 18. Satake K, Nishiwaki H, Yokomatsu H, et al: Surgical curability and prognosis for standard versus extended resection for T1 carcinoma of the pancreas. Surg Gynecol Obstet 175:259-265, 1992 19. Sindelar WF: Clinical experience with regional pancreatectomy for adenocarcinoma of the pancreas. Arch Surg 124:127-132, 1989 20. Tashiro S, Uchino R, Hiraoka T, et al: Surgical indication and significance of portal vein resection in biliary and pancreatic cancer. Surgery 109:481-487, 1991 21. Tsuchiya R, Noda T, Harada N, et al: Collective review of small carcinomas of the pancreas. Ann Surg 203:77-81, 1986 22. Tsuchiya R, Oribe T, Noda T: Size of the tumor and other factors influencing prognosis of carcinoma of the head of the pancreas. Am J Gastroenterol 80:459-462, 1985 23. Tsuchiya R, Tsunoda T, Yamaguchi T: Operation of choice for resectable carcinoma of the head of the pancreas. Int J Pancreatol 6:295-306, 1990 24. Tsukasa T, Ura K, Eto T, et al: UICC and Japanese stage classifications for carcinoma of the pancreas. Int J Pancreatol 8:205-214, 1991 25. Zirngibl H, Gall FP: Cancer of the pancreas: Extensive lymph node dissection. Presented at Cancer of the Pancreas; Molecular Biology and Progress in Diagnosis and Treatment, Meersburg, Germany, June 1993 Address reprint requests to

Howard A. Reber, MD Department of Surgery UCLA Medical Center CHS 72-215 10833 Le Conte Avenue Los Angeles, CA 90024