We reviewed our 25-year experience on cPD for patients of locally advanced colon cancer. A nationwide survey in the Netherlands reported that the most common indication of cPD is first, locally advanced pancreatic head cancer (23) that directly invades colon or mesocolon, followed by locally invaded colon cancer at the duodenum and/or pancreatic head. The procedure of cPD, is rarely performed in gastrointestinal surgery. This is due to its complexity, difficulty and high risks. In certain acute situations, cPD is the efficacious path forward. According to surgeons at the Memorial Sloan Kettering Cancer Center, New York, the most complex surgical procedures in cancer surgery are esophagectomy, hepatectomy, pancreatectomy and total pelvic exenteration (45).
The above procedures, are recommended to be best performed at well-experienced medical centers. From our present study, we found similar postoperative courses in both cPD and PD. Therefore, a successful emergent cPD may be similarly done like a successful emergent PD. Emergent PD is associated with high postoperative morbidity and mortality, reaching 30% and 50% (30-33).
Gulle et al. (29) reported their operation on 10 patients with emergent PD to treat complex pancreaticoduodenal trauma with zero death. However, their complication rate is high (80%). All their patients were relatively young and healthy without challenging co-morbidities. Emergent PD for non-trauma cases has risks higher than trauma cases, because of their often unrecognized preoperative poor conditions, and co-existed inflammation or organ dysfunction that lead to failed emergent PD (30, 31, 33).
Managing postoperative complications after emergent cPD is also an important issue after emergent PD. Performing cPD is itself a challenging, involving high levels of skills and long operation times. It requires meticulous and experienced care during the evaluation of perioperative period to minimize complications and deaths. Thus, the diagnosis and evaluation of preoperative general conditions of these patients should be well-surveyed to prevent occurrences of potential postoperative adverse events. Therefore, proper preoperative selection of patients is critical for the success of cPD. Despite high complication rates of emergent cPD in our patients, their rates of early and long term survivals rate appear acceptable.
Abdominal CT scan plays an important role in the preoperative diagnosis of such advanced cancer. However, in patients presented with acute abdomen, CT scan facilitates the visualization of tumors due to marked intestinal edema or marked intraperitoneal free air. When a right colon cancer is loosely adhered to the duodenal wall, it may be regarded as a duodenal invasion. Such cases were observed in our current study. Likewise, when a relatively small right colon cancer directly invades the pancreatic head, even with a small invaded area, the condition may be regarded as “no invasion”. Therefore, definite diagnosis of colon cancer with duodenum or pancreas invasion could be comfirmed only after “exploratory laparotomy”. Moreover, damage-controlled procedures can be given to those patients not fulfilling our criteria of emergent cPD.
Patients with advanced age, poorly controlled comorbidities, unstable vital signs, or obesity, are typically at risk of PD. They were therefore excluded from our emergent cPD when treating bowel obstruction or perforation. Moreover, if the perforation time is long (> 6 hours), severe intra-abdominal contamination could lead to edematous and fragile conditions. Long periods of generalized peritonitis may destroy sutures in cPD early postoperative period. These patients are therefore not recommended for emergent cPD.
Acute massive bleeding from gastrointestinal malignancy is very rare but the sequel is grave. Once occurred, emergent resection to stop the bleeding is most likely the only life-saving option. The aforementioned patient selection criteria and managements for obstruction and perforation are not be applicable in bleeding cases. Trans-arterial embolization may be temporarily helpful for hemostasis (29, 46). However, due to abundant vascular collaterals in pancreaticoduodenal regions, total hemostasis is difficult. The resection of a bleeding tumor resection is still mandatory after embolic control of hemostasis.
Tsai et al. reported that in emergent PD, intraperitoneal infections have outcomes worse than bleeding (33). In our series, we recommended 2 stage pancreaticojejunostomy after cPD.
Pancreaticojejunostomy has been considered as the “Achilles tendon” of PD. For a successful cPD, the management of the anastomosis is also crucial. During the early period of our study [before March, 1996], we had a case of grade C catastrophic pancreatic leakage with bleeding (disruption of pancreaticojejunostomy with massive internal bleeding). The patient fortunately survived after our timely and appropriate managements.
Intraoperative management for cPD adverse events, related or not related to techniques, is of great importance to reduce chances of operative mortality.
Staged pancreaticojejunostomy was routinely used for all PD after cases when the pancreatic duct size is small (<2mm), pancreatic parenchyma is soft or associated with large vessel resection or controlled troublesome comorbidity.
Staged pancreaticojejunostomy was firstly proposed by Japanese surgeons, Miyagawa and Makuuchi in 1994 (39). They covered the common hepatic each proper hepatic artery, and gastroduodenal artery stump using a sheet of pedicled greater omentum or liver falciform ligament (35, 36). A thin plastic tube was then inserted into the main pancreatic duct to totally exteriorize the pancreatic juice. Pancreatic juice was fed into intestine lumen through another tube jejunostomy, with the seromusculature sutured with the posterior wall of the pancreatic stump. The anastomosis was performed three months later by inserting the aforementioned plastic tube into the neighboring jejunal lumen.
The pedicled falciform ligament of the liver, or greater omentum is capable of covering the transected stump of gastroduodenal artery [the most common site of postpancreatectomy hemorrhage after PD](39). This vessel can be protected from erosion by the leaked out pancreatic juice. Actually, no catastrophic complication had occurred or even minor leakage occurred after staged pancreaticojejunostomy. This quarantined further the safety of cPD.
The experience of treating acute necrotizing pancreatitis is also helpful for treating pancreatic leakage after PD (46). An appropriate and timely management of complications after PD could improve healing and prevent operative death. The procedure of pancreatectomy for disruptured pancreatic anastomosis (28) has a high death rate (28, 40). It should therefore be avoided [43]. Even with the development of chronic pancreatic fistula, treatment by fistulojejunostomy can be effective without undesirable sequelae (44).
Delayed gastric empty is also a problematic adverse event of both PD and cPD. The event is likely related to the destruction of upper abdominal autonomic nerve plexus during the lymphadenectomy. The condition often requires prolonged hospitalization, long-term nasogastric decompression and total parenteral nutritional support. These management protocols could cause other systemic problems, like catheter sepsis, electrolyte imbalance, trace elements deficiencies, aspiration pneumonia and hepatic dysfunction. Some of these can be fatal. To avoid such severe complications, efforts to preserve the upper abdominal vagus nerve and sympathetic nerve plexus, could minimize delayed gastric emptying. Upper abdominal lymphadenectomy, which is typically carried out for periampullary or pancreatic cancer, is not mandatory for colon cancer patients.
Total removal of locally advanced colon cancer is essential to obtain a microscopic R0 resection. In our experience, R1 resection for colon cancer [cancer cells microscopically after operation] was actually palliative as no patient from the R1 resection had survived for more than 3 years.
The 5-years OS for locally advanced colorectal cancer is 51%. The reported 5-year OS rate after cPD for advanced colon cancer is 50 to 60% (26). Our non-e group patients showed prognosis comparable to the literatures (4-11). Because of the high incidence of lymph node metastasis, prognosis of the e-group patients is often poor. Nevertheless, both groups had similar OS. Histological TNM staging, lymph involvement and cancer cell differentiation are prognostic factors (5-12, 14, 15, 17-21, 23-25). The cancer conditions of our patients receiving cPD were similar to literatures on colon cancer patients. Development of new target agents or chemotherapeutic drugs is helpful to prolong survival.
Several limitations of the current study are as follows.
First, our comparison of two groups was not flawless. Because, patient selection or the choice of emergency procedure was based on perceived co-morbidities of patients, by the operating team, and that could be quite arbitrary. Nevertheless, no differences were found in other variables.
Second, the current study is a longitudinal observational cohort study. It was not a randomized-controlled trial because of the small sample size. Colon cancer that involves duodenum and /or pancreas is a distinctively unique presentation. During the 25-year duration of our study, diagnostic tools, surgical techniques, operative equipment, and peri-preoperative assessments have advanced markedly. Thus, our initial case-selection criteria in the e-group could have been too conservative. For example, the patient age can be extended as the life expectancy of the general population has increased by 5 years over the course of this study period (47). It is reasonable to assume that the safety range of emergent cPD could have also been extended.
Third, treatment strategies for locally advanced colon cancer were decided by our experienced colorectal and hepato-pancreatico-biliary surgeons as well as oncologists. Over the study period, some of the involved members of this research study had retired or shifted on to other projects. Although a senior surgeon (CCW) had led the treatment strategy of individual patients in a constant manner, there is an inherent discrepancy in the continuity of management for this complex disease. For example, the adjuvant therapies after elective or emergent cPD may have changed across different postoperative courses.
In the current study, we found no patients experiencing ileosolostomy leakage. Most postoperative complications were related to ileosotomy pancreatic or colon leakage, which could be managed by diverting proximal presence of ileostomy necrosis. Nevertheless, abscesses could be treated by percutaneous drainage. These complications should be diagnosed early and promptly treated.
Fourth, although staged pancreaticojejenectomy may improve safety of cPD of patients, additional admission and operations are needed for a complete pancreaticojejunostomy. Given new safe-guarded techniques on dealing with pancreatico anastomosis, the safety of cPD in single operation could be developed, reducing both hospital costs and anesthetic risks.
Despite high complication rates, our reviewed experiences supported that emergent cPD is a feasible procedure on the highly selected patients with locally advanced colon cancer presenting acute abdomen. The long-term outcomes after emergent and non-emergent cPD are comparable.