Survival following pancreaticoduodenectomy with resection of the superior mesenteric-portal vein confluence for adenocarcinoma of the pancreatic head
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
British Journal of Surgery 1998, 85, 611–617
Survival following pancreaticoduodenectomy with resection of the
superior mesenteric–portal vein confluence for adenocarcinoma of the
pancreatic head
S. D. LEACH, J. E. LEE, C. CHARNSANGAVEJ*, K. R. CLEARY†, A. M. LOWY,
C . J . F E N O G L I O , P . W . T . P I S T E R S and D . B . E V A N S
Pancreatic Tumor Study Group, Departments of Surgical Oncology, *Diagnostic Radiology and †Pathology, Box 106, University of Texas M. D.
Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77 030, USA
Correspondence to: Dr D. B. Evans
Background The survival of patients who underwent pancreaticoduodenectomy with or without en
bloc resection of the superior mesenteric–portal vein (SMPV) confluence for adenocarcinoma of
the pancreatic head was compared.
Methods To be considered for surgery, patients were required to fulfil the following computed
tomography criteria for resectability: (1) absence of extrapancreatic disease, (2) no evidence of
tumour extension to the superior mesenteric artery (SMA) or coeliac axis, and (3) a patent SMPV
confluence. Tumour adherence to the superior mesenteric vein (SMV) or SMPV confluence was
assessed at operation and en bloc venous resection was performed when necessary to achieve
complete tumour extirpation.
Results Seventy-five consecutive patients underwent pancreaticoduodenectomy, 44 without venous
resection and 31 with en bloc resection of the SMPV confluence. There were no perioperative
deaths in either group; late (more than 6 months) occlusion of the reconstructed SMPV confluence
contributed to the death of two patients. Median survival in the 31 patients who required venous
resection at the time of pancreaticoduodenectomy was 22 months, and that for the 44 control
patients was 20 months (P 0·25).
Conclusion Patients with adenocarcinoma of the pancreatic head who require venous resection
during pancreaticoduodenectomy for isolated tumour extension to the SMV or SMPV confluence
(in the absence of tumour extension to the SMA or coeliac axis) have a duration of survival no
different from that of patients who undergo standard pancreaticoduodenectomy. These data
suggest that venous involvement is a function of tumour location rather than an indicator of
aggressive tumour biology.
The management of tumour adherence to the lateral wall operative time and potential patient morbidity involved in
of the superior mesenteric vein (SMV) or superior a complicated venous resection and reconstruction are
mesenteric–portal vein (SMPV) confluence represents the wasted.
most challenging technical aspect of pancreaticoduo- Several studies have demonstrated three fundamental
denectomy. This finding, whether unexpected or suggested principles of vascular resection at the time of pan-
by preoperative imaging studies, is usually not visible until creaticoduodenectomy. (1) Tumour involvement of the
after pancreatic and gastric transection, a point in the proximal SMA or coeliac axis, in contrast to involvement
operation at which the surgeon has committed to of the SMPV confluence, usually includes extensive
resection1. If tumour is left on the lateral wall of the SMV involvement of the mesenteric neural plexus, making it
or SMPV confluence, local recurrence and short survival impossible to achieve a negative retroperitoneal margin of
are to be expected. Recent data have clearly excision even with radical resection8–10. (2) In contrast to
demonstrated that patients who undergo pan- arterial resection, segmental resection of the SMV or
creaticoduodenectomy with a positive margin of SMPV confluence can be performed safely with no
resection2–6 have a survival duration similar to that of increase in perioperative morbidity or mortality rate
patients who have locally advanced disease treated non- compared with standard pancreaticoduodenectomy11,12. (3)
surgically with 5-fluorouracil–based chemotherapy and Tumour involvement of the SMPV confluence in the
radiation7. In contrast, the intraoperative finding of absence of tumour extension to the proximal SMA or
tumour adherence to the SMV or SMPV confluence can coeliac axis is not associated with histopathological
be managed by partial or segmental venous resection and variables predictive of a poor prognosis; tumour invasion
reconstruction. However, this allows resection of all gross of the SMV or SMPV confluence appears to be a function
disease only when tumour adherence is limited to the of tumour location rather than an indicator of biological
SMV or SMPV confluence. If tumour extension into the aggressiveness11.
retroperitoneum with encasement of the superior Despite recent reports supporting the safety of venous
mesenteric artery (SMA) is found, the additional resection at the time of pancreaticoduodenectomy12,13, the
overall poor prognosis of patients with pancreatic cancer
and reports suggesting even shorter survival in patients
Paper accepted 18 August 1997 who undergo venous resection6,14 have convinced many
© 1998 Blackwell Science Ltd 611612 S . D . L E A C H , J . E . L E E , C . C H A R N S A N G A V E J e t a l .
surgeons that isolated involvement of the SMV or SMPV reconstruction, many patients received electron-beam
confluence is a contraindication to pancreaticoduo- intraoperative radiation therapy; 10 Gy was delivered to the bed
denectomy. In an attempt to prospectively study this of the resected pancreas in a dedicated radiation therapy–
controversy, the authors’ previous report outlined the surgical suite, obviating patient relocation19. All patients who
underwent venous resection were begun on aspirin within 24 h of
anatomical rationale for venous resection and its safety in surgery and were requested to take low-dose aspirin 80 mg daily
carefully selected patients11. The present study compares indefinitely. For determining length of hospital stay, the day of
the survival of patients who underwent pancreatico- surgery was counted as day 1.
duodenectomy with or without en bloc resection of the For pathological analysis, the operative specimen was oriented
SMV or SMPV confluence for adenocarcinoma of the and dissected by the surgeon and pathologist in a pathology suite
pancreatic head. Importantly, all patients underwent in the operating room complex. The retroperitoneal margin was
standardized preoperative imaging using state-of-the-art defined as the soft tissue margin directly adjacent to the proximal
contrast-enhanced computed tomography (CT). In 3–4 cm of the SMA. The margin was evaluated by frozen-section
contrast to all previously published reports outside of microscopic examination of a 2–3-mm full-face (en face) section
of the margin and interpreted as positive if tumour was seen on
Japan, prospective evaluation of the retroperitoneal this section. Tumour size was calculated following surgical
margin of excision was performed in all resected resection by measuring the greatest transverse diameter of the
specimens. tumour. In patients who had received preoperative chemo-
radiation this was often difficult, and in some specimens gross
tumour could not be demarcated from uninvolved pancreatic
parenchyma. Beginning in late 1993, a standardized system for
Patients and methods the pathological evaluation of the pancreaticoduodenectomy
Data on all patients who underwent pancreaticoduodenectomy specimen was performed to include histology of the segment of
for adenocarcinoma of the pancreatic head or uncinate process resected vein to determine the presence or absence of tumour
from June 1990 to August 1995 were recorded prospectively in cell infiltration of the vein wall; infiltration was defined as the
the pancreatic tumour database. Standardized preoperative presence of neoplastic cells within the tunica adventitia and/or
imaging and operative techniques were applied to all patients. tunica media of the vein wall20. On completion of pathological
Preoperative evaluation included physical examination, chest analysis, all cases were reviewed by a single histopathologist
radiography and contrast-enhanced CT. Angiography and (K.R.C.).
laparoscopy were used selectively. To be considered for Following the completion of all treatment, patients were
operation, all patients were required to fulfil the following CT evaluated by physical examination, chest radiography and thin-
criteria for resectability15: (1) absence of extrapancreatic disease, section, contrast-enhanced CT every 3–4 months. The venous
(2) no evidence of tumour encasement of the SMA or coeliac phase of the most recently obtained CT scan was used to assess
axis as defined by the presence of a normal fat plane between the SMPV patency in patients who underwent venous resection. The
tumour and these arterial structures, and (3) a patent SMPV reconstructed SMPV confluence was judged to be patent if the
confluence. The study group comprised all patients who required shape, contour and enhancement of the vessel were uniform
resection (tangential or segmental) of the SMV or SMPV from the level of the hepatic hilum to the level of the tributaries
confluence during pancreaticoduodenectomy. The control group of the SMV. The SMPV confluence was judged to be patent but
comprised all patients who underwent pancreaticoduodenectomy stenotic if the vessel contour was irregular and enhancement
that did not require any type of venous resection or recon- visible, but not uniform, throughout the region of venous
struction. Patients who underwent pancreaticoduodenectomy for reconstruction. The SMPV confluence was considered occluded
diagnoses other than ductal adenocarcinoma were excluded. Also if the vein could not be visualized or thrombus could be
excluded were all patients who underwent operations other than identified in the region of venous reconstruction. When the SMV
pancreaticoduodenectomy (i.e. total pancreatectomy or distal or SMPV confluence was occluded, enlarged venous tributaries
pancreatectomy). were often present in the small bowel mesentery and the
The majority of patients received adjuvant chemotherapy and mesocolon. All scans were interpreted by a single diagnostic
radiation therapy (chemoradiation). Radiation therapy was radiologist (C.C.).
delivered with 18-MeV photons either before operation to a total Statistical comparisons between groups were by 2 analysis of
dose of 50·4 Gy (standard fractionation) or 30·0 Gy (rapid two-way frequencies and Student’s two-tailed t test of differences
fractionation), or after operation to a total dose of 50·4 Gy. between means. Overall survival analysis was by Kaplan–Meier
Concurrent with either standard- or rapid-fractionation radiation product-limit estimation, and comparison of overall survival rates
therapy, 5-fluorouracil was given by continuous infusion at a dose between groups was performed using log rank analysis. P 0·05
of 300 mg per m2 per day, 5 days per week, through a central was considered statistically significant.
venous catheter16,17.
All operations were performed using a standardized
technique18. Tumour adherence to the SMV or SMPV
confluence was assessed during surgery, and venous resection Results
was performed if the surgeon was unable to fully mobilize the Seventy-five consecutive patients with adenocarcinoma
SMPV confluence from the pancreatic head or uncinate process
because of presumed tumour adherence. Reconstruction of the of pancreatic head origin underwent either pancreatico-
SMPV confluence was performed using one of four techniques. duodenectomy (44) or pancreaticoduodenectomy with en
Patients requiring tangential resection of less than one-third of bloc resection of the SMV or SMPV confluence (31).
the circumference of the SMPV confluence underwent repair There were 42 men and 33 women with a median age of
with an autologous saphenous vein patch. Patients requiring 65 (range 42–81) years. The two treatment groups were
segmental resection of the SMV or SMPV confluence had similar with respect to sex, age, number of patients who
reconstruction with either an autologous internal jugular vein underwent reoperative pancreaticoduodenectomy follow-
interposition graft, a GORE-TEX interposition graft (W. L. ing a non-therapeutic laparotomy before referral to this
Gore, Flagstaff, Arizona, USA), or with a primary anastomosis institution, and the number of patients who received
without interposition grafting. The majority of patients who
underwent primary anastomosis without interposition grafting chemoradiation either before or after pancreatico-
required splenic vein ligation and division. Inflow occlusion was duodenectomy (Table 1). A significantly larger proportion
performed routinely during the period of venous resection and of patients in the control group (36 of 44) than in the
reconstruction; systemic heparinization was used at the discretion venous resection group (14 of 31) received intraoperative
of the operating surgeon. Following tumour resection and venous radiation (P 0·01), possibly because of an initial
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617PANCREATICODUODENECTOMY FOR CARCINOMA OF THE PANCREATIC HEAD 613
reluctance to utilize intraoperative radiation following for histological evidence of tumour invasion of the wall of
reconstruction of the SMV or SMPV confluence. the SMV or SMPV20. Adenocarcinoma was found to
Operative and pathological characteristics are listed in invade the vein wall in 13 of 18 specimens. Neoplastic
Table 2. There were no perioperative treatment-related cells penetrated into or through the tunica media in 12 of
deaths, and all patients were successfully discharged from 13 specimens and were confined to the tunica adventitia
the hospital after a median hospital stay of 16 days for in only one specimen.
those who underwent venous resection and 15 days for In patients who underwent venous resection, left
those in the control group. The need for venous resection internal jugular vein interposition grafting was the
was associated with a significant increase in operative preferred method of reconstruction (Table 3). The
blood loss (P 0·001). Median tumour size was infrequent use of tangential excision with saphenous vein
significantly larger in patients who underwent venous (patch) repair suggests that most patients had apparent
resection (P 0·02). No patient was found to have a tumour involvement of greater than one-third of the
grossly positive retroperitoneal margin of excision, and circumference of the vein wall. As demonstrated in Table
there was no difference in the frequency of micro- 3, patency rates were quite high when autologous tissue
scopically positive retroperitoneal margins between was used for reconstruction. Of the 11 patients who
treatment groups. Following the adoption of a underwent internal jugular vein interposition grafting,
standardized pathological examination, the most recent nine were evaluable for graft patency. Three of the nine
18 operative specimens from patients who required patients are dead from disease (at 13, 13 and 24 months
segmental venous resection were evaluated pathologically from diagnosis), and six are alive with follow-up ranging
from 12 to 37 months; all grafts were patent on the most
recent contrast-enhanced CT scans.
Table 1 Patient demographics A total of seven patients were documented to have
occlusion of the reconstructed SMV or SMPV confluence
No. of patients (Table 3). Local tumour recurrence as a cause of venous
Venous resection Control
occlusion could not be documented by CT or autopsy in
(n 31) (n 44) any of these patients. Occlusion of the SMV or SMPV
confluence was asymptomatic in five patients but directly
Sex related to the cause of death in two patients (8 and
Male 19 (61) 23 (52) 16 months following pancreaticoduodenectomy). Both of
Female 12 (39) 21 (48) these patients developed intractable ascites (cytology
Median age (years) 66 64 negative) and nutritional depletion, and one required
Reoperative pancreaticoduodenectomy emergency splenectomy for upper gastrointestinal
Yes* 9 (29) 10 (23) haemorrhage secondary to sinistral portal hypertension.
No 22 (71) 34 (77)
Adjuvant therapy
Preoperative chemoradiation 22 (71) 24 (55)
Postoperative chemoradiation 6 (19) 14 (32)
No adjuvant therapy 3 (10) 6 (14) Table 3 Method of venous reconstruction and patency at last
Intraoperative radiation therapy follow-up
Yes 14 (45) 36 (82)
No 17 (55) 8 (18) Status of SMPV
confluence
Values in parentheses are percentages. *Patients who had
undergone abdominal laparotomy for planned pancreatic Total Patent with
resection before referral, excluding patients who had undergone Method of no. of probable
recent abdominal or biliary surgery for reasons other than reconstruction patients Patent stenosis Occluded
resection of a pancreatic head tumour and patients who had
undergone minilaparotomy for biopsy alone Saphenous vein 4 (13)* 2 0 1
patch
Primary end-to-end 8 (26) 3 2 3
anastomosis with
Table 2 Operative and pathological characteristics
ligation and
division of the
Venous
splenic vein
resection Control
Primary end-to-end 7 (23) 4 1 2
Variable (n 31) (n 44) P
anastomosis without
ligation of the
No. of perioperative 0 0
splenic vein
deaths*
Internal jugular vein 11 (35)*† 7 2 0
Median operative blood 1700 900 0·001
interposition graft
loss (ml)
GORE-TEX 1 (3) 0 0 1
Median tumour size 3·5 3·0 0·02
interposition graft
(cm)
No. with microscopically positive 4 (13) 7 (16) 0·72 Total 31 16 (57) 5 (18) 7 (25)
retroperitoneal margin
No. with positive 13 (42) 23 (52) 0·38 Values in parentheses are percentages. *Includes one patient
lymph nodes who could not be evaluated because follow-up radiographic
Median hospital 16 15 0·93 images (performed elsewhere) were inadequate for assessment of
stay (days) graft patency. †Includes one patient who could not be evaluated
who died from gastrointestinal haemorrhage secondary to a
Values in parentheses are percentages. *In-hospital or within presumed marginal ulcer before undergoing adequate follow-up
30 days of surgery radiographic imaging. SMPV, superior mesenteric–portal vein
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617614 S . D . L E A C H , J . E . L E E , C . C H A R N S A N G A V E J e t a l .
1·0
Proportion surviving
0·8
0·6
0·4
0·2
0 12 24 36 48 60 72 84
Length of follow-up (months)
Fig. 1 Actuarial survival for 31 patients who underwent
pancreaticoduodenectomy with venous resection () and 44
control patients who underwent standard pancreatico-
duodenectomy (). P 0·25 (log rank test)
Median follow-up was 17 months; minimum follow-up
for all patients was 12 months. Overall median survival for
the 75 patients was 21 months. Median survival in the 31
patients who required venous resection at the time of
pancreaticoduodenectomy was 22 months and that for the Fig. 2 Illustration of tumours in the pancreatic head involving
44 control patients was 20 months (P 0·25) (Fig. 1). the superior mesenteric vein (SMV) with and without invasion of
the superior mesenteric artery (SMA). Attempted pancreatico-
duodenectomy with resection of the SMV in the presence of
Discussion tumour extension to the SMA will result in short patient survival;
this is related to the surgeon’s inability to achieve a negative
The survival time of patients with adenocarcinoma of retroperitoneal margin of resection and probably unrelated to
the pancreatic head who required en bloc resection of whether or not a segment of SMV is removed
the SMV or SMPV confluence at the time of pancreatico-
duodenectomy was no different from that of patients who
underwent standard pancreaticoduodenectomy. This is
consistent with the hypothesis that tumour adherence to data2–6 also support a short survival time if tumour
or invasion of the SMPV confluence is a function of remains in the retroperitoneum along the proximal SMA
tumour location and possibly tumour size but not a following pancreaticoduodenectomy, regardless of
prognostic factor associated with early tumour recurrence whether a segment of SMV or portal vein was removed.
and short patient survival11. However, in contrast to In the absence of information about prospective
previous reports on venous resection during pancreatico- evaluation of the retroperitoneal margin (along the
duodenectomy, all patients in the present study underwent proximal SMA), reports of venous resection during
high-quality contrast-enhanced CT before tumour pancreaticoduodenectomy are impossible to interpret.
resection to ensure the presence of a normal fat plane The lack of accurate evaluation of the retroperitoneal
between the tumour and the proximal SMA and coeliac margin of resection and the high frequency of positive
axis origin. This vital tumour–vessel relationship must be margins in patients who underwent venous resection
assessed before operation. When assessed at the time of (Table 4)4,6,12–14,24–26 suggest that the majority of these
surgery, the relationship of the tumour to the proximal patients had locally advanced tumours with arterial
SMA can be examined directly only during the final step encasement and were poor candidates for pancreatico-
in tumour resection, following gastric and pancreatic duodenectomy. For example, Roder and colleagues14
transection. Furthermore, if the SMV or SMPV concluded that resection of the SMV or portal vein in
confluence is inseparable from the tumour, the proximal patients with pancreatic adenocarcinoma was associated
SMA cannot be accessed in the absence of interposition with a poor prognosis as demonstrated by a median
grafting (and medial retraction of the grafted SMPV survival of only 8 months. They concluded appropriately
confluence), or division of the splenic vein and lateral that their short patient survival was due to an incomplete
retraction of the specimen and attached SMPV resection (positive margin) in 15 of 22 patients. However,
confluence1,21. In either case, the surgeon has committed they provided no data to support the implication that the
to venous resection and completion of pancreatico- need for venous resection predicts margin positivity; that
duodenectomy even if tumour encasement of the SMA is would be true only for patients with tumour extension to
found, in which case gross tumour will be left behind the proximal SMA, a finding that can be accurately
along the proximal SMA. A fundamental principle of assessed before operation with contrast-enhanced helical
venous resection at the time of pancreaticoduodenectomy CT. Such patients should not undergo attempted
is to operate only on patients who have no evidence of pancreaticoduodenectomy; their locally advanced disease
arterial involvement (Fig. 2). The Japanese experience is currently best treated by chemoradiation, systemic
with regional pancreatectomy clearly demonstrates that gemcitabine or phase II investigational agents.
patients who have positive resection margins following In this study, objective radiographic (CT) criteria were
even extended resection experience early tumour used for operation based on the belief that there is a
recurrence and short survival22,23. American and European fundamental difference between tumour involvement of
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617PANCREATICODUODENECTOMY FOR CARCINOMA OF THE PANCREATIC HEAD 615
Table 4 Reports of venous resection during defined in this report, their results were exactly as one
pancreaticoduodenectomy or total pancreatectomy (excluding would predict based on the assumption that patients with
the Japanese experience) angiograms demonstrating type IV or V disease probably
had tumour extension to the SMA and probably
Percentage
with underwent incomplete resections. A working knowledge
positive of the three-dimensional anatomy of the SMA and SMV
Mortality Median retro- is critical for the pancreatic surgeon; tumours that extend
No. of rate survival peritoneal to the left (medial) of the SMV on the venous phase of
Reference Year patients (%) (months) margin the SMA angiogram will virtually always involve the SMA.
The proximity of the SMA to the posterior–medial aspect
Sindelar26 1989 20 20 12 n.a. of the SMV is a constant anatomical finding easily seen
Trede et al.4 1990 12 0 n.a. n.a. on high-quality CT (Fig. 3).
Launois et al.25 1993 9 0 6·1* n.a. In the present report, histological evidence of tumour
Allema et al.24 1994 20† 15 8 20‡
Yeo et al.6 1995 10 n.a. n.a.§ n.a. invasion into the resected vein was seen in 13 of 18
Fortner et al.13 1996 51 9 n.a. —¶ specimens examined. This is in agreement with the
Harrison et al.12 1996 50 6 13 24** findings of previous authors23,30 and suggests that the
Roder et al.14 1996 22 0 8 68** surgeon’s intraoperative assessment that venous resection
Current article 1997 31 0 22 13 is necessary to achieve complete tumour extirpation is
correct in the majority of cases. Importantly, a previous
*Mean; †Eight of 20 had adenocarcinoma of the distal bile duct report demonstrated that in 83 per cent of patients,
or ampulla of Vater; ‡17 of 20 had positive resection margins contrast-enhanced CT correctly predicted the need for
overall; §3-year survival rate was 13 per cent; ¶patients with venous resection, as indicated by the presence of direct
positive margins were excluded; **site of positive margin not
defined. n.a., Data not available adherence of the low-density tumour to the lateral wall of
the SMV or SMPV confluence11.
The technical details of segmental resection of the
SMPV confluence have been reviewed previously18,21.
the SMPV confluence and tumour involvement of the When tumour invasion of the SMPV confluence prevents
SMA11. The SMV is typically in direct contact with the mobilization and medial retraction of the SMPV
pancreatic head and uncinate process and lacks an confluence from the pancreatic head and uncinate
investing sheath of perivascular neural tissues. This allows process, access to the SMA origin and completion of the
direct extension to and invasion of the SMV or SMPV retroperitoneal dissection can be achieved in one of two
confluence by strategically located tumours even in the ways: ligation and division of the splenic vein or venous
absence of tumour extension to the SMA. In contrast, the resection and reconstruction. Division of the splenic vein
SMA lies posterior and medial to the SMV and is at its junction with the SMPV confluence allows access to
surrounded by the mesenteric neural plexus, a 2–4-mm the origin of the SMA medial to the SMV and provides
investing sheath that extends laterally to the posterior increased mobility of the portal vein, enabling a primary
aspect of the pancreatic head and uncinate process, and venous anastomosis to be constructed without tension31.
superiorly to the right and left coeliac ganglia27. However, splenic vein ligation occasionally results in
Involvement of the SMA by a tumour in the pancreatic gastrointestinal haemorrhage due to sinistral portal
head or uncinate process defines locally advanced disease; hypertension; one such case occurred in this study.
the propensity for pancreatic adenocarcinoma to spread Maintaining an intact splenic vein–portal vein junction
along perineural planes makes negative-margin resection has two important consequences in terms of surgical
unlikely in such cases even with arterial resection. strategy. First, the intact splenic vein–portal vein junction
Resection of the SMPV confluence should, therefore, be significantly limits the mobilization of the portal vein and
undertaken only in highly selected patients predicted to prevents primary anastomosis between the SMV and
have venous but not arterial involvement based on portal vein unless excision of the SMV is limited to less
preoperative contrast-enhanced CT. A high rate of than 2 cm. Second, maintenance of an intact splenic vein
negative-margin resections and the possibility of long-term prevents direct access to the proximal SMA, making
survival may be anticipated only in this group. completion of the retroperitoneal dissection impossible in
Ishikawa and colleagues28 recently published a system most patients. This difficulty is circumvented by
for angiographic typing of the degree of SMPV invasion performing venous resection and reconstruction with
by adenocarcinoma of the pancreas. They classified the autologous internal jugular vein before completion of the
contour of the SMPV confluence on the venous phase of retroperitoneal dissection and removal of the specimen
the SMA angiogram based on the extent of presumed (Fig. 4). Placement of the interposition graft effectively
tumour invasion: type I, no invasion (normal venous restores a mobile length of SMV, which can then be
phase); types II and III, abnormalities suggesting tumour retracted medially (to the patient’s left), allowing
invasion limited to the right lateral wall of the SMV or retroperitoneal dissection along the origin of the SMA to
SMPV confluence; and types IV and V, bilateral invasion be completed in the usual manner. Inflow occlusion
of the SMV or SMPV confluence. Patients with unilateral during the time of venous resection and reconstruction is
(semicircular) venous invasion (types I, II and III) had a critical to prevent small bowel oedema.
3-year survival rate of 59 per cent, whereas patients with In summary, these data suggest that venous resection
bilateral venous invasion (types IV and V) all died within during pancreaticoduodenectomy for adenocarcinoma of
18 months of pancreatectomy and had a median survival the pancreatic head is associated with a survival duration
similar to that of patients who did not undergo pancreatic no different from that of patients who undergo standard
resection. Similar results were subsequently reported by pancreaticoduodenectomy. However, this conclusion is
Nakao and colleagues29. While these authors did not based on a programme of standardized preoperative
examine the retroperitoneal margin of resection as imaging (staging) and operative technique. Future studies
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617616 S . D . L E A C H , J . E . L E E , C . C H A R N S A N G A V E J e t a l .
Fig. 3 a Venous phase of the superior mesenteric artery
angiogram demonstrating bilateral narrowing of the superior
mesenteric vein (SMV) (arrows). b Contrast-enhanced helical
computed tomography in the same patient demonstrating
tumour encasement (small arrows) of both the SMV (large
arrow) and the superior mesenteric artery (long arrow). Bilateral Fig. 4 Technique of venous resection and reconstruction. a
narrowing of the SMV or superior mesenteric–portal vein Segmental resection of the superior mesenteric vein (SMV) is
confluence on angiography should always raise the possibility of performed while the specimen is still attached to the superior
tumour extension to the proximal superior mesenteric artery or mesenteric artery (SMA) before completion of the
coeliac axis. Note the dilatation of the middle colic vein due to retroperitoneal dissection. b An internal jugular vein
significant narrowing of the SMV interposition graft enables medial retraction of the reconstructed
superior mesenteric–portal vein confluence, allowing access to
the retroperitoneum for standard dissection of the tumour from
the lateral wall of the SMA
examining this question must include similar objective
criteria for operation combined with a standardized
approach to pathological evaluation of the resected
specimen. References
1 Cusack JC Jr, Fuhrman GM, Lee JE, Evans DB. Managing
unsuspected tumor invasion of the superior mesenteric–
Acknowledgements portal venous confluence during pancreaticoduodenectomy.
Am J Surg 1994; 168: 352–4.
The authors thank Melissa G. Burkett, Kathleen Wagner and 2 Klempnauer J, Ridder GJ, Bektas H, Pichlmayr R. Surgery
Pam Stanford for assistance in the preparation of this for exocrine pancreatic cancer — who are the 5- and 10-year
manuscript. survivors? Oncology 1995; 52: 353–9.
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617PANCREATICODUODENECTOMY FOR CARCINOMA OF THE PANCREATIC HEAD 617
3 Nitecki SS, Sarr MG, Colby TV, van Heerden JA. Long-term 17 Staley CA, Lee JE, Cleary KR et al. Preoperative
survival after resection for ductal adenocarcinoma of the chemoradiation, pancreaticoduodenectomy, and intra-
pancreas. Is it really improving? Ann Surg 1995; 221: 59–66. operative radiation therapy for adenocarcinoma of the
4 Trede M, Schwall G, Saeger H. Survival after pancreatic head . Am J Surg 1996; 171: 118–25.
pancreaticoduodenectomy. 118 consecutive resections without 18 Evans DB, Lee JE, Pisters PWT. Pancreaticoduodenectomy
an operative mortality. Ann Surg 1990; 211: 447–58. (Whipple operation) and total pancreatectomy for cancer. In:
5 Willett CG, Lewandrowski K, Warshaw AL, Efird J, Nyhus LM, Baker RJ, Fischer JF, eds. Mastery of Surgery. 3rd
Compton CC. Resection margins in carcinoma of the head of ed. Boston, Massachusetts: Little, Brown, 1997: 1233–49.
the pancreas. Implications for radiation therapy. Ann Surg 19 Evans DB, Termuhlen PM, Byrd DR, Ames FC, Ochran TG,
1993; 217: 144–8. Rich TA. Intraoperative radiation therapy following pan-
6 Yeo CJ, Cameron JL, Lillemoe KD et al. Pancreatico- creaticoduodenectomy. Ann Surg 1993; 218: 54–60.
duodenectomy for cancer of the head of the pancreas: 201 20 Staley CA, Cleary KR, Abbruzzese JA et al. The need for
patients . Ann Surg 1995; 221: 721–33. standardized pathologic staging of pancreaticoduodenectomy
7 The Gastrointestinal Tumor Study Group. A multi- specimens . Pancreas 1996; 12: 373–80.
institutional comparative trial of radiation therapy alone and 21 Evans DB, Lee JE, Leach SD, Fuhrman GM, Cusack JC Jr,
in combination with 5-fluorouracil for locally unresectable Rich TA. Vascular resection and intraoperative radiation
pancreatic carcinoma . Ann Surg 1979; 189: 205–8. therapy during pancreaticoduodenectomy: rationale and
8 Kayahara M, Nagakawa T, Konishi I, Ueno K, Ohta T, technique . Adv Surg 1996; 29: 235–62.
Miyazaki I. Clinicopathological study of pancreatic carcinoma 22 Nagakawa T, Konishi I, Ueno K et al. The results and
with particular reference to the invasion of the problems of extensive radical surgery for carcinoma of the
extrapancreatic neural plexus . Int J Pancreatol 1991; 10: head of the pancreas . Jpn J Surg 1991; 21: 262–7.
105–11. 23 Takahashi S, Ogata Y, Tsuzuki T. Combined resection of the
9 Nagakawa T, Kayahara M, Ohta T, Ueno K, Konishi I, pancreas and portal vein for pancreatic cancer. Br J Surg
Miyazaki I. Patterns of neural and plexus invasion of human 1994; 81: 1190–3.
pancreatic cancer and experimental cancer. Int J Pancreatol
24 Allema JH, Reinders ME, van Gulik TM et al. Portal vein
1991; 10: 113–19.
resection in patients undergoing pancreaticoduodenectomy
10 Nagakawa T, Mori K, Nakano T et al. Perineural invasion of
for carcinoma of the pancreatic head . Br J Surg 1994; 81:
carcinoma of the pancreas and biliary tract . Br J Surg 1993;
80: 619–21. 1642–6.
11 Fuhrman GM, Leach SD, Staley CA et al. Rationale for en 25 Launois B, Franci J, Bardaxoglou E et al. Total pan-
bloc vein resection in the treatment of pancreatic createctomy for ductal adenocarcinoma of the pancreas with
adenocarcinoma adherent to the superior mesenteric–portal special reference to resection of the portal vein and
vein confluence. Pancreatic Tumor Study Group . Ann Surg multicentric cancer. World J Surg 1993; 17: 122–7.
1996; 223: 154–62. 26 Sindelar WF. Clinical experience with regional pan-
12 Harrison LE, Klimstra DS, Brennan MF. Isolated portal vein createctomy for adenocarcinoma of the pancreas . Arch Surg
involvement in pancreatic adenocarcinoma. A contra- 1989; 124: 127–32.
indication for resection? Ann Surg 1996; 224: 342–9. 27 Nakao A, Harada A, Nonami T, Kaneko T, Takagi H.
13 Fortner JG, Klimstra DS, Senie RT, Maclean BJ. Tumor size Clinical significance of carcinoma invasion of the
is the primary prognosticator for pancreatic cancer after extrapancreatic nerve plexus in pancreatic cancer. Pancreas
regional pancreatectomy. Ann Surg 1996; 223: 147–53. 1996; 12: 357–61.
14 Roder JD, Stein HJ, Siewert JR. Carcinoma of the 28 Ishikawa O, Ohigashi H, Imaoka S et al. Preoperative
periampullary region: who benefits from portal vein indications for extended pancreatectomy for locally advanced
resection? Am J Surg 1996; 171: 170–5. pancreas cancer involving the portal vein . Ann Surg 1992;
15 Evans DB, Abbruzzese JL, Rich TA. Cancer of the pancreas. 215: 231–6.
In: De Vita VT, Hellman S, Rosenberg SA, eds. Cancer, 29 Nakao A, Harada A, Nonami T, Kaneko T, Inoue S, Takagi
Principles and Practice of Oncology. 5th ed. Philadelphia, H. Clinical significance of portal invasion by pancreatic head
Pennsylvania: JB Lippincott, 1997: 1054–87. carcinoma . Surgery 1995; 117: 50–5.
16 Spitz FR, Abbruzzese JL, Lee JE, Pisters PWT et al. 30 Tashiro S, Uchino R, Hiraoka T et al. Surgical indication and
Preoperative and postoperative chemoradiation strategies in significance of portal vein resection in biliary and pancreatic
patients treated with pancreaticoduodenectomy for cancer. Surgery 1991; 109: 481–7.
adenocarcinoma of the pancreas . J Clin Oncol 1997; 15: 31 Fortner J. Technique of regional subtotal and total pan-
928–37. createctomy. Am J Surg 1985; 150: 593–600.
© 1998 Blackwell Science Ltd, British Journal of Surgery 1998, 85, 611–617You can also read
