Surgical Palliation of Advanced Pancreatic Cancer

In about 80% of patients with pancreatic cancer surgical resection is not feasible at the time of diagnosis. Therefore, palliative treatment plays a key role in the treatment of pancreatic cancer. The defined goals of palliative treatment are: reduc-tion of symptoms, reduction of in-hospital stays, and an adequate control of pain. In patients with nonresectable pancreatic carcinoma the leading goal of palliative strategies should be the con¬trol of biliary and duodenal obstructions such as jaundice-associated pruritus or sustained nausea and vomiting due to gastric outlet obstruction. Although the role of endoscopy for palliation has been increasing, operative palliation is still indicated in selected cases. Obstructive jaundice is found in approximately 70% of patients suf-fering from carcinoma of the pancreatic head at diagnosis and has to be eliminated to avoid progressive liver dysfunction and liver failure. In up to 50% of patients with pancreatic cancer, clinical symptoms such as nausea and vomiting occur. For the treatment of malignant biliary ob-structions in patients with pancreatic carcinoma, endoscopic biliary drainage is the option of first choice. In case of persistent stent-problems such as occlusion or recurrent cholangitis, a hepatico-jejunostomy should be considered. The role of a prophylactic gastroenterostomy is still under dis¬cussion. In patients with combined biliary and gastric obstruction a combined bypass should be performed to avoid a second operation. The sig¬nificance of laparoscopic biliary bypass is not yet clear. A surgical, minimally invasive approach for treating bile duct obstruction is not the stan¬dard nowadays. The role of surgical pain relief is mostly negligible today. Computed tomography (CT)- or EUS-guided celiac plexus neurolysis has replaced surgical intervention today. The significance of palliative resections is currently a controversial topic. However, beyond controlled randomized studies, a palliative pancreaticoduo-denectomy in patients with advanced pancreatic carcinoma cannot be recommended at this time.

13.1    Introduction
Considerable advances in the treatment of pa-tients with pancreatic cancer have been reached during recent decades and surgical results after pancreatic head resection have clearly improved in the majority of patients, yet the disease is di-agnosed too late for a curative surgical approach [1]. This fact and the aggressiveness of the pan-creatic adenocarcinoma are the reasons for the poor overall 5-year survival rate, which has only moderately increased from less than 5% to ap-prox. 7% nowadays. In about 80% of the patients coming to diagnosis with pancreatic carcinoma, palliative therapy is the only treatment option.
Palliative strategies in patients with pancre¬atic carcinoma focus on three symptoms: pain, duodenal obstruction, and obstructive jaundice, whereby the palliative treatment of these symp-toms is primarily directed at reducing the clini¬cal symptoms, reducing the hospital stay, and last but not least ensuring as much overall quality of life as possible. Currently both nonoperative endoscopic procedures and surgical techniques are available to provide palliation of the leading symptoms, and the principal goal of a palliative treatment plan should be tailored to most effec¬

tively suit the patients' clinical presentation and their overall physical and mental condition; it should also incorporate the estimated prognosis.
In the past, surgery has been the only treat-ment option for effective palliation in patients with pancreatic carcinoma. But with the rapid advances in the development of endoscopic tech-niques, the significance of surgical palliation has declined. The main consideration for avoiding surgical palliation is the morbidity and mortal¬ity associated with surgical procedures such as gastroenterostomy, hepaticojejunostomy, and even laparotomy, although the latter occurs only in advanced cancer patients. It remains, however, unclear as to how the value of surgical palliation vs endoscopic palliation may be appropriately judged, and neither life quality investigations nor purely outcome-centered evaluations have so far succeeded in establishing a useful therapeutic al-gorithm for this severely ill patient population. Hence, decision-making on a palliative strategy for an individual patient remains difficult and may have to be based on an interdisciplinary dis-cussion between the patient, oncologist, thera-peutic radiologist, and surgeon.

13.2    Palliative Therapy of Biliary
Obstruction: Stent or Surgery?
The surgical options for palliative therapy in biliary obstruction include operative placement of biliary drains such as a T-drainage, choledo-choduodenostomy, hepaticojejunostomy, or, in rare cases, a peripheral or distal hepaticojejunos-tomy (Fig. 13.1). It is important to note that the often-discussed argument that the placement of a T-drainage is a very small surgical procedure with a low complication rate is, on one hand, correct; however, it has to be considered that this procedure creates an external biliary fistula with all its possible complications and implications for life quality. It should therefore be considered only when other measures of palliative treatment of obstructive jaundice have failed or cannot be undertaken for technical reasons. In addition, the amount of bile loss through such an external biliary fistula may lead to a profound electrolyte and fluid imbalance; hence, in such cases we ad-vocate a simple anastomosis between jejunum and the gallbladder to provide relief of biliary obstruction. The cholecystogastrostomy de¬scribed in earlier reports may lead to bile gastri¬tis, increased gastrin release, and secondary acid hypersecretion, as well as food entry into the bili-ary system and subsequent recurrent obstruction or cholangitis (or both) [2]. We therefore have abandoned this procedure along with most other centers experienced in pancreatic surgery.
Hence, the cholecystojejunostomy remains the standard surgical procedure for palliation when the surgical dissection of the hepatoduo-denal ligament has to be avoided. If cholecys-tojejunostomy is chosen, the cystic duct ought to attach a common bile duct and the distance to the tumor mass needs to be at least 2-3 cm to prevent early reobstruction by continuing tumor growth. Several trials have compared cholecystojejunostomy and hepatic enterostomy to evaluate whether the risk of bile duct injuries due to resection of the hepatoduodenal ligament may be avoided. Watanapa et al. [36] found that cholecystoenterostomy yielded a success rate of 89%, which was not significantly different from a success rate of 97% in patients receiving a hepaticoenterostomy. In addition, the authors found that cholangitis and recurrent jaundice were observed in 20% of the cholecystojejunos-tomy cases, whereas the complication rate in the group with hepaticoenterostomy was higher. The authors concluded that there may be a slightly increased risk of surgical complications when dissecting the hepatoduodenal ligament for he-paticojejunostomy. Furthermore, other authors have indicated that the possible troublesome dissection of the hepatoduodenal ligament may often be avoided when the common bile duct is transected in the middle or lower section and a side-to-side choledochoenterostomy is performed rather than the standard end-to-side hepaticojejunostomy.
Some authors have evaluated the choledocho-duodenostomy, which has been proved to be an effective surgical method for treating obstructive jaundice in benign conditions and also has been used in selected cases for biliary reconstruc¬tion after orthotopic liver transplantation [3]. However, in the case of pancreatic carcinoma, which has led to jaundice, many surgeons today feel that, in the advanced stages of pancreatic carcinoma (when patients are receiving pallia¬tive surgical treatment), an anastomosis in close proximity to the tumor may lead to early reste-nosis and occurrence of jaundice. Additionally, the peritumoral inflammation usually leads to stiff duodenum, which will not allow attention-free anastomosis, thereby increasing the risk for anastomotic leakage. However, other authors have utilized the choledochoduodenostomy rou¬tinely and have shown that the procedure is as¬sociated with a lower complication rate, a short length of postoperative hospital stay, and a very low recurrence rate of obstructive jaundice (be¬low 2%). Therefore, since the overall complica¬tions rate for the other methods of biliary bypass were higher, they advocate the choledochojeju-nostomy as the standard method for surgical palliation in obstructive jaundice caused by ad¬vanced pancreatic carcinoma [4].
The introduction and development of endo-scopic methods of biliary reconstruction reach¬ing from papillotomy to placements of intra-ductal stents in patients has revolutionized the palliative treatment of patients with obstructive jaundice due to pancreatic cancer. Today endo¬scopic placement of biliary stents is accepted as a standard treatment in patients with unresectable pancreatic carcinoma. However, the controversy regarding the abdominal palliative treatment— stent or surgery—is still ongoing and undecided. Several prospective randomized trials have com¬pared nonoperative biliary stenting with opera¬tive procedures such as hepaticojejunostomy or others. The study by Shepherd et al. did not show a significant difference in complication rate, 30-day mortality rate, incidence of postoperative gastric outlet obstruction, or median survival [5]. However, the rate of recurrent jaundice was significantly higher after biliary stenting com-pared to the surgical bypass procedure (43% vs. 0%). Furthermore, in a randomized trial Smith et al. demonstrated that recurrent jaundice oc¬curred more often in patients after stent place¬ment than in patients after surgical biliary bypass [6] (Table 13.1).
The main argument for surgical bypass is that the surgical procedure is thought to be a defini-tive treatment avoiding the regular endoscopic procedures for changing of stents or treating stent complications, which are frequent in this patient population. In addition, many surgeons feel that the definitive palliative surgical procedure is more cost-effective for the same reason. However, a re¬cent study by Artifon and coworkers shows that endoscopic biliary drainage carries lower costs and provides better quality of life when compared to palliative surgical procedures [7]. Again, as in most other studies comparing surgical endo-scopic bypass, no difference in the median sur¬vival of the investigated groups was found.
Taken together it is still unclear whether there is a standard treatment algorithm to be advo-cated, since all studies carry the problem of bias in patient selection and the lack of acceptable and validated quality-of-life data. Nevertheless, endoscopic biliary drainage has become the gold standard for palliation of malignant bile duct ob-structions in patients with pancreatic carcinoma, and the numbers of surgical palliative procedures have clearly declined. However, surgical options still carry significance. For example, in the case of refractory stent problems such as stent occlusion or recurrent cholangitis, operative stent with¬draw and hepaticojejunostomy may be indicated. Furthermore, primary hepaticojejunostomy should be performed in cases of endoscopically impassable tumor masses; finally, if an advanced pancreatic tumor is judged to be nonresectable at laparotomy, a prophylactic hepaticojejunos-tomy should be considered in patients with ob¬structive jaundice or in the case of threatening obstructive jaundice. To decide on the optimal treatment strategy for a patient with biliary ob¬struction due to pancreatic carcinoma, a close collaboration between the surgeon, the endos-copist, and the oncologic specialist is necessary; complicated cases should be managed based on interdisciplinary approaches.

13.3    Palliative Surgery for Gastric Outlet Obstruction Alone or in Combination with Biliary Bypass?
The standard palliative surgical procedure for gastric outlet obstruction due to upper abdomi-nal malignancies is a retro- or antecolic end-to-side or side-to-side gastrojejunostomy. While  this is normally a simple surgical procedure due to the often marginal clinical condition of the patients in advanced tumor stages, the gastroje-junostomy shows high morbidity and mortality rates [8]. The question of whether a prophylac¬tic gastroenterostomy is rational and should be performed when a normal resectable situation in patients with pancreatic head carcinoma is found at laparotomy is undecided as of yet and under discussion. A recent study by Egrari et al. shows that the mean time to obstruction was 15.7 months compared to a mean overall survival of approx. 13 months in patients with advanced pancreatic carcinoma. The authors demonstrated that due to the rapid natural progression of pan¬creatic adenocarcinoma, most patients do not survive long enough to obstruct and therefore do not need a prophylactic gastroenterostomy
[9].
Today many investigators feel that, due to possible morbidity and mortality, prophylac¬tic gastroenterostomy is unnecessary, and only a selective use of gastroenterostomy should be exercised in the case of present or impending duodenal obstruction that has already led to clinical symptoms. A second area of discussion is the question whether a combination of bili¬ary and gastric bypass is reasonable and profit¬able for the patient with pancreatic carcinoma in the palliative situation. A French study analyzing 2,493 patients with unresected cancer of the pan¬creas demonstrated that the mortality in patients with a combination of biliary and gastric bypass was similar [10]. However, they also observed that 16% of the patients undergoing biliary by¬pass alone developed a gastric obstruction. This finding was confirmed by other groups [11, 12]. Therefore, a number of authors concluded that a combination of biliary and gastric bypass as the initial procedure should be performed, since it minimizes the risk of reoperation and provides definitive palliation [10, 13]. To create a gastro-jejunostomy in addition to a surgical bypass of biliary obstruction is not a technical challenge for experienced general surgeons and today is as¬sociated with low morbidity and mortality rates. However, the decision for an initial combination of biliary and gastric bypass depends on several factors such as preexisting gastric outlet obstruc¬tion at the time of operation, imminent gastric obstruction, the overall condition of the patient, tumor stage, and tumor biology (Fig. 13.2).
As stated above, the best therapeutic strategy and the surgical method chosen for an individual patient should be discussed with consideration for all clinical factors defining the individual patient; when surgical options are considered, it seems important to underline that for patients with unresectable pancreatic cancer who present clinically manifest gastric obstruction at admis-sion, the median survival often may be as little as 4 weeks, even when newer oncologic treatment concepts are initiated [14]. This may be an im-portant argument for an initial combined biliary and gastric bypass to ensure that such patients have the chance to leave the hospital with imme-diately effected palliation.

13.4   Minimally Invasive Procedures for Surgical Palliation
Throughout the last few decades, minimally in-vasive procedures for palliative surgery have been reported in increasing numbers. This holds true also for palliative biliary and gastric bypass procedures, and today a considerable number laparoscopic gastric and biliary bypasses for periampullary carcinomas have been reported in the literature [15, 16]. In these studies, lapa-roscopic techniques are either performed as cho-lecystojejunostomies or hepaticojejunostomies to create a biliary bypass. The possible advan¬tages of minimally invasive surgical approaches seem obvious since especially in those severely ill patients the trauma of the surgical procedure and the time of hospital stay are very important factors. However, some possible disadvantages of the minimally invasive procedures have to be considered. The mean operating time seems to be significantly longer compared to standard open surgical procedures, thereby increasing the surgical trauma. Furthermore, special laparo-scopic expertise is required to ensure a low com¬plication rate, since both a hepaticojejunostomy and gastroenterostomy are considered advanced laparoscopic procedures. Today, endoscopic stenting via endoscopic retrograde cholangiogra-phy is the gold standard for palliation in patients with malignant bile duct obstruction due to car¬cinoma of the pancreatic head, and although several authors reported results for laparoscopic biliary bypass in single patients, no prospective randomized study comparing laparoscopic sur¬gery vs stenting has yet been reported. The first choice of treatment in patients with bile duct obstruction due to pancreatic cancer should therefore be the endoscopic stenting. If indicated (based on repeated stent occlusions or recurrent cholangitis), surgical intervention regardless of the surgical technique should be discussed. Dur¬ing this discussion it has to be considered that the laparoscopic biliary bypass is not a standard minimally invasive procedure and only experts in the field of laparoscopy should perform such operations in this very ill patient population.
In a small group of patients, Kazanjian et al. demonstrated that laparoscopic gastrojejunos-tomy is a safe and effective palliation for patients with gastric outlet obstruction due to pancreatic carcinoma. In their analysis it was especially sig¬nificant in a group of patients with a very limited survival [17]. In addition, a group from Norway compared open vs laparoscopic gastrojejunos-tomy for palliation in advanced pancreatic can¬cer retrospectively and found that laparoscopic gastrojejunostomy in advanced cases offered a reduced estimated blood loss and a shortened hospital stay when compared to open gastrojeju-nostomy [18]. Hence, at this time minimally in¬vasive procedures using standardized techniques should be considered for relief of gastric outlet obstruction due to pancreatic carcinoma when the laparoscopic expertise is present. In this situ¬ation a low complication rate can be ensured and the minimally invasive techniques might be a vi¬able alternative for open surgical procedures, es-pecially in patients with a very limited prognosis
[19].

13.5    Role of Surgical Pain Relief
The surgeons treating patients with advanced pancreatic head carcinoma have to keep in mind that quality of life is the most important factor for these patients who have such a dire prognosis. In this context, pain is the most feared symptom for a majority of the patients and for many of them pain constitutes a clinically significant problem until death; pain management is troublesome.
Pain fibers from the pancreatic gland (the ce-liac ganglion) run within the major and minor splanchnic nerves to the spinal column. An in-terruption of this pathway can provide pain re¬lief, and such a disruption can be accomplished by either targeting the abdominal or the thoracic cavity. The first intraoperative chemical splanch-nicectomy was introduced by Copping and col¬leagues in 1969 [20]. In their clinical experience reported almost 10 year later, approximately 90% of patients with pain at diagnosis experienced significant relief after intraoperative chemical splanchnicectomy [21]. Since then many in¬vestigators have utilized this method, and Lil-lemoe and coworkers reported in a randomized controlled trial that intraoperative chemical splanchnicectomy with 50% alcohol significantly reduced or prevented pain in patients with unre-sectable pancreatic cancer [22]. In contrast, van Geenen and coworkers from Amsterdam could not confirm these findings. In their randomized study, patients were divided into three groups: (1) palliative bypass surgery receiving intraop-erative celiac plexus blockade, (2) palliative by¬pass surgery without celiac plexus blockade but followed by high-dose conformal radiotherapy, and (3) palliative bypass surgery with both (ce-liac plexus blockade, followed by high-dose con-formal radiotherapy). They concluded that celiac plexus blockade for pain management did not result in an increase to pain medication-free sur¬vival and therefore presumed that celiac plexus blockade could not demonstrate a positive effect on pain management for the patients with ad¬vanced pancreatic carcinoma [23]. To disrupt the pain neuropathway, splanchnic nerves within the thorax could also be interrupted. This can be ac-complished either via thoracotomy or via video-assisted thoracoscopy (VATS) [24, 25]. In recent times, the plexus blockade has been reached via nonsurgical interventions, namely using endo-scopic ultrasound (EUS) or wild-guided tech¬niques. A prospective study of the EUS-guided celiac plexus neurolysis for pain treatment in patients with advanced pancreatic head can¬cer showed that the technique is safe and yields pain control [26]. In light of such nonsurgical alternatives for celiac plexus blockage, the role of surgical pain relief seems to be marginal nowa¬days. However, in selected patients, namely who are not responding to noninvasive methods of plexus blockade, a surgical intervention may still be indicated.

13.6    Palliative Resection: Does It Play a Role?
Adenocarcinoma of the pancreatic head is con-sidered one of the gastrointestinal malignancies with the worst prognosis. If no standardized op-erative procedures have been established (classi-cal Kausch-Whipple resection, pylorus preserv¬ing pancreatic head resection) the overall 5-year survival rate of patients with pancreatic head carcinoma today is estimated to be approx. 5%, and the 5-year survival rate after curative resec-tion reaches approx. 20% in specialized centers around the world [27, 28]. However, differing from earlier reports, the perioperative mortal¬ity has decreased significantly during the last few decades, and today morbidity rates around 15% and mortality rates below 3% for standard¬ized pancreatic head resection have been reached in high-volume centers. This fact and the con-sideration that the patient with pancreatic head carcinoma in the majority of cases presents in stage 3, in which advanced disease is present and undetected further tumor spread has to be ex-pected, many investigators today believe that real curative resections are rare events. Further argu-ments on this line are supported by the results of recent multicenter trials in adjuvant chemother-apy after R0 or R1 resections for pancreatic head carcinoma showing that even the patients with R1 resection profit considerably from postopera¬tive adjuvant therapy [29]. Therefore, oncologists have long proposed that the pancreatic head re-section for a defined tumor no longer be termed a curative resection, since in most of the cases advanced stages of the disease are present. Many investigators today believe that all resections for pancreatic head carcinoma are, in principle, pal-liative. Following this argument, the goal of the resection may change much more toward reach¬ing quality of life for the patients, as with other palliative procedures.
At present no prospective data are available in which a palliative resection was investigated in a randomized fashion. Several authors have re-ported about retrospective data comparing pal-liative procedures (biliary and gastric bypasses combined or alone) with pancreatic resections [30]. Lillemoe et al. investigated the role of pallia-tive resection compared to combined biliary and gastric bypass, showing a significant improved overall survival for patients undergoing pallia¬tive pancreaticoduodenectomy. All patients were patients in which, after transecting the pancreas (passing the point of no return during pancreatic head surgery), nonresectability was found in the retro-pancreatico-duodenal plane [31]. However, in this study no further subdivisions regard¬ing the R-status was accomplished. In another study, Reinders et al. compared patients after a microscopically nonradical pancreaticoduode-nectomy with patients after surgical bypass [32]. Both studies neither evaluated the quality of life nor the long-term follow-up criteria and only showed that the so-called palliative pancreati-coduodenectomy procedures yield significantly better results and longer survival than ones in which patients received surgical bypass, leaving their primary tumor mass in place. In one recent study the investigators compared patients in a palliative situation undergoing double loop by¬pass surgery with patients undergoing palliative pancreaticoduodenectomy. Special emphasis was laid on the investigation of quality of life in this study. All patients undergoing bypass were sub-grouped into those with locally advanced disease and those with metastasized diseased. The 1-year survival was 25% in the palliative resected group vs 20% in the locally advanced and 15% in the metastasized disease group. The quality-of-life data were favorable for the patients after bypass surgery; however, the morbidity and mortality rates in patients after palliative resection were elevated [33].
These results prompted us to propose a study in which the role of the palliative resection it¬self should be evaluated. Following extensive interdisciplinary discussions with gastroenter-ologists and oncologists, we derived a protocol in which patients with carcinomas of the pan¬creatic head that had already metastasized into the liver at diagnosis and revealed a resectable situation were randomized into two groups. One group would receive standard gemcitabine che¬motherapy until tumor progression, whereas the other group would receive a pancreatic head re¬section with or without liver resection and sub¬sequently standard gemcitabine treatment until tumor progression. The liver resection was only to be performed when resections could provide a significant tumor mass reduction because the additional surgical risk to the pancreatic head procedure was to be avoided This study is now underway (and hopefully open for recruitment) and we will be able to analyze the results in the near future.
Palliative pancreatic head resections outside of accepted study protocols should not be per-formed since the significant additional clinical risk of complications for morbidity and mortality is not acceptable; they must only be performed in the framework of randomized prospective trails.

13.7    Summary
For the treatment of malignant biliary obstruc-tions in patients with pancreatic carcinoma, en-doscopic biliary drainage is the option of first choice. In case of persistent stent-problems such as occlusion or recurrent cholangitis, a hepatico-jejunostomy should be considered. The role of a prophylactic gastroenterostomy is still under dis¬cussion. In selected patients with duodenal ste¬nosis present at the time of operation, or patients with impending duodenal obstruction, a prophy¬lactic gastroenteric bypass may be indicated. The same should be considered for patients showing a duodenal stenosis during an operation for biliary obstruction. In such patients an initial combined biliary and gastric bypass should be performed to avoid a second operation for gastric outlet ob¬struction. The significance of laparoscopic bili¬ary bypass is not yet clear. A surgical, minimally invasive approach for treating bile duct obstruc¬tion is not the standard nowadays, and it should be reserved for experts in the field of laparos-copy. Otherwise, laparoscopic gastrojejunostomy is a standardized surgical procedure that offers significant advantages in regards to morbidity and mortality compared to open surgical tech¬niques. The role of surgical pain relief is mostly negligible today. Computed tomography (CT)-or EUS-guided celiac plexus neurolysis have re¬placed surgical interventions. The significance of palliative resections is a controversial topic now¬adays. However, beyond controlled randomized studies, a palliative pancreaticoduodenectomy in patients with advanced pancreatic carcinoma cannot be recommended at this time.