Current status on the place of FOLFIRINOX in metastatic pancreatic cancer and future directions

Abstract

Pancreatic cancer (PC) incidence rates are rapidly increasing in developed countries, with half the patients being metastatic at diagnosis. For decades, fluorouracil, then gemcitabine regimens were the preferred palliative first-line options for fit patients with metastatic PC. FOLFIRINOX (a combination of bolus and infusional fluorouracil, leucovorin, irinotecan and oxaliplatin) was introduced to clinical practice in 2010 due to the results of the phase II/III trial (PRODIGE 4/ACCORD 11) comparing FOLFIRINOX with single-agent gemcitabine as first-line treatment for patients with MPC. Median overall survival, progression-free survival, and objective response rate were superior with FOLFIRINOX over gemcitabine and there was prolonged time to definitive deterioration in quality of life. Although FOLFIRINOX was also associated with increased toxicity, mainly febrile neutropenia and diarrhea, there has been rapid uptake of this regimen. This review closely examines optimal management and prevention of toxicities, international recommendations for first-line treatment, and use of modified FOLFIRINOX protocols. In this review, we also look at the potential benefit of FOLFIRINOX in selected groups of patients: second-line therapy, adjuvant chemotherapy, induction therapy in patients with borderline resectable and locally advanced PC. Robust validation of the FOLFIRINOX regimen in these settings requires confirmation in further randomized trials.

Introduction: actual situation

Pancreatic ductal adenocarcinoma is a major cause of cancer-related mortality in western countries and projected to emerge as the second leading cause of cancer-related deaths in the US by 2030.¹ Its prognosis remains extremely poor, with a 5-year survival rate of just 8%.² The only potential curative therapy is surgical resection; however, as little as 10–15% of patients with pancreatic cancer (PC) are eligible. At diagnosis, 52% of PC cases are stage IV.² These patients may be candidates for systemic chemotherapy with the aim of prolonging survival, reducing symptoms and either maintaining or increasing quality of life (QoL). In this review, we aimed to update knowledge about the use of the FOLFIRINOX regimen in routine practice taking into account pharmacoeconomic data and future developments. Finally, we tried to clarify the best options in first-line and second-line in patients with metastatic PC (MPC).

For decades, fluorouracil was the most common chemotherapy agent in advanced PC until the 1997 approval of gemcitabine. Single-agent gemcitabine has since been the standard palliative treatment as it significantly increased the ‘clinical benefit response’ compared with bolus fluorouracil in a landmark randomized phase II study.³ Gemcitabine also slightly increased median survival from 4.4 to 5.7 months (p = 0.0025). Moreover, gemcitabine was well tolerated, with few severe effects. Thereafter, many targeted or cytotoxic agents in combination with gemcitabine failed to show any improvement in either overall survival (OS) or QoL.⁴ The only exception being erlotinib, which provided a statistically significant, although clinically irrelevant, median OS benefit of 10 days in combination with gemcitabine.⁵

The FOLFIRINOX regimen (a combination of bolus and infusional fluorouracil, leucovorin, irinotecan and oxaliplatin) has since emerged to significantly advance treatment of MPC. Both the safety and efficacy of FOLFIRINOX were first indicated in a phase II study performed by the UNICANCER gastrointestinal group in patients with locally advanced or MPC.⁶ This study reported a response rate (RR) of 26% [95% confidence interval (CI), 13–39%] and a promising median OS of 10.2 months, with no toxic deaths. However, 52% of patients experienced a severe neutropenia and 4% a grade 4 febrile neutropenia. The EORTC QLQ-C30 assessed QoL data revealing a 10-point improvement in either a functional or symptom scale in 42–69% of patients. Responders demonstrated a major improvement in their global QoL scores (>25 points).⁶

Phase III trials in metastatic pancreatic cancer

Following this encouraging data, we conducted a phase II/III trial (PRODIGE 4/ACCORD 11) comparing FOLFIRINOX with single-agent gemcitabine as first-line treatment of patients with MPC in France.⁷ The trial used OS as primary objective, and progression-free survival (PFS), QoL, RR and toxicity as secondary endpoints. A total of 342 patients were enrolled. Trial entry was restricted to patients with Eastern Cooperative Oncology Group (ECOG) performance status (PS) 0 or 1 given one previous trial, suggesting a possible benefit from combination chemotherapy was limited to patients with a Karnofsky score of 90 or 100.⁸ Other eligibility criteria included age under 76, bilirubin below 1.5 times the upper limit of normal, measurable MPC and adequate organ function. Treatment of 6 months was recommended for responding patients. No granulocyte-colony stimulating factor (G-CSF) was permitted as primary prophylaxis of neutropenia. Assessment of tumor response occurred every 8 weeks, using a CT scan with contrast.

The trial began as a randomized phase II study using RR as a primary endpoint. Following an independent review confirming an objective RR of 34.1% in the FOLFIRINOX group and 11% in the gemcitabine group, the trial then proceeded to phase III. The study design enabled an 80% power for detecting an increase in median OS from 7 to 10 months (HR = 0.70, α = 0.05). The final analysis required 360 patients to reach 250 events. At the preplanned interim analysis after 167 events, the study achieved its primary endpoint with a p-value < 0.001; therefore the independent data monitoring committee recommended termination of accrual.

A total of 342 patients from 48 French centers had enrolled into the study. Compared with treatment with gemcitabine, the FOLFIRINOX regimen resulted in a superior objective RR (9.4 versus 31.6 %; p < 0.001). The median number of months on treatment was 5 (range: 0.5–23.5) in the FOLFIRINOX group and 3 (range: 0.5–13) in the gemcitabine group (p < 0.001). Analysis also revealed a median duration of response of 5.9 and 3.9 months in the FOLFIRINOX and gemcitabine groups, respectively (p = 0.57). The FOLFIRINOX group reported a median OS of 11.1 months compared with 6.8 months for the gemcitabine group (p < 0.001). The FOLFIRINOX regimen homogeneously affected OS across all prognostic subgroups: age, sex, PS, albumin level, primary tumor site, synchronous or metachronous metastases, number of metastatic sites, presence or absence of hepatic metastases, carbohydrate antigen 19-9 (CA 19-9), and carcinoembryonic antigen dosages. Multifactorial analysis revealed favorable prognostic factors of the FOLFIRINOX group as: age ⩽ 65 years, normal albumin level, metachronous metastases, and the absence of hepatic metastases. The FOLFIRINOX regimen demonstrated greater, yet manageable, toxicities, including grade 3–4 neutropenia in 45.7% patients and grade 3 febrile neutropenia in 5.4%. FOLFIRINOX also significantly increased grade 3–4 diarrhea (12.7% versus 1.2%) and grade 2 alopecia (11.4% versus 1.2%).⁷ Out of 67 patients (39.2%) with a pancreatic head tumor in the FOLFIRINOX group, 27 (40.2%) had a biliary stent. During the course of the trial, investigators observed neither single cholangitis nor mortality attributed to FOLFIRINOX. In our experience, we observed responses to FOLFIRINOX rechallenge in patients progressing after a discontinuation of FOLFIRINOX due to either a treatment pause (originally planned after 12 courses in the pivotal trial) or oxaliplatin-induced sensory neuropathy. In the pivotal trial, a stop-and-go strategy reintroduced FOLFIRINOX in 7.6% of patients of the FOLFIRINOX group and 7.0% of the gemcitabine group to the same regimen.

The secondary endpoint of the PRODIGE 4/ACCORD 11 trial utilized longitudinal QoL. Compliance for questionnaire completion achieved 92–95% at baseline, although decreased over time to 40–50%.⁹ EORTC QLQ-C30 domain scores revealed no differences over time, except for higher diarrhea scores during the first 2 months of FOLFIRINOX treatment. Global QoL scores improved in the FOLFIRINOX group (p < 0.001) and emotional functional scores improved (p < 0.001) in both arms between baseline and the end of 12 cycles. Both treatments decreased pain, insomnia, anorexia and constipation. The FOLFIRINOX group demonstrated a significantly longer time until definitive deterioration for the 15 domains of the QLQ-C30. At 6 months, 31% of patients in the FOLFIRINOX group showed a definitive degradation in QoL versus 66% of the gemcitabine group [hazard ratio (HR), 0.47; 95% CI, 0.30–0.70; p < 0.001]. Multivariate analysis identified significant poor prognostic factors for OS as: age > 65 years, low serum albumin and three QoL domains including physical functioning, and surprisingly, constipation and dyspnea. Patients with a severely compromised baseline physical function (despite PS 0–1 according to investigators) in the lower tertile (33.3/100) from both groups correlated with a short median OS of 2.1 months, compared with 9 months for those patients with a physical functioning score > 66.6. Despite the inclusion criteria stating ECOG PS should be 0 or 1, 30 patients (18.4%) in the FOLFIRINOX group responded either ‘quite a bit’ or ‘very much’ to the baseline QoL question: ‘Do you need to stay in bed or a chair during the day?’ These patients demonstrated a worse prognosis than the overall population with a median OS of 6.7 months and 4.6 months in the FOLFIRINOX and gemcitabine arms, respectively.⁹

The main results from a confirmatory phase III trial performed in India with the same study design are available as an abstract.¹⁰ A total of 310 patients with an ECOG PS score of 0 or 1 were randomized to receive either FOLFIRINOX or gemcitabine using the Burris schedule. They reported a median OS of 10.8 months in the FOLFIRINOX group compared with 7.4 months in the gemcitabine group (HR, 0.48; p < 0.001). Median PFS was 5.6 months in the FOLFIRINOX group versus 3.1 months in the gemcitabine group (HR, 0.44; p < 0.001). The FOLFIRINOX group reported an objective RR of 29.6% compared with 8.3% for the gemcitabine group. At 6 months, 29% of patients in the FOLFIRINOX group showed a definitive degradation in their QoL versus 59% of the gemcitabine group (HR, 0.45; p < 0.001).

Real-world view

Many prospective and retrospective studies have confirmed the ACCORD 11 data in terms of OS, PFS and safety.^11–14 Consequently, the FOLFIRINOX regimen is recommended by the main international guidelines as the standard of care in those patients who are fit, with a good PS and no major comorbidities. Therefore, the National Comprehensive Cancer Network, American Society of Clinical Oncology (ASCO) and European Society for Medical Oncology (ESMO) guidelines recommend FOLFIRINOX for metastatic pancreatic adenocarcinoma (MPA) patients with either 0 or 1 ECOG PS and a favorable comorbidity profile.^15–20 All guidelines state the mandatory assessment of patients in terms of their condition and comorbidities in order to offer appropriate treatment with a safe toxicity profile. Despite this, the FOLFIRINOX regimen exhibits a specific profile of toxicity among distinct populations. Okusaka et al. conducted a phase II study among chemotherapy-naïve Japanese patients²¹ with RR as the primary endpoint. A total of 36 patients enrolled between June 2011 and September 2012, with a median age of 61.5 years and an ECOG of 0 or 1. The trial involved a median number of eight cycles (1–25) and reported a median OS of 10.7 months, a median PFS of 5.6 months and an RR of 38.9% (95% CI, 23.1–56.5). The toxicity profile of FOLFIRINOX appeared less favorable in this study, especially in terms of hematological side effects: 28 patients (77.8%) experienced grade 3 or 4 neutropenia with febrile neutropenia observed in eight patients (22.2%). Consequently, dose reduction or cycle delay occurred in 89% of the patients. The use of full-dose FOLFIRINOX regimen therefore requires careful consideration in specific Japanese populations. However, in 2013, this combination therapy was approved for unresectable PC in Japan.

Also of concern in daily practice is care of the elderly. The median age at enrollment was 61 (25–76) in the ACCORD 11 FOLFIRINOX arm, however, Baldini et al. focused on the specific population of elderly patients.²² Between 2008 and 2015, their retrospective chart review enrolled 42 patients with a median age of 73 (70–79), good PS (ECOG PS 0–1: 93%) and few comorbidities (median Charlson index = 10). Most patients (88%) presented with metastatic disease. Dose reductions occurred at the discretion of the treating physician, so from the first cycle, 57% of the patients received a primary dose reduction. The FOLFIRINOX regimen was most frequently prescribed after 2011 (37/42 patients, 88%). This study reported data consistent with the ACCORD 11 trial, with a median OS of 12.6 months. Moreover, reducing the primary dose appeared to have no significant effect on OS: 11.7 months (6.9–16.4) compared with 16.6 months (0.37–32.8), p = 0.69. Although toxicity was manageable, 54% of all grade neuropathy reached grade 3 for almost half of patients. The sepsis and septic shock rates appeared quite low, consistent with primary dose reduction and primary G-CSF prophylaxis used in 33% of patients. Currently, an ongoing phase II prospective French trial, PAMELA-70, is assessing the safety and efficacy of dose-adjusted FOLFIRINOX in elderly patients with MPA [ClinicalTrials.gov identifier: NCT02143219]. The irinotecan dose will be adjusted according to the UGT1A1 status and the 5-fluorouracil dose according to the dihydropyrimidine dehydrogenase pharmacogenetic status.

First-line regimen options for patients with metastatic pancreatic cancer

The current standard of care for MPC patients with a good PS is either FOLFIRINOX or gemcitabine plus nanoparticle albumin-bound (nab)-paclitaxel (GnP).¹⁷ To our knowledge, no head-to-head clinical trials comparing these two new regimens have been published and interstudy comparisons are inherently limited. The main results of the two pivotal trials are presented in Table 1. Some authors argue that different populations in the phase III pivotal trials of these combinations may have influenced more favorable results of FOLFIRINOX.^23,24 The main difference between the trials is PS in the inclusion criteria: the FOLFIRINOX trial registered patients with an ECOG PS of 0 or 1, whereas the MPACT trial allowed a Karnofsky PS of 70 corresponding to a PS of 2.²⁵ Patients classified as Karnofsky 70 (7.6%)²⁶ demonstrated a median OS of 3.9 months in the GnP group and 2.8 months in the gemcitabine group.²⁷ In both randomized trials, the two treatment groups showed similar demographic and clinical characteristics at baseline. Although the MPACT and ACCORD 11 trials reported similar median ages, no specified age limit for inclusion existed in the MPACT trial, with 10% of patients older than 75-years old. Both trials included identical control arms: gemcitabine as monotherapy 1000 mg/m² prescribed weekly for 7 weeks followed by treatment on days 1, 8 and 15, every 4 weeks. Both randomized trials reported identical median OS in the gemcitabine group of 6.8 months⁷ and 6.7 months,²⁶ suggesting that both study populations consisted of similar prognostic factors.

Table 1.

PRODIGE 4/ACCORD 11 and MPACT trials: indirect comparison of main results, demographics, and most common adverse reactions.

	Experimental arms		Control arms
	FOLFIRINOX7	nab-Paclitaxel + gemcitabine26	Gemcitabine PRODIGE 4/ACCORD 117	Gemcitabine MPACT26
Number of patients	171	431	171	430
PS	ECOG 0-1	Karnofsky index ⩾ 70	ECOG 0-1	Karnofsky index ⩾ 70
Pancreatic head tumor location	39.2%	44%	36.8%	42%
Median age (range)	61 (25–76)	62 (27–86)	61 (34–75)	63 (27–86)
RR [investigator assessment]	31.6% [95% CI: 24.7–39.1]	29% [95% CI: 25–34]	9.4% [95% CI: 5.4–14.7]	8% [95% CI: 5–11]
Discontinuation of treatment for unacceptable toxicities	3.8%	20%	10%	7%
Progression-free survival	6.4 months	5.5 months	3.3 months	3.7 months
Median OS	11.1 months	8.5 months	6.8 months	6.7 months
Hazard ratio for death versus gemcitabine	0.57 [95% CI: 0.45–0.73; p < 0.001]	0.72 [95% CI: 0.58–0.82; p < 0.001]
Grade 3–4 toxicities
Fatigue	23.6%	18%	17.8%	9%
Neutropenia	45.7%	38%	21%	27%
Febrile neutropenia	5.4%	3%	1.2%	1%
Diarrhea	12.7%	6%	1.8%	1%
Peripheral neuropathy	9%	17%	0%	1%
Vomiting	23.6%	6%	8.3%	4%
Growth factor use	42.5%	26%	5.3%	15%
Grade 2–3 alopecia	11.4%	50%	1.2%	5%
Second-line therapy	56.7%	38%	60.3%	42%

PS, performance status; RR, response rate; OS, overall survival; CI, confidence interval.

Gemcitabine is an appropriate treatment option for patients who choose to receive less-toxic therapy, and those who have either an ECOG PS of 2, or a comorbidity profile, precluding the use of combination chemotherapy.¹⁷

Benefit–risk balance and pharmacoeconomic analysis

The benefit–risk balance has previously been assessed using a generalized pairwise comparison based on prioritized outcomes: OS, PFS and grade 3–4 adverse events. The main analysis of benefit risk balance, as well as all sensitivity analyses, revealed a strong preference towards the FOLFIRINOX regimen.²⁸ In Canada, the cost effectiveness of first-line FOLFIRINOX was compared with first-line gemcitabine in MPC patients to submit to Ontario public decision makers. Results of the Markov analytical decision model revealed FOLFIRINOX as cost effective, especially with the reducing drug costs of using generic drugs.²⁹ A second Markov model compared single-agent gemcitabine with gemcitabine–capecitabine, gemcitabine–erlotinib and FOLFIRINOX.³⁰ Tariffs from nine different countries revealed differences in quality-adjusted life years (QALY) and incremental cost-effectiveness ratio (ICER) between countries with a willingness-to-pay threshold of $50,000. Gemcitabine showed the highest probability as a cost-effective treatment for all countries with a higher willingness-to-pay threshold (>$150,000), whereas FOLFIRINOX appeared as the treatment most likely to be cost effective using any of the country indices.³⁰

A Markov model also analyzed GnP, FOLFIRINOX, and gemcitabine as first-line treatment from an economic payer perspective in the US.²³ Indirect comparisons suggest that GnP and FOLFIRINOX have statistically similar OS efficacy (HR, 0.79; 95% CI, 0.59–1.05). FOLFIRINOX prevailed in PFS efficacy (HR, 0.68; 95% CI, 0.51–0.91). Calculations also showed an incremental cost/life year of $358,069 associated with FOLFIRINOX, based on the US costs of $860 for 100 mg of oxaliplatin and $221 for 100 mg of irinotecan. To highlight cost differences between countries that may impact on results of the pharmacoeconomic evaluations: 100 mg doses of generic irinotecan and generic oxaliplatin cost $5.60 and $6.30 in France, respectively.

In 2014, although total drug costs per cycle remained high, at $1633.21 for each cycle of FOLFIRINOX,³¹ another Canadian analysis revealed FOLFIRINOX as cost effective, with an ICER below commonly quoted ICER thresholds for oncology products. Since then, the cost of oxaliplatin has considerably reduced, particularly as generic oxaliplatin is available almost worldwide. This must be weighed up by the fact that most North American physicians use G-CSF or pegfilgrastim prophylaxis³² despite a febrile neutropenia risk far below the 20% cutoff of ASCO guidelines for primary prophylaxis, and no evidence-based DFS or OS benefit in secondary prophylaxis as compared with reduced doses or treatment delay; thus was not included into most of the economic models.

Furthermore, using the 2015 Canadian reduced costs ($549 for FOLFIRINOX and $4431 for GnP over 4 weeks), FOLFIRINOX is cost effective compared with GnP.³³ Overall, this analysis concluded FOLFIRINOX as the optimal regimen on the basis of cost and QALY. On the other hand, we must consider that drug prices account for only a part of the cost of delivering chemotherapy. The time spent in an infusion chair and the added cost of the ambulatory infusion pump for delivery of 5-FU must be considered. For example, the chair time for GnP is substantially less but at the price of less efficiency. Finally, the attractive price of the FOLFIRINOX regimen must be counterbalanced by the cost of supportive care (antiemetics, antidiarrheals, etc.) sometimes required. Those factors were integrated in the Goldstein et al. review. As expected, the treatment cost variability depends on the use of both GnP and growth factor.³⁴ Only a real-world claims database study might shed light on this issue.

Maintenance therapy after first-line FOLFIRINOX

Oxaliplatin neurotoxicity is often limiting and certain patients in the ACCORD 11 study received prolonged treatment without oxaliplatin over the planned 12 cycles. A maintenance strategy provides another option to delay deterioration of QoL in patients with controlled disease. A study of 30 patients retrospectively evaluated a continuous maintenance strategy with capecitabine until progression after four-to-eight cycles of FOLFIRINOX. This single-institution study suggested partial efficacy in selected patients treated with capecitabine, with a median PFS of 5 months and a median OS of 17 months. FOLFIRINOX reintroduction occurred in 45% of patients.³⁵ The PRODIGE 35/PANOPTIMOX phase II randomized trial [ClinicalTrials.gov identifier: NCT02352337] assessed the role of maintenance chemotherapy with fluorouracil and leucovorin (LV5FU2 regimen). The three treatment arms included: (i) FOLFIRINOX until progression; (ii) eight cycles of FOLFIRINOX followed by LV5FU2 until progression, or (iii) 2 months of FOLFIRINOX followed by 2 months of gemcitabine until progression. This trial recently completed accrual and we await results.

FOLFIRINOX as second-line treatment

A randomized trial assessing chemotherapy versus best supportive care has shown that certain patients benefit from a second-line treatment.³⁶ FOLFIRINOX as a second-line therapy in patients with MPA has been evaluated in a few small retrospective studies involving patients treated with a gemcitabine-based regimen. For example, in an exploratory analysis of the MPACT trial, FOLFIRINOX after GnP resulted in the longest OS of all second-line treatments analysed. However, FOLFIRINOX was given to patients with better PS.³⁷ A French study including 27 patients showed that the FOLFIRINOX regimen could be safely used as second-line therapy, with partial anticancer activity: 5/27 (19%) achieved partial responses, with a median PFS and OS of 5.4 months (0.5–25.5) and 8.5 months (0–26), respectively.³⁸ Another survey revealed an objective RR of 27.8% and a disease control rate of 55.6%,³⁹ however, the median PFS was 2.8 months. This short PFS suggests that FOLFIRINOX may be more appropriate as first-line chemotherapy. A single-institution experience Japanese study including 13 patients with an ECOG PS of 0 or 1 and no major comorbidities, compared pretreatment with either gemcitabine or S-1, or a combination of both. In the study, 4 out of 13 patients (30.8%) achieved a partial response and data revealed a median PFS and median survival of 4.7 months and 6 months, respectively.⁴⁰ In our opinion, it would be unethical to launch a strategic phase III trial comparing two successive lines of chemotherapy, especially if determining the best sequence of gemcitabine as a single agent versus FOLFIRINOX, as the PS of patients can rapidly decline after tumor progression in PC. Moreover, the controversial data on oxaliplatin/fluorouracil as second-line treatment perhaps suggests that the benefit from oxaliplatin-based chemotherapy is suitable only as first-line therapy.⁴¹ Until now, there is no clear standard treatment strategy for second-line therapy in advanced PC.⁴²

Second-line treatment after first-line FOLFIRINOX

There are very little data and no established evidence regarding second-line chemotherapy in patients with MPC progressing on first-line FOLFIRINOX and therapy options are limited. A retrospective survey involving 20 patients reported data on gemcitabine as second-line therapy after progression on FOLFIRINOX. It reported a median PFS and OS of 2.0 months and 5.7 months, respectively. These results are very similar to those obtained with gemcitabine as first-line treatment.⁴³ To our knowledge, there are no published data on the use of nanoliposomal irinotecan in combination with fluorouracil in this setting. A single-institution retrospective study involving 28 patients evaluated second-line GnP and reported a median PFS and median OS of 2.8 months and 5.3 months, respectively. Five patients (17.9%) achieved a partial response. Toxicities were mild. The authors suggest no effect of first-line FOLFIRINOX on limiting the use or compromising efficacy of second-line GnP.⁴⁴ GnP has also been evaluated in a prospective study involving 57 patients. Objective responses occurred in 10 patients (17.5%) with a disease control of 58%. Median PFS was an encouraging 5.1 months, with a median OS of 8.8 months.⁴⁵ According to the ASCO clinical practice guidelines, GnP may be an appropriate option for second-line therapy in patients who experience disease progression following first-line treatment with FOLFIRINOX.^17,45 Single-agent gemcitabine or inclusion into a clinical trial may also be considered for patients with a well-preserved PS.⁴⁶

How important are modifications to the FOLFIRINOX regimen?

In the phase III ACCORD-11 trial,⁷ FOLFIRINOX (oxaliplatin 85 mg/m², irinotecan 180 mg/m², leucovorin 400 mg/m², and fluorouracil given as a bolus of 400 mg/m² followed by a 46-h continuous infusion of 2,400 mg/m² given every 2 weeks) significantly increased grade 3 or higher neutropenia compared with single-agent gemcitabine (45.7% versus 21%), as well as febrile neutropenia (5.4 versus 1.2%), thrombocytopenia, diarrhea, and neuropathy.

Modified FOLFIRINOX (mFOLFIRINOX) typically consists of reduced irinotecan or suppression of the bolus fluorouracil, or sometimes both. Although many retrospective and thorough prospective pilot studies have been reported, no randomized trial evaluating mFOLFIRINOX has been conducted. Major recent studies are reported in Table 2.

Table 2.

First-line FOLFIRINOX in pancreatic cancer: treatment regimen, toxicity and efficacy data.

Study	Regimen	Febrile neutropenia prophylaxis	Population	Grade 3–4 neutropenia; febrile neutropenia	Fatigue	Vomiting	Diarrhea	OS (months) PFS (months)
Conroy et al.7	FOLFIRINOX	Filgrastim not recommended as primary prophylaxis	MPA	75/164 (45.7%) 9/166 (5.4%)	39/165 (23.6%)	24/166 (14.5%)	21/165 (12.7%)	11.1 6.4
Hosein et al.62	FOLFIRINOX	Pegfilgrastim 6 mg for 89% of the patients	LAPC, BRPC	4/18 (22%) 3/18 (17%)	2/18 (11%)	0	2/18 (11%)	100% at 1 year 95% at 1 year
Faris et al.63	FOLFIRINOX	Pegfilgrastim 6 mg	LAPC	4/22 (18%) 0/22	NAπ	NAπ	4/22 (18%) (all grades)	7% at 3 years 11.7
Gunturu et al.49	Dose modifications at physician discretion; 6/35 received full doses	Pegfilgrastim 6 mg	MPA, LAPC	4/35 (11.4%) 1/35 (2.9%)	2/35 (5.7%)	1/35 (2.9%)	1/35 (2.9%)	11.2 (MPA) 9.9 (MPA)
Mahaseth et al.64	mFOLFIRINOX no bolus 5-FU	Pegfilgrastim 6 mg	MPA, LAPC, BRPC	2/60 (3.3%) 0/60	8/60 (13.3%)	5/60 (8.3%)	8/60 (13.3%)	9.0 8.5
Alessandretti et al.65	mFOLFIRINOX (no bolus 5-FU and reduced dose of at least one agent since first cycle)	Pegfilgrastim 6 mg	LAPC	4/19 (21%) 3/19 (15.7%)	3/19 (15.7%)	1/19 (5.2%)	1/19 (5.2%)	NA
Umemura et al.40	20% dose reduction of FOLFIRINOX, and 3 weeks interval.	NA§	MPA, LAPC$	5/13 (38.5%) 0/13	NA	NA	NA	6 4.7
James et al.66	mFOLFIRINOX with 25% dose reductions of irinotecan and bolus 5-FU	Pegfilgrastim 6 mg	MPA, LAPC	10/62 (16.2%) 3/62 (4.8%)	7/62 (11.3%)	NA	NA	NA
Ghorani et al.67	mFOLFIRINOX irinotecan 135 mg/m2; no bolus 5-FU	Filgrastim for 7 days	LAPC	0/18 1/18 (5.6%)	1/18 (5.6%)	5/18 (27.8%)	3/18 (16.7%)	9.3 7.2
Nanda et al.68	mFOLFIRINOX no bolus 5-FU	Pegfilgrastim 6 mg	LAPC	NA	NA	NA	NA	65.5% at 1 year 49.2% at 1 year
Blazer et al.55	mFOLFIRINOX irinotecan 165 mg/m2; no bolus 5-FU	Pegfilgrastim 6 mg	LAPC, BRPC	0/43	4/43 (9.3%)	0	6/43 (14%)	21.2 18 (resection) 8 (no resection)
Stein et al.69	mFOLFIRINOX irinotecan 135 mg/m2; 300 mg/m2 bolus 5-FU	Pegfilgrastim on day 3 or 4	MPA, LAPC	9/74 (12.2%) 3/74 (4.1%)	9/74 (16.2%)	2/74 (2.7%)	12/74 (16.2%)	10.2 (MPA) 26.6 (LAPC) 6.1 (MPA) 17.8 (LAPC)

OS, overall survival; PFS, progression-free survival; MPA, metastatic pancreatic cancer; LAPC, locally advanced pancreatic cancer; BRPC, borderline resectable pancreatic cancer; NA, nonavailable; 5-FU, fluorouracil.

Regardless of the efficacy related to OS and PFS, the FOLFIRINOX regimen is associated with less toxicity, especially hematological toxicities. However, there are still some patients who experience febrile neutropenia, although the main toxicities are easily manageable in routine practice. Despite the heterogeneity of dose-reduction protocols, the great majority of these studies show comparable efficacy and OS data with the phase III ACCORD 11 trial. Alternative mFOLFOXIRI regimens using irinotecan first (instead of oxaliplatin first in the FOLFIRINOX regimen) have been tested in advanced PC. In their observational cohort study, the GONO group showed a comparable RR of 38.6% and OS of 10.8 months using FOLFORIXI and mFOLFOXIRI regimens in locally advanced and metastatic PC and concluded that it was a suitable alternative to FOLFIRINOX.⁴⁷

Management of toxicities

Dosage adjustment guidelines for FOLFIRINOX toxicities are available in the supplementary appendix of the pivotal study online.⁷ FOLFIRINOX should be considered as a highly emetogenic chemotherapy due to oxaliplatin and warrants optimal prophylaxis of nausea/vomiting. As in routine practice, reducing a patient’s drug dose is a standard approach in cases of toxicity. Retrospective single-institution series from highly selected academic centers have confirmed the safety of FOLFIRINOX in patients with indwelling biliary stents and successful biliary drainage.⁴⁸ Recommendations for the management and prevention of the most frequent toxicities are shown in Table 3. There are no differences in the use of oxaliplatin, irinotecan and fluorouracil in MPC and colorectal cancer, for either FOLFIRI or FOLFOX common regimens. Transient dysarthria and blepharospasm are rare toxicities of irinotecan infusion, first recognized during the FOLFIRINOX infusion by Gunturu et al.⁴⁹ Japanese investigators observed no cases of dysarthria in clinical trials of FOLFOXIRI in advanced colorectal cancer.⁵⁰ In this regimen, irinotecan is administered prior to oxaliplatin. They suggested that the FOLFIRINOX-induced dysarthria may be associated with the sequence of drug administration: intravenous infusion of oxaliplatin, immediately followed by leucovorin, followed by 30 min irinotecan. However, in vitro data suggest that cytotoxicity was maximal when oxaliplatin was administered prior to SN-38,⁵¹ the active metabolite of irinotecan; therefore we do not recommend the use of irinotecan prior to oxaliplatin except in cases of dysarthria.

Table 3.

Management of FOLFIRINOX toxicity.

Toxicity	Management
Neutropenia	Delay chemotherapy until neutrophils ⩾ 1.5 × 109/l; At the next cycle, remove bolus 5-FU and reduce irinotecan to 150 mg/m²; Use G-CSF or pegfilgrastim if febrile neutropenia
Thrombocytopenia	Consider cycle postponement until platelets > 75 × 109/l; At the next cycle, reduce oxaliplatin dose to 60 mg/m² and infusional 5-FU to 25%
Anemia	Evaluate etiology (e.g. cancer-induced inflammation, iron deficiency); Consider red blood cell transfusion; Erythropoietin is poorly efficient in oxaliplatin-induced anemia
Vomiting	Consider FOLFIRINOX as a highly emetogenic regimen and use corticosteroids, an anti-5HT3 antagonist plus anti-NK1 antagonist to prevent acute emesis; prepare outpatient prescription for delayed emesis; patients with diabetes receiving steroids may require regular checks of blood glucose and adjustment of diabetes medication
Grade 3–4 diarrhea	Early diarrhea or abdominal cramps result from hypercholinergic reactions (see below); Delayed diarrhea: use 2 mg of loperamide (two capsules), then 2 mg every 2 h (outpatient supply); see the supplementary appendix of the pivotal study for details;7 At the next cycle, reduce irinotecan dose to 150 mg/m² and remove the bolus 5-FU dose
Sensory neuropathy	At the next cycle, reduce oxaliplatin to 65 mg/m² for grade 3 neurotoxicity and stop oxaliplatin if persistent neurotoxicity occurs between cycles
Transient dysarthric speech, blepharospasm	Use prophylactic subcutaneous atropine 0.25–0.50 mg; Hold infusion until symptoms resolve; Give a warm drink; Check electrolytes levels for hypokalemia; Consider use irinotecan first, then oxaliplatin for next cycles
Laryngo-pharyngeal dysesthesia Laryngo-pseudospasm throat discomfort	Check oxygen saturation: if normal, an anxiolytic may be given; Slow oxaliplatin infusion rate (6 h) and warm drink; Add intravenous calcium gluconate plus magnesium sulphate (1 g of each) in 100 ml of 5DW over 30 min before and after oxaliplatin
Grade 3–4 mucositis	Reduce bolus and infusional 5-FU to 25%
Hypercholinergic reaction with cramping and sweating	Slow infusion rate; Premedicate with subcutaneous atropine 0.25–0.50 mg
Immediate hypersensitivity reactions to oxaliplatin	Refer patient to an allergy unit for skin testing; Slow down the run rate; Consider use of an antihistamine with or without a steroid; Use desensitization protocol using serial dilutions; Possible discontinuation if severe reaction

G-CSF, granulocyte-colony stimulating factor; 5-FU, 5-fluorouracil; 5DW, solution with 5% dextrose in water.

Future directions of FOLFIRINOX use

In patients with resectable PC, the current standard of care is curative surgery followed by 6 months of adjuvant chemotherapy with gemcitabine or fluorouracil plus leucovorin. However, despite macroscopically curative resection, the long-term results remain poor with a median disease-free survival (DFS) and OS of 14 months and 23 months, respectively. The corresponding 5-year OS rate is 20–24%.²⁰ Due to the survival benefit of FOLFIRINOX over gemcitabine in MPC, we launched the PRODIGE 24/NCIC CTG PA.6 randomized phase III study in 2012, comparing 6-month adjuvant chemotherapy with gemcitabine versus mFOLFIRINOX in patients with resected PC [ClinicalTrials.gov identifier: NCT01526135]. The objective of the trial is to demonstrate a 10% increase in DFS after 3 years. The trial is closed after recruitment of 493 patients. Presently, there are no data to support use of adjuvant FOLFIRINOX in patients with resected PC outside clinical trials.

At diagnosis, borderline resectable PC (BRPC) includes tumors involving the mesenteric vasculature to a limited extent, and those eligible for resection yet likely to be compromised by positive surgical margins (R1) in the absence of induction treatment. Although radiographically localized, they are often associated with subclinical metastases. Induction treatment of BRPC aims to reduce tumor volume prior to surgery in order to improve the chances of radical (R0) resection and reduce the rate of lymph node positivity and recurrences. The FOLFIRINOX regimen as upfront chemotherapy has been evaluated in this setting in many case series⁵² and appears to be a well-tolerated and promising treatment for patients with BRPC. A meta-analysis, specifically on induction of FOLFIRINOX-based chemotherapy in 87 patients with BRPC, showed a promising 63% R0 resection rate without grade 4 toxicity.⁵² Studies have shown a significantly decreased rate of lymph node involvement^53,54 and decreased postoperative morbidity, including a very low rate of pancreatic fistula complications.^53–55 Some series also suggest an improvement in survival.^53,54 Based on these results, three prospective randomized trials are ongoing to validate the use of FOLFIRINOX as induction therapy. The ESPAC-5F [EudraCT number: 2013-003932-56] randomized phase II trial aims to compare upfront surgery with neoadjuvant chemotherapy regimens (gemcitabine plus capecitabine or FOLFIRINOX) in BRPC. The NeoLAP [ClinicalTrials.gov identifier: NCT02125136] and NeoPAN phase III trials [EudraCT number: 2013-005087-25] are currently investigating preoperative GnP compared with FOLFIRINOX in locally advanced PC (LAPC) or BRPC and consecutive tumor conversion to resectability. The US Intergroup Alliance tested the sequence of preoperative mFOLFIRINOX (four cycles) followed by chemoradiotherapy in 23 BRPC patients. Toxicities were manageable and reversible and did not preclude subsequent resection. Pancreatectomy and R0 resection occurred in 65% of patients.⁵⁶ The Alliance Intergroup is running another randomized trial testing mFOLFIRINOX with or without hypofractionated radiation therapy before surgery with BRPC of the head of the pancreas [ClinicalTrials.gov identifier: NCT02839343]. The ongoing PRODIGE 44/PANDAS randomized trial [ClinicalTrials.gov identifier: NCT02676349] is investigating the effectiveness of preoperative concomitant chemoradiotherapy after induction of mFOLFIRINOX in patients with BRPC.

The prognosis of patients with LAPC has historically been poor with a reported median survival of 6–13 months for gemcitabine-treated patients. Induction of FOLFIRINOX may downstage unresectable LAPC to such an extent that it becomes eligible for resection. Patients with BRPC or LAPC showed a promising resection rate of 43% (92 R0) following induction treatment with FOLFIRINOX.⁵² Retrospective studies of patients with both BRPC and LAPC have suggested an RR of approximately 30% with FOLFIRINOX, associated with the ability to perform resection.^57,58 Resection is linked to improved median OS.^58–60 In a monoinstitutional retrospective analysis of 51 patients with BRPC and LAPC, a promising median OS of 35.4 months (95% CI, 25.8–45) was achieved. Increasing number of full-dose cycles was significantly associated with increased survival.⁶⁰

A recent publication outlined a patient-level meta-analysis comprising 315 patients with LAPC treated with FOLFIRINOX.⁵⁷ Resection following FOLFIRINOX treatment occurred in 28% of patients, with an R0 rate of 74%. Patients showed a median PFS of 15.0 months (95% CI, 13.7–16.3) and a promising median OS of 24.2 months (95% CI, 21.7–26.8). However, it remains unclear whether such improvement in OS is related to tumor downstaging and subsequent resection, selection of fit patients or due to more effective treatment of micrometastases. Despite the lack of reported data from controlled trials, the ASCO guidelines recommend combination regimen chemoradiotherapy as initial systemic therapy in patients with ECOG PS 0 or 1 and a favorable comorbidity profile. Randomized studies are currently assessing the role of radiation in patients who have responded to induction. Authors of the CONKO-007 trial [ClinicalTrials.gov identifier: NCT01827553] aim to examine the effectiveness of chemoradiotherapy compared with chemotherapy alone after induction chemotherapy with either gemcitabine or FOLFIRINOX in patients with LAPC. A phase III study at Stanford University [ClinicalTrials.gov identifier: NCT01926197] is allocating patients with LAPC to receive mFOLFIRINOX alone versus mFOLFIRINOX followed by stereotactic body radiotherapy.

In patients with resectable PC, curative surgery followed by adjuvant chemotherapy is currently the standard of care. However, long-term results remain poor, with a median DFS and OS of 14 months and 23 months, respectively, and a corresponding 5-year OS rate of 20%. Neoadjuvant chemotherapy with FOLFIRINOX in patients with resectable PC offers advantages over upfront surgery: patients receive early systemic therapy to eradicate subclinical metastases; tumor downstaging potentially facilitates complete resection (R0) and futile surgery is avoided in patients with evolving metastases. However, no prospective study in resectable PC has been reported yet. Three controlled trials are ongoing to assess the feasibility and potential benefits of neoadjuvant FOLFIRINOX. PANACHE-01 [ClinicalTrials.gov identifier: NTC02959879] is a randomized phase II study designed to assess the safety and efficacy of two neoadjuvant regimens, FOLFIRINOX and FOLFOX, compared with surgery and adjuvant chemotherapy. SWOG S1505 [ClinicalTrials.gov identifier: NCT02562716] is an ongoing randomized phase II study of perioperative mFOLFIRINOX versus gemcitabine/nab-paclitaxel to select the better therapy for resectable pancreatic adenocarcinoma. The Nepafox trial [ClinicalTrials.gov identifier: NCT02172976] is a randomized phase II/III study comparing surgery plus adjuvant gemcitabine with six cycles of neoadjuvant FOLFIRINOX followed by surgery, then six cycles of adjuvant FOLFIRINOX. The primary endpoint is OS and the trial requires the recruitment of 310 patients during the phase III step. Presently, the use of neoadjuvant therapies in patients with resectable PC is recommended only in the context of a multidisciplinary approach, and within clinical trials.

Predictive biomarkers to guide FOLFIRINOX use

Nowadays, no conclusive evidence has yet been obtained on predictive or monitoring tools to optimize PC chemotherapy. Candidates’ micro RNA have been evaluated to predict the sensitivity/resistance to the drugs of FOLFIRINOX and GnP regimens, but data are still controversial.⁶¹

Conclusion

Since 2011, the landscape of systemic treatment in MPC has significantly changed. FOLFIRINOX has emerged as the optimum initial approach for fit patients with a well-preserved PS. At present, the choice of initial chemotherapy regimen is still based on careful patient selection as no biomarkers in the tumor or blood can reliably predict patient response to FOLFIRINOX. Several publications report FOLFIRINOX as a potentially attractive regimen for induction treatment of BRPC patients, especially for cases with no option of entering a clinical trial. The R0 resection rates and survival rate observed in patients with LAPC treated with the FOLFIRINOX regimen are promising, marking the beginning of a new era in the treatment strategy of PC in appropriately selected patients.