Pancreatic Cancer Surgery Treatment

Pancreatic cancer continues to pose a major public health concern and clinical challenge. The incidence of the disease is nearly equivalent to the death rate associated with the diagnosis of pancreatic cancer. Thus, there exists a need for continued improvement in the diagnostic, therepeutic and palliative care of these patients.

Surgeons play an integral role in the management of pancreatic cancer patients, with surgery providing the only potentially curative intervention for the disease. Specialized centers have reported improved hospital morbidity, mortality and survival after pancreaticoduodenectomy; however, disease-specific survival after surgical resection remains dismal. An emphasis therefore has been placed upon the accurate preoperative staging of patients in order to identify those patients who would benefit from a complete surgical resection. Surgical staging that incorporates the use of laparoscopic techniques now complements non-surgical methods of staging, including helical CT scans. While there is no defined preoperative staging approach, it is imperative that centers identify areas of expertise and experience with available modalities in any combination to effect accurate staging.

Once patients have been accurately staged and deemed resectable, there exist various methods for resection of pancreas lesions, which include the standard “Whipple procedure,” pylorus-preserving pencreaticoduedenectomy, regional pancreatectomy, total pancreatectomy, and en bloc vascular resection, where appropriate. Reconstructive techniques have been explored and include methods of pancreaticojejunostomy and pancreaticogastrostomy with or without pancreatic ductal stents and intreoperatively placed closed suction drains.

Perioperative mortality following pancreaticoduodenectomy for cancer has a general reported incidence of 1% to 4% at high volume centers experienced with the operation. Morbidity however still remains high with that of delayed gastric emptying, pancreatic anastomotic leak or fistula, intraabdominal abscess, and hemorrhage as the leading reported complications. Researchers have investigated several agents and strategies to decrease or prevent the potential morbidity of these complications including the use of octreotide, drainage of the pancreatic bed and institution of early enteral feeding.

Unfortunately, the majority of patients with pancreatic cancer present with either locally advanced or metastatic disease that precludes operative cure. The expected survival for these patients is usually less than six months from diagnosis. Therefore, a goal of therapy should be adequate palliation of symptoms of pain, biliary or duodenal obstruction and improvement of remaining quality of life with the least degree of morbidity possible.

Pancreatic cancer continues to pose a major public health concern and clinical challenge. In 2000, an estimated 28,300 new cases of pancreatic cancer were diagnosed, accounting for 2% of all new cancers and 5% of all cancer-related mortality in the United States. This lethal tumor ranks fourth in males and fifth in females as a leading cause of death behind lung, breast, prostate, colorectal, and ovarian cancer in the United States. Worldwide, more than 185,000 cases occur annually, with a death: incidence ratio of 0.99. Surgery remains the only potentially curative intervention for those patients who present with localized diseased. Although pancreatic resection has previously been associated with high operative morbidity and mortality, significant advances in surgical technique, perioperative care, and patient selection have ted to a decline in associated morbidity and mortality. In those patients who present with locally advanced or unresectable disease, palliative surgery may play a significant role in terms of quality of life. This review focuses on the advances made in the surgical therapy of patients with pancreatic adenocarcinoma encompassing applied operative staging, therapeutic, and palliative techniques.
PANCREATIC RESECTION: HISTORICAL TO MODERN DAY

Walter Kausch, a German surgeon from Berlin, is credited with performing the first successful pancreaticoduodenectomy (PD) in two stages in 1912. Two years later, Hirschel reported the first successful one-stage PD³; during the early 1900s, however, most periampullary cancers were managed by a transduodenal approach originally described by Halsted in 1899. It wasn't until the 1930s to mid-1940s that Whipple and colleagues popularized the procedure that was initially performed as a two-stage procedure for ampullary cancers. Throughout Whipple's career, the procedure evolved into the one-stage operation that now bears his name. In 1937, Brunschwig from the University of Chicago extended the indications for PD to include those patients with cancer of the head of the pancreas, and the procedure was subsequently applied to patients with periampullary cancers.[ 8] In the early series of 1960–1970, reported pancreatic resections carried a high operative mortality (> 20%) and morbidity (> 50%), leading some authors to suggest that the operative approach be abandoned.

More recently, specialized centers have reported improved hospital morbidity, mortality, and survival after PD with mortality rates of < 4%, approaching 1% in selected series and associated morbidity of < 25%. In spite of these improvements, however, the results of PD with respect to disease-specific survival remain dismal. Overall, approximately 90% of patients present with locally advanced or disseminated disease, and over 95% of patients succumb to the disease, most within the first year. Of the patients undergoing exploration for a potentially curative resection, only 20%–25% actually have resectable disease, with resulting 5-year survival rates of 15%–20%.

RATIONALE FOR SURGICAL THERAPY

The traditional approach to patients with pancreatic cancer has been to explore all patients with the intent of determining resectability—a determination obtained only through a hands-on intraoperative examination by the surgeon, with the exploration procedure accompanied in many cases by biopsy for a tissue diagnosis. In those patients who were deemed inoperable, palliative bypass procedures were performed, thereby justifying exploration.

This practice has now been brought to question with the advent of non-operative palliative techniques that have proven efficacious in patients with inoperable, disseminated disease. An emphasis has been placed upon the accurate preoperative staging of patients in order to identify the subgroup of patients who would in fact benefit from a complete surgical resection. The goals of clinical staging should reliably define the extent of disease, thereby avoiding unnecessary intervention in a safe and cost effective manner. Patients who stand to derive the greatest benefit of a selective approach to surgical intervention are the > 50% of patients who present with advanced disease and who have the poorest prognosis. Accurate staging avoids unnecessary morbidity, mortality, and diminished quality of life in a patient population with very limited survival.
STAGING PANCREATIC CANCER: IMPLEMENTATION OF A MINIMALLY INVASIVE APPROACH

Pancreatic cancer patients have much to gain from accurate clinical staging. Multiple modalities have been developed towards this goal and have been utilized in different combinations by specializing centers treating the disease. While there is no defined preoperative staging approach, it is imperative that centers identify areas of expertise and experience with available modalities in any combination to effect accurate staging.

Even with the advent of many sophisticated staging modalities, the accurate pre-operative assessment of disease remains difficult. Dynamic, contrast-enhanced, high resolution CT scan is the principal radiological study for the preoperative evaluation of pancreatic cancer patients. Published reports of the ability of high quality CT scan to predict the resectability of patients preoperatively range from 53% to 89%. The wide range of reported resectability rates are likely due to poor quality scanning techniques and differences in definitions of radiological criteria for resectability. When high quality CT scans are employed preoperatively and appropriately interpreted with prospectively defined criteria for resection, resectability rates approach 80%, and approximately 5% to 20% of patients may be spared laparotomy.[ 30] This constitutes the small, yet defined, subgroup of patients determined to be resectable by high quality preoperative CT scan who may benefit from further staging via laparoscopy performed at the time of surgical resection.

Laparoscopy has emerged as a powerful tool for staging pancreatic cancer. The addition of laparoscopic ultrasound has furthermore overcome, at least in part, the lack of tactile sensation in a standard two-dimensional laparoscopic examination, improving upon the ability to detect subhepatic metastases and tumor-vessel involvement. The combination of techniques allows the direct visualization of the primary tumor, determination of its association with vascular structures, detection of the presence of regional nodal disease, and identification of small volume peritoneal or liver metastases not appreciated on preoperative imaging.

At Memorial Sloan-Kettering Cancer Center, an aggressive surgical approach has been adopted utilizing a combination of high resolution CT scan and multi-port, extended laparoscopy with laparoscopic ultrasound to determine resectability. Conlon and colleagues have demonstrated the usefulness of laparoscopy in the identification of occult metastatic disease, with improved resectability rates translating into a reduction in the hospital stay for those patients who underwent laparoscopic staging. In our hands, laparoscopic assessment in conjunction with helical CT scan provided a positive predictive index of 100%, a negative predictive index of 91%, and an overall accuracy of 94% The addition of laparoscopic ultrasound improved the accuracy of determining resectability from 94% to 98%. Contraindications to resection include confirmed hepatic, serosal, peritoneal, or omental metastases, and/or invasion or encasement of the perihilar portal vein, celiac axis, hepatic artery, or superior mesenteric artery. Relative contraindications to resection include tumor extension outside the pancreas or nodal disease. The introduction of laparoscopic staging, coupled with improvements in preoperative imaging and better patient selection, has led to a change in the requirements for open surgery and the type of procedure performed for pancreatic cancer at our institution. Unnecessary surgery has been avoided in those patients with unresectable disease, yet potentially curative surgery has been performed in those who would otherwise benefit from surgical intervention.

Despite these reported results and similar results of other investigators, laparoscopic staging of patients with pancreatic cancer continues to provoke considerable debate. Critics have argued that open exploration with assessment of local extension and vascular encasement is the most accurate way to determine resectability.Others offer that even with a 10%–20% benefit of laparoscopic staging over conventional radiological modalities, the additive value of laparoscopic staging is minimal and not cost effective. Still others believe that there are few patients who will be spared eventual open operation for palliation, and therefore the impact of laparoscopy in avoiding an operation is limited. On the other hand, subsequent palliation of patients after stagging laparoscopy was reported by Espat and colleagues, who showed that in a prospective cohort of 155 patients with pancreatic cancer, only 3% of patients eventually required an operation for palliation of symptoms.

All together, reported clinical experience demonstrates that laparoscopy prevents unnecessary laparotomy in patients who are presumed to have resectable disease by high resolution CT scan. The selective use of this staging modality at the time of planned laparotomy in patients considered for resection is appropriate, especially in those patients at high risk for occult metastatic disease.[ 30]
PREOPERATIVE PREPARATION: THE ROLE OF BILIARY DRAINAGE

Patients with pancreatic cancer who have jaundice at initial presentation are also at risk for associated coagulopathy, malabsorption, malnutrition, and immune dysfunction. The development of transhepatic and endoscopic stents for biliary drainage led to the hypothesis that preoperative biliary drainage may decrease the morbidity and mortality related to pancreatic surgery. Several small prospective trials have failed to show a benefit with preoperative biliary decompression.More contemporary, larger series have likewise shown no advantage to preoperative biliary decompression. On the contrary, Povoski and colleagues reported the results of 240 patients undergoing PD, of which 126 were managed with preoperative biliary decompression. In patients undergoing preoperative biliary drainage, overall complication rates (55% vs 39%), operative morbidity related to infectious complications (41% vs 25%) and mortality (8% vs 3%) were significantly increased. Sohn and colleagues likewise reported an increase in the incidence of wound infection and pancreatic fistulas in 567 patients undergoing PD of which 408 had biliary decompression procedures. In contrast, Martignoni and colleagues reported a consecutive series of 257 patients, of which 99 patients (38%) underwent preoperative biliary drainage, and found no effect on operative morbidity or mortality in treated patients.

These data suggest that as a routine, preoperative biliary decompression is not warranted, especially in the absence of documented cholangitis or other severe complications of jaundice that would preclude a safe resection.
TECHNICAL CONSIDERATIONS Operative Management of Head of Pancreas Lesions

Four main surgical options exist for patients with pancreatic adenocarcinoma located within the head or uncinate process of the gland. They include what most accept as the standard “Whipple” PD, pylorus-preserving PD (PPPD), regional pancreatectomy, and total pancreatectomy. Controversies exist regarding the performance and techniques of pancreatic resection, including the extent of resection as well as the methods of reconstruction.

The “Whipple” Procedure The standard operative approach to pancreatic resection for head and uncinate process lesions incorporates many aspects of the traditional Whipple procedure. The initial exploration is dedicated to thorough evaluation of patients for evidence of locoregional or metastatic disease. Once a patient is deemed resectable, the pancreatic resection proceeds in a systematic fashion. An extensive Kocher maneuver is performed followed by cholecystectomy and portal dissection. The distal common hepatic duct and the gastroduodenal artery are divided. The anterior surface of the portal and superior mesenteric veins are cleared and transected. Traditionally, the antrum of the stomach is divided, however the pylorus may be preserved transecting the duodenum 2 cm distal to the pylorus. The pancreatic neck is divided and the jejunum transected beyond the ligament of Treitz. The specimen is dissected from the superior mesenteric and portal veins medially followed by complete mobilization of the superior mesentric vein and removal of all tissue to the fight of the superior mesenteric artery constituting the retroperitoneal margin. Gastrointestinal continuity is typically restored via an end-to-end or end-to-side pancreaticojejunostomy, hepaticojejunostomy, and duodeno-or gastrojejunostomy.

Standard PD versus PPPD Traverso and Longmire re-popularized the PPPD in 1978, and this operation is the most typical variation of the standard Whipple procedure performed. Its use has come into favor due to the potential for preservation of gastric reservoir function and more normal gastrointestinal physiology with no compromise to an adequate tumor operation.

Researchers have prospectively addressed the hypothesis that a pylorus-sparing operation may be superior to standard PD with respect to outcome measures of nutrition, morbidity and mortality. Seiler and colleagues prospectively compared standard PD to PPPD in 77 patients with periampullary cancer. Patients who underwent PPPD demonstrated significantly shorter operative times, reduced blood loss, had fewer blood transfusions, and showed no difference in the incidence of delayed gastric emptying or overall mortality with respect to the standard PD group. Furthermore, there were no differences in tumor recurrence or long-term survival at a median follow-up on 1.1 years. Lin and colleagues also reported a small prospective study comparing standard PD to PPPD in 30 randomized patients with resectable periampullary adenocarcinoma and found no difference in operative blood loss, mean operative time, median transfusions or pancreatic leaks between the two study groups. There was a suggestion of a higher incidence of delayed gastric emptying in the PPPD, however the study groups were too small to accurately analyze the predictive effect of procedure type on outcome measures.

Large series reviewing outcomes of PPPD for pancreatic cancer have failed to demonstrate any deleterious effects of pylorus preservation on outcome. In the Johns Hopkins series, more than 80% of patients underwent successful resection with pylorus-preservation. The pylorus was sacrificed mainly in cases of tumor involvement or ischemia of the duodenum.

There appears to be little difference between standard PD and PPPD in terms of operative time, blood loss, overall morbidity and mortality. PPPD does not appear to compromise long-term survival or recurrence rates although may be associated with an increase in delayed gastric emptying. Prospective trials as yet are underpowered to definitively address these issues.

Regional Pancreatectomy Fortner employed a regional pancreatectomy procedure in the early 1970s, which included an en bloc total or subtotal pancreatectomy with radical lymph node dissection and in some cases portal vein resection. Two prospective randomized trials have recently attempted to further define the utility of a more extended procedure for pancreatic adenocarcinoma.

The first trial, reported by Pedrazzoli and colleagues, was a prospective, randomized, multi-center trial comparing standard (N = 40) versus extended (N = 41) lymphadenectomy with pancreatic resection for cancer. The extended procedure included resection of hepatic hilar and periaortic lymph nodes, extending from the diaphragmatic hiatus to the inferior mesenteric artery and laterally to the renal hila, in addition to the standard peripancreatic lymph nodes. The authors found no difference in transfusion requirements, morbidity, mortality, and overall survival between the two groups. Subgroup analysis suggested an associated survival benefit to node-positive patients (Stage III) undergoing an extended regional lymphadenectomy although the 4 year overall survival of all patients with potentially curative pancreatic cancer was only 7.4%.

A prospective, randomized single-institution trial was conducted by investigators at Johns Hopkins comparing standard versus extended PD in patients with periampullary cancer to determine outcome measures of post-operative morbidity, mortality and survival. Fifty-six patients underwent standard PD (86% pylorus-preserving resection) and 58 underwent an extended resection (100% included a distal gastrectomy). There were no statistical differences noted in operative time, blood loss, transfusion requirements, positive lymph node or margin status, complication rates, or 1-year actuarial survival. Of note, however, is that the radical resection group had a higher incidence of delayed gastric emptying, as well as a higher number of mean lymph nodes resected, compared to the standard resection group. Only 10% of patients in this group, however, had metastatic disease detected in retroperitoneal lymph nodes, and no patient had the retroperitoneal nodes as the only site of nodal metastasis.

These studies demonstrate that extended PD performed at centers with experience may be conducted safely with similar morbidity and mortality to standard resections, although without apparent improvement in survival. A potential bias, however, is that both trials may be considered underpowered to detect a difference in outcome. Nonetheless, regional nodal metastases portend a grave prognosis; whether or not removal of these nodal basins imparts a benefit to the patient remains an area of investigation.

Total Pancreatectomy The rationale for total pancreatectomy for pancreatic adenocarcinoma stems from the theoretical advantages of eliminating multifocal disease, achieving wider resection margins, and avoiding spillage of tumor cells and the complications of postoperative pancreatic anastomotic leak. Several large series, however, have demonstrated no advantage of total pancreatectomy compared to standard PD for pancreatic cancer. In the Johns Hopkins series of 201 patients undergoing pancreatic resections for head of pancreas lesions, median survival was 10 months in the total pancreatectomy group, compared to 16 months in the PD group.[ Likewise, researchers at Memorial Sloan-Kettingting Cancer Center have shown that in patients undergoing pancreatectomy for cancer, total pancreatectomy had a significantly lower overall survival and high morbidity compared to a contemporary cohort of patients undergoing PD and distal pancreatectomy. Interestingly, in this selected group of patients, there was no apparent effect of number of positive nodes, positive margins, or tumor differentiation on outcome.

In summary, total pancreatectomy for adenocarcinoma offers no survival advantage over subtotal resections. In addition, positive margins at the time of PD should not encourage the surgeon to perform a total pancreatectomy, as morbidity is greater and outcomes are not improved over standard resections.

Vascular Resection The technique of portal vein (PV) resection was initially reported by Fortner in a series of type I regional pancreatectomy that included PV resection. The rationale for this procedure was to improve resectability rates and negative resection margins. Since his original report, portal venous resection has undergone considerable refinement with respect to the group of patients with isolated tumor involvement of the PV or superior mesentetic vein (SMV), with acceptable associated morbidity and mortality. Local involvement of venous structures is present in almost one third of patients with pancreatic cancer and is the only barrier to successful resection. Isolated tumor adherence to SMV or PV structures is generally not considered a contraindication to resection. Long segment involvement with tumor and thrombosis of the SMV or PV, determined either by clinical signs or by radiologic evidence of clot or varices, constitute conditions of unresectability.Several authors have demonstrated that invasion of the PV or SMV is not associated with margin and lymph node positivity, histopathologic prognostic features that suggest a poor prognosis. Compared to pancreatectomy without PV resection, however, invasion of the PV or SMV is associated with greater blood loss, transfusion requirements, and operative time, and with larger tumors. Most important, several reports have confirmed that the need for PV resection at the time of PD does not affect overall patient survival, reflecting acceptable morbidity and mortality (< 5%) with this more extensive procedure.The preferred technique for resection should be tailored to the degree of vessel involvement and the experience of the surgeon. Primary reconstruction with an end-to-end anastomosis is usually possible. Some authors advocate preservation of the splenic vein, when possible, and reconstruction with patch or interposition grafts utilizing harvested saphenous, internal jugular, or renal veins. In all, suspected isolated PV or SMV involvement frequently does not preclude operability and, by itself, should not be considered a contraindication for pancreatic resection.
Operative Management of Distal Pancreatic Lesions

Typically, patients with adenocarcinoma of the body and tail of the gland present with advanced disease because tumors do not obstruct the bile ducts and patients are not diagnosed earlier due to symptomatic jaundice. Tumors of the body and tail therefore tend to be larger and are associated with a higher incidence of metastatic disease at presentation. A large series of patients with body and tail lesions was recently reported, and its authors concluded that patients with body and tail pancreatic cancers have similar survival rates stage-to-stage compared to head of gland lesions. Overall resection rates, however, were lower for distal lesions. Operative mortality for distal pancreatectomy was less than that of PD performed for cancer. In general, patients with lesions in the body and tail should be approached in a similar fashion to patients with head of gland cancers. These patients, in particular, should be staged aggressively and should be considered for operative resection without evidence for vascular invasion or metastatic disease.

While splenic preservation is not indicated in patients with adenocarcinoma, it may be considered in patients with low-grade malignancies without compromising survival. Investigators have addressed this topic in a retrospective review of patients with benign or low-grade malignant lesions undergoing distal pancreatectomy with (N = 46) or without (N = 79) splenic preservation. Interestingly, the incidence of infectious complications, blood loss, and length of hospital stay were reduced in patients who had splenic preservation (Shoup, 2001, unpublished data).
Management of the Pancreatic Remnant

There are multiple methods available for reconstruction of the gastrointestinal tract. Controversy continues to exist regarding the optimal pancreatic anastomosis, the importance of duct-to-mucosa sutures, and the use of pancreatic duct stents. Because pancreatic fistula is the most ominous complication of PD and is the leading cause of surgical morbidity and mortality in 10% to 20% of patients, efforts have been concentrated on developing the safest pancreatic anastomosis.

Pancreaticojejunostomy versus Pancreaticogastrostomy Pancreaticogastrostomy (PG) has been investigated prospectively as an alternative drainage procedure for the pancreatic remnant based on several retrospective series reporting lower rates of pancreatic fistula in patients reconstructed with PG versus pancreaticojejunostomy (PJ). The first prospective, randomized series was reported by the Johns Hopkins group and demonstrated similar rates of pancreatic fistula in groups with PG and PJ reconstruction. Independent predictors of pancreatic fistula were the surgeon's volume of cases and ampullary or duodenal disease in the resected specimen. The occurrence of pancreatic fistula significantly prolonged hospital stay and was associated with other perioperative morbidity such as delayed gastric emptying, intra-abdominal abscess, and choleangitis. Notably, pancreatic stents were not utilized in this series, and all pancreatic duct anastomoses were performed in two layers.

A non-randomized, prospective series, reported by a Japanese group, found decreased pancreatic fistula rates (0% vs 13%, P = 0.014) and decreased intraabdominal abscess rates (0% vs 6%, P = 0.036) in patients undergoing PG versus PJ, respectively. Pancreatic anastomoses were carried out in two layers with pancreatic duct stents. Two hospital deaths occurred in the PJ group (3%) and no hospital mortality occurred in the PG group.

Method of Pancreaticojejunostomy Other technical modifications have been explored to reduce the rate of pancreatic fistula. Chou and colleagues prospectively compared two methods of PJ reconstruction, including end-to-end (invaginating/telescoping) and end-to-side (duct-to-mucosa with stent) techniques in patients undergoing standard PD for periampullary cancers. There was a trend toward decreased fistula rates with the end-to-side technique utilizing stents compared to the end-to-end, invaginating technique (4% vs 15%, P = 0.09, respectively). Other associated morbidity/mortality factors were not significantly different between the two groups.

These clinical data indicate that there is likely no difference in reconstruction methods with respect to pancreatic fistula rates; thus, pancreatic anastomosis should be carried out according to the surgeon's expertise. Although the Japanese data suggest an advantage to PG and, in particular, stenting of the pancreatic duct, these results must be interpreted within the context of non-randomized data. In addition, an end-to-side/duct-to-mucosa technique with stenting for PJ reconstruction can be performed safely with a lower associated rate of anastomotic dehiscence compared to an end-to-end/invaginating technique.
PERIOPERATIVE MANAGEMENT AND COMPLICATIONS

Perioperative mortality following PD for cancer has a general reported incidence of 1% to 4% at high volume centers experienced with the operation. Morbidity still remains high, however, with delayed gastric emptying, pancreatic anastomotic leak or fistula, intra-abdominal abscess, and hemorrhage as the leading reported complications. Researchers have investigated several agents and strategies to decrease or prevent the potential morbidity of these complications, including the use of octreotide, drainage of the pancreatic bed, and institution of early enteral feeding.
Role of Octreotide Following Pancreatic Resection

Several prospective, randomized European trials have been conducted to evaluate the role of prophylactic octreotide in patients undergoing elective pancreatic surgery. A subsequent meta-analysis of the European trials found that prophylactic use of octreotide significantly reduced the rate of pancreatic fistula formation, although the extent of resection was not considered in the analysis of this heterogeneous group of patients with both benign and malignant histology.

The limitations inherent in the European trials prompted researchers at Johns Hopkins Hospital and M.D. Anderson Cancer Center to reevaluate the role of octreotide in patients specifically undergoing PD for malignancy. Lowy and colleagues randomized 11.0 patients to treatment with either octreotide or placebo in postoperative pancreatectomy patients and found no significant difference in pancreatic fistula rates, length of hospitalization, or mortality between the study groups. Reoperative PD was an independent predictor of pancreatic disruption on multivariate analysis. Yeo et al randomized 211 patients to a similar treatment with either octreotide or placebo in a similar group of pancreas cancer patients and likewise found no significant difference in pancreatic fistula rates, morbidity, or mortality Interestingly, complications were significantly more common with soft glands, which were also found to be an independent predictor for fistula formation on multivariate subset analysis.

Thus, octreotide does not appear to benefit patients with pancreatic cancer undergoing PD in terms of pancreatic fistula formation, overall morbidity, or mortality. There may exist subsets of patients who should be considered for prophylactic octreotide therapy They include those patients who are undergoing reoperative procedures or those who are found to have soft glands at operative exploration.
Drainage of the Pancreatic Bed

Closed suction drains after PD are theoretically used to drain potential collections and anastomotic leaks thereby decreasing associated morbidity and mortality. Heslin and colleagues[ 79] reviewed 89 consecutive patients undergoing PD for cancer according to whether or not an intra-abdominal drain was placed at the time of surgery. They found no significant difference in pancreatic fistula rates, intra-abdominal abscess formation, need for CT-guided drainage of collections, reoperation for bleeding, or length of hospital stay between the two groups.

This finding prompted a subsequent recent prospective, randomized trial conducted by Conlon and colleagues, who failed to demonstrate a reduction in morbidity or mortality with the addition of surgically placed, closed-suction drains following pancreatic resection. For the subgroup of 88 patients randomized to drainage, the presence of drains failed to reduce either the need for subsequent interventional radiological drainage or surgical exploration for intra-abdominal sepsis. Based on the results of this study, the authors concluded that drainage should not be considered mandatory or standard following pancreatic resection.
Early Delayed Gastric Emptying

In most series, delayed gastric emptying occurs in up to 30% to 40% of PD patients. It is broadly defined as the need for prolonged postoperative nasogastric decompression (>10 days) leading to prolonged hospital stay and cost. The pathogenesis of delayed gastric emptying is multi-factorial.

Researchers studied the effects of intravenous erythromycin, a motilin agonist, in a prospective, randomized, placebo-controlled trial in patients undergoing PD for cancer, in an attempt to reduce the incidence of this complication. Erythromycin significantly reduced the incidence of early delayed gastric emptying (19%) versus the control (30%) and was associated with improvement in both solid and liquid emptying of the stomach, resulting in a 37% reduction in the incidence of delayed gastric emptying. Many support its prophylactic routine use in patients undergoing PD.

Nutrition

Investigators at Memorial Sloan-Kettering Cancer Center conducted a prospective, randomized trial of early enteral feeding with an immune-enhancing formula (containing arginine, RNA, omega-3 fatty acids) in patients with upper gastrointestinal malignancies. The rationale was based on the improved outcomes noted with early enteral feeding in trauma and critical care patients. Patients were randomized to early enteral feeding via jejunostomy tube or control (standard intravenous fluid support until resumption of oral intake). The two treatment groups showed no significant difference in number of major, minor, or infectious wound complications. Hospital stay (11 days) and overall mortality (2.5%) also did not differ between the two groups. Authors concluded that early enteral nutrition does not provide any benefit over postoperative crystalloid support in generally well-nourished patients and cannot be recommended routinely to patients undergoing major upper gastrointestinal surgery.
PALLIATIVE SURGERY

As described above, the majority of patients with pancreatic cancer present with either locally advanced or metastatic disease that precludes operative cure.[ 83] The expected survival for these patients is usually less than six months from diagnosis. Therefore, a goal of therapy should be to adequately palliate symptoms of pain, biliary or duodenal obstruction and to improve the quality of remaining life with the least degree of morbidity possible.
Pain Control: Role of Chemical Splanchnicectomy

Lillemoe and colleagues conducted a prospective, randomized, double-blind trial comparing intraoperative chemical splanchnicectomy to placebo in 139 patients with unresectable pancreatic carcinoma.[ 84] In patients receiving alcohol splanchnicectomy, pain scores were significantly lowered and narcotic requirement decreased compared to controls. Patients in the alcohol-treated group required subsequent percutaneous celiac axis block 12 months after chemical splanchnicectomy postoperatively, versus 4 months for control patients. For patients that were pain-free at presentation, chemical splanchnicectomy more than doubled the mean number of pain-free months compared to controls (7.2 months vs 3.0 months, P = 0.001). For patients who initially presented with pain, alcohol injection provided longer, though less durable, pain relief (3.3 months vs 0.8 months, P = 0.05). Significant pain was present in 65% of alcohol-treated patients and 100% of control patients at the time of death. Interestingly, patients treated with alcohol injection had a significantly longer survival than control patients did, an unexpected finding not explained by this study In general, chemical splanchnicectomy is advocated for patients presenting with unresectable pancreatic cancer and can be performed at the time of open surgery or as a percutaneous procedure in patients who are otherwise staged non-operatively.

Biliary Endoprosthesis versus Surgical Bypass for Obstructive Jaundice

Biliary obstruction is present in up to 7596 of patients who present with unresectable pancreatic cancer and, if untreated, leads to progressive pruritus, cholangitis, and decreased survival. Options for palliation include percutaneous transhepatic and endoscopic biliary stenting, and surgical bypass. Surgical bypass is known to be effective, with large operative series reporting recurrent jaundice in only 4% of bypassed patients prior to death. Biliary endoprostheses, however, have been challenged in multiple prospective randomized trials comparing their utility to operative bypass procedures.

In general, clinical data indicate that overall morbidity, mortality, and survival are similar for patients treated with percutaneous transhepatic endoprostheses, endoscopic stenting or surgical bypass in the setting of malignant biliary obstruction. Surgery carries a higher early morbidity compared to stenting, but stenting has a higher long-term failure rate due to recurrent jaundice from stent blockage and gastric outlet obstruction. Therefore, current recommendations support the use of non-operative biliary decompression in medically poor-risk patients or those with widespread metastatic disease and associated ascites and carcinomatosis. Patients who require operative bypass for duodenal obstruction and are otherwise expected to live longer than six months are best treated with operative biliary bypass.
Role of Prophylactic for Unresectable Periampullary Cancer

The appropriate palliation of duodenal obstruction is the subject of considerable debate. In the pre-CT era, exploratory laparotomy was the only way to diagnose and stage patients with pancreatic cancer. At the time of exploration, if the tumor was unresectable, a biliary bypass and gastric bypass could be performed as therapy or prophylaxis of symptoms without associated increase in mortality. Retrospective series and prospective non-operative trials suggest that duodenal obstruction eventually develops in approximately 10% to 20% of patients, and many series have both supported and condemned the use of prophylactic gastric bypass.

In an effort to evaluate the role of prophylactic gastric bypass in patients with unresectable periampullary cancer, a prospective trial was conducted randomizing patients to retrocolic gastrojejunostomy or no gastric bypass at the time of laparotomy for pancreatic resection.[ 85] Eighty-seven patients who were determined by the operating surgeon not to be at significant risk for gastric outlet obstruction were randomized. There were no significant differences between the two groups with respect to operative mortality and length of hospital stay. Gastric outlet obstruction eventually developed in 19% of non-bypassed patients and required intervention approximately two months after initial exploration. Patients who underwent initial gastric bypass procedures, however, had no subsequent obstructive symptoms, and only 2% of patients experienced delayed gastric emptying. In contrast, a contemporary prospective, non-randomized series collected during the current era of CT and laparoscopic staging demonstrated that only 3% of patients required operative intervention for gastric outlet obstruction after initial determination of unresectability at laparoscopy and subsequent non-operative management.[ 39]

Despite these data, controversy continues to exist regarding the appropriate timing and selection of patients for surgical bypass procedures. The ability to define unresectability has improved considerably with the application of spiral CT-angiography and laparoscopic staging and the need for open exploration in adequately staged patients has become less common. The introduction of minimally invasive techniques has broadened the range of solutions available to the clinician and patient. Non-operative means to palliate biliary obstruction and pain have emerged and experience with laparoscopic bypass techniques for duodenal and or biliary obstruction is increasing. The choice between these procedures should be weighed against the cost to the patient in terms of morbidity, mortality, and the quality of the patient's remaining life.

Last updated Jan 2/07