Adjuvant Therapy for Pancreatic Cancer—GITSG Results Confirmed


Synopsis: Patients undergoing pancreaticoduodenectomy for cancer of the pancreas who received postoperative adjuvant radiotherapy plus 5FU-based chemotherapy experienced an improvement in median and two-year survival compared to patients who elected not to have adjuvant therapy.

Source: Yeo CJ, et al. Ann Surg 1997;225:621-636.

Cancer of the pancreas is the fifth leading cause of cancer death in the United States. A curative surgical procedure, which is possible in only 20% of patients, results in a five-year survival rate of 15-20%. Thus, 5% or less of all patients will be "long-term" survivors. Effective adjuvant therapy following potentially curative resection could improve this outcome. The GITSG has reported a positive trial of postoperative radiation therapy and 5-FU, but for a variety of reasons this approach has not been generally adopted. Between 1991 and 1995, all patients undergoing elective pancreaticoduodenectomy at Johns Hopkins Hospital for cancers of the pancreatic head, neck, and uncinate process were followed prospectively after being offered one of two different postoperative treatment options. Because of the authors’ belief in the efficacy of adjuvant therapy, all patients were offered some adjuvant therapy—either a regimen modeled after the GITSG study that they called standard therapy, or a more aggressive regimen. Patients who were unwilling to receive either treatment, regardless of the reason, were followed in a separate group.

One hundred seventy-four patients underwent surgery. The goal was to avoid an extensive retroperitoneal dissection, spare the pylorus, and perform only a partial pancreatic resection if possible. "Standard therapy" consisted of radiation therapy (4000-4500 cGy) to the tumor bed and paraaortic nodes with 5-FU 500 mg/m2 on days 1-3, 29-31, and then weekly bolus 5-FU for four months. This differs from the GITSG study primarily in the length of therapy; weekly treatment with 5-FU was continued for one year in the GITSG study. "Intensive" therapy consisted of radiation therapy to a higher total dose (5040-5760 cGy)—addition of hepatic irradiation (2340-2700 cGy) with 5-FU given by continuous infusion at 200 mg/m2 in combination with oral leucovorin 5 mg/m2 five days out of every seven for four months.

Of the 174 patients enrolled, pylorus-sparing procedures were possible in 84%, and a partial pancreatectomy was performed in 94% of patients. The resection margin was positive in 29% of patients, and in 75% the lymph nodes were positive. In-house mortality was 0.6%. Standard therapy was selected by 99 patients—21 chose intensive therapy, and 53 refused treatment.

With a median follow-up of 19 months, the median survival for all patients was 19 months, and the two-year and four-year survival rates were 36% and 24%, respectively. Univariate analysis revealed that the following factors had no impact on survival: age, gender, race, status of margin after resection, and lymph node status. However, operative blood loss of greater than 700 cc, blood transfusion, and tumor size greater than 3 cm all had a negative impact on survival. The median and two-year survival rates were 13.5 months and 30%, respectively, for patients observed after surgery vs. 19.5 months and 39% for either type of adjuvant therapy (P = 0.003). Patients treated with standard adjuvant therapy had a median survival rate of 21 months and two-year survival rate of 44%, compared to 17.5 months and 22% for patients receiving intensive therapy (P = NS).

Although the study was prospective, it was not randomized and is therefore subject to bias. For example, patients who received adjuvant therapy were significantly younger, had fewer complications and shorter lengths of stay in the hospital than did patients who elected not to undergo adjuvant therapy. However, the authors" analysis of this issue suggest that these factors did not affect survival. Following multivariate analysis, adjuvant therapy was found to be the strongest predictor for survival.


The poor prognosis of patients with cancer of the pancreas has led to an extremely nihilistic approach to their management. Some patients are not even evaluated for surgical resection, and those who are, often do not receive any adjuvant therapy following their potentially "curative" operation. Although nihilism in regard to the treatment of pancreatic cancer is certainly justified, if we approach every person as if their disease is incurable, this will become a self-fulfilling hypothesis. This study "warts and all" offers additional evidence in support of the findings, first described by the GITSG, that adjuvant 5-fluorouracil and standard radiation therapy benefit patients who have undergone potentially curative resections of the pancreas. The initial GITSG trial1 did not include patients with positive margins following surgery, included patients with cancers of the body and tail and used a more prolonged chemotherapy regimen. Despite the small size of the study, Yeo et al were able to detect a statistically significant survival advantage. A follow-up study from GITSG, in which all subsequent patients were treated with adjuvant therapy in a similar fashion, confirmed their initial results.2

The median survival for patients reported in this Johns Hopkins study are comparable to those reported by GITSG for their treated group. This trial finds similar benefits for patients despite positive margins, and with a duration of therapy of four months as opposed to more than a year in the GITSG trial. Although the trial was not randomized and there are major problems with having patients "choose" therapies in studies of this nature, we feel that the data, when taken together, support the use of combined modality therapy for patients with resectable cancers of the pancreas. Despite the contrary claims of internists and gastroenterologists, who in our opinion do much to heighten the nihilistic approach to this disease, fear of toxicity is not a legitimate reason to withhold therapy; less than 10% of patients required hospitalization or had to have therapy permanently discontinued because of intolerable side effects. In addition, 5FU is one of the least expensive chemotherapy drugs, so cost is not a huge consideration.

The rationale for testing a more aggressive adjuvant therapy protocol for these patients was based on the enhanced activity of leucovorin-modulated 5-fluorouracil in a variety of gastrointestinal tumors and the high frequency of hepatic metastases. Similarly, previous RTOG trials found a reduction of hepatic metastases for patients treated with prophylactic liver irradiation. Unfortunately, the use of these two modifications does not necessarily translate into any improvement in local control or survival.3 Significant morbidity is found for these patients as well. This study did not find any advantage for the more intensive adjuvant therapy regimen, and although it would be difficult to argue that this study adequately tested this approach, we agree with the authors’ decision to recommend standard postoperative radiation therapy and 5FU for patients undergoing curative pancreatic resection.

The introduction of gemcitabine has raised expectations regarding its potential value in the adjuvant setting. Thus far, early trials of gemcitabine in combination with radiation therapy have noted severe toxicity, and a safe dose and schedule for efficacy testing have not yet been established.


1. Kalser MH, Ellenberg SS. Arch Surg 1985;120:889-903.

2. Gastrointestinal Tumor Study Group. Cancer 1987;59:2006-2010.

3. Komaki R, et al. Cancer 1992;69:2807-2812.