Dr Castellanos et al have provided a very comprehensive review of the multimodality therapy of localized pancreatic cancer, with an emphasis on adjuvant and neoadjuvant therapies. They have accurately stated some of the major challenges in interpreting existing data from clinical trials: The use of neoadjuvant therapy in patients with potentially resectable pancreatic cancer remains experimental and is largely limited to academic centers and primarily driven by surgeons who subspecialize in pancreatic cancer. Its role, especially in patients with borderline resectable disease, is being better defined by advancements in staging and an agreement of expert opinion as to what constitutes a potentially resectable disease. In order to be an effective component of the multimodality therapy of localized pancreatic cancer, neoadjuvant therapy must be considered as a treatment strategy for improving survival, rather than just improving resectability.
Current studies testing neoadjuvant therapy in patients with resectable pancreatic cancer are based on the use of systemic therapy and radiation, and existing data support the benefit of systemic therapy in patients with resected pancreatic cancer, when used in the adjuvant setting. However, the role of adjuvant radiation therapy in these patients still needs to be clarified. As would be expected, radiation reduces the frequency of local disease progression, however its impact on survival has yet to be definitively quantified. No study has primarily asked a radiation question in a randomized setting: the often-cited GITSG study that became the basis of the current “standard of care” did not have a radiation-alone arm, and the experimental arm of radiation was combined with 5-fluorouracil (5-FU). The ESPAC-1 study from Europe, despite its limitations with respect to design and quality control, accelerated the move away from radiation therapy in many parts of Europe. Newer generation phase III trials in Europe (eg, CONKO-1, ESPAC-3, ESPAC-4) did not include radiation therapy; these trials reflect the European standard of care in treating resectable pancreatic cancer.[3,4] Nevertheless, in the United States, radiation therapy remains widely used by oncologists treating patients with localized pancreatic cancer. The argument for radiation therapy after an R0 or R1 pancreatic resection is based on the frequency of positive margins (documented and undocumented), the high frequency of loco-regional failures after surgery, and the relative ineffectiveness of the systemic therapies. Given the cost and morbidity associated with radiation therapy, it is very important that the current RTOG 0848 study that addresses the radiation question be completed in a timely fashion so that we may obtain a definitive answer as to whether radian therapy is necessary in all patients with resected pancreatic cancer. It is possible that the benefit of radiation is seen more in patients who do not experience early disease progression at distant sites. The design of RTOG 0848 includes the delivery of radiation after the completion of six cycles of gemcitabine (a practice based on CONKO-1—and one not commonly seen in the United States) and only after clinical and radiological proof of lack of disease progression.
The development of new therapies in resectable patients, whether before or after surgery, must focus on discovering better systemic agents to treat the occult metastatic disease that is present in all but a few patients with early-stage pancreatic cancer. Traditionally such efforts follow the successful testing of drugs in patients with advanced disease. However, one can make the case for testing novel agents in early-stage disease given the dismal outcome of those patients and the possibility of disease biology that is different from that of metastatic disease. 5-FU or gemcitabine is used in patients with resectable pancreatic cancer in both the adjuvant and the neoadjuvant setting.
The conclusion from the RTOG 9704  and CONKO-3  studies was that gemcitabine offered either no benefit or a marginal survival benefit over 5-FU but had a better toxicity profile; hence it is now favored as the backbone of chemotherapy. None of the gemcitabine-based doublets tested in advanced disease showed a survival advantage over gemcitabine alone. Consequently, none of these regimens were tested in resectable patients in a phase III setting, with the exception of the ESPAC-4 study of gemcitabine alone vs the gemcitabine/capecitabine doublet (without radiation), which is currently accruing patients. Testing of targeted agents in resectable patients has attracted limited enthusiasm, with the exception of the inclusion of erlotinib in the RTOG 0848 study. Nevertheless, a recent finding of the superiority of the FOLFIRINOX (5FU, leucovorin, oxaliplatin, irinotecan) combination over gemcitabine in patients with metastatic disease generated interest in extending its role to earlier-stage disease. In this phase III study from Europe (PRODIGE 4/ACCORD 11), there was notable improvement in median survival time from 6.8 months to 11.1 months in the gemcitabine and FOLFIRINOX arms, respectively. FOLFIRINOX was clearly an active regimen when used in patients who had favorable performance status (0 or 1) and optimal liver function, conditions that are also seen in the majority of patients with resectable disease.
It would be beneficial to be able to deliver adequate systemic therapy (and radiation) to “all” patients prior to surgery, given the post-operative morbidities that may limit therapy in the adjuvant setting. Moreover, neoadjuvant treatment does not appear to interfere with resectability of pancreatic tumors. Despite the arguments favoring neoadjuvant therapy, results of pilot phase II trials (mostly from single institutions) did not indicate a major improvement in survival when favorable patient selection was also taken into consideration.[7,8] This is largely explained either by the limited effectiveness of existing systemic therapies in pancreatic cancer or the manner in which they were used in these studies. In the search for better regimens, it is noteworthy that FOLFIRINOX tripled the objective response rate over that of gemcitabine in patients with metastatic disease  and may offer better pre-operative downsizing of the pancreatic tumor in addition to improved systemic effects. It is therefore very likely that FOLFIRINOX will prove useful in the neoadjuvant setting as well as in the adjuvant setting.
In conclusion, with the establishment of a role for adjuvant therapies in resectable pancreatic cancers, there is much interest in developing pre-operative therapies for this disease. Studies must be designed using newer agents, especially studies combining agents with tissue, blood, or image-based correlative science. At some point we must address the relative benefit of neoadjuvant vs adjuvant therapy. A large-scale effort to determine the role of neoadjuvant therapy in pancreatic cancer may be dependent upon the development of more effective systemic therapies, and emphasis must continue to be placed on the creation of well-designed pilot trials that test novel treatment regimens.
Financial Disclosure: Dr. Philip acts as a consultant for Curis, sanofi-aventis, AstraZeneca, Eli Lilly, and Bristol Myers Squibb; he has research grants from Bristol Myers Squibb, Novartis, and Amgen; and he is a speaker for Roche, Amgen, and Bayer.