A Pilot Study of Pre- and Post-surgery Chemotherapy With mFOLFIRINOX in Localized, Resectable Pancreatic Adenocarcinoma

Pancreatic ductal adenocarcinoma is a highly lethal disease. It is estimated that in 2010 there were approximately 43,140 new cases in the United States and 36,800 deaths which ranks fourth among cancer related deaths . At diagnosis, roughly 17,000 of the initial 43,000 patients have disease which appears to be localized to the pancreatic bed, but unfortunately, only a minority of these patients (15-20%) are immediately operable (on imaging studies their disease appears to be both localized to the pancreas and technically resectable with clear margins, and at surgery no unexpected findings preclude an R0 resection ). It should be noted that of those patients deemed immediately operable on clinical grounds, between 20 and 57% are found to have inoperable disease on exploration depending on the series and on the extent and nature of preoperative staging. The majority of patients have locally advanced, borderline resectable or unresectable disease owing to involvement of critical structures, most particularly sentinel blood vessels such as the SMA, celiac axis, hepatic artery, SMV, or portal vein. Furthermore, despite recent advances in systemic therapy, even those fortunate few able to undergo immediate surgery remain largely incurable, with a five year survival of slightly less than 20%.

The current standard of care for adjuvant therapy in resectable pancreatic cancer is based on the CONKO-001 study which randomized patients to 6 cycles of postoperative adjuvant therapy with gemcitabine versus observation alone. Results favored adjuvant gemcitabine in both disease free survival (13.4 vs 6.9 months, p<0.001) and overall survival (24.2 months vs 20.5 months p=0.02) strongly supporting the use of adjuvant gemcitabine in the setting of both R0 and R1 resections. An RTOG study - 97-04 - concluded that gemcitabine was probably superior to 5FU when used pre and postoperatively in combination with 5FU/RT, with a HR of 0.82. The role of radiation therapy in this setting remains controversial, with studies such as the GITSG trial (pro) and ESPAC 1 (con) criticized either for questionable design, outdated chemotherapy or unconventional radiation therapy. Current studies in progress examining postoperative therapy include RTOG 0848 (phase III looking at adjuvant gemcitabine versus gemcitabine plus erlotinib plus/minus chemo/RT using fluorouracil), ACOSOG Z5041 evaluating gemcitabine plus erlotinib in the pre and postoperative setting and ESPAC 4 evaluating gemcitabine versus gemcitabine plus capecitabine.

Many studies have explored the use of neoadjuvant therapy in initially resectable, borderline resectable and unresectable disease. The majority of these have been single arm, single institution phase I/II studies, and results have been mixed. Potential benefits of a neoadjuvant approach include: downstaging of disease with an increased percentage of margin negative and lymph node negative resections; no delay in systemic therapy aimed at eradicating micrometastatic disease; the detection of biologically aggressive tumors, as evidenced by early progression/metastases during this phase of therapy, thereby avoiding inappropriate surgery; and the greater likelihood of completing all intended therapy as opposed to postoperative treatment, where fully 22 - 35% of patients do not complete their intended program.

A comprehensive meta-analysis and systematic review of neoadjuvant therapy in both resectable and unresectable pancreatic cancer has recently been published. The conclusion reached in resectable patients was that resection frequency and survival following neoadjuvant therapy was similar to that in patients undergoing primary resection followed by adjuvant therapy. In patients with initially unresectable disease, fully one third had resectable tumors following neoadjuvant therapy with survival comparable to initially resectable patients. In aggregate, these observations indicate that a neoadjuvant approach is feasible and effective, and that this sequence does not compromise resectability or survival, even in those patients with the best prognosis.

With respect to specific studies in resectable disease, investigators at MD Anderson Cancer Center have published their most recent results. In their first study, 86 patients with resectable disease in the head of the pancreas received radiation therapy (30 Gy in 10 fractions over 2 weeks) plus 7 weekly infusions of gemcitabine at 400 mg/m2/week. 85% of these patients were taken to surgery and 74% were able to undergo the intended pancreaticoduodenectomy. Median survival for patients whose disease was resected was 34 months, but it was only 7 months for those whose disease could not be resected. In a second study of patients with resectable adenocarcinoma of the pancreatic head, 90 patients were enrolled with the goal of administering chemotherapy alone for eight weeks, using gemcitabine and cisplatin, followed by combined low dose gemcitabine and radiation therapy (30 Gy in 10 fractions). Ultimately, 79 (88%) patients completed the full course of preoperative therapy and 62 of these (78%) patients were taken to surgery. 52 (66%) had their disease resected with a median survival of 31 months for those who had surgery, versus 10.5 months for those who did not. They concluded that initial combination chemotherapy with gemcitabine and cisplatin followed by chemotherapy/RT did not improve on the results achieved with chemotherapy/RT alone. This result is perhaps not surprising as gemcitabine, both alone and in doublet combinations, has simply not been active enough to materially impact on the outcome of this disease. Interestingly, the longer preoperative interval did not result in local tumor progression.

These findings are similar to those of other major centers reporting studies of neoadjuvant gemcitabine based chemoradiation for potentially resectable disease, and suggest that this is a valid strategy for further study in this setting.

Recently, investigators in Europe reported their results with a novel combination of docetaxel 30 mg/m2 weekly and RT of 45 Gy in 34 patients with resectable disease. 32% had progression, 59% stable disease and 9% partial remission. 50% of the original cohort had pancreaticoduodenectomy with 100% R0 resection and a median survival of 32 months. The numbers are small and the overall resection rate disappointingly low, but the R0 resection rate and median survival for those resected is equivalent to patients receiving gemcitabine based regimens.

In locally advanced, unresectable disease, results have recently been reported from Austria with neoadjuvant gemcitabine/oxaliplatin without RT (39% of patients undergoing resection of disease with 69% R0 and 22 months median survival), Italy with neoadjuvant PEFG/PEXG (cisplatin, epirubicin, 5-fluorouracil (F)/capecitabine (X), gemcitabine) or PDXG (docetaxel substituting for epirubicin) followed by RT plus X, F or G (14% resected with median survival 16.2 months)and Japan with neoadjuvant gemcitabine 1000 mg/m2/wk and RT 50 Gy in T3 disease followed by postoperative 5FU liver perfusion (82% of patients with disease resected, 43% 5 year survival). In each of these studies, there is a mixture of borderline resectable and unresectable disease with the result that the interpretation of outcomes is problematic. However, it is clear that a neoadjuvant approach is feasible and active.

It is clear from an examination of the NCI CTEP database that the neoadjuvant approach has been widely embraced for future study. More than 40 active protocols are listed, including the following select few: a UVA study of hypofractionated RT plus chronomodulated capecitabine in resectable and borderline resectable disease; a UT Southwestern phase I study of SBRT or SBRT plus gemcitabine; an Emory phase I study of FOLFIRINOX and SBRT; a Fred Hutchinson study of GTX and oxaliplatin with IMRT and adjuvant gemcitabine; a UF phase II study of risk adapted gemcitabine plus abraxane; a European randomized phase II study of gemcitabine plus oxaliplatin pre and gemcitabine postoperatively versus gemcitabine postoperatively only; a Memorial Sloan Kettering phase II study of gemcitabine plus oxaliplatin preoperatively plus gemcitabine postoperatively; and an ACOSOG study of gemcitabine plus erlotinib pre and postoperatively.

Looking to the future there have been a number of recent innovations in chemotherapy. In an ongoing effort to discover non-gemcitabine based chemotherapy for those who have progressed on gemcitabine, and also a new regimen with more efficacy than those currently used in patients with pancreatic cancer, the combination of 5FU, leucovorin, oxaliplatin and irinotecan has been tested. This combination was initially studied in colorectal cancer, in a regimen known as FOLFOXIRI. It was established that this combination was both tolerable and effective in this setting. Subsequently, the regimen was modified slightly to the current FOLFIRINOX format (oxaliplatin 85mg/m2, irinotecan 180 mg/m2, leucovorin 400 mg/m2 , 5FU 400 mg/m2 on day 1, then 5FU 2400 mg/m2 as a 46 hour continuous infusion) and tested in a phase II study in pancreatic cancer. 47 chemotherapy-naïve patients with metastatic disease were enrolled and 46 were treated. Confirmed response rate was 26% with 4% complete responses. Median time to progression was 8.2 months and median overall survival was 10.2 months. Grade 3/4 toxicities included neutropenia (52%), nausea (20%), vomiting (17%), diarrhea (15%), and neuropathy (15%). No toxic death occurred. FOLFIRINOX was then tested after failure of previous gemcitabine therapy in metastatic disease and was deemed to be promising. 13 patients were treated, with 9 evaluable for response - 6 had stable disease with a mean time to progression of 6.6 months, and 3 progressed.

As a consequence of the previously mentioned phase II study in chemo-naïve patients, a randomized phase II/phase III study comparing gemcitabine (G) to FOLFIRINOX (F) as first line treatment of metastatic pancreatic cancer was conducted. This study was terminated prematurely by the study IDMC as it was determined that additional patient accrual would not add to the statistical power of the study. 342 patients were accrued with roughly one third of the primary disease involving the head of the pancreas. Overall objective response rate was 32% for F versus 9.4% for G. PFS was 6.4 months versus 3.4 months and overall survival 11.1 months versus 6.8 months, all in favor of F.

The overall survival rate of 11.1 months is the best result achieved thus far in a randomized phase III study of chemotherapy in metastatic pancreatic cancer. Notable toxicities of at least grade III/IV, which were all worse with F, were neutropenia (45.7 vs 18.7%), febrile neutropenia (5.4 vs 0.6%), fatigue (23.7 vs 14.2%), vomiting (14.5 vs 4.7%) and diarrhea (12.7 vs 1.2%). These results indicate that this is a notably active regimen with an encouraging response rate. However, the potential toxicities are significant, and it is a regimen that should be offered only to patients with ECOG 0-1 performance status and excellent supportive care. In this regard, in an attempt to ameliorate these toxicities, modifications to the published regimen have already been proposed by the French group and others. In their forthcoming study of FOLFIRINOX in the adjuvant setting, the French will omit the bolus of 5FU which contributes significantly to the myelosuppression but which is thought to have minimal impact on the therapeutic efficacy. In addition, most physicians now incorporate the routine use of neulasta with each treatment cycle. This study will similarly incorporate these modifications and the regimen will be named mFOLFIRINOX.

Thus, in the context of perioperative therapy, we have identified a regimen - FOLFIRINOX - with the best results to date in the treatment of metastatic disease and by inference, promise of improved outcome in those patients with resectable disease. If successful, this has the potential to improve DFS and overall survival (until now no better than 15-20% at 5 years) in these patients, and may establish a new paradigm for future studies.