A Randomized Phase III Study of Gemcitabine in Combination with Radiation Therapy Versus Gemcitabine Alone in Patients with Localized, Unresectable Pancreatic Cancer: E4201

Reviewer: Eric Shinohara MD, MSCI
Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 3, 2008

Presenter: Loehrer P.J.
Presenter's Affiliation: Indiana University Melvin and Bren Simon Cancer Center
Type of Session: Scientific

  • There are approximately 232,000 new cases of pancreatic cancer diagnosed annually worldwide. It is the fourth leading cause of cancer death in the United States and in 2007, over 37,000 cases were diagnosed, with over 33,000 deaths documented [Jemal, CA Cancer J Clin, 2007].
  • Surgery is the only curative treatment; however, most patients are unresectable at the time of diagnosis (~80%), and even in those who undergo complete resection, there are high rates of local recurrence (50-80%).
  • Prior studies which compared 5-FU plus radiation versus 5-FU alone have suggested that radiation adds toxicity and does not improve outcomes (Klaassen D.J. et al., JCO, 1985).
  • A prior phase III study (Burris H.A., et al., JCO 1997) demonstrated that single agent gemcitabine significantly improved survival and clinical benefit compared with fluorouracil in the treatment of advanced pancreatic cancer.
  • Gemcitabine has been shown to be a potent radiosensitizer in vitro, and the combination of gemcitabine and radiation may be better than fluropyrimidine-based chemotherapy with radiation.
  • Gemcitabine is approved for use in treatment of pancreatic adenocarcinoma, but has been associated with severe toxicity when used as a radiosensitizer in phase I and retrospective trials [Crane CH, IJROBP, 2002]. Conversely, a recent phase II trial has suggested that full dose gemcitabine given with concurrent radiation is well tolerated and has activity (Small W. et al, JCO 2008).
  • The present study was designed to further examine the role of concurrent therapy with either radiation and gemcitabine followed by adjuvant gemcitabine, or gemcitabine alone, in patients with unresectable, non-metastatic pancreatic cancer.
Materials and Methods
  • Inclusion and exclusion criteria included:
o Patients had to have histologically confirmed, locally unresectable pancreatic adenocarcinoma. o Patients could not have metastatic disease. o Patients had to have a performance status of <2. o Patients could not have been treated with prior chemotherapy or radiation.
  • Patients were stratified by:
o Performance status: 0 versus 1 o Weight loss in the prior 6 months: greater than 10% versus 10% or less
  • Eligible patients were then randomized to either gemcitabine alone (arm A) or concurrent gemcitabine with radiation followed by adjuvant gemcitabine (arm B).
o Arm A: Gemcitabine: 1000 mg/m2 weekly x3 every 4 weeks for seven cycles. o Arm B: Gemcitabine: 600 mg/m2 weekly x6 concurrently with radiation followed by five cycles of 1000 mg/m2 weekly x3 every 4 weeks. Radiation: patients were treated with 3D conformal radiation. The initial fields were treated in 180 cGy fractions to 3960 cGy followed by a boost to 5040 cGy. All radiation plans were centrally reviewed.
  • The planned sample size for this study was 316 patients, based on a power of 88% to detect a 50% improvement in median OS (from 8 to 12 months) using a one-sided long rank test with significance set at 0.025.
  • Quality-of-life (QOL) was assessed prospectively using the functional assessment of cancer therapy-hepatobiliary (FACT-Hep) measure of QOL.
  • The primary endpoint was overall survival (OS).
  • Secondary endpoints were response rate (RR), QOL, and progression-free survival (PFS).
  • 74 patients were enrolled onto the present study from April 2003 to December 2005. The study was halted early due to slow patient accrual. Ultimately, 38 patients were randomized to arm A, of which 37 where eligible; 36 patients were randomized to arm B, of which 34 were eligible.
  • The three patients who were deemed ineligible were found to have metastatic disease. Five patients who were enrolled in the study did not receive assigned therapy for the following reasons:
    • two patients had metastic disease on CT scan for radiation planning
    • two patients had significant clinical deterioration
    • one patient withdrew consent
    • all five of these patients were included in survival analyses, but not in toxicity analysis
  • The demographics between the two arms were well balanced.
  • Outcomes between the two treatment arms are as follows:

Arm A

Arm B


Grade IV toxicity




Objective response

2.7% (95% CI=0.09%-14.1%)

8.8% (95% CI= 1.9%-23.7%)


6.1 months

6.3 months


Median OS

9.2 months (95% CI= 7.8-11.4 months)

11.0 months (95% CI= 8.4-15.5 months)


  • HR for OS was 0.574
  • OS over time:

6 months

12 months

18 months

24 months

Arm A





Arm B





  • Response:

Arm A

Arm B




Stable Disease









  • Most patients were inevaluable due to clinical deterioration or because they had scans obtained at times which were outside the protocol rules.
  • Relapse:
o Time to relapse: 6.7 months in arm A versus 6.0 months in arm B o Local recurrence: 41% in arm A versus 23% in arm B o Distant recurrence: 14% in arm A versus 23% in arm B o Both local and distant recurrence: 5% in arm A versus 9% in arm B o No documentation: 41% in arm A versus 44% in arm B
  • It is unclear why there was no difference in PFS, while there was a difference in OS. The authors feel that this may have occurred due to the difficulty in assessing progression of disease radiologically (see above relapse and response data, large number of unevaluable patients). Clinical symptoms of progression are common in patients with pancreatic cancer, which may have also made assessment of true progression difficult.
  • Toxicity
o Overall toxicity was 82% in arm A and 93% in arm B; this was found to be non-significant. o Grade IV toxicity was predominantly GI and hematologic. o All 71 patients had died at last analysis. There were two deaths on treatment, one in each arm.
  • The cause of the slow accrual in this study was multifactorial. There were several competing trials being run simultaneously. Some physicians did not want to enroll patients because the dosages of gemcitabine were not equal between the two arms. Some physicians felt that giving radiation was unethical, while others felt that holding radiation was unethical.
Author's Conclusions
  • The present study demonstrated a modest increase in survival with the addition of radiation to gemcitabine compared with gemcitabine alone, but this was a significant difference.
  • This was a small study with poor accrual.
  • Toxicity was common but manageable. Specific toxicity data is to be reported at a later time.
Clinical/Scientific Implications
  • The present study is an interesting one showing that the combined use of gemcitabine and radiation significantly improved OS compared with gemcitabine alone. While the improvements are modest, the benefits seen in this study are comparable to improvements seen in many recent trials of biological agents.
  • There are some potential difficulties in the analysis of this study. It is a small, multi-institutional trial, and as noted by the authors, some physicians believed in radiation therapy while others did not. This could result in an institutional bias regarding which patients were and were not enrolled onto this study, thereby affecting results. The lack of an improvement in PFS is puzzling, and it appears from the authors' explanations that current radiological evaluation may not be an adequate way to assess response. Perhaps PET scans could help, both by eliminating patients with occult metastatic disease discovered at the beginning of the trial, as well as by detecting recurrence or progression during follow up.
  • Overall, it appears that radiation prolonged survival, but ultimately most patients still die, based on the overall survival over time presented by the authors. Weight loss, QOL and toxicity were recorded but not reported at this time. Aggressive supportive care is essential, and we await reports of toxicity. With aggressive supportive care, it may be possible to escalate the dose of chemotherapy or radiation used in order to improve outcomes. It appears that although local relapse was reduced with radiation in the present study, distant relapse increased (though with the large number of patients with missing data, it is difficult to make any definitive conclusions). This is not likely due to the improved survival, but rather the lower gemcitabine dose given concurrently. If full-dose gemcitabine could be safely delivered concurrently with radiation, this might further improve outcomes. Proton therapy delivers less dose to normal tissues, and may allow radiation and chemotherapy doses to be escalated.
  • Unfortunately, after nearly three decades of research, the role of radiation in pancreatic cancer is still not clear at present. These results do appear to affirm that radiation and gemcitabine are more effective than gemcitabine alone, but due to the above limitations, a larger study is still needed.

View Patient Summary


Lucky Number Seven
by Rodney Warner, JD
April 3, 2010

Stay informed with the latest information from OncoLink!   Subscribe to OncoLink eNews
View our newsletter archives