Abstract
Background: We hypothesized that perioperative FOLFOX (5-fluorouracil, leucovorin, oxaliplatin) might be used as an alternative to standard FLOT (docetaxel, 5-fluorouracil, leucovorin, and oxaliplatin) in patients with locally advanced oesogastric adenocarcinomas (OGA), particularly those with frailties. Patients and Methods: We reviewed the charts of 61 consecutives patients treated with FOLFOX for resectable OGA to estimate overall survival, recurrence-free survival, and safety. Results: The median follow-up was 69.7 (range=3.6-97.9) months. Few patients experienced grade 3 adverse events during the preoperative (n=6; 10%) and postoperative (n=6; 16%) phases. One patient experienced a fatal grade 5 adverse events (cardiogenic shock). Median overall survival was 51.7 months [95% confidence interval (CI)=31.6-93.2 months] and the 5-year survival rate was 44.4% (95% CI=30.3%-57.5%). Conclusion: Regarding its comparable efficacy and its favourable toxicity profile, perioperative FOLFOX is a reasonable alternative to FLOT for frail patients with resectable OGA.
Gastric cancer remains a major public health concern worldwide. It is the fifth most frequently diagnosed cancer, with over 1 million new cases each year, and the fifth leading cause of premature death, with over 780,000 deaths from gastric cancer in 2018 (1). Adenocarcinomas of the lower oesophagus (LO) and gastroesophageal junction (GEJ), which have shown a relative increase in incidence and mortality rate in Western countries in recent years (2), share common international treatment guidelines as oesogastric adenocarcinoma (OGA) (3, 4).
Surgery still remains the cornerstone of the treatment of non-metastatic gastric cancer (4), however, most patients will experience relapse within 5 years despite R0 resection (5). This poor outcome has led to multimodal treatment protocols such as perioperative chemotherapy (i.e. before and after surgery) (5). In 2006, the MAGIC study, which compared the ECF regimen [intravenous epirubicin, cisplatin, and 5-fluorouracil (5-FU)] with surgery alone, reported a significant 5-year overall survival (OS) benefit compared with surgery alone (36% vs. 23%, respectively) in stage II/III resectable OGA (6). The Fédération Nationale des Centres de Lutte Contre le Cancer/Fédération Française de Cancérologie Digestive multicenter randomized study combined surgery with preoperative chemotherapy (cisplatin and 5-FU) and showed an improvement in a 5-year OS in the perioperative chemotherapy arm (38% vs. 24%) (7). Perioperative chemotherapy has thus become the gold standard for locally advanced OGA (4). Given the good efficacy of docetaxel in metastatic gastric cancer (8), the German AIO group compared the ECF regimen with the experimental FLOT regimen (5-FU, leucovorin, oxaliplatin and docetaxel given intravenously as four preoperative and four postoperative cycles) (9, 10). The OS in the FLOT arm was significantly longer than in the ECF arm with 50 vs. 35 months and projected 5-year survival of 45% vs. 36% [hazard ratio=0.77, 95% confidence interval (CI)=0.63-0.94)], respectively. Consequently, FLOT has been proposed as the new gold standard of care for patients with locally advanced OGA (11, 12). However, the FLOT protocol raised some concerns in real-life practice (12) as even in a trial-selected population (9, 10), FLOT led to substantial toxicities, with frequent grade 3-4 adverse events [AEs; leuconeutropenia (51.0%), severe infection (18.0%), peripheral neuropathy (7.0%), and diarrhoea (10%)]. This should be considered particularly for unselected patients as these toxicities can have a huge impact on treatment completion and quality of life. Thus, use of a more manageable and tolerated alternative regimen to FLOT is needed in OGA, especially when considering patients in real-life practice, who frequently exhibit comorbidities and frailties. The combination of 5-FU, leucovorin, and oxaliplatin (the FOLFOX regimen) has been shown to be a tolerable and effective regimen in older patients (5, 13, 14).
In this monocentric retrospective study, we aimed to evaluate the efficacy and toxicity of FOLFOX-based perioperative chemotherapy in consecutive patients with locally advanced OGA.
Patients and Methods
Patients. All consecutive patients who were treated with the perioperative FOLFOX regimen and underwent surgery for treatment of resectable locally advanced OGA at the Cancer Institute of Montpellier (France) between January 2010 and December 2017 were retrospectively enrolled in this study. The inclusion criteria were patients over 18 years old who had never received any treatment (i.e., chemotherapy, radiotherapy, or surgery) for OGA; an Eastern Cooperative Oncology Group Performance Status (ECOG-PS) of ≤2; histologically confirmed gastric, LO or GEJ adenocarcinoma; and no evidence of distant metastases. Follow-up data were collected until December 2020.
The study was reviewed by the Institutional Review Board (registration number: 2020-118). All patients provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki guidelines for Good Clinical Practice, as defined by the International Conference on Harmonization.
Study treatment. Enrolled patients were treated with perioperative chemotherapy with modified FOLFOX6 (mFOLFOX6: 100 mg/m2 folinic acid, 85 mg/m2 oxaliplatin, and 400 mg/m2 5-FU bolus on day 1, followed by 2,400 mg/m2 5-FU over 46 h) and radical surgery (gastrectomy/oesophagectomy). Preplanned treatment consisted of four preoperative and six postoperative cycles (adjusted according to global health status, toxicity, frailty, pTNM staging, and physician’s decision). The dose of chemotherapy was initially (i.e., prior to treatment) i) reduced by 20% when the patient was considered frail or with comorbidities or according to toxicity (mainly if grade ≥3 AEs occurred), ii) stopped when intolerable toxicity occurred or upon physician’s or patient’s decision. Surgery was performed 4 weeks after the last preoperative chemotherapy treatment. Follow-up was carried out with computed tomographic scan and clinical examinations every 6 months for 5 years unless relapse or death occurred. Granulocyte colony-stimulating factor (G-CSF) was administered only when febrile neutropenia occurred (i.e., as secondary prophylaxis).
Data collection. Medical records were reviewed to obtain patient’s demographic, clinicopathological, and surgery-related variables including age, sex, height, weight, body mass index (BMI), ECOG PS, presence of malnutrition, comorbidities, alcohol use, tobacco use (defined as active consumption), toxicity [by the National Cancer Institute-Common Terminology Criteria for Adverse Events version 4.0 (15)], dosage, unplanned hospitalization, reasons for dosage reduction or discontinuation, administration of G-CSF, tumour location, cancer type by the Lauren classification, human epidermal growth factor receptor 2 status, TNM status, grade of differentiation, surgery type, resection status, staging, presence of vascular and perineural invasion, length of hospitalization, morbidity and mortality within 30 days, cause of death and date and type of recurrence. Malnutrition was considered as moderate or severe. Moderate malnutrition was defined as loss of 5% of usual body weight in the previous month or 10% within the previous 6 months and severe malnutrition as loss >10% in the previous month or 15% in the previous 6 months. The baseline biological data on whole blood counts with neutrophil/lymphocyte ratio (NLR), albumin, creatinine clearance, carcinoembryonic antigen, and carbohydrate antigen 19.9 (CA 19.9) were also retrieved.
Study endpoints. The primary endpoint was OS (calculated from the start of the preoperative mFOLFOX6 treatment to the date of death for any reason). Secondary endpoints were recurrence-free survival (RFS; calculated from the start of the preoperative mFOLFOX6 treatment to recurrence or death from any reason), therapy tolerance, AEs, resection margin, and predictive factors of the whole therapy (perioperative mFOLFOX6 plus surgery), and RFS.
Statistical analysis. Survival curves were estimated using the Kaplan– Meier method. Therapy tolerance was defined as the completion of the whole treatment encompassing perioperative mFOLFOX6 chemotherapy (at least one postoperative cycle) and surgery.
Cox proportional hazards model was used for univariate and multivariate analyses of RFS and OS. Twenty-five variables: Age (<65 vs. ≥65 years), gender, BMI (<25 vs. ≥25 kg/m2), severe malnutrition (yes vs. no), comorbidities (diabetes mellitus, multiple organ dysfunction), alcohol and tobacco status, ECOG PS (0 vs. ≥1), haemoglobin count (<13 vs. ≥13 g/dl), NLR (<2.5 vs. ≥2.5), platelet/lymphocyte ratio (<150 vs. ≥150), albumin (<25 vs. ≥25 g/l), creatinine clearance (<80 vs. ≥80 ml/min), carcinoembryonic antigen (≤5 vs. >5 ng/ml), CA 19.9 (≤37 vs. >37 U/ml), tumour localization (GEJ vs. stomach), cancer type (intestinal vs. non intestinal), human epidermal growth factor receptor 2 status (positive vs. negative), vascular/perineural invasions (yes vs. no), cancer staging (0 vs. I/II vs. III/IV), ypT status (T0 vs. T1/T2 vs. T3/T4), ypN status (N− vs. N+), resection margin (R0 vs. R1/R2), grade 3/4 AEs (yes vs. no), and postoperative chemotherapy (yes vs. no) that were significantly associated (p<0.2) with OS and RFS were subsequently included in multivariate analysis.
Univariate analyses were performed for therapy tolerance using Cox proportional hazard regression model according to: Age, gender, BMI, severe malnutrition, comorbidities, alcohol and tobacco use, ECOG PS, haemoglobin, platelet, neutrophils, and lymphocytes count, NLR, platelet/lymphocyte ratio, albumin, creatinine, CAE, CA 19.9, and tumour location. All factors with p<0.2 in univariate analysis were applied for multivariate analysis. Summary statistics were used to describe the data. The median (range) was used to describe continuous variables. Categorical variables are presented as frequencies (percentages).
Results
Patient characteristics. Between 2010 and 2017, a total of 61 consecutive patients were treated for a potentially curative OGA (Figure 1). The baseline data are shown in Table I. The majority of patients had an ECOG PS of 0 (n=36; 59.0%) or of 1 (n=23; 37.7%). Overall, 17 (27.9%) of tumours were localized in the LO (including Siewert type 1 GEJ cancer), 17 (27.9%) at the GEJ (Siewert type 1 excluded), and 27 (44.3%) in the stomach.
Study flow chart.
Patient and tumour baseline characteristics.
Treatment feasibility and tolerance. Of the 61 patients enrolled, all received at least one cycle of preoperative chemotherapy, 58 (95.0%) underwent radical surgery, 38 (63%) received postoperative chemotherapy, and 22 (36.0%) completed the whole treatment. The mean time from diagnosis to data cut-off was 69.7 (range=3.6-97.9) months. The mean time interval from cancer diagnosis to the first cycle of preoperative chemotherapy was 40 (range=8-84) days.
A total of 248 preoperative and 190 postoperative chemotherapy cycles were performed, with a median number of 4 (range=2-7) and 5 (range=1-8) cycles, respectively. The median number of perioperative chemotherapy cycles was 7 (range=2-12). Overall, 52 (85.2%) and 33 (86.8%) patients received at least 4 cycles during the preoperative and postoperative phases, respectively. The administered dose was modified (i.e., 20% reduction) during preoperative chemotherapy before the first cycle in seven (11.4%) patients and after at least 1 cycle in six (9.8%) patients. Neoadjuvant chemotherapy was stopped in eight (13.1%) patients, after 2-3 cycles (Table II) due to toxicity (n=7) or at the patient’s request (n=1). Dose adjustment (80% of the planned dose) prior to postoperative chemotherapy was needed for nine (23.6%) patients and after at least 1 cycle for 14 (36.8%) patients. Postoperative chemotherapy was stopped for 16 (42.1%) patients (AEs: n=12, patient’s request: n=3, death: n=1), after 1-5 cycles (Table II). Grade 3 AEs were observed in six (9.8%) patients during the preoperative phase and in six (15.7%) patients during the postoperative phase (Table III). Only one patient experienced grade 4-5 AE (cardiogenic shock) during the postoperative phase.
Neoadjuvant and adjuvant chemotherapy treatments.
Treatment-emergent adverse events.
Three (4.9%) out of 61 patients who started treatment did not undergo radical surgery because of progressive disease (Table IV). In 15 (25.9%) patients oesophagectomy was performed, seven (12.1%) had partial gastrectomy, and 36 (62.0%) had total gastrectomy (with/without splenectomy). The median length of hospitalization was 23.4 (range=8-65) days, with postoperative morbidity (i.e., AEs within 30 days post-surgery) occurring in 30 (51.7%) patients, mainly due to an anastomotic leak (63.3%). Nine (15.5%) patients needed admission to the Intensive Care Unit, and two out of these died. Margin-negative (R0) resection was possible in 54 (93.1%) patients and a pathological complete response was observed in four (6.9%).
Postoperative outcomes and pathology results.
RFS and OS analyses. The last follow-up visit occurred in December 2020. Median RFS was 26.9 (95% CI=13.8-46.2) months and the estimated 5-year RFS rate was 34.5% (95% CI=22.6%-46.8%; Figure 2A). Disease recurrence was observed in 35 (57.3%) patients [local in six, locoregional in four, and distant metastasis in 25 (peritoneal in 11, liver in 11, bone in one, ovary in one, and adrenal gland in one)]. Thirty-four (55.7%) patients died, of whom five from causes other than OGA or treatment. Median OS was 51.7 (95% CI=31.6-93.2) months, with a 5-year survival rate of 44.4% (95% CI=30.3%-57.5%; Figure 2B). In patients with diffuse-type gastric cancer according to the Lauren classification (n=12), the median RFS and OS were 33.4 and 68.7 months, respectively.
Kaplan–Meier survival curves of recurrence-free (A) and overall survival (B).
In univariate analysis, age ≥65 years, carcinoembryonic antigen >5 ng/ml, LO localization, presence of vascular or perineural invasion, III/IV cancer stage, ypT3/T4, ypN+, R1/R2 resection, and absence of postoperative chemotherapy were correlated with a shorter RFS (p<0.005). In multivariate analysis, age ≥65 years (p=0.001), LO localization (p<0.001), R1/R2 resection (p<0.001), III/IV stages (p<0.001) and complete lack of postoperative chemotherapy were associated with a shorter RFS (p=0.004).
In univariate analysis, age ≥65 years, presence of diabetes mellitus, LO location, presence of vascular or perineural invasion, stage III/IV, ypT3/T4, ypN+, and absence of postoperative chemotherapy were associated with shorter OS (p<0.05). Age ≥65 years (p=0.002), LO location (p<0.001), R1/R2 resection (p=0.001), stages III/IV (p<0.05), and no postoperative chemotherapy were reported as prognostic factors (p=0.001).
Discussion
Following the phase III FLOT4-AIO trial, which compared FLOT and ECF perioperative chemotherapy regimens in patients with resectable gastric cancer (9), the FLOT-based regimen has been proposed as the gold standard for resectable OGA. However, the high toxicity profile of FLOT has raised concerns about its applicability in real-life practice (10, 11). It has been shown that it can be administered to older patients, albeit only in those who are deemed fit. During the period encompassed by our study (2010 to 2017), the FLOT regimen was not considered the standard treatment in patients with resectable OGA and, among other 5-FU and platinum salt-containing regimens, FOLFOX was commonly used by us and others (5). The median OS observed in our study is similar to that reported with the standard FLOT regimen treatment (51.7 vs. 50 months, respectively). Moreover, we showed that FOLFOX appears suitable for patients with disabilities or frailties because of its favorable toxicity profile.
Our results are in accordance with the scarce literature published so far in this setting (16, 17). The only data on use of the perioperative FOLFOX regimen in resectable OGA come from two studies, the FOLFOX-AGEO (n=109) retrospective cohort and the FOLFOX-PUMCH (n=73) phase II trial (16, 17). In the FOLFOX-AGEO cohort (16), the median RFS was 40.3 months, which is longer than in our study (26.9 months). However, in our study, at 51.7 months, the median OS was longer than in the aforementioned study (40.3 months). These differences might be explicable by patient and treatment characteristics, including fewer N+ patients and more postoperative cycles in our series. In addition, unlike in our series where patients were analysed with an intent-to-treat purpose, patients in the AGEO group were selected and received at least 3 preoperative cycles. In the prospective FOLFOX-PUMCH trial of Asian patients (17), mFOLFOX6 showed an excellent tolerance profile and clear survival benefit (median RFS and OS: 56.0 months and 76.0 months, respectively) in a fitter population (PS 0-1) with gastric cancer only.
Our study considered a range of parameters that resulted in a few relevant differences between our and the FLOT4-AIO trial. In the present study, the population was quite heterogeneous, reflecting real-life practice in gastrointestinal oncology. In fact, the majority of patients were ≥65 years old (59.0%) and the median age was 67 (range=42-79) years vs. 62 (range=54-69) years in the FLOT4-AIO trial, which reflects an older population in our study. Furthermore, 47.5% of our patients displayed at least one organ dysfunction (with cardiovascular disease and diabetes mellitus present in 28.0% and 26.0% of the cases, respectively) and 45.9% exhibited moderate to severe malnutrition upon diagnosis. The absence of a precise description of these parameters in the FLOT4-AIO study did not allow for cross-study comparison; meanwhile, medical operability was considered as an inclusion criterion and several indicators of chronic conditions (i.e., creatinine clearance ≤50 ml/min, left ventricle ejection fraction ≤50%, active congestive heart disease, severe internal accompanying disease, severe liver dysfunction, chronic inflammable intestinal disease) were exclusion criteria, reflecting a far more selected population. It is worth mentioning that as opposed to FLOT4-AIO, in our study, there were fewer N+ patients (39.4%, vs. 78.0%) and patients with cT3 tumours (45.9%, vs. 75.0%).
We are aware that cross comparison between different series or clinical trials is challenging. Notably, however, the median OS observed for our cohort (51.7 months) appears comparable to that reported in the FLOT trial (50 months) (10). In addition, the 5-year survival rate of 44.4% (95% CI=30.3-57.5%) with FOLFOX observed in our study appeared to be as high as the 45.0% (95% CI=38-51%) rate for with the FLOT regimen (10). Furthermore, 26.9-month median RFS and 93.1% R0 resection rate for FOLFOX are in the same range as these reported in the FLOT4-AIO study (30 months and 85%, respectively). Nevertheless, the pathological complete response rate was lower in our study than in the FLOT4-AIO study (6.9% vs. 15.0%). Considering the safety results, the FOLFOX regimen appears to be a potent alternative to FLOT. Notably, FOLFOX was associated with less diarrhoea (5.2% vs. 10.0%), severe infection (2.6% vs. 18.0%), and neutropenia (13.1% vs. 51.0%). In the FLOT4-AIO trial, G-CSF prophylaxis and unplanned hospitalization occurred more often with FLOT than with FOLFOX (8.3% vs. 34.0% and 11.4% vs. 25.0%, respectively). Similarly, the 28.0% rate reported of grade 2 alopecia with the FLOT regimen, although not life-threatening, was significantly higher than in our study (there was no single case reported), which might be relevant concerning treatment acceptability and patient quality of life (9, 10).
Our study has some limitations. It was a monocentric, retrospective, and uncontrolled study with a limited number of patients. The retrospective nature of our study may have led to underestimation of collected data such as AEs. Moreover, baseline tumour staging was mainly performed through computed tomographic scan, which is not the most accurate method for GEJ and gastric cancer (notably for assessing the nodal status).
In conclusion, our study showed that in terms of its efficacy and tolerance, FOLFOX can be considered as a reasonable alternative to FLOT in patients with resectable OGA, especially the elderly and those with comorbidities in the perioperative setting.
Acknowledgements
The Authors acknowledge the editorial assistance of Magdalena Benetkiewicz, PhD.
Footnotes
Authors’ Contributions
Conception and design: AA, ES, MY; Collection and assembly of data: SQ, ES, AM, LQ, FP, TM, MY; data analysis and interpretation: SQ, ST, AA; primary draft writing: SQ, AA; Final approval of manuscript: All Authors.
Conflicts of Interest
None for all co-authors.
- Received November 5, 2021.
- Revision received November 25, 2021.
- Accepted November 30, 2021.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.