Controversies in the Adjuvant Treatment of Pancreas Cancer: A Lack of Knowledge or Simply a Lack of Will?

“The difference between a successful person and others is not a lack of strength, not a lack of knowledge, but rather a lack of will.”

—Vince Lombardi

The article by Garofalo et al that appears in this issue of Gastrointestinal Cancer Research (see page 182) is a well-written, comprehensive review of the major adjuvant trials in pancreatic cancer.¹ Unfortunately (as noted by the authors), the optimal adjuvant therapy for pancreatic cancer is still debated, despite several decades of studies in this area. Even more discouraging is that despite the progress that has been made in cancer therapy in recent years, the prognosis for pancreas cancer remains largely unchanged. The majority of patients present with disease too advanced to resect. For the few who are able to undergo surgical resection, most experience disease recurrence and ultimately die of pancreas cancer.

WHAT HAVE WE LEARNED?

Chemotherapy

With regard to chemotherapy, all randomized trials of adjuvant therapy vs. surgery alone included chemotherapy that was mostly 5-fluorouracil (5-FU) based. Three trials looked at chemotherapy alone vs. no chemotherapy. The Norwegian trial, studying doxorubicin, mitomycin C, and 5-FU, established a better median survival with chemotherapy at the cost of significant toxicity, but did not improve long-term survival.²

The ESPAC-1 (European Study Group for Pancreatic Cancer) trial was designed as a 2 × 2 multifactorial study.³ In the Garofalo article, the final data set of 289 patients who participated in the 2 × 2 design is called a subset analysis, which is not entirely the case. The original statistical design called for just this set of patients, and the statistical design was based on only them. The other patients reported in the Lancet article were “extras” added to the study for unclear reasons.⁴ These extra patients have helped to raise the level of confusion about the ESPAC-1 study. With a 2 × 2 design it is possible that one of the randomizations can be readily doubted while the other is accepted. Such is the case with ESPAC-1. Difficulties with the proper administration of protocol therapy only strengthened the fact that chemotherapy resulted in a significant improvement in survival. Therefore, it can be accepted that 5-FU improves survival as adjuvant therapy.

The Charité Onkologie (CONKO-001) study compared gemcitabine adjuvant therapy after surgery to surgery alone.⁵ The primary end point was disease-free survival. The study also called for the standard intent-to-treat analysis and a “qualified analysis” consisting of only those on the treatment arm who received at least one cycle of therapy and those on the control arm who did not receive therapy. The results of this trial were positive, with a near doubling of disease-free survival for both the intent-to-treat analysis and the qualified analysis. However, median survival was significantly improved on the gemcitabine arm for only the qualified analysis. By intent-to-treat, patients treated with gemcitabine had a median survival of 22.1 months, with 22.5%alive at 5 years vs. 20.2 month median survival for the control arm, with 11.5%alive at 5 years (log-rank P = .06).

Overall, gemcitabine or 5-FU can improve the results of surgery for pancreas cancer. The fact that both drugs resulted in doubling the percentage of survivors at 5 years over surgery alone suggests that chemotherapy can lead to improved longterm survival. However, in light of the fact that chemotherapy plus surgery results in barely more than 20% of patients alive at 5 years, we still have a long way to go in pancreas cancer treatment.

What are the chemotherapy questions that exist? The most obvious need is a better drug or combination of drugs than 5-FU or gemcitabine. With a series of negative trials in the metastatic setting, hope is waning for finding a magic bullet in the immediate future. However, the adjuvant setting is not the metastatic setting. The current Intergroup trial is a randomized phase II study comparing the addition of bevacizumab to chemoradiation vs. the addition of cetuximab to chemoradiation. This study may now represent the only chance to evaluate these two agents in the adjuvant setting.

The results of ESPAC-3, which randomly assigns patients to either adjuvant gemcitabine or adjuvant 5-FU, will help to clarify if one of these two agents is better in the adjuvant setting. The recently reported Radiation Therapy Oncology Group (RTOG) 97-04 trial randomly assigned patients to either gemcitabine or 5-FU prior to and after chemoradiation with concurrent 5-FU.⁶ Patients in the gemcitabine arm experienced a statistically significant improvement in overall survival compared to the 5-FU–only arm. This suggests that gemcitabine may be superior in terms of the chemotherapy, but there is a small possibility that some of the benefit seen in the gemcitabine arm came from the patients receiving two drugs (5-FU during radiation therapy and gemcitabine before and after radiation therapy) rather than just one drug (5-FU in the control arm). Therefore, the results of ESPAC-3 remain crucial.

Chemoradiotherapy

Four of the reported randomized trials included radiation therapy. All four used concurrent chemotherapy with radiation, or “chemoradiotherapy.” Three of the four trials combined chemotherapy with splitcourse radiation, meaning that patients had a preplanned break of 2 weeks in the middle of the chemoradiation. As in the Garofalo article, radiation oncologists note that this is a suboptimal method for administering radiation therapy. To review, the first trial was conducted by the Gastrointestinal Study Group (GITSG).^7,8 With only 43 patients enrolled over 8 years there was a statistically significant improvement in survival for the chemoradiotherapy arm. This arm included both chemoradiotherapy and subsequent chemotherapy. First, it is not possible to separate the relative contributions of the chemotherapy and the chemoradiotherapy. Second, it is the one study that the chemoradiotherapy arm of the ESPAC-1 trial disproves.³ This trial was a very important trial in its time; right or wrong, it set the standard of care in the United States (although nobody actually uses the regimen from the GITSG trial). The time has come, however, to stop discussing this study in anything other than a historical context.

The second trial, conducted by the European Organisation for Research and Treatment of Cancer (EORTC), randomly assigned patients to chemoradiotherapy vs. no chemoradiotherapy.⁹ There was no postradiation chemotherapy in this trial, thereby isolating the effect of chemoradiotherapy. However, the results from this trial were negative. Admittedly, as noted by Garofalo et al, subset analysis suggests a trend for improved survival for pancreas cancer patients.¹⁰ However, a subset analysis is not definitive proof, and this suggests that further study of chemoradiotherapy is appropriate. Nothing more should be taken from these data except that other periampullary tumors should be excluded from pancreas cancer adjuvant therapy trials.

Finally, as with all papers that discuss ESPAC-1, the Garofalo paper notes the many flaws in the chemoradiotherapy arm. What can be learned from the chemoradiotherapy randomization in ESPAC-1? First, that the results of the GITSG trial probably reflect the small numbers of patients on that trial. Second, that giving chemoradiotherapy administered in an inferior, uncontrolled manner is not beneficial to patients and may even be harmful. Nothing more can be stated.

The last of the four trials using chemoradiotherapy is the RTOG 97-04 trial, a very important American trial for several reasons.⁶ But what does it teach us about chemoradiotherapy? It teaches us that infusional 5-FU in combination with standard radiation, instead of split-course radiation, can be administered safely. Second, it informs us that when using chemoradiotherapy, gemcitabine should be administered prior to or after radiation. What did it fail to tell us? It failed to answer one very important question: “What is the role of chemoradiotherapy in adjuvant therapy of pancreas cancer?”

Increasingly, the important questions such as these are not being asked in American clinical trials. We have left this question to be answered by the EORTC. However, without government funding of the EORTC, their randomized trial comparing chemoradiotherapy to chemotherapy alone is in jeopardy of not moving beyond a randomized phase II trial to a randomized phase III trial.

It can be readily assumed that in pancreas cancer, recurrence—either local or metastatic—will result in eventual death. Therefore, control of all disease is important. In ESPAC-1, 35% of patients had local recurrence alone, so local control is important. However, 61% of patients had either combined local and metastatic recurrence or metastatic recurrence alone. Without better control of systemic disease, the effect of chemoradiotherapy on local control may not contribute sufficiently to improve overall survival. In a randomized trial comparing chemoradiotherapy to chemotherapy alone, a negative result would eliminate the use of radiation. If, however, better chemotherapy regimens were developed, the role of radiation may become more important again, possibly necessitating reevaluation. In the interim, the role of chemoradiotherapy remains unproven.

ADJUVANT TREATMENT: WHAT ARE THE STANDARDS?

Since there are no data refuting the use of chemoradiotherapy, Garofalo et al suggest that this is a standard regimen. The gemcitabine-containing arm of RTOG 97-04 using chemotherapy with chemoradiotherapy should remain a standard arm in developing new clinical trials and for those physicians who wish to use radiation as part of the adjuvant therapy of pancreas cancer, but the data are not good enough to state unequivocally that chemoradiotherapy ought to be a standard component of pancreas cancer adjuvant therapy. A meta-analysis of adjuvant trials demonstrated that chemotherapy was effective in the adjuvant therapy of pancreatic cancer but that adjuvant chemoradiotherapy was not effective.¹¹ Subgroup analyses estimated that chemoradiation was more effective in patients who had marginpositive disease. Therefore, it is more than reasonable to consider either 6 months of 5-FU as given on ESPAC-1 or 6 months of gemcitabine as administered on CONKO-001 as the true standards of care for adjuvant treatment of resected pancreas cancer.

FUTURE DIRECTIONS

Garofalo et al suggest asking the question of what is the optimal timing for radiotherapy. This is very reasonable. Other studies of new combinations of chemotherapy or chemotherapy with targeted therapy may also be reasonable, using either a chemotherapy or chemoradiotherapy standard as the building block. The regimen of cisplatin, interferon, and 5-FU used with radiotherapy by Picozzi et al resulted in remarkable long-term survival, with 49% of patients alive at 5 years.¹² The American College of Surgeons Oncology Group (ACOSOG) repeated this regimen in a phase II trial, for which results are pending. If ACOSOG Z05031A2 confirms the phase II results reported by Picozzi et al, further study would definitely be warranted in carefully selected patients, despite the excessive toxicity seen in the trial.

Questions about neoadjuvant chemoradiotherapy or neoadjuvant chemotherapy are also reasonable. However, before moving forward, it is time to ask a few other important questions. First, how can we have spent 3 decades answering almost no questions? We owe our patients better designed trials built to answer the questions that really matter. Second, can we improve quality control and standard definitions? For example, ESPAC-1 did not assess radiation ports. This is absolutely essential to all trials involving radiation therapy. In addition, the evaluation of margins is not consistently performed. This requires close communication between the surgeon and pathologist. Finally, it is time to consider separating the margin-positive from the margin-negative patients. As we move toward more uniform practice in designing our trials, we may be able to move toward both a single standard and, hopefully, a better overall standard of care for patients with resected pancreas cancer.

After years of research we owe it to our patients to develop trials that can yield results. Then we must have the will to support and complete them, or knowledge will always remain just beyond our collective grasp.