535
Aust. N . Z . J . Sirrg. 1992.62, 545-549
LONG SURVIVORS AFTER PANCREATODUODENECTOMY FOR PANCREAS HEAD CARCINOMA KOJI YAMAGUCHI,* KAZUYOSHI NISHIHARA PIOTR , ~ KOLODZIEJCZYK,.~ A N D MASAZUMI TSUNEYOSH? Departments of *Surgery I and 'PatholoRy I I , KxLishu Universiry Fuculh
of Medicine, Fukuoka, JapLlrl
Twelve Japanese patients with pancreas head carcinoma w h o survived 3 years or more after a pancreatoduodenectomy and 50 who survived less than 12 months were reviewed clinicopathologically.The 12 patients who survived for 2 3 years exhibited more favourable prognostic factors: a higher incidence of jaundice; a smaller mass; a higher prevalence of an earlier stage tumour and adenocarcinoma of differentiated type; and a lower incidence of venous invasion, lymph node metastasis. and cancer cells at the surgical margins. However the difference was not significant. Univariate log-rank analysis regarding 13 prognostic variables showed that histologic type was a significant factor but multivariate Cox regression analysis failed to reveal an independent significant parameter. Nine of the 12 long-term survivors showed lymph node metastasis and six of the 12 revealed cancer cells at the surgical margins. Six of the I2 long-term survivors died from local recurnnce and/or distant metastasis 37-78 months after operation. Only two patients survived more than 5 years after the operation. At the time of writing. one of them was still alive and another was dead 78 months after the operation. Pancreatoduodenectomy for pancreas head carcinoma infrequently offers a permanent cure for the patients with pancreas head carcinoma but sometimes produces a worthwhile long-term survival, even if the resected margins were affected by malignant cells or the lymph node metastasis was evident.
.
Key words: long-term survivor, pancreas carcinoma, pancreatoduodenectomy
Introduction Despite the recent advances in the diagnostic and therapeutic modalities, the clinical course of patients with pancreas carcinoma remains gloomy. Some surgeons abandon a radical resection for pancreas carcinoma. Nevertheless, surgical resection does lead to a long-term survival or even to a cure for a few patients. Only surgical resection gives patients with pancreas carcinoma an opportunity to live long and to be cured permanently. It is important to identify this small subpopulation in whom the risks of pancreatoduodenectomy would be justified. It may be equally important to identify those patients with resectable tumours who might survive only a few months after pancreatectomy. In these patients, the radical resection could be avoided and a less dangerous, palliative procedure chosen. The clinicopathological and immunohistochemical features of 96 patients with pancreas carcinoma have been reported p r e v i o ~ s l yIn . ~ the present study, 12 patients with pancreas head carcinoma who surCorrespondence: Koji Yamaguchi. MD, Department of Surgery I. Kyu5hu University Faculty of Medicine. 3- I - I Maidaahi, Higashi-
ku. Fukuoka X12, Japan. Accepted for puhlication 13 February 1992
vived 3 years after a pancreatoduodenectomy are compared with SO who survived < 12 months in order to introduce a potential prognostic variable.
Methods A total of 86 Japanese patients with pancreas head carcinoma who had undergone a pancreatoduodenectomy at one of 25 Japanese institutions including Kyushu University Hospital were studied. Only ordinary adenocarcinoma of ductal origin in the head of the pancreas was included. Carcinoma of the body or tail of the pancreas. carcinoma of the ampulla of Vater. carcinoma of the bile duct, cystic neoplasm of the pancreas, papillary cystic neoplasm of the pancreas. islet cell tumour. sarcoma and all other malignant lesions were excluded in order to achieve a histogenetically pure group of carcinomas. The clinical charts were available for all 86 patients. Macroscopically, the tumours were divided into three groups according to tumour margin: expansive, infiltrative and intermediate type. Histopathologically, the tumours were classified as well, moderately, and poorly differentiated adenocarcinoma and others (adenosquamous carcinoma and undifferentiated carcinoma). The histopathologic tumour margin (INF) was divided into three
546
YAMAGUCHI
types: well defined and expansive (alpha), intermediate (beta) and infiltrative (gamma). The tuniours were classified into four stages according to the American Joint Committee on Cancer.5 The pancreatoduodenectomy specimens were examined by step-wise tissue sections at S mm intervals. All sections were stained with haematoxylin and eosin and were reviewed by one of the authors (KY). The clinical follow-up was current as of 30 May 1991 and the data were available for 85 of the 86 patients excluding one patient who was lost at clinical followup. Fifty of the 85 patients died within 1 year of surgery. Twelve patients survived > 3 years after surgery. Twenty patients were either doing well or were dead 1-3 years after surgery and three patients died within 1 month after surgery. These 23 patients were excluded from the current study in order to obtain two dichotomous groups of longterm and short-term survivors with no overlapping. The current study is comprised of the 12 long-term survivors and the SO short-term survivors. Six of the 12 long-term survivors are doing well at clinical follow-up 37-78 months after surgical intervention
and the remaining six patients died from local recurrence or distant metastases between 37 and 78 months after the operation. The mean values of age and size were examined by Student’s t-test and the distribution of prognostic factors was measured by Chisquared test. Univariate log-rank and multivariate Cox regression analyses were calculated for 13 potential factors to find a significant prognostic variable.
Results C L I N I C A L FINDINGS
The 12 long-term survivors comprised five men and seven women ranging from 37 to 76 years of age with a mean of 62 years (Tables 1.2). The SO shortterm survivors comprised 31 men and 19 women who ranged from 34 to 77 years old with a mean age of 62 years. Eleven (92%) of the twelve longterm survivors developed icterus, compared with 35 (70%) of the SO short-term survivors. The 12 tumours ranged from 25 to 37 mrn with a mean of 31 mm, whereas the 50 tumours were from 20 to
Table 1. Clinicopathologic features of long- and short-term survivors after pancreatoduodenectomy for pancreas head carcinoma
Age (years)
Macroscopic type Expansive Mixed Infiltrating
Size
Sex
(MIF) lcterus
(mm)
Long-term survivor 62. I k 10.2 5/7
11/12 30.7 i 9.7 8 3 (67%) (25%) (0.7) (92%) Short-term survivor 61.9 ? 10.1 31/19 35/50 34.9 t 8.7 38 8 (76%) (16%) I 1 = 50 ( I .6) (70%) I7
1
II
111
IV
3 0 9 0 (25%) ( 0 % ) ( 7 5 % ) ( 0 % ) 7 0 4 3 0 (14%) (0%) (86%) (0%)
1
(8%) 4
= 12
Histopathologic stage
(8%)
There i \ no ugnificant difference between the two group.;.
Table 2. List of 12 patients with pancreas head carcinoma who survived more than 3 years after a pancreatoduodenectomy ~
~
Agelhex
Case no ~
I 2 3 4 5 6 7 8 9 10 11
12
Chief complaint
Size (mm)
~~
56/M 661F 64lF
S8/M 67lM 76lM 75/F 37lF 6YF 60lF 65/F 561M
icterus DM icterus icterus icterus icterus
icterus icterus
icterus icterus icterus icterus
27 34 35 25 25 37 36 35 37 35 35 37
Histologic N*
-
+ -
-
+ + + + + + + +
EWi
type ~~
~
ADSQ
MOD MOD WELL WELL WELL WELL WELL WELL WELL WELL MOD
Clinical follow-up Duration (months) Outcome
_ _ 37 39 41
dead dead dead
44
dead dead
47 78 37 41 48 48 55 78
dead
alive alive alive: alive alive
alive
*N:lymph node rnctastavs. tEW: cancer cell\ at the resected margins. :;:Alive with lung metastasis. DM. diahctcs rnellitu\: WELL: uell differentiated adenocarcinoma; MOD: moderately differentiatcd adcnocarcinoma. ADSQ. adenosquarnous carcinoma.
517
PANCREATODUODENECTOMY FOR PANCREAS CARCINOMA CI. I N IC A I
65 mm with a mean of 35 mm. Five (7 1YO)of seven long-term survivors were diabetic compared with nine (47'/0) of 19 short-term survivors. Three (43%) of seven long-term survivors showed elevation of pre-operative serum CEA levels in comparison with 14 (58%) of 24 short-term survivors. Two (33%) of the six long-term survivors showed elevated pre-operative serum CA 19-9 concentrations, compared with 13 (87%) of 15 short-term survivors. None of the 12 long-term survivors had radiation therapy but eight of them had chemotherapy including 5-fluorouracil (5-FU), adriamycin and mitomycin C and/or immunotherapy such as Krestin (a protein-bound polysaccharide isolated from Coriolus versicolour, a mushroom belonging to the Basidiomycetes).
FOLLOW - U P
The overall cumulative I year, 2 year, 3 year, 4 year and 5 year survival rates of 85 patients with pancreas head carcinoma that underwent a pancreatoduodenectomy were 40.2, 20. I , 17. I , 8.6 and 8.6'%, respectively. The 50 short-term survivors died within 12 months after the operation as a result of local recurrence and/or distant metastases. The 12 long-term survivors were alive for more than 3 years after the surgical intervention, but six of the 12 long-term survivors died within 37-78 months as a result of local recurrence and distant metastases including the liver, lung, and/or lymph nodes. One long-term survivor was alive 48 months after the operation but lung metastasis was evident on chest X-ray film. Only two patients, one still alive and another dead at the time of writing, survived more than 5 years after the operation. Univariate log-rank analysis concerning the 13 prognostic factors showed that histologic type was a significant factor. Multivariate Cox regression analysis failed to reveal an independent prognostic variable.
PATHOLOGIC FEATURES
Macroscopically, of the 12 tumours in the longterm survivor group, the tuniour margin was expansive in eight, intermediate in three and infiltrative in one. Of the SO tumours in the short-term survivor group, the tumour margin was expansive in 38, intermediate in eight and infiltrative in four (Table 3). Eleven (92%) of the 12 tumours were well or moderately differentiated adenocarcinoma, whereas 35 (70%) of the 50 cases were adenocarcinoma of differentiated type. The tumours in each group invaded lymphatic channels in 83 and 88%, venous spaces in 50 and 68%, and perineural spaces in 100 and 96%, respectively. The tumours metastasized the lymph nodes in nine (75%) of the 12 long-term cases, compared with 43 (86%) of the 50 short-term cases. The surgically resected margin was affected by malignant cells in five (42%) of the 12 long-term survivors and 31 (62%) of the 50 short-term survivors. The 12 long-term survivors included three (25%) in stage I and nine (75%) in stage 111, while the 50 short-term survivors included seven (14%) in stage I and 43 (86%) in stage 111. The two groups were not significantly different clinicopathologicall y .
Discussion The clinical course of patients with pancreas carcinoma remains gloomy.' Gudjonsson reviewed 37 OOO cases of pancreas carcinoma in 1985 and reported that 4100 patients underwent resection and of these only 157 (3.8%) were alive after 5 years.'The cure rate for pancreatic carcinoma is admittedly dismal, but this does not warrant surgical nihilism. Morbidity and mortality after pancreatoduodenectomy has rapidly decreased and pancreatoduodenectomy has been carried out safely.'-' Surgical excision continues to be the only possible cure and results are emerging which suggest that combined adjuvant radiation and chemotherapy after radical resection prolong survival. Mannel e t a / . analysed long- and short-term survivors after pancreatic resection for pancreas carcinoma and reported a significant association of
Histopathologic features of long- and short-term survivors after pancreatoduodenectoniy for pancreas head carcinoma Table 3. ~
~~
~
~
~
~
~
~
Histopathologictype TUB1 TUB2 POR ADSQ UNDIF
~
~
INF
LY
V
PN
N
EW
alpha beta gamma
Long-ternsurvivor
10112 6/12 12/12 9/12 5/12 4 7 I 3 0 I 0 8 (67%)(25%)( 0%) (9%) (0%) (83%) (50%) (100%) (75%) (42"/0) (33'/0) (58%) ( 9 Y o ) Short-term survivor 18 17 7 2 6 44/50 34/50 48/50 43/50 31/50 19 26 5 ( n = 50) ( 3 6 % ) (34%)(18%) (4"/0) (7%) (88"/u) (68%) (96%) (86%) (62Yo) (38'%) (52%) (10%)
(n = 12)
There is no significant difference between the two groups. TUB I : well differentiated tubular adenocarcinomd; TUB2: moderately differentiated tubular adenocarcinoma. POR: poorly differentiated adenocarcinoma: ADSQ: adenosquamous carcinoma; UNDIF: undifferentiated carcinorna; LY: lymphatic permeation; V venous invasion; PN. perineurel infiltration: N: lymph node metastask. EW: cancer cells at the resected margin\.
s4x
Broder's Grade I11 and IV in the primary tumour, a round-cell infiltrate at the tumour margin and atypia of the pancreatic ductal epithelium with short-term survival. l o The association of steatorrhea with long-term survival was significant ( P < 0.05) and the association of back pain with short-term survival showed a trend toward significance. Trede er al. said that nine of 1 1 long-term survivors belonged to stage I (TINOMx) but it was worth noting that three of 13 stage 111 patients survived more than 3 years, in spite of lymph node metastases.' Furthermore, one of the latter was still alive 1 I years after total pancreatoduodenectomy including portal vein resection. Kairaluoma et a / . reported that no patient who underwent a palliative resection survived for 3 years." Lymph node metastasis and cancer invasion to the pancreatic capsule crucially affect the survival of the patients who undergo a pancreatoduodenectomy . 12- l 4 In this study. longterm survivors harboured preferable factors: a high incidence of icterus, a small mass, a stage I tumour, and adenocarcinoma of differentiated type, but the difference was not significant. It is also noteworthy that, in more than a half of the long-term survivors, the resection margins were affected by malignant cells and/or the lymph nodes were metastasized by malignant cells. This may be either because lymph node dissection was complete or because tumour tissue left behind was very limited. Early or small pancreas carcinoma has now been detected clinically. "-I7 Tsuchiya et a / . reported a collective review of 106 small pancreas cancers from 441 Japanese institutions and mentioned that even small pancreatic carcinoma showed frequent lymph node involvement, capsular invasion, retroperitoneal infiltration, and vascular permeation. " The postoperative cumulative 5 year survival rate was 30.3% and small pancreatic carcinoma did not always mean early pancreatic carcinoma. Manabe et a / . reported that small carcinoma of the pancreas was not always curable, but that a small, localized lesion without any extratumoral extension could be resected with a chance of cure. In particular, pancreas carcinoma of less than I cm in diameter showed limited extension. I x In this study, 12 of the 86 patients with pancreas head carcinoma survived 3 3 years after a pancreatoduodenectomy . This neither directly provides a strong support for a radical resection of pancreas head carcinoma nor means that 3 year survivors are patients who had been cured, because six of the 12 long-term survivors died from local recurrence and/ or distant metastases. Also, the present study does not represent the entire clinical course of patients with pancreas head carcinoma, because most patients with pancreas head carcinoma could not undergo laparotomy or resection. This study was composed only for fortunate patients who could
YAMAGLK'HI
undergo a pancreatoduodenectomy . Only a minority of patients with pancreas head carcinoma survived 3 3 years after a pancreatoduodenectomy. However the present findings may encourage surgeons to perform radical resections of pancreas carcinoma. It is reported that combined use of radiation therapy and fluorouracil as adjuvant therapy after a curative pancreas resection is effective and is preferred to no adjuvant therapy.'"."" Douglas Jr reported that adjuvant combined modality therapy may offer the potential of doubling median survival and possibly tripling the cure rate (possibly to 20% at 5 years) for patients with resectable diseases.' The Gastrointestinal Tumor Study Group published the results of a randomized series in which patients treated by 5-fluorouracil (5-FU) and radiotherapy had a 48% survival at 2 years versus the anticipated 18% survival for surgery alone.'"." This study was a collective series of cases from more than 20 Japanese institutions and no uniform regimens of adjuvant therapy were applied. Therefore, no conclusive relationship between the long-term survivor and the adjuvant therapy could be produced. Kairaluoma er al. reported that the longest survival time achieved after a palliative resection was about 2 years, the median survival being somewhat longer after a palliative resection than after a palliative bypass procedure, but the quality of life during the remaining period was much better after palliative resection." It is the relative balance of operative mortality versus the gain in short-term survival and quality of life that must be considered by each individual surgeon when operating on a patient with ductal adenocarcinoma of the pancreas. Contrary to current opinion, pancreatic resection is justified if the operative mortality is low enough, even though it may result in palliation for only 1 or 2 years. The curative resection continues to be the only approach likely to offer any chance of a cure. The only way to obtain long-term survivors is either to detect pancreas cancer in its early stages or to resect a pancreas mass followed by multidisciplinary therapy, even if some turnour tissues are left behind.
Acknowledgements The authors thank the following 24 institutions for the use of their cases: National Fukuoka Higashi Hospital, Fukuoka; National Fukuoka Central Hospital, Fukuoka; National Kokura Hospital, Kitakyushu; National Shimonoseki Hospital, Shimonoseki; National Nakatsu Hospital, Nakatsu; Fukuoka Red Cross Hospital, Fukuoka; Matsuyama Red Cross Hospital, Matsuyama; Yamaguchi Red Cross Hospital, Yamaguchi; Hamanomachi Hospital, Fukuoka; Kyushu Central Hospital, Fukuoka; Shin-Kokura Hospital, Kitakyushu; Kosei Nennkin Hospital,
PANCREATODUODENECTOMYFOR PANCREAS CARCINOMA
5-19
resections without an operative mortality. Ann. S~cr~q. 211,447-58. L. H.. ~\ A N HtEROtN prefectural central ~ ~ ~M ~ i ~~ ~ ~l 10. , MANNELL ~ A , . WEILAND ~ ~ J. A. & ~ ILSTRLP D. M. (1986) Factors influencing survival Gunshi lshikai Hospital, Usa; lzuhara Hospital, lki; after resection for ductal adenocarcinoma of the panFukuoka Municipal Hospital, Fukuoka; Kitakyushu creas. Ann. Surg. 203,403-7. 11. K A I R A U I ~ M.M I., A STAHLHERC; M.. K I V I N ~H. ~M &I City Kitakyushu; Kitakyushu City Wakamatsu Hospital, Kitakyushu; Eiko HosHAUKIPUKO K. ( 1989) Results of pancreatoduodenecpital. Fukuoka; Kimura Surgery Hospital, Fuktomy for carcinoma of the head of the pancreas. Hepafogustroenterology36, 4 12-8. uoka; Hofu Gastroenterology Hospital, Hofu; and 12. CRIST D. W. & CAMERON J . L. (1989) Current status Sada Surgery Hospital, Fukuoka. Th e authors also of pancreatoduodenectomy for periampullary carcithank Mr Brian T. Quinn (Kyushu University) for noma. Heparogustroentrrolog~36, 478-85. his critical reading of this manuscript. 13. MA-ISUNO S. & SATO T. (1986) Surgical treatment of carcinoma of the pancreas. Experience in 272 patients. Am. J . Surg. 152, 499-504. References s. (1978) 14. S A r O T.. SAI-IOH Y . , No-ro N. & MATSUNO 1 . DOUGLASS H. 0. JR (1987) Pancreatic cancer: Nihilism Factors influencing the late results of operation for is obsolete! Pmcreus 2, 230-2. carcinoma of the pancreas. Am. J . Surg. 136, 582-6. 2. GUOJ~NSSON B. (1987) Cancer of the pancreas. Cuncer 15. MWSAA. R. & LEVIN 8. (1981) The diagnosis of 60, 2284-303. 'early' pancreatic cancer: The University of Chicago A. L. & SWANSON R. S. (1988) Pancreatic 3. WAKSHAW experience. Cuncer 47. 1688-97. cancer in 1988. Possibilities and probabilities. Ann. 16. MANABE T., MIYASHIIA T., OHSHIO G. er crl. (1988) Surg. 208, 541-53. Small carcinoma of the pancreas. Clinical and patho4. YAMAGL~CHI K. & ENJWIM. ( I 989) Carcinoma of the logical evaluation of 17 patients. Cancer 62. 135-41. pancreas. A clinicopathologic study of 96 cases with 17. TSGCHIYA R . , TOMIOKA T., I L ~ W K. A ef a / . (1986) immunohistochemical observation. Jpn J . Clin. Collective review of small carcinomas of the pancreas. Oncol. 19, 14-22. Ann. Surg. 203, 77-8 I . 5 . AMEKICAN ON CANCER ( 1 988) Manual JOINT COMMITTEL 18. NAGAI H., K U R ~ OA. A & MOKIOKA Y. (1986) for staging of cancer, 3rd edn. J. B . Lippincott ComLymphatic and local spread of TI and T2 pancreatic pany, Philadelphia. cancer. A study of autopsy material. A m . Surg. 204, 6 . CONNONLY P. J . , MICHELASSI F., M. M . , DAWSON 65-7 I . MWSA A. R. & LowtNsrm F. (1987) Survival in 19. FUNOVICS J . M . . KARMEKJ . . PKATSCHNER T. H. & 1001 patients with carcinoma of the pancreas. A m . FRISCH A. ( 1989) Current trends in the management Surg. 206, 366-73. of carcinoma of the pancreatic head. Heparogusrr
Aust. N . Z . J . Sirrg. 1992.62, 545-549
LONG SURVIVORS AFTER PANCREATODUODENECTOMY FOR PANCREAS HEAD CARCINOMA KOJI YAMAGUCHI,* KAZUYOSHI NISHIHARA PIOTR , ~ KOLODZIEJCZYK,.~ A N D MASAZUMI TSUNEYOSH? Departments of *Surgery I and 'PatholoRy I I , KxLishu Universiry Fuculh
of Medicine, Fukuoka, JapLlrl
Twelve Japanese patients with pancreas head carcinoma w h o survived 3 years or more after a pancreatoduodenectomy and 50 who survived less than 12 months were reviewed clinicopathologically.The 12 patients who survived for 2 3 years exhibited more favourable prognostic factors: a higher incidence of jaundice; a smaller mass; a higher prevalence of an earlier stage tumour and adenocarcinoma of differentiated type; and a lower incidence of venous invasion, lymph node metastasis. and cancer cells at the surgical margins. However the difference was not significant. Univariate log-rank analysis regarding 13 prognostic variables showed that histologic type was a significant factor but multivariate Cox regression analysis failed to reveal an independent significant parameter. Nine of the 12 long-term survivors showed lymph node metastasis and six of the 12 revealed cancer cells at the surgical margins. Six of the I2 long-term survivors died from local recurnnce and/or distant metastasis 37-78 months after operation. Only two patients survived more than 5 years after the operation. At the time of writing. one of them was still alive and another was dead 78 months after the operation. Pancreatoduodenectomy for pancreas head carcinoma infrequently offers a permanent cure for the patients with pancreas head carcinoma but sometimes produces a worthwhile long-term survival, even if the resected margins were affected by malignant cells or the lymph node metastasis was evident.
.
Key words: long-term survivor, pancreas carcinoma, pancreatoduodenectomy
Introduction Despite the recent advances in the diagnostic and therapeutic modalities, the clinical course of patients with pancreas carcinoma remains gloomy. Some surgeons abandon a radical resection for pancreas carcinoma. Nevertheless, surgical resection does lead to a long-term survival or even to a cure for a few patients. Only surgical resection gives patients with pancreas carcinoma an opportunity to live long and to be cured permanently. It is important to identify this small subpopulation in whom the risks of pancreatoduodenectomy would be justified. It may be equally important to identify those patients with resectable tumours who might survive only a few months after pancreatectomy. In these patients, the radical resection could be avoided and a less dangerous, palliative procedure chosen. The clinicopathological and immunohistochemical features of 96 patients with pancreas carcinoma have been reported p r e v i o ~ s l yIn . ~ the present study, 12 patients with pancreas head carcinoma who surCorrespondence: Koji Yamaguchi. MD, Department of Surgery I. Kyu5hu University Faculty of Medicine. 3- I - I Maidaahi, Higashi-
ku. Fukuoka X12, Japan. Accepted for puhlication 13 February 1992
vived 3 years after a pancreatoduodenectomy are compared with SO who survived < 12 months in order to introduce a potential prognostic variable.
Methods A total of 86 Japanese patients with pancreas head carcinoma who had undergone a pancreatoduodenectomy at one of 25 Japanese institutions including Kyushu University Hospital were studied. Only ordinary adenocarcinoma of ductal origin in the head of the pancreas was included. Carcinoma of the body or tail of the pancreas. carcinoma of the ampulla of Vater. carcinoma of the bile duct, cystic neoplasm of the pancreas, papillary cystic neoplasm of the pancreas. islet cell tumour. sarcoma and all other malignant lesions were excluded in order to achieve a histogenetically pure group of carcinomas. The clinical charts were available for all 86 patients. Macroscopically, the tumours were divided into three groups according to tumour margin: expansive, infiltrative and intermediate type. Histopathologically, the tumours were classified as well, moderately, and poorly differentiated adenocarcinoma and others (adenosquamous carcinoma and undifferentiated carcinoma). The histopathologic tumour margin (INF) was divided into three
546
YAMAGUCHI
types: well defined and expansive (alpha), intermediate (beta) and infiltrative (gamma). The tuniours were classified into four stages according to the American Joint Committee on Cancer.5 The pancreatoduodenectomy specimens were examined by step-wise tissue sections at S mm intervals. All sections were stained with haematoxylin and eosin and were reviewed by one of the authors (KY). The clinical follow-up was current as of 30 May 1991 and the data were available for 85 of the 86 patients excluding one patient who was lost at clinical followup. Fifty of the 85 patients died within 1 year of surgery. Twelve patients survived > 3 years after surgery. Twenty patients were either doing well or were dead 1-3 years after surgery and three patients died within 1 month after surgery. These 23 patients were excluded from the current study in order to obtain two dichotomous groups of longterm and short-term survivors with no overlapping. The current study is comprised of the 12 long-term survivors and the SO short-term survivors. Six of the 12 long-term survivors are doing well at clinical follow-up 37-78 months after surgical intervention
and the remaining six patients died from local recurrence or distant metastases between 37 and 78 months after the operation. The mean values of age and size were examined by Student’s t-test and the distribution of prognostic factors was measured by Chisquared test. Univariate log-rank and multivariate Cox regression analyses were calculated for 13 potential factors to find a significant prognostic variable.
Results C L I N I C A L FINDINGS
The 12 long-term survivors comprised five men and seven women ranging from 37 to 76 years of age with a mean of 62 years (Tables 1.2). The SO shortterm survivors comprised 31 men and 19 women who ranged from 34 to 77 years old with a mean age of 62 years. Eleven (92%) of the twelve longterm survivors developed icterus, compared with 35 (70%) of the SO short-term survivors. The 12 tumours ranged from 25 to 37 mrn with a mean of 31 mm, whereas the 50 tumours were from 20 to
Table 1. Clinicopathologic features of long- and short-term survivors after pancreatoduodenectomy for pancreas head carcinoma
Age (years)
Macroscopic type Expansive Mixed Infiltrating
Size
Sex
(MIF) lcterus
(mm)
Long-term survivor 62. I k 10.2 5/7
11/12 30.7 i 9.7 8 3 (67%) (25%) (0.7) (92%) Short-term survivor 61.9 ? 10.1 31/19 35/50 34.9 t 8.7 38 8 (76%) (16%) I 1 = 50 ( I .6) (70%) I7
1
II
111
IV
3 0 9 0 (25%) ( 0 % ) ( 7 5 % ) ( 0 % ) 7 0 4 3 0 (14%) (0%) (86%) (0%)
1
(8%) 4
= 12
Histopathologic stage
(8%)
There i \ no ugnificant difference between the two group.;.
Table 2. List of 12 patients with pancreas head carcinoma who survived more than 3 years after a pancreatoduodenectomy ~
~
Agelhex
Case no ~
I 2 3 4 5 6 7 8 9 10 11
12
Chief complaint
Size (mm)
~~
56/M 661F 64lF
S8/M 67lM 76lM 75/F 37lF 6YF 60lF 65/F 561M
icterus DM icterus icterus icterus icterus
icterus icterus
icterus icterus icterus icterus
27 34 35 25 25 37 36 35 37 35 35 37
Histologic N*
-
+ -
-
+ + + + + + + +
EWi
type ~~
~
ADSQ
MOD MOD WELL WELL WELL WELL WELL WELL WELL WELL MOD
Clinical follow-up Duration (months) Outcome
_ _ 37 39 41
dead dead dead
44
dead dead
47 78 37 41 48 48 55 78
dead
alive alive alive: alive alive
alive
*N:lymph node rnctastavs. tEW: cancer cell\ at the resected margins. :;:Alive with lung metastasis. DM. diahctcs rnellitu\: WELL: uell differentiated adenocarcinoma; MOD: moderately differentiatcd adcnocarcinoma. ADSQ. adenosquarnous carcinoma.
517
PANCREATODUODENECTOMY FOR PANCREAS CARCINOMA CI. I N IC A I
65 mm with a mean of 35 mm. Five (7 1YO)of seven long-term survivors were diabetic compared with nine (47'/0) of 19 short-term survivors. Three (43%) of seven long-term survivors showed elevation of pre-operative serum CEA levels in comparison with 14 (58%) of 24 short-term survivors. Two (33%) of the six long-term survivors showed elevated pre-operative serum CA 19-9 concentrations, compared with 13 (87%) of 15 short-term survivors. None of the 12 long-term survivors had radiation therapy but eight of them had chemotherapy including 5-fluorouracil (5-FU), adriamycin and mitomycin C and/or immunotherapy such as Krestin (a protein-bound polysaccharide isolated from Coriolus versicolour, a mushroom belonging to the Basidiomycetes).
FOLLOW - U P
The overall cumulative I year, 2 year, 3 year, 4 year and 5 year survival rates of 85 patients with pancreas head carcinoma that underwent a pancreatoduodenectomy were 40.2, 20. I , 17. I , 8.6 and 8.6'%, respectively. The 50 short-term survivors died within 12 months after the operation as a result of local recurrence and/or distant metastases. The 12 long-term survivors were alive for more than 3 years after the surgical intervention, but six of the 12 long-term survivors died within 37-78 months as a result of local recurrence and distant metastases including the liver, lung, and/or lymph nodes. One long-term survivor was alive 48 months after the operation but lung metastasis was evident on chest X-ray film. Only two patients, one still alive and another dead at the time of writing, survived more than 5 years after the operation. Univariate log-rank analysis concerning the 13 prognostic factors showed that histologic type was a significant factor. Multivariate Cox regression analysis failed to reveal an independent prognostic variable.
PATHOLOGIC FEATURES
Macroscopically, of the 12 tumours in the longterm survivor group, the tuniour margin was expansive in eight, intermediate in three and infiltrative in one. Of the SO tumours in the short-term survivor group, the tumour margin was expansive in 38, intermediate in eight and infiltrative in four (Table 3). Eleven (92%) of the 12 tumours were well or moderately differentiated adenocarcinoma, whereas 35 (70%) of the 50 cases were adenocarcinoma of differentiated type. The tumours in each group invaded lymphatic channels in 83 and 88%, venous spaces in 50 and 68%, and perineural spaces in 100 and 96%, respectively. The tumours metastasized the lymph nodes in nine (75%) of the 12 long-term cases, compared with 43 (86%) of the 50 short-term cases. The surgically resected margin was affected by malignant cells in five (42%) of the 12 long-term survivors and 31 (62%) of the 50 short-term survivors. The 12 long-term survivors included three (25%) in stage I and nine (75%) in stage 111, while the 50 short-term survivors included seven (14%) in stage I and 43 (86%) in stage 111. The two groups were not significantly different clinicopathologicall y .
Discussion The clinical course of patients with pancreas carcinoma remains gloomy.' Gudjonsson reviewed 37 OOO cases of pancreas carcinoma in 1985 and reported that 4100 patients underwent resection and of these only 157 (3.8%) were alive after 5 years.'The cure rate for pancreatic carcinoma is admittedly dismal, but this does not warrant surgical nihilism. Morbidity and mortality after pancreatoduodenectomy has rapidly decreased and pancreatoduodenectomy has been carried out safely.'-' Surgical excision continues to be the only possible cure and results are emerging which suggest that combined adjuvant radiation and chemotherapy after radical resection prolong survival. Mannel e t a / . analysed long- and short-term survivors after pancreatic resection for pancreas carcinoma and reported a significant association of
Histopathologic features of long- and short-term survivors after pancreatoduodenectoniy for pancreas head carcinoma Table 3. ~
~~
~
~
~
~
~
~
Histopathologictype TUB1 TUB2 POR ADSQ UNDIF
~
~
INF
LY
V
PN
N
EW
alpha beta gamma
Long-ternsurvivor
10112 6/12 12/12 9/12 5/12 4 7 I 3 0 I 0 8 (67%)(25%)( 0%) (9%) (0%) (83%) (50%) (100%) (75%) (42"/0) (33'/0) (58%) ( 9 Y o ) Short-term survivor 18 17 7 2 6 44/50 34/50 48/50 43/50 31/50 19 26 5 ( n = 50) ( 3 6 % ) (34%)(18%) (4"/0) (7%) (88"/u) (68%) (96%) (86%) (62Yo) (38'%) (52%) (10%)
(n = 12)
There is no significant difference between the two groups. TUB I : well differentiated tubular adenocarcinomd; TUB2: moderately differentiated tubular adenocarcinoma. POR: poorly differentiated adenocarcinoma: ADSQ: adenosquamous carcinoma; UNDIF: undifferentiated carcinorna; LY: lymphatic permeation; V venous invasion; PN. perineurel infiltration: N: lymph node metastask. EW: cancer cells at the resected margin\.
s4x
Broder's Grade I11 and IV in the primary tumour, a round-cell infiltrate at the tumour margin and atypia of the pancreatic ductal epithelium with short-term survival. l o The association of steatorrhea with long-term survival was significant ( P < 0.05) and the association of back pain with short-term survival showed a trend toward significance. Trede er al. said that nine of 1 1 long-term survivors belonged to stage I (TINOMx) but it was worth noting that three of 13 stage 111 patients survived more than 3 years, in spite of lymph node metastases.' Furthermore, one of the latter was still alive 1 I years after total pancreatoduodenectomy including portal vein resection. Kairaluoma et a / . reported that no patient who underwent a palliative resection survived for 3 years." Lymph node metastasis and cancer invasion to the pancreatic capsule crucially affect the survival of the patients who undergo a pancreatoduodenectomy . 12- l 4 In this study. longterm survivors harboured preferable factors: a high incidence of icterus, a small mass, a stage I tumour, and adenocarcinoma of differentiated type, but the difference was not significant. It is also noteworthy that, in more than a half of the long-term survivors, the resection margins were affected by malignant cells and/or the lymph nodes were metastasized by malignant cells. This may be either because lymph node dissection was complete or because tumour tissue left behind was very limited. Early or small pancreas carcinoma has now been detected clinically. "-I7 Tsuchiya et a / . reported a collective review of 106 small pancreas cancers from 441 Japanese institutions and mentioned that even small pancreatic carcinoma showed frequent lymph node involvement, capsular invasion, retroperitoneal infiltration, and vascular permeation. " The postoperative cumulative 5 year survival rate was 30.3% and small pancreatic carcinoma did not always mean early pancreatic carcinoma. Manabe et a / . reported that small carcinoma of the pancreas was not always curable, but that a small, localized lesion without any extratumoral extension could be resected with a chance of cure. In particular, pancreas carcinoma of less than I cm in diameter showed limited extension. I x In this study, 12 of the 86 patients with pancreas head carcinoma survived 3 3 years after a pancreatoduodenectomy . This neither directly provides a strong support for a radical resection of pancreas head carcinoma nor means that 3 year survivors are patients who had been cured, because six of the 12 long-term survivors died from local recurrence and/ or distant metastases. Also, the present study does not represent the entire clinical course of patients with pancreas head carcinoma, because most patients with pancreas head carcinoma could not undergo laparotomy or resection. This study was composed only for fortunate patients who could
YAMAGLK'HI
undergo a pancreatoduodenectomy . Only a minority of patients with pancreas head carcinoma survived 3 3 years after a pancreatoduodenectomy. However the present findings may encourage surgeons to perform radical resections of pancreas carcinoma. It is reported that combined use of radiation therapy and fluorouracil as adjuvant therapy after a curative pancreas resection is effective and is preferred to no adjuvant therapy.'"."" Douglas Jr reported that adjuvant combined modality therapy may offer the potential of doubling median survival and possibly tripling the cure rate (possibly to 20% at 5 years) for patients with resectable diseases.' The Gastrointestinal Tumor Study Group published the results of a randomized series in which patients treated by 5-fluorouracil (5-FU) and radiotherapy had a 48% survival at 2 years versus the anticipated 18% survival for surgery alone.'"." This study was a collective series of cases from more than 20 Japanese institutions and no uniform regimens of adjuvant therapy were applied. Therefore, no conclusive relationship between the long-term survivor and the adjuvant therapy could be produced. Kairaluoma er al. reported that the longest survival time achieved after a palliative resection was about 2 years, the median survival being somewhat longer after a palliative resection than after a palliative bypass procedure, but the quality of life during the remaining period was much better after palliative resection." It is the relative balance of operative mortality versus the gain in short-term survival and quality of life that must be considered by each individual surgeon when operating on a patient with ductal adenocarcinoma of the pancreas. Contrary to current opinion, pancreatic resection is justified if the operative mortality is low enough, even though it may result in palliation for only 1 or 2 years. The curative resection continues to be the only approach likely to offer any chance of a cure. The only way to obtain long-term survivors is either to detect pancreas cancer in its early stages or to resect a pancreas mass followed by multidisciplinary therapy, even if some turnour tissues are left behind.
Acknowledgements The authors thank the following 24 institutions for the use of their cases: National Fukuoka Higashi Hospital, Fukuoka; National Fukuoka Central Hospital, Fukuoka; National Kokura Hospital, Kitakyushu; National Shimonoseki Hospital, Shimonoseki; National Nakatsu Hospital, Nakatsu; Fukuoka Red Cross Hospital, Fukuoka; Matsuyama Red Cross Hospital, Matsuyama; Yamaguchi Red Cross Hospital, Yamaguchi; Hamanomachi Hospital, Fukuoka; Kyushu Central Hospital, Fukuoka; Shin-Kokura Hospital, Kitakyushu; Kosei Nennkin Hospital,
PANCREATODUODENECTOMYFOR PANCREAS CARCINOMA
5-19
resections without an operative mortality. Ann. S~cr~q. 211,447-58. L. H.. ~\ A N HtEROtN prefectural central ~ ~ ~M ~ i ~~ ~ ~l 10. , MANNELL ~ A , . WEILAND ~ ~ J. A. & ~ ILSTRLP D. M. (1986) Factors influencing survival Gunshi lshikai Hospital, Usa; lzuhara Hospital, lki; after resection for ductal adenocarcinoma of the panFukuoka Municipal Hospital, Fukuoka; Kitakyushu creas. Ann. Surg. 203,403-7. 11. K A I R A U I ~ M.M I., A STAHLHERC; M.. K I V I N ~H. ~M &I City Kitakyushu; Kitakyushu City Wakamatsu Hospital, Kitakyushu; Eiko HosHAUKIPUKO K. ( 1989) Results of pancreatoduodenecpital. Fukuoka; Kimura Surgery Hospital, Fuktomy for carcinoma of the head of the pancreas. Hepafogustroenterology36, 4 12-8. uoka; Hofu Gastroenterology Hospital, Hofu; and 12. CRIST D. W. & CAMERON J . L. (1989) Current status Sada Surgery Hospital, Fukuoka. Th e authors also of pancreatoduodenectomy for periampullary carcithank Mr Brian T. Quinn (Kyushu University) for noma. Heparogustroentrrolog~36, 478-85. his critical reading of this manuscript. 13. MA-ISUNO S. & SATO T. (1986) Surgical treatment of carcinoma of the pancreas. Experience in 272 patients. Am. J . Surg. 152, 499-504. References s. (1978) 14. S A r O T.. SAI-IOH Y . , No-ro N. & MATSUNO 1 . DOUGLASS H. 0. JR (1987) Pancreatic cancer: Nihilism Factors influencing the late results of operation for is obsolete! Pmcreus 2, 230-2. carcinoma of the pancreas. Am. J . Surg. 136, 582-6. 2. GUOJ~NSSON B. (1987) Cancer of the pancreas. Cuncer 15. MWSAA. R. & LEVIN 8. (1981) The diagnosis of 60, 2284-303. 'early' pancreatic cancer: The University of Chicago A. L. & SWANSON R. S. (1988) Pancreatic 3. WAKSHAW experience. Cuncer 47. 1688-97. cancer in 1988. Possibilities and probabilities. Ann. 16. MANABE T., MIYASHIIA T., OHSHIO G. er crl. (1988) Surg. 208, 541-53. Small carcinoma of the pancreas. Clinical and patho4. YAMAGL~CHI K. & ENJWIM. ( I 989) Carcinoma of the logical evaluation of 17 patients. Cancer 62. 135-41. pancreas. A clinicopathologic study of 96 cases with 17. TSGCHIYA R . , TOMIOKA T., I L ~ W K. A ef a / . (1986) immunohistochemical observation. Jpn J . Clin. Collective review of small carcinomas of the pancreas. Oncol. 19, 14-22. Ann. Surg. 203, 77-8 I . 5 . AMEKICAN ON CANCER ( 1 988) Manual JOINT COMMITTEL 18. NAGAI H., K U R ~ OA. A & MOKIOKA Y. (1986) for staging of cancer, 3rd edn. J. B . Lippincott ComLymphatic and local spread of TI and T2 pancreatic pany, Philadelphia. cancer. A study of autopsy material. A m . Surg. 204, 6 . CONNONLY P. J . , MICHELASSI F., M. M . , DAWSON 65-7 I . MWSA A. R. & LowtNsrm F. (1987) Survival in 19. FUNOVICS J . M . . KARMEKJ . . PKATSCHNER T. H. & 1001 patients with carcinoma of the pancreas. A m . FRISCH A. ( 1989) Current trends in the management Surg. 206, 366-73. of carcinoma of the pancreatic head. Heparogusrr
