FOLFIRINOX De‐Escalation in Advanced Pancreatic Cancer: A Multicenter Real‐Life Study

Abstract

Background

Our study describes the feasibility and efficacy of a first‐line FOLFIRINOX (5‐fluorouracil [5FU], folinic acid, irinotecan, and oxaliplatin) induction chemotherapy (CT) followed by de‐escalation as a maintenance strategy for advanced pancreatic cancer.

Materials and Methods

This multicenter retrospective study was conducted from January 2011 to December 2018. FOLFIRINOX de‐escalation was defined as stopping oxaliplatin and/or irinotecan after at least four cycles of FOLFIRINOX, without evidence of disease progression. Maintenance schedules were fluoropyrimidine monotherapy (intravenous or oral [capecitabine]), FOLFOX (5FU, oxaliplatin), or FOLFIRI (5FU, irinotecan). Primary endpoint was overall survival (OS). Secondary endpoints were first progression‐free survival (PFS1), second progression‐free survival (PFS2), and toxicity.

Results

Among 321 patients treated with FOLFIRINOX, 147 (45.8%) were included. Median OS was 16.1 months (95% confidence interval [CI], 13.7–20.3) and median PFS1 was 9.4 months (95% CI, 8.5–10.4). The preferred maintenance regimen was FOLFIRI in 66 (45%) patients versus 5FU monotherapy in 52 (35%) and FOLFOX in 25 (17%) patients. Among 118 patients who received maintenance CT with FOLFIRI or 5FU, there was no difference in PFS1 (median, 9.0 vs. 10.1 months, respectively; p = .33) or OS (median, 16.6 vs. 18.7 months; p = .86) between the two maintenance regimens. Reintroduction of FOLFIRINOX was performed in 20.2% of patients, with a median PFS2 of 2.8 months (95% CI, 2.0–22.3). The rates of grade 3–4 toxicity were significantly higher with FOLFIRI maintenance CT than with 5FU (41% vs. 22%; p = .03), especially for neuropathy (73% vs. 9%).

Conclusion

5FU monotherapy maintenance appeared to be as effective as FOLFIRI, in a FOLFIRINOX de‐escalation strategy, which is largely used in France.

Implications for Practice

FOLFIRINOX de‐escalation and maintenance is a feasible strategy in advanced pancreatic cancer that decreases chemotherapy toxicity to improve both survival and quality of life. Survivals in patients with maintenance therapy are clinically meaningful. Fluoropyrimidine monotherapy maintenance seems to be as efficient as FOLFIRI and should be a reference arm in future pancreatic cancer maintenance trials.

Introduction

Pancreatic adenocarcinoma is expected to become the second leading cause of cancer‐related death in the U.S. and Europe by 2030 [1, 2]. Despite recent progress, prognosis remains poor, with a 5‐year overall survival (OS) rate, all stages combined, of 5% to 7% [3]. In 2011, substantial progress in survival was made with the use of FOLFIRINOX (5‐fluorouracil [5FU] combined with folinic acid, irinotecan, and oxaliplatin) as a first‐line treatment, after the results of the PRODIGE4/ACCORD11 trial in patients with metastatic pancreatic cancer (aPC) [4]. FOLFIRINOX was compared with gemcitabine and showed an improvement in median OS of 4.3 months (11.1 months vs. 6.8 months; p < .001) as well as in the quality of life of the patients [5]. However, triplet chemotherapy is associated with a higher burden of toxicities, including grade 3–4 neutropenia (45.7%), vomiting (14.5%), diarrhea (12.7%), and peripheral neuropathy (9%) [4]. Thus, in patients who achieve longer survival, the challenge of cytotoxic treatments is to reach a compromise between quality of life and disease control. Modified doses of FOLFIRINOX (bolus removal and reduced dose of irinotecan) did not decrease survival but resulted in fewer toxicities [6]. This protocol is the preferred first‐line regimen in France, where access to gemcitabine‐nab‐paclitaxel, the alternative active first‐line regimen, is limited because of reimbursement issues [7, 8].

The concept of maintenance generally covers the strategies of (a) therapeutic de‐escalation (continuation maintenance) and (b) introducing a different molecule (switch maintenance) after a maximum response to the induction chemotherapy [9]. This concept is part of a therapeutic top‐down objective, which aims to decrease the amount and therapeutic intensity while maintaining efficacy. To date, this strategy has been underevaluated in aPC but is used in other cancers such as colon [10], lung [11], and head and neck cancers [12], making it possible to maintain antitumoral pressure while reducing toxicities [9]. A few studies have addressed the maintenance in aPC: Reni et al. [13] sought to show the benefit of maintenance with sunitinib after chemotherapy, whereas Petrioli et al. [14] demonstrated that maintenance with gemcitabine after doublet chemotherapy with gemcitabine and nab‐paclitaxel was feasible in older patients.

The first prospective phase II trial PRODIGE35‐PANOPTIMOX investigating the feasibility of a de‐escalation strategy in aPC demonstrated the feasibility of maintenance with 5‐fluorouracil leucovorin (LV5FU2) after an induction strategy of eight cycles of FOLFIRINOX, without compromising survival (OS, 11.2 vs. 10.1 months) [15]. However, the study population had been selected for a clinical trial and differed from that of the clinical routine. Currently, there are no real‐life data on therapeutic de‐escalation practices in aPC.

We conducted a retrospective multicenter study whose main objective was to provide a descriptive overview of the feasibility and efficacy results of therapeutic de‐escalation of FOLFIRINOX in aPC.

Materials and Methods

Study Design and Population

We performed a retrospective study in five French centers: three university hospitals (Lille University Hospital, St. Vincent de Paul Hospital in Lille, Besançon University Hospital) and two comprehensive centers (Oscar Lambret Centre in Lille and Eugène Marquis Centre in Rennes). The study population included all consecutive patients with aPC (locally advanced or metastatic) who received FOLFIRINOX between January 2011 and December 2018, and for whom the protocol was reduced after at least four cycles of FOLFIRINOX. De‐escalation was performed using oral (capecitabine) or intravenous (LV5FU2) fluoropyrimidine, FOLFIRI (LV5FU2, irinotecan), or FOLFOX (LV5FU2, oxaliplatin). Patients under 18 years of age, those who had received less than four cycles of FOLFIRINOX, and those who had a progression disease on FOLFIRINOX were excluded. As the number of patients included in the FOLFOX group or those who had received treatment other than 5FU monotherapy or FOLFIRI was low, we focused our attention on patients who had received de‐escalation with fluoropyrimidine or FOLFIRI. We investigated whether de‐escalation that was performed after partial response (according to RECIST version 1.1) or stable disease (according to RECIST version 1.1) under FOLFIRINOX was sufficient to consider therapeutic decrementation. The primary endpoint was OS, and the secondary endpoints were first progression‐free survival (PFS1), second progression‐free survival (PFS2) in the event of FOLFIRINOX reintroduction, and toxicity.

Treatment efficacy was evaluated by computed tomography scan of the thorax, abdomen, and pelvis every 3 months. The data collected included the general characteristics of the population, metastatic or nonmetastatic status at diagnosis and at the different lines of treatment, type of treatment received, date of introduction and progression, presence and type of toxicities, notion of de‐escalation, if applicable the presence of a FOLFIRINOX reintroduction, and the date of death or date of last news. A search for prognostic factors for maintenance was also performed.

The French data protection authority (Comission Nationale de l'Informatique et des Libertés agreement no. 1595361) provided a waiver of informed consent for this retrospective study and permitted the publication of anonymized data.

Statistical Analysis

Median value (interquartile range) and frequency (percentage) were provided for the description of continuous and categorical variables, respectively. Medians and proportions were compared using Student's t test and chi‐square test (or Fisher's exact test, if appropriate), respectively. OS was calculated from the date of the first administration of first‐line therapy to date of death from any cause, or the date of the last follow‐up, at which point data were censored. PFS1 was defined as the time between the start of the first cycle of FOLFIRINOX and the first objective progression (RECIST version 1.1) of the tumor or death, whichever occurred first. PFS2 was defined as the time from reintroduction of FOLFIRINOX after maintenance therapy to objective tumor progression or death, whichever occurred first. Survival data were censored at the last follow‐up. OS and progression‐free survival were estimated using the Kaplan‐Meier method, described using median or rate at specific time points with 95% confidence intervals (CIs), and compared using the log‐rank test. Follow‐up time was estimated using a reverse Kaplan‐Meier estimation when feasible. Objective tumor response was determined according to RECIST version 1.1. Toxicity was evaluated according to the National Cancer Institute Common Terminology Criteria version 4.0.

Cox proportional hazard models were performed to estimate hazard ratio and 95% confidence interval for factors associated with OS. The association of baseline parameters with OS was first assessed using univariate Cox analyses, and then parameters with p values of less than .05 were entered into a final multivariable Cox regression model. All analyses were performed using SAS software version 9.4 (SAS Institute, Cary NC, USA). A value of p < .05 was considered statistically significant; all tests were two‐sided.

Results

Population Characteristics

We included 147 (46%) patients treated with FOLFIRINOX as first‐line therapy, who received therapeutic de‐escalation after at least four cycles of FOLFIRINOX (Fig. 1). The median age was 60.0 years (53.1–35.7). At the initiation of FOLFIRINOX, 32 (21.8%) patients had locally advanced pancreatic cancer, and 115 (78.2%) had metastatic pancreatic cancer. The median total number of cycles of induction chemotherapy was 9.0 (6.0–11.0). Of these 147 patients, 66 (44.9%) received oral (capecitabine) or intravenous (LV5FU2) 5FU, 52 (35.4%) received FOLFIRI (5FU, folinic acid, and irinotecan), 25 (17%) received FOLFOX (5FU, folinic acid, and oxaliplatin), and 4 (2.7%) received other maintenance protocols, mainly olaparib in clinical trials. The population of the FOLFIRI group was older, and performance status (PS) was higher than that of the 5FU group (Table 1).

De‐Escalation Strategy

In the de‐escalation strategy population, median OS was 16.1 months (95% CI, 13.7–20.3) and median PFS1 was 9.4 months (95% CI, 8.5–10.4) (Fig. 2). There was no statistically significant improvement in OS and PFS1 depending on whether maintenance was started after 12 cycles of FOLFIRINOX or earlier (median OS, 20.5 vs. 15.0; p = .2362; median PFS1, 13.2 vs. 8.8 months; p = .4234) (supplemental online Fig. 1). Patients who received maintenance with FOLFIRI and 5FU had similar survivals (median OS, 18.7 vs. 16.6 months; p = .8678; median PFS1, 9.0 vs. 10.1, respectively; p = .3327) (supplemental online Fig. 2). On the other hand, there appeared to be a decrease in OS and PFS1 when de‐escalation was performed with FOLFOX, compared with FOLFIRI or 5FU (median OS, 11.8 vs. 18.7 and 16.6 months; p = .5590; PFS1, 6.7 vs. 9.0 and 10.1 months; p = .0265) (Fig. 3). PFS1 was similar whether there was a response or stability under FOLFIRINOX, regardless of the chemotherapy regimen (5FU or FOLFIRI) (p = .5857) (Fig. 4).

Discontinuation of de‐escalation therapy was mostly due to disease progression (n = 108 [74%]). Six (4.1%) patients stopped the treatment because of grade 3–4 toxicities, and 32 (21.9%) stopped treatment for other reasons, such as altered general condition or in relation to the oncologist's assessment (Table 1).

Adverse Events

In de‐escalation population, 53 (37.6%) patients had grade 3–4 toxicities, most of which were digestive (n = 21 [39.6%]) and neurological (n = 16 [30.2%]). Eight (15.1%) patients had hematological toxicity (Table 2).

Among the 118 patients who received maintenance with FOLFIRI or 5FU, 37 (31.4%) had grade 3–4 toxicities, including 26 (41.3%) in the FOLFIRI group and 11 (22%) in the 5FU group. Toxicities in the FOLFIRI maintenance group were mainly neurological (n = 19 [73.1%]). In the 5FU group, toxicities were hematological (n = 5 [45.5%]) and digestive (n = 3 [27.3%]) (Table 2).

FOLFIRINOX Reintroduction

After progression under maintenance therapy by 5FU or FOLFIRI, reintroduction by triplet (FOLFIRINOX) or doublet of chemotherapy was performed in 28.1% of patients; that is, 5 (7.6%) received de‐escalation with FOLFIRI, and 27 (51.9%) received de‐escalation with 5FU. In the FOLFIRI regimen, four patients had reintroduction by FOLFIRINOX, and one had intensification by FOLFIRI‐3 (irinotecan 100 mg/m² Day 1 and Day 3, folinic acid 400 mg/m² Day 1, continuous 5FU 2,000 mg/m² Day 1‐Day 2). In the 5FU maintenance group, 19 patients (70.4%) had reintroduction by FOLFIRINOX, 6 (22.2%) by FOLFIRI, and 2 (7.4%) by FOLFOX (Table 1). The median PFS2 in the 5FU maintenance group was 2.8 months (95% CI, 2.0–20.5). Data were not available in the FOLFIRI group (p = .2934) (Fig. 5).

Prognostic Factors

The search for prognostic factors was carried out by univariate analysis on the 118 patients who received de‐escalation with 5FU or FOLFIRI. Demographic parameters and tumor characteristics at diagnosis, whether clinical, radiological, or biological, were not associated with increased survival (supplemental online Table 1). Similarly, the number of FOLFIRINOX cycles received, best response to FOLFIRINOX, and the presence of grade 3 or 4 toxicities were not significant prognostic factors.

Discussion

We aimed to describe the conditions of maintenance therapy in advanced pancreatic cancer in France. In our study, a large proportion of patients received therapeutic de‐escalation after at least four cycles of FOLFIRINOX, showing that this strategy is widely used by French oncologists. Considering the limitations of a retrospective study, 5FU maintenance seems to be as effective as FOLFIRI. Previously, Reure et al. [16] showed that de‐escalation of FOLFIRINOX after four to eight cycles with capecitabine was feasible. The median OS was 17 months and median PFS1 was 5 months. Franck et al. [17] analyzed survival in patients who received a maintenance strategy with FOLFIRI after 2 to 6 months of treatment with FOLFIRINOX regimen. In this cohort of 22 patients, the median PFS1 (considering FOLFIRINOX induction and subsequent FOLFIRI maintenance therapy) was 11 months. Another retrospective study published by Hann et al. [18] showed a PFS1 of 10.6 months (95% CI, 6.7–14.4) and an OS of 18.3 months (95% CI, 14.8–21.8) in a cohort of 13 cases in which patients received de‐escalation treatment with 5FU after FOLFIRINOX regimen. Our results were obtained in a real‐life population with inclusion starting before the presentation of the first results of the PRODIGE35 trial [15]. In this phase II trial, patients were randomized into three arms: 12 cycles of FOLFIRINOX (arm A), 8 cycles of FOLFIRINOX followed by maintenance with 5FU and leucovorin (LV5FU2) with the possibility of reintroducing FOLFIRINOX at disease progression (arm B), and sequential treatment with gemcitabine and FOLFIRI‐3 (arm C). Progression‐free survival at 6 months in arms A and B (47% and 44%) and median OS (10.1 and 11.2 months) were similar, whereas arm C appeared inferior. However, the neurotoxicity rate was higher in arm B after 6 months of treatment, mainly because of the higher number of oxaliplatin cycles received by the patients in this arm with FOLFIRINOX reintroduction. We observed different results in our study, with a significantly higher grade 3–4 toxicity rate with FOLFIRI maintenance than that with 5FU (41% vs. 22%; p = .03), especially for the neuropathy (73% vs. 9.1%; p = .03). These toxicities must be associated with FOLFIRINOX induction chemotherapy, especially with oxaliplatin for neuropathy.

In colorectal cancer, de‐escalation of LV5FU2 treatment in responder patients after six cycles of FOLFOX reduced toxicities in OPTIMOX trials. This strategy also improved progression‐free survival compared with patients in whom treatment was suspended after six to eight cycles [19, 20]. In our study, the median OS (from the beginning of FOLFIRINOX) for all de‐escalation regimens (i.e., 5FU, FOLFIRI, FOLFOX) was 16.4 months (95% CI, 13.7–20.3), and the median PFS1 was 8.8 months (95% CI, 8.3–9.7). These survivals were greater than those presented in the PRODIGE4/ACCORD11 trial as well as in the PRODIGE35‐PANOPTIMOX trial [15] and similar to those shown in the retrospective study by Reure et al. [16]. The major limitation was the exclusion of early progressing patients, who were not able to receive a de‐escalation regimen. Furthermore, our study included patients with both locally advanced and metastatic aPC (vs. only patients with metastatic aPC in PRODIGE4/ACCORD11 and PRODIGE35‐PANOPTIMOX), whereas the OS of locally advanced pancreatic cancer was expected to be more favorable (even if this was not observed in our study), which introduces a new bias for the interpretation of OS [21].

An interesting finding was the no obvious difference in survival between the FOLFIRI and 5FU maintenance groups, although patients’ characteristics were not in favor of FOLFIRI (older and higher PS). Oral or intravenous 5FU is classically better tolerated than a FOLFIRI regimen, which is an additional argument to encourage oncologists to consider a therapeutic de‐escalation by 5FU. There was more reintroduction in 5FU group than in FOLFIRI group (51.9% vs. 7.6%; p < .0001) suggesting that this schedule was better tolerated than FOLFIRI. However, the higher reintroduction rate was not associated with higher survival. We also observed that patients with stable disease and those with objective response had similar survival outcomes, suggesting that FOLFIRINOX de‐escalation with 5FU or FOLFIRI was appropriate whatever the tumor response, once disease control has been achieved after at least four cycles of induction chemotherapy. Finally, we did not find any prognostic factors that would allow better patient selection; however, these prognostic and predictive factors of response to maintenance should be studied prospectively by conducting ancillary studies of robust clinical trials such as PRODIGE35. Nevertheless, these interesting data from clinical practice support the development of further prospective maintenance studies, either de‐escalation or switch maintenance, in order to improve therapeutic strategies for patients with aPC, maintaining tumor control while reducing toxicities. Thus, the 5FU arm may be a reasonable reference arm in future randomized maintenance trials in aPC [22].

Conclusion

We have shown that the de‐escalation and maintenance strategy in aPC is currently widely accepted by French oncologists. In this trial, 5FU monotherapy de‐escalation under FOLFIRINOX appeared to have similar results as those of FOLFIRI and may be an option in clinical routine and as a reference arm in maintenance trials. Maintenance trials should be encouraged in aPC to establish this therapeutic strategy in order to improve both therapeutic efficacy and quality of life of patients.

Acknowledgments

We would like to thank Editage (www.editage.com) for English language editing.

Author Contributions

Hortense Chevalier, Angélique Vienot, Astrid Lièvre, Julien Edeline, Farid El Hajbi, Charlotte Peugniez, Dewi Vernerey, Aurélia Meurisse, Pascal Hammel, Cindy Neuzillet, Christophe Borg, Anthony Turpin

Conception/design: Hortense Chevalier, Angélique Vienot, Astrid Lièvre, Dewi Vernerey, Aurélia Meurisse, Pascal Hammel, Cindy Neuzillet, Christophe Borg, Anthony Turpin

Provision of study material or patients: Julien Edeline, Farid El Hajbi, Charlotte Peugniez, Anthony Turpin

Collection and/or assembly of data: Hortense Chevalier, Angélique Vienot

Data analysis and interpretation: Hortense Chevalier, Angélique Vienot, Aurélia Meurisse, Anthony Turpin

Manuscript writing: Hortense Chevalier, Angélique Vienot, Astrid Lièvre, Julien Edeline, Aurélia Meurisse, Pascal Hammel, Cindy Neuzillet, Anthony Turpin

Final approval of manuscript: Hortense Chevalier, Angélique Vienot, Astrid Lièvre, Julien Edeline, Farid El Hajbi, Charlotte Peugniez, Dewi Vernerey, Aurélia Meurisse, Pascal Hammel, Cindy Neuzillet, Christophe Borg, Anthony Turpin

Disclosures

Dewi Vernerey: GERCOR, HalioDX, Incyte, CellProthera (C/A); Cindy Neuzillet: Servier, AstraZeneca, Bristol‐Myers Squibb, Amgen, Merck, Merck Sharp & Dohme, Novartis, Incyte, Nutricia, Baxter, Fresenius Kabi, Kiplin (C/A), OSE Immunotherapeutics, Roche, AstraZeneca (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

Author notes