postchemoradiation). Although the study showed no overall difference in aggregate survival, when

pancreatic head lesions only were considered (eliminating study results from resected lesions in the

pancreatic body or tail), both median survival (16.7 vs. 18.8 months) and overall survival at 3 years

(21% vs. 31%) favored the gemcitabine arm (p = 0.047). The study concluded that the addition of

adjuvant gemcitabine to postoperative 5-FU chemoradiation was superior to the addition of 5-FU.

Figure 55-8. Survival of patients with pancreaticoduodenectomy based on tumor size (A), lymph node status (B), margin status

(C), histologic grade (D), and historical context (E). From Winter JM. Cameron JL. Campbell KA, et al. 1423

pancreaticoduodenectomies for pancreatic cancer: a single-institution experience. J Gastrointest Surg 2006; 10:1199–210, with

permission.

TREATMENT

Table 55-14 Randomized Prospective Trials of Adjuvant Therapy for Pancreatic

Cancer

1418

The CONKO-1 trial,42 conducted in Germany and Austria, represented a randomization of 368 patients

following R0 or R1 resection to either observation or an experimental arm of gemcitabine. After a

median follow-up time of 53 months, the median disease-free survival was 13.9 months in the

gemcitabine arm versus 6.9 months in the observation arm (p <0.001). There was no difference in

overall survival for the gemcitabine arm versus the control group – median survival was 22 versus 20

months. Although survival was not different, the authors concluded that postoperative gemcitabine

significantly delayed the development of recurrent disease after complete resection of pancreatic cancer

compared with observation alone and, thus, was supported as adjuvant therapy in resectable pancreatic

cancer.

At present, many centers are utilizing preoperative neoadjuvant chemoradiation for the treatment of

pancreatic cancer (Table 55-15). Neoadjuvant therapy offers several potential benefits including: (1)

delivery of treatment to well-oxygenated tissue which enhances efficacy of chemoradiation, (2)

downstaging can enhance ability to achieve a negative-margin resection and thereby reduce local

recurrence, and (3) avoidance of surgery in patients with rapidly progressive disease. Neoadjuvant

therapy can be completed without increasing the subsequent morbidity and mortality of surgical

resection. The group from the M.D. Anderson Cancer Center reported on the multimodality treatment of

142 consecutive patients with localized adenocarcinoma of the pancreatic head.43 A subset of 41 patients

treated by preoperative chemoradiation and pancreatoduodenectomy were compared with 19 patients

receiving pancreatoduodenectomy and postoperative adjuvant chemoradiation. Surgery was not delayed

for any patient who received preoperative chemoradiation because of chemoradiation toxicity, but 24%

of the eligible patients did not receive their intended postoperative chemoradiation because of delayed

recovery following PD. The patients treated with rapid fractionation were reported to have a

significantly shorter duration of treatment (median, 62.5 days) than patients who received

postoperative chemoradiation (median, 98.5 days). In early follow-up, no patient who received

preoperative chemoradiation experienced a local recurrence, and peritoneal recurrence developed in

only 10% of these patients. Local or regional recurrence developed in 21% of patients who received

postoperative chemoradiation. The overall survival curves were similar for both cohorts.

Wolff et al.46 examined 86 patients treated with weekly gemcitabine at a dose of 400 mg/m2 and 30

Gy of radiation. Sixty-one patients ultimately underwent resection (71%). The median survival in the

resected patients was 36 months which is significantly longer than those seen in regimens using 5-FU or

paclitaxel as the radiation sensitizer. Analysis of the specimens revealed two pathologic complete

responses and more than 50% nonviable tumor cells in 36 (59%). A gemcitabine-based regimen was also

used in a multi-institutional study of 20 patients reported by Talamonti et al.47 This group used full-dose

gemcitabine and limited field radiation to 36 Gy (2.4 Gy/fraction). The authors described 14 patients as

resectable and six as borderline resectable. Ultimately, all patients were explored and 17 resected

(85%), again representing a very high rate of resectability. A single pathologic complete response was

observed and, in 24% of tumors, greater than 90% of the tumor cells were felt to be nonviable. Also

notable was the low incidence (6%) of margin positivity in this trial. The median survival in the

resected patients was 26 months. Based on the results of these initial trials, gemcitabine-based

neoadjuvant regimens remain of considerable interest.

TREATMENT

Table 55-15 Selected Neoadjuvant Trials for Potentially Resectable Pancreatic

Cancer

1419

PALLIATION

7 Unfortunately, it has been the experience nationwide that only a minority of patients with carcinoma

of the pancreas can undergo resection for possible cure at the time diagnosis is made. Therefore, the

optimal palliation of symptoms to maximize quality of life is of primary importance in most patients

with pancreatic cancer. Both operative and nonoperative options are available for the palliation of

pancreatic cancer.

Jaundice

Obstructive jaundice is present in most patients who have pancreatic cancer. If left untreated, it can

result in progressive liver dysfunction, hepatic failure, and early death. In addition, the pruritus

associated with obstructive jaundice can be debilitating and usually does not respond to medication.

When patients undergo exploration for possible cure and are found to have unresectable disease, a

biliary bypass should be performed.

Traditionally, surgeons have performed either choledochojejunostomy or cholecystojejunostomy for

the relief of malignant biliary obstruction. Both procedures are effective in relieving jaundice, but it

appears that the rate of recurrent jaundice after cholecystojejunostomy is approximately 10%.

Therefore, our preference for the palliation of obstructive jaundice is a hepaticojejunostomy or

choledochojejunostomy reconstructed with a Roux-en-Y limb of jejunum. The surgical palliation of

jaundice can be accomplished safely, with a mortality rate of less than 3% and an overall morbidity rate

of 30% to 40%.50 In recent years, nonoperative palliation has become available as an option for

managing patients who are deemed unresectable by preop staging. Plastic or metal stents can be placed

across the biliary obstruction by either an endoscopic or a percutaneous technique. For pancreatic

cancer, the endoscopic approach is usually preferred. The overall morbidity rate for endoscopic stenting

ranges up to 35%, but the rate of major procedure-related morbidity is less than 10%. Early

complications include cholangitis, pancreatitis, and bile duct or duodenal perforation. The major late

complications of stent placement are cholecystitis, duodenal perforation, and stent migration. Stent

occlusion can result in episodes of cholangitis and recurrent jaundice. For most patients, an exchange of

stents is required every 3 to 6 months. The newer metal stents appear to remain patent for longer

periods.

Nonoperative palliation appears to be associated with lower complication rates, lower procedurerelated mortality rates, and shorter initial periods of hospitalization in comparison with surgical

palliation. However, the rate of recurrent jaundice is higher. No advantage with respect to long-term

survival has been noted for either approach. Therefore, nonoperative palliation should be offered to

patients with advanced disease or poor performance status. Surgical palliation should be considered for

patients with an anticipated life expectancy of at least 6 months.

Duodenal Obstruction

At the time that pancreatic cancer is diagnosed, approximately one-third of patients have symptoms of

nausea or vomiting. Although true mechanical obstruction of the duodenum seen by radiologic or

endoscopic examination is much less frequent, duodenal obstruction develops in almost 20% of patients

before they die as the disease progresses.51 Duodenal obstruction can be caused in the C-loop by cancers

of the head or at the ligament of Trietz by cancers of the body and tail. In patients with evidence of

duodenal obstruction or impending obstruction, a gastrojejunostomy is indicated for palliation. This is

typically performed as a retrocolic, isoperistaltic loop gastrojejunostomy with a loop of jejunum 20 to

30 cm distal to the ligament of Trietz.

In patients with unresectable pancreatic cancer who do not have symptoms of gastric outlet

obstruction, whether or not to perform a prophylactic gastric bypass at the time of biliary bypass is a

1420