Meta-analysis of FOLFIRINOX-based neoadjuvant therapy for locally advanced pancreatic cancer

Supplemental Digital Content is available in the text


Introduction
Pancreatic cancer (PC) is among the most malignant cancers, characterized by rapid progression, poor prognoses, high postoperative recurrence rates, and 5-year survival rates lower than 5%. [1] PC accounted for 4.5% of all cancer-related deaths in 2018, [2] and is expected to become the second leading cause of cancer-related deaths by 2030. [3] The diagnosis of this disease in the early stages is difficult and only 15% to 20% of patients undergo surgery at first diagnosis. [4,5] However, even after radical surgery, the 5-year survival rate associated with the disease is only 15% to 25%. [6] Neoadjuvant chemotherapy for PC, a hot topic in clinical research, has been shown to improve the prognoses of patients, through reductions in the development rate of tumor lesions, increases in the R0 resection rate, reductions in the rates of vascular invasion and micrometastasis, and decreases in the incidence of postoperative complications. [7][8][9] The combination of 5-fluorouracil, leucovorin, irinotecan, and oxaliplatin (FOL-FIRINOX) is increasingly gaining importance in the treatment of metastatic pancreatic cancer (MPC). A randomized controlled trial (RCT) from the United States (PROD-IGY 4/ ACCORD 11) showed that MPC patients treated with FOLFIRINOX had a better objective response rate (ORR) (31.6% vs 9.4%, P < .001) and median survival duration (11.1 vs 6.8 months, P < .001) than those with gemcitabine monotherapy. [10] In addition, FOLFIR-INOX can improve the quality of life of MPC patients compared with gemcitabine alone. [11] Given its efficacy in MPC treatment, FOLFIRINOX is now widely used for patients with LAPC, as it has the potential to reduce the disease stage, transform it to borderline resectable or even resectable pancreatic cancer, and improve the associated surgical resection rate. [12] Furthermore, a meta-analysis also indicated that FOLFIRINOX had an advantage in improving the median overall survival (OS) of patients with LAPC. [13] However, most previous studies on the topic had an insufficient sample size, owing to which definitive conclusions about the efficacy and safety of FOLFIRINOX in LAPC patients could not be drawn.
The aim of the present meta-analysis was to assess the efficacy of FOLFIRINOX as a first-line chemotherapy regimen for patients with LAPC.

Search strategy
This meta-analysis was PRISMA-compliant. The ethics committee (Medical Ethics Committee of Yongchuan Hospital, Chongqing Medical University) approved the study. We systematically searched PubMed, EMBASE, and the Cochrane Library for relevant studies published from database establishment to January 1, 2020, including RCTs, clinical controlled studies, and cohort studies, among others. The search terms included "FOLFIRINOX," "fluorouracil," "irinotecan," "oxaliplatin," "pancreatic cancer," "pancreatic neoplasm," "drug combination," and relevant variants thereof. Only studies published in English were included. There were no restrictions on the population, race, or publication date. Potential eligible studies were identified by searching the references of the selected studies. The search was completed independently by 3 authors. Specific details on the search strategy are presented in Appendix 1, Supplemental Content (http://links.lww.com/MD/F496).

Selection criteria and data extraction
All the included studies focused on the effectiveness of FOLFIR-INOX in patients with LAPC. The following selection criteria were applied for the included studies: study design: RCT, cohort study, clinical controlled study, etc, presence of at least 1 patient with LAPC, accurate LAPC diagnosis by imaging or pathology prior to chemotherapy, and use of FOLFIRINOX as the first-line chemotherapy regimen for LAPC. In addition, studies were excluded if they met the following criteria: study design: case report, review, conference abstract, and republication, lack of data on resection rate or R0 resection rate, and insufficient data on the outcome of interest, or impossibility of the calculation of the outcome of interest. Two reviewers assessed the titles and abstracts independently for eligibility, and the full tests were further assessed to check if they met the selection criteria. Disagreements were resolved through discussions with a third reviewer. A predefined data collection form was used for the extraction of data from the selected studies. The primary outcomes were the rates of resection and R0 resection after first-line FOLFIRINOX treatment for LAPC. The secondary outcomes were the rates of response, median OS, median progression-free survival (PFS), and grade 3 to 4 adverse events. Other collected information included that on the first author, year of publication, type of study, total sample size, number of patients treated with FOLFIRINOX, median age, performance status, treatment regimen, tumor stage, median chemotherapy duration, and follow-up duration. The characteristics of the selected studies are shown in Table 1.

Assessment of the methodological quality of the included studies
The methodological quality of the included studies was evaluated independently according to the Newcastle-Ottawa scale (NOS) by 2 reviewers. [14] The NOS comprises 3 factors: patient selection, comparability of the study groups, and evaluation of results, including 8 items with a full score of 9. A score lower than 4 is indicative of a low study quality and that higher than 7 reflects a high quality. The NOS scores of the included studies are shown in Table 1. Details on the NOS scale are presented in Appendix 2, Supplemental Content (http://links.lww.com/MD/ F496), and the specific NOS scores of the included studies are presented in Appendix 3, Supplemental Content (http://links. lww.com/MD/F496).

Statistical analysis
We used the meta package of R 3.6.2 software for the data analysis. The heterogeneity of the included studies was assessed using the x 2 -based Q test and I2 test. [15,16] The level of heterogeneity was considered to be significantly different at I2 > 50% or P < .1 in the Q test. According to Cochrane review guidelines, the level of heterogeneity was considered significant at I2 > 50%, and the random effects model was selected. Otherwise, the fixed effects model was used to evaluate the 95% confidence intervals (CIs). Additionally, sensitivity analyses of the surgical resection rates were performed by the removal of each study for the assessment of the quality and stability of the results. P < .1 was considered statistically significant. Publication bias was assessed using funnel plots (See Appendix 7, Supplemental Content, http://links.lww.com/MD/F496).

Literature search and study characteristics
A total of 1302 studies were identified through the database search; 248 duplicate references were removed and 999 references were excluded after the titles and abstracts were read. The remaining 57 full-text articles were further assessed for eligibility. Finally, 21 articles were included in the meta-analysis.  A flow chart of the literature search is shown in Figure 1. The characteristics of the included studies are shown in Table 1. A total of 1013 patients were included from the 21 studies, including 75 patients with borderline resectable pancreatic cancer (BRPC), 270 with MPC, 653 with LAPC, and 15 with tumor recurrence. A majority of the patients had an Eastern Cooperative Oncology Group (ECOG) performance status score of 0 or 1. Overall, 51.0% of the patients were male, 44.6% were female, and 4.4% had an unknown gender. Nearly half of the patients (44.7%) were from the United States. All the studies reported the median age of the patients, which ranged from 58 to 67 years, with the youngest patient aged 34 years and oldest 81 years. One study was a phase II multicenter study, 3 were prospective cohort studies, and 17 were retrospective cohort studies. Eight studies focused on LAPC, 5 on BRPC and LAPC, 4 on LAPC and MPC, 3 on BRPC, LAPC, and MPC, and 1 on LAPC, MPC, and tumor recurrence. LAPC was defined by National Comprehensive Cancer Network (NCCN) criteria in 9 studies, [17] Americas Hepato-Pancreato-Biliary Association/Society of Surgical Oncology/Society for Surgery of the Alimentary Tract (AHPBA/SSO/SSAT) criteria in 3 studies, [18] and other criteria in 9 studies. Sixteen studies reported the median number of FOLFIRINOX cycles administered to patients with LAPC, which ranged from 4.9 to 11.5. And of these studies, 12 reported the median number of FOLFIRINOX cycles was 6 or higher. In the 12 studies, 280 patients received additional radiation after chemotherapy, and in 6 of these, 146 patients received it in combination with gemcitabine, oxaliplatin, capecitabine, or 5fluorouracil as a radiosensitize, Thirty-one patients from 2 studies were administered stereotactic radiotherapy, of whom 12 patients received a dose of 36 Gy in 3 fractions and the other 19 received a dose of 33 Gy in 5 fractions. Furthermore, 6 patients in 1 study received intraoperative radiotherapy.

Discussion
In our meta-analysis, which included 21 studies involving 648 patients with LAPC who received FOLFIRINOX as first-line chemotherapy, we observed an ORR of 34% (95% CI = 25%-43%, I2 = 56%), which was significantly higher than the value associated with gemcitabine treatment (9.4%). [10] In addition, previous studies have shown that objective efficiency can improve patients' prognoses and survival, suggesting that the FOLFIR-INOX approach rather than the gemcitabine-based approach may have potential benefits, in terms of survival outcomes. [40,41] PC is systemic in nature and up to 85% of those with the disease are diagnosed with tumors that involve local arteries or distant metastases [42,43] ; therefore, palliative chemotherapy has become the mainstay in the treatment of advanced PC. In 1997, a randomized trial by Burris et al [44] confirmed that patients receiving gemcitabine monotherapy had a slight advantage in terms of median OS over those receiving 5-fluorouracil monotherapy in LAPC and MPC settings (5.6 vs 4.4 months, P < .001). In the years that followed, gemcitabine became the standard treatment for MPC and LAPC. Chauffert et al [45] evaluated gemcitabine as a first-line treatment for LAPC, and showed a median OS duration of 6 to 13 months. However, in Table 3 G3 to G4 adverse events * of nonhematologic.   G3 to G4 adverse events * of hematologic.
Author Total number  Neutr-openia  Febrile neutropenia  Anemia  Throm-bocytopenia  Infections   Hosein  12  4  3  2  3  -Peddi  17  12 3 2011, an RCT showed that FOLFIRINOX treatment improved the PFS (6.4 vs 3.3 months) and OS (11.1 vs 6.8 months) durations in MPC patients to a greater degree than gemcitabine monotherapy. [10] As a majority of LAPC patients show better performance rates than MPC patients, a growing number of studies are now using FOLFIRINOX as the first-line chemotherapy regimen for patients with LAPC. Suker et al [13] conducted a patient-level meta-analysis that evaluated the role of FOLFIR-INOX in LAPC patients, including 11 studies involving 315 patients, and demonstrated a median OS duration of 24.0 months (95% CI = 21.7-26.8 months). Subsequently, Suker et al [39] conducted a cohort study that contrasted with the gemcitabine scheme investigated by Chauffert  PC is associated with high mortality values, and most patients with early-stage disease tend to die. Surgical resection is the only chance for cure in PC, although few patients are eligible for surgery. In a previous meta-analysis of PC patients who underwent surgical resection, a survival duration of 3 to 5 years was observed, which was longer than that noted among those who did not undergo surgical resection. [46] Therefore, for a large number of patients with LAPC, surgical resection yields the highest long-term values. [47] In the past, only 1% to 5% of LAPC patients underwent complete surgical resection after single-drug neoadjuvant chemotherapy. [48,49] Fortunately, a growing number of studies are now reporting that LAPC patients who receive neoadjuvant FOLFIRINOX may potentially be able to undergo surgical resection as a result of tumor downstaging. All the 21 studies included in the current meta-analysis reported surgical resection rates ranging from 6.3% to 61.1% in LAPC patients with neoadjuvant FOLFIRINOX chemotherapy. The surgical resection rates reported across different studies show an obvious degree of heterogeneity. At the same time, in our sensitivity analysis of the rate of surgical resection, we found that when 2 studies were excluded (Ulusakarya et al and Lakatos et al), the level of heterogeneity of the results decreased, but there was no directional change. One reason for this difference may be the lack of consensus regarding the resectability criteria after neoadjuvant therapy. Therefore, for the performance of more accurate comparisons, future studies may need to reach a consensus on the resectability criteria. In the present meta-analysis, the surgical resection rate in the LAPC patients with FOLFIRINOX as first-line chemotherapy was 26% and 88% of these patients underwent R0 resection. The R0 resection rate was even higher than that reported by Gillen et al [50] in patients with a resectable status (88% vs 80%). This indicates that the use of the FOLFIRINOX regimen significantly increases the chance of R0 resection in LAPC patients. In addition, we performed subgroup analyses of the surgical resection rates and R0 resection rates using the median number of FOLFIRINOX cycles as a grouping factor to evaluate whether the number of FOLFIRINOX cycles influences the rates of surgical resection and R0 resection; however, no statistically significant differences were observed. This may be attributed to the small sample size of the included studies. In addition, Janssen et al [51] showed that a median number of FOLFIRINOX cycles were greater than or equal to 6, with a median OS of 21.4 months (95% CI = 16.7-36.0 months), compared with those observed of 21.7 months (95% CI = 15.0-28.4 months) in a study with a median number of chemotherapy cycles lower than 6, with no statistical difference. Therefore, further studies are required to investigate whether the median number of FOLFIRINOX cycles is an important factor in the determination of the rates of surgical resection and median OS.
In addition, in terms of the rate of local tumor progression, we found that 28% of the patients received additional radiotherapy or chemoradiotherapy after FOLFIRINOX treatment. There was no significant difference in the degree of resectability between patients who received radiotherapy or chemoradiotherapy and those who did not. A meta-analysis of 14 phase II clinical trials showed that the rates of surgical resection and R0 resection were 32% and 20%, respectively, in LAPC patients after neoadjuvant chemotherapy, with or without radiotherapy. [52] Meanwhile, in another recently published study, 26% of 215 patients with LAPC who received chemoradiotherapy underwent surgical resection and 10% achieved R0 resection. [53] Thus, although the use of radiotherapy or chemoradiotherapy may be convincing for local control, the unresectable status persists after treatment in many patients. Therefore, the role of these treatments in locally advanced disease needs further clarification in future studies. [54] Although the effect of FOLFIRINOX is superior to that of gemcitabine in patients with metastatic PC, its toxicity somewhat hinders its clinical application. In our study, the most commonly observed grade 3 to 4 adverse event was neutropenia, which showed an incidence that was similar to that reported by Suker et al (20% vs 19.6%). Granulocyte-colony stimulating factor is widely used in the prevention of this hematological toxicity. In addition, in a majority of the studies we included, the dose of the chemotherapeutic drugs was reduced based on the patients' tolerance level, with the aim of reducing the rates of adverse events and improving the efficacy of chemotherapy. It has been demonstrated that the modified FOLFIRINOX regimen yields satisfactory results in patients with an ECOG performance status score of 0 to 1. [55] This study has some limitations that should be considered. First, the sample size of the included studies was relatively small, with a high level of heterogeneity, and a majority of the studies had a retrospective design (17/21, 81%), which may challenge the accuracy of the study results and lead to their overestimation. Second, different criteria were used in the diagnosis of LAPC. Most of the studies referred to NCCN or AHPBA/SSO/SSAT criteria, while some used other criteria, affecting the determination of the outcome indicators. Third, a majority of the studies did not clearly report the implementation details of the FOLFIRINOX-based treatment regimens and involved reductions in the dose of the chemotherapeutic drugs. The presence of heterogeneity across the studies may have biased the results. Fourth, the included studies did not report on the ethnicities and dietary habits of the study populations, owing to which the comprehensiveness of the results is low.
In conclusion, FOLFIRINOX-based neoadjuvant chemotherapy can improve the rates of resection, R0 resection, and median OS in LAPC. These results require further validation in large, high-quality RCTs.