Trial Information

A Multi-Institutional Phase II Study of Neoadjuvant Gemcitabine and Oxaliplatin With Radiation Therapy in Patients With Pancreatic Cancer

Pancreatic cancer remains incurable in the great majority of patients afflicted with the
disease 1. Most patients present with tumor that is localized but unresectable. Median
survival in this group of patients averages 6-10 months, with some suggestion of improvement
for patients treated with combinations of radiation and chemotherapy 2, 3. Surgical
treatment, when possible, is provided with curative intent. In those who have undergone
surgery, however, median survival is only 13-20 months. Following resection, patients may
benefit from the addition of adjuvant radiation and/or chemotherapy 4-6. Even in this most
favorable group of patients, the 5 year survival is less than 30% in single institution
series 7, 8. The vast majority of patients treated with surgery, with or without adjuvant
therapy, fail with hepatic metastases. It is clear that a local treatment modality will
have a limited impact on outcome in these patients. More effective treatment for pancreatic
cancer must simultaneously address both local and distant sites of failure.

In advanced pancreatic cancer, gemcitabine is the standard for systemic therapy 9. Emerging
data indicate that gemcitabine combinations, with additional cytotoxics or targeted agents,
increase response rates, time to progression and overall survival in patients with advanced
disease 10, 11. While the impact on survival is limited in the advanced disease setting,
the potential for more active combination therapy to control or eradicate micro-metastatic
disease as pre- or post-operative adjuvant treatment might be expected, as has been observed
in other diseases 12, 13. Since a large majority of patients with pancreatic cancer relapse
with a component of distant disease, investigation and development of adjuvant systemic
combination treatments are warranted.

In resectable pancreatic cancer, surgery is most often the initial treatment. The adequacy
of that resection is questionable, however, in that a significant minority of patients have
positive margins and some patients have incomplete resection (R2 resections) 14, 15. Median
survivals in patients with R2 resections or positive margins are no different than that
observed with non-operative therapy. A reasonable strategy to address the limitations of
surgery is neoadjuvant therapy 16. Potential advantages of neoadjuvant therapy also include
an improved tolerance of combined modality treatment preoperatively, and a greater
proportion of patients receiving all components of multimodality therapy completed over a
shorter time interval. Recognizing a role for multimodality therapy in resectable
pancreatic cancer, as many as 30 % of initially resected patients do not receive
post-operative adjuvant therapy due to inadequate recovery from surgery or refusal.
Treatment delays following surgery may also impact efficacy of adjuvant treatment. In the
recently reported ESPAC adjuvant trial, the median time to initiation of post operative
chemotherapy was 46 days and for combined modality treatment 61 days 5. Neoadjuvant
treatment provides a more timely systemic therapy. Furthermore, patients presenting with
borderline resectable lesions may become operable because of neoadjuvant treatment. We have
successfully and safely resected patients following gemcitabine based chemoradiotherapy 17,
18. Finally, patients that progress during neoadjuvant therapy likely have a biologically
aggressive disease and are unlikely to benefit from surgery. Patients who develop
metastases during pre-operative treatment avoid a major operation which provides no benefit.

There are, admittedly, barriers to neoadjuvant treatment. These include biliary obstruction
and a requirement for decompression of the biliary system, as well as a need for a
pathologically confirmed diagnosis prior to treatment. Practical barriers to neoadjuvant
therapy in this disease also include emotion, a desire by the patient and caregivers to have
the tumor removed "before it's too late." Similar circumstances and obstacles have been
overcome in other malignancies in which neoadjuvant treatment has become established.

The prognosis for resected pancreatic cancer remains poor. Clinical-pathologic parameters
which correlate with survival are limited to lymph node and margin status. Additional
potential prognostic parameters, including pathologic response to neoadjuvant treatment and
functional imaging using positron emission tomography (PET) scans and standard uptake value
(SUV) estimation in pancreatic cancer, are poorly characterized.

Conventional morphologic imaging modalities after neoadjuvant treatment of pancreatic cancer
such a CT and MRI often have difficulty distinguishing viable tumor from inflammatory
changes, necrotic debris or scar tissue. Functional imaging modalities such as
2-fluoro-2-deoxy-D-glucose (FDG) PET which rely on the metabolic activity in tumor may be
able to discern these differences 19. FDG-PET imaging has been studied as a predictive tool
in evaluating response to preoperative therapy in solid tumors 19-22. Decreases in the SUVs
have allowed a semi-quantified evaluation that correlates with tumor histopathologic
response and survival 21, 22. Additionally, preoperative FDG-PET imaging after neoadjuvant
therapy has been useful in detecting unsuspected metastases avoiding radical surgical
intervention in the patient with metastatic disease 22. These data in pancreatic cancer
have been limited thus far 23. We propose to incorporate pre- and post-therapy FDG-PET
imaging to further evaluate its utility in determining resectability, pathologic response to
therapy and its relationship to survival.

In 1997, a phase 1 gemcitabine based chemotherapy and radiation therapy protocol was
initiated at the University of Michigan that differed from other gemcitabine based combined
modality regimens in two important ways. First, a standard dose or full dose of gemcitabine
was used (as opposed to a "radiosensitizing" dose), considering the clinical benefit
associated with standard dose gemcitabine as a systemic agent 9. The use of a standard dose
was also consistent with laboratory data that demonstrate maximum radiosensitization when
cytotoxic concentrations of drug are used 24. However, considering prior clinical
experience indicating that substantial normal tissue radiosensitization could occur, use of
full dose gemcitabine required reduction and investigation of the radiation dose. This
approach differed from the more traditional approach of chemotherapy dose escalation with a
fixed dose of radiation. Secondly, a decision was made to irradiate the primary tumor
alone, without the inclusion of normal appearing regional lymph nodes. This was based on
the assumption that the majority of the benefit from radiation would result from control of
the primary tumor, rather than control of subclinical disease in these nodes. Regional
nodes could potentially be controlled by standard doses of systemic therapy, as would
distant sites of microscopic disease. Concerns for excess normal tissue toxicity that might
occur with the use of more conventional treatment volumes contributed to this decision. This
strategy required more accurate identification of the primary tumor and 3D radiation
treatment planning.

In the ensuing phase I trial, thirty-seven patients with unresectable (34) or incompletely
resected pancreatic cancer (3) were treated 25. Suspected or confirmed metastatic disease
was identified at the time of enrollment in fourteen. Gemcitabine was administered as a 30
minute intravenous infusion at a dose of 1000 mg/m2 on day 1, 8 and 15 of a 28 day cycle.
Radiation therapy was initiated on day 1 and directed at the primary tumor alone as
indicated above. The starting radiation dose was 24 Gy in 1.6 Gy fractions. Escalation was
achieved by increasing the fraction size in 0.2 Gy increments, keeping the duration of
radiation constant at 3 weeks. A second cycle of gemcitabine alone was given following a
one week rest. Two patients experienced dose limiting toxicity (DLT) at the final dose
level (42 Gy), consisting of grade 4 vomiting and gastric/duodenal ulceration. Two
additional patients at this dose level experienced late gastrointestinal toxicity which
required surgical repair. Hematological toxicity was not substantively different than with
gemcitabine alone. Median survival for the study population was 11.6 months. Isolated
local or regional progression, a possible consequence of lower dose and limited field
radiation, was observed in one patient only.

Based on these results, we conducted a multi-institutional phase II trial using full dose
gemcitabine with 36 Gy in 2.4 Gy fractions for patients with resectable (20) and
unresectable (21) pancreatic cancer17, 26. Protocol therapy included 3 cycles of
gemcitabine with radiation therapy during the second cycle. Patients who remained or became
resectable based on CT scanning were surgically explored 4-6 weeks following the last
gemcitabine infusion. A total of 41 patients were enrolled at 6 institutions. Grade > 3
non-hematologic toxicity was seen in 19.5% of patients including five with grade 3
gastrointestinal toxicity, 2 with grade 3 fatigue and one unexplained death. A total of 20
patients underwent surgical exploration, with 17 patients resected. Those resected included
16 patients with clear margins, 1 pathologic complete response and 3 with only microscopic
foci of tumor remaining. The complication rate was 24% with average length of stay 13.5
days and there was no 30 day mortality. After a median follow-up of 12 months, ten of 17
patients were alive without recurrence17. The development of liver metastasis in some of
these patients following successful resection describes the need for a more active systemic
treatment than gemcitabine alone.

In an effort to increase local and systemic treatment effects of combined modality therapy
in pancreatic cancer, we have investigated combination chemotherapy with radiation therapy
27, 28. Our initial trial incorporated cisplatin with gemcitabine, based on activity of
this combination in advanced disease as well as therapeutic interactions with radiation
therapy 27. While activity was observed and we were able to give full doses of gemcitabine
and cisplatin (40 mg/m2 every other week), the overall burden of treatment was felt to be
too great for further development of this combination. Oxaliplatin, as compared to
cisplatin, has less gastrointestinal and constitutional toxicities, and may be better suited
for combined modality therapy in pancreatic cancer.

Oxaliplatin is a platinum salt with a 1,2-diaminocyclohexane ring in its structure.
Oxaliplatin has efficacy in many cell lines in vitro and has been shown to be active in
gastrointestinal malignancies. In a randomized phase III study, the combination of
gemcitabine and oxaliplatin (GEMOX) was superior to gemcitabine alone in patients with
advanced pancreatic cancer with regard to response rate (26.8% vs. 17.3%, P=.04),
progression-free survival (5.8 vs. 3.7, P=.04) and clinical benefit (38.2% vs. 26.9%, P=.03)
although the median overall survival difference was not statistically significant (9.0 vs.
7.1, P=.13) 10. Oxaliplatin appears to be synergistic with radiation therapy in vivo 29.
In clinical trials, oxaliplatin concurrent with radiation therapy has been reported to be
tolerable and efficacious in rectal and esophageal cancer 30, 31.

With these considerations, we began a phase I dose escalation trial in 2003 incorporating
oxaliplatin into our full dose gemcitabine and radiation therapy approach 28. In dose levels
1 to 4, oxaliplatin was given days 1 and 15 of a 28 day cycle, beginning at a 40 mg/m2 and
escalating to a targeted dose of 85 mg/m2. Gemcitabine was given as 1000mg/m2 IV over 30
minutes on days 1, 8, and 15 of a 28 day cycle. In dose levels 5 and 6, oxaliplatin was
continued at target dose 85 mg/m2, but gemcitabine 1000mg/m2 was given with infusion times
increased to 65 and 100 minutes, respectively, based on reports that fixed dose rate
infusion of gemcitabine may increase efficacy and toxicity 32. Radiation was delivered to
27 Gy in fifteen, 1.8 Gy fractions. Eligible patients were those with untreated resectable
or unresectable pancreatic cancer including low volume metastatic disease. The objective of
this study is to determine the dose level that was associated with DLT in the first two
cycles in less than 20% of patients. The planned accrual is 40 patients evaluable for DLT.

The trial has been completed and 40 patients are evaluable for DLT 33. Nine patients have
experienced 10 DLTs (dose levels 2, 3 (3), 4(2) and 5(3)) including four developing grade 4
thrombocytopenia, four gastrointestinal toxicity (GI bleed (2), nausea/anorexia, weight
loss), and two developing a decline in performance status. The estimated probability of DLT
is 21% for dose level 3 and 23% for dose level 4 with 12 patients treated at each dose
level. Twelve patients were enrolled as resectable, all 12 were explored, and 7 underwent
resections with clear margins. There was one pathologic complete response and two
additional patients with marked treatment effect. In preparation for this neoadjuvant
trial, our phase I study was amended to provide 30 Gy of radiation in the first 3 weeks and
the last 7 patients received this modified radiation dose. Two patients with unresectable
disease experienced grade 3 gastrointestinal toxicity. Subsequently, 3 additional patients
have been treated off trial and have tolerated this neoadjuvant therapy well without
experiencing DLT.

Based on tolerability seen with the addition of oxaliplatin, the anticipated benefit of
combination chemotherapy in augmenting local and systemic disease control and the pathologic
responses in resected specimens in our current trial, we are interested in implementing a
phase II study in patients with resectable or borderline resectable pancreatic cancer using
data from the current study to select dosing. We have chosen dose level 4 for the
neoadjuvant trial despite a DLT estimate of 23% based on the following observations. We
expect that a resectable patient population will have better tolerance of combined modality
therapy than those with unresectable disease (the majority in our current trial). In our
phase II gemcitabine and radiation trial, only one of 20 patients (5%) with resectable
disease experienced grade 3 non-hematologic toxicity in contrast to 33% of those with
unresectable disease17, 26. In our current trial, all DLT have been observed in patients
with unresectable or metastatic disease. We believe that this poorer tolerance is related
to a lower baseline PS and the larger radiation volume needed to encompass unresectable
disease. Furthermore, in the current trial, four of the DLTs were grade 4 platelets which
resolved without consequence. In the proposed trial, oxaliplatin will be dose reduced for a
platelet count < 75K, in contrast to the current trial where no dose reduction was
performed.

We are hopeful that the application of this novel, multimodality neoadjuvant approach will
increase margin negative resections in this patient population and ultimately survival. In
addition to further development and characterization of the combined modality therapy, we
are interested in defining the value of FDG-PET scans in determining resectability,
prognosis and response to neoadjuvant therapy in pancreatic cancer. Furthermore, the
relationship between histopathologic response to neoadjuvant treatment and prognosis will be
studied. It is anticipated that the data from this trial will be used in the design of a
larger phase III cooperative group trial.