The BEVATOMOX phase II trial: raltitrexed/oxaliplatin/bevacizumab vs mFOLFOX6/bevacizumab in 2nd-line metastatic colorectal cancer.

Introduction
Colorectal cancer is the third most common cancer, with over 1.9 million new cases and about 904,000 deaths reported in 2022 worldwide.1 Approximately 25% of patients have metastatic disease at diagnosis, and nearly 50% will develop metastases over the course of their disease.
First-line treatments for metastatic colorectal cancer (mCRC) depend on RAS/BRAF mutational status, tumor location, presence of microsatellite instability (dMMR/MSI), patient age, performance status, and comorbidities. European guidelines recommend doublet chemotherapies (FOLFOX or FOLFIRI) for pMMR/MSS tumors, combined with biological agents targeting the epidermal growth factor receptor (anti-EGFR) (eg, cetuximab or panitumumab) or the vascular endothelial growth factor (eg, bevacizumab).2,3 Triplet chemotherapy (FOLFOXIRI) alone or combined with bevacizumab is also an option for fit patients.3 The PARADIGM, STRATEGIC-1, or IMPROVE trials endorse the use of anti-EGFR with doublet chemotherapy for patients with RAS wt/BRAF wt and left-sided tumors.4–6 Phase III clinical trials encourage combining doublet or triplet chemotherapy with bevacizumab for patients with RAS-mutated (mt) or BRAF mt tumors, and also for those with right-sided tumors.7–9
Several studies have also investigated second-line treatments. After the failure of first-line irinotecan-based chemotherapy, a combination of FOLFOX4 and bevacizumab improved overall survival (OS) (12.9 vs 10.8 months, HR = 0.75), progression-free survival (PFS) (7.3 vs 4.7 months, HR = 0.61), and objective response rate (ORR) (22.7% vs 8.6%, P < .0001) compared to FOLFOX alone.10 Continuing bevacizumab while switching to doublet chemotherapy (FOLFOX6 or FOLFIRI) after the failure of a first-line treatment showed improved PFS (5.7 vs 4.1 months, HR 0.68, 95% CI 0.59-0.78 P < .0001) and OS (11.2 months vs 9.8 months, HR 0.81, 95% CI 0.69-0.94; P = .0062) than chemotherapy alone.11 The PRODIGE18 study showed a non-significant difference but favored continuation of bevacizumab with chemotherapy crossover for patients with RAS wt metastatic colorectal cancer that progressed with first-line bevacizumab plus chemotherapy over switching to cetuximab.12 The FIRE-3 trial, which analyzed subsequent lines of therapy after failure of cetuximab plus FOLFIRI vs bevacizumab plus FOLFIRI in 414 patients with KRAS wt mCRC, favored FOLFIRI-cetuximab followed by FOLFOX6-bevacizumab compared to FOLFIRI-bevacizumab followed by irinotecan-cetuximab sequence.13
Raltitrexed is a specific thymidylate synthase inhibitor that does not metabolize into fluorodeoxyuridine monophosphate, and its mechanism of action does not involve the dihydropyrimidine dehydrogenase (DPD). It has an extended terminal plasma half-life of 148-379 hours, allowing for administration at an extended dosing interval of 3 weeks.14 Four randomized phase III trials have evaluated the efficacy of raltitrexed monotherapy compared to different 5-fluorouracil (5-FU) schedules (bolus or continuous infusion) in patients with mCRC. The results showed comparable efficacy of raltitrexed, with a median OS ranging from 10.3 to 12.2 months.15–18 Phase II studies combining raltitrexed and oxaliplatin (TOMOX) every 3 weeks showed good outcomes with manageable toxicities for these patients in first- or second-line settings. Raltitrexed generally had an acceptable tolerability profile and was associated with less leukopenia and mucositis/stomatitis than 5-FU/LV.19–23 However, no randomized study has evaluated the combination of TOMOX and bevacizumab in second-line metastatic treatment.
The aim of the BEVATOMOX study was to assess the efficacy and safety of the combination of bevacizumab with TOMOX as a second-line treatment after the failure of a first-line irinotecan-based chemotherapy in patients with mCRC. The mFOLFOX6 plus bevacizumab arm was used as an internal control.

Methods

Study design and objectives
The BEVATOMOX study was a non-comparative, prospective, randomized, open-label, multicentric phase II trial that assessed the efficacy of the TOMOX and bevacizumab combination as a second-line treatment for patients with mCRC. Secondary objectives included ORR, treatment toxicity, and OS. Patients were randomized (1:2) to receive either mFOLFOX6-bevacizumab (control arm) or TOMOX-bevacizumab (experimental arm).
The study was conducted following the principles of the Declaration of Helsinki and the International Conference on Harmonization and Good Clinical Practice guidelines. The study protocol (ClinicalTrials.gov identifier: NCT01532804; EudraCT identifier: 2010-023447-15), including all amendments, was reviewed and approved by the French ethics committee (Comité de Protection des Personnes Sud Méditerranée III, ref 2010.11.04 bis) on February 16, 2011, and the French National Agency for the Safety of Health Products (ANSM) on February 14, 2011.

Patient eligibility
Eligible patients were ≥18 years old, with histologically confirmed diagnosis of colorectal adenocarcinoma and asymptomatic tumor, unresectable metastatic disease with at least one measurable target lesion according to RECIST (v1.1) criteria, an Eastern Cooperative Oncology Group (ECOG) performance status ≤2, and disease progression after a first-line irinotecan-based chemotherapy. Irinotecan may have been administered as first-line treatment as monotherapy or in combination with 5FU, oral fluoropyrimidines, anti-EGFR (cetuximab or panitumumab), or anti-VEGF (bevacizumab).
Main exclusion criteria were previous treatment with raltitrexed, brain metastases or carcinomatous meningitis, uncontrolled hypertension (systolic pressure> 150 mmHg or diastolic pressure > 100 mmHg), peripheral neuropathy of grade ≥2 according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 4, myocardial infarction, pulmonary embolism or severe vascular disease within the six months prior to randomization. Additionally, prior treatment with oxaliplatin was prohibited if administered for metastatic disease but permitted as adjuvant therapy, provided it had been completed more than six months earlier.

Study procedures
Eligible patients were randomized and stratified based on prior use of bevacizumab (yes vs no) and number of metastatic sites (1 vs >1). In the control arm, patients were scheduled to receive twelve cycles of the mFOLFOX6-bevacizumab combination. This regimen consisted of intravenous (IV) bevacizumab 5 mg/kg followed by mFOLFOX6 (oxaliplatin 85 mg/m2 IV infusion over 2 hours, leucovorin 200 mg/m2 IV infusion over 2 hours, 5-FU bolus 400 mg/m2, followed by 5-FU IV infusion 2400 mg/m2 over 46 hours) on day 1 every 2 weeks (day 1 = day 15). In the experimental arm, patients were intended to undergo eight cycles of the TOMOX-bevacizumab combination. This regimen consisted of IV bevacizumab 7.5 mg/kg and raltitrexed 3 mg/m2 administered according to creatinine clearance, followed by oxaliplatin 130 mg/m2 every 3 weeks (TOMOX, day 1 = day 21).
Patients were eligible for repeated treatment cycles in the absence of disease progression, unexpected adverse events, investigator decision, or withdrawal of consent. During the study treatment, all patients had visits every 2 weeks in the control arm or every 3 weeks in the experimental arm to assess if treatment could be continued until disease progression. In case of toxicity, dose reductions and treatment delays were implemented.

Outcome measures
The primary endpoint was the six-month progression-free survival (6PFS) rate, (percentage at 6 months of PFS defined as the time from randomization until disease progression or death from any cause, whichever occurred first). Secondary endpoints included PFS (time from randomization to the first documented tumor recurrence or death), ORR (percentage of patients with complete response or partial response as best overall response), OS (time from the date of randomization until death of any cause), DCR (percentage of patients with stable disease as best overall response), toxicities regardless of causality according to the Common Terminology Criteria for Adverse Events (CTC-AE) v4.0, and length of hospital stay (days).
Efficacy endpoints were evaluated locally according to RECIST criteria v1.1 by CT scan at baseline and then every 8 and 9 weeks from treatment initiation until disease progression in the control and experimental arm, respectively. All patients had an end-of-treatment visit and long-term follow-up visits every 2 months until progression and every 6 months after. Safety was evaluated based on the occurrence of all adverse events, regardless of the causality, graded according to the NCI-CTCAE v4.0.
At the time of the study, only the KRAS exon 2 and exon 3 statuses were determined in the tumor samples.

Statistical considerations
The sample size calculation was estimated using the one-step Fleming procedure. With 85% power (β = 15%) and an α risk = 10%, 57 patients were required in each arm (P0 = 0.50 and P1 = 0.65). Considering 10% of patients not evaluable, and a 1:2 randomization ratio, a total of 92 patients were required, with 30 patients in the control arm and 62 patients in the experimental arm. No formal statistical comparison was performed. Success, that is, the efficacy of the experimental treatment, was defined as at least 34 out of 57 patients showing no progression at 6 months, which correspond to an estimated median PFS of 8.0 months (90% confidence interval [CI], 6.1-11.0). The efficacy population included all treated patients who had at least one post-baseline tumor evaluation, analyzed according to the treatment actually received. The safety population comprised all patients who received at least one cycle of treatment, also analyzed according to the treatment actually received.
Categorical variables were summarized using contingency tables, and continuous variables were reported as medians and ranges. Categorical variables were analyzed using Pearson’s chi-square test or Fisher’s exact test, and continuous variables were analyzed using the Kruskal–Wallis test or Student’s t-test. Survival data were estimated using the Kaplan–Meier method and presented as medians and survival rates with 95% confidence intervals (95% CIs). The median follow-up time was estimated using the reverse Kaplan–Meier method and reported with 95% CI. Patients who were alive and event-free at the end of the study were censored at the last known date of survival. Differences between treatment arms in survival analyses were evaluated using the log-rank test. All statistical analyses were two-sided, with statistical significance set at 5% (P < .05). The analyses were performed with STATA version 16.1 (StataCorp).

Results

Patients’ characteristics
Between July 2011 and May 2016, 83 patients were included in the study across 12 French centers. Thirty-three patients were randomized to the control arm and 50 patients to the experimental arm (Figure 1). All patients were evaluable for tolerance, but four patients were excluded from the primary endpoint analysis. In the control arm, one patient was excluded due to a protocol deviation (no RECIST lesion at inclusion) identified during the first monitoring. In the experimental arm, one patient died before the first assessment, another discontinued treatment after two cycles due to a grade 3 staphylococcal infection and a grade 3 hypotension, and a third completed four cycles but was withdrawn from the study because the tumor was not measurable at the first tumor assessment. The study was prematurely terminated due to recruitment failure, which prevented the inclusion of patients according to the 1:2 ratio. Baseline characteristics were well balanced between the treatment arms except for previous treatment with oxaliplatin-based chemotherapy in adjuvant setting (9% in the control arm and 34% in the experimental arm) (Table 1). The median age of the patients was 66 (48-82) years, and 63.9% were male. Forty-seven patients (57.3%) had an ECOG performance status of 0. Primary tumors (66.3%) were left colon-sided, and 54.2% were KRAS mt. All patients had previously been exposed to irinotecan, 97.6% had received 5-FU, 66.7% oxaliplatin (in adjuvant setting), 89.2% bevacizumab, and 7.2% cetuximab (Table 1).

Efficacy
The median follow-up duration was 30.6 months (95% CI, 18.5-not reached). The 6PFS rates were 51.5% (n = 17, 90% CI, 36-67) in the control arm and 38% (n = 19, 90% CI, 26-51) in the experimental arm, respectively. The median PFS was 6.0 months (95% CI, 4.2-7.6) and 4.3 months (95% CI, 3.2-6.1) (P = .372) in the control arm and the experimental arms, respectively. The median OS was 11.1 months (95% CI, 9.5-16.4) in the control arm vs 9.3 months (95% CI, 5.7-11.6) (P = .543) in the experimental arm, (Figure 2A and B).
No significant difference in PFS was observed based on tumor location (P = .15) or KRAS status (P = .72) (Table 2). KRAS status also did not significantly affect OS outcomes (P = .37). Interestingly, OS was significantly longer for patients with left-sided tumors (11.6 months) compared to right-sided tumors (4.6 months, P < .001) in the experimental arm (Table 2 and Figure 2C).
In the ITT population, 4.8% of patients were not evaluable for ORR. Among the 79 evaluable patients, 15.2% achieved a partial response and 60.6% had a stable disease in the control arm, compared to 8% and 58%, respectively, in the experimental arm. The ORR was 15.6% (90% CI, 6-30) in the control arm vs 8.5% (90% CI, 3-18) in the experimental arm. The DCR was 75.8% (60.5-87.3) in the control arm vs 66.0% (53.5-77.0) in the experimental arm (Table 3).

Compliance and safety
The median number of treatment cycles was 8 (4-12) in the control arm and 4 (1-8) in the experimental arm (Table 4). The mean relative dose intensity (RDI) was 92.1% in the control arm and 76.9% in the experimental arm related to a lower RDI of bevacizumab and raltitrexed. Delayed administration occurred in 12 patients (36.4%) in the control arm and 14 patients (28.0%) in the experimental arm. Dose modifications occurred in 20 patients (60.6%) in the control arm and 32 patients (64.0%) in the experimental arm. In the control arm, among the 12 patients with treatment delays, the primary reasons were hematologic toxicities (33.3%), non-hematologic toxicities (16.7%), intercurrent illness (8.3%), or other reason (41.7%). In the experimental arm, 14 patients experienced treatments delays due to hematologic toxicities (21.4%), non-hematologic toxicities (35.7%), intercurrent illness (7.1%), or other reason (35.7%). Reasons for dose adjustments were non-hematologic toxicity (67.5%), hematologic toxicity (6.3%), intercurrent illness (3.2%), and other causes (23.0%). The overall rate of premature discontinuation due to disease progression was 42.2% and similar in two arms but the rate related to toxicity was lower in the control arm, 3.0%, than in the experimental arm, 20.0%. The median days of hospitalization for treatment administration were 2 (1-14) and 1 (1-6), respectively.
In the whole cohort, 45.8% of patients (n = 38/83) experienced grade 3 toxicity, and 15.7% (n = 13/83) had grade 4 toxicity. No statistically significant differences in grade 3-4 toxicities were observed between the treatment arms, as reported in Table 5. The only grade 5 toxicity was due to a suicide. The rate of maximum observed severe toxicities was 60.6% (n = 20/33) in the control arm and 64% (n = 32/50) in the experimental arm (Table 5). Major grade 3-4 toxicities were mucositis (12.1% vs 12.5%), paresthesia (0% vs 6.3%), hand-foot syndrome (3.0% vs 0%), neutropenia/febrile (6.1% vs 8.2%), and diarrhea (0% vs 6.3%) (Table 5). More patients reported grade 2 hand-foot syndrome (12.1% vs 0%) and mucositis rates (15.2% vs 8.0%) in the control arm than in the experimental arm.

Discussion
The BEVATOMOX study was the first randomized trial to evaluate the combination of raltitrexed, oxaliplatin, and bevacizumab in second-line treatment of patients with mCRC. Our results demonstrated the feasibility of this regimen when used after the failure of first-line irinotecan-based chemotherapy. However, because the trial was discontinued before reaching the target number of patients (83 instead of 92), no definitive conclusion can be drawn regarding the efficacy of the experimental treatment.
In the experimental group, we observed a 6PFS rate of 38% (95% CI, 26-51), a median PFS of 4.3 months (95% CI, 3.2-6.1), a median OS of 9.3 months (95% CI, 5.7-11.6), an ORR of 8.5% (90% CI, 3-18), and a disease control rate of 66.0%, which is lower than typically observed with 2nd-line regimens that include bevacizumab.10,11 Few studies have evaluated second-line treatments with raltitrexed combined to oxaliplatin, with or without bevacizumab, in patients with mCRC who had failed fluoropyrimidine-based chemotherapy. A non-randomized phase II trial assessed a combination of raltitrexed-based chemotherapy with bevacizumab as second-line treatment.24 The authors reported a median PFS of 8.4 (6.2-11) months (11.6 months for the patients treated with TOMOX-bev) and a median OS of 17.6 (15.2-22) months. While their results are better than those observed in our study, only 6% of patients (6/94) were treated with TOMOX-bevacizumab, the majority receiving a combination with irinotecan. Moreover, our findings may be partly explained by the lower relative dose intensity of TOMOX-bevacizumab compared to the control arm (76.9% vs 96.1%) as well as differences in prior exposure to oxaliplatin. Scheitauer et al. reported a median PFS of 6.5 (1.2-14) months and an ORR of 33.3 (18.6-51) in 36 patients treated with raltitrexed combined to oxaliplatin without bevacizumab.21 Nonetheless, this study was a single-arm phase II trial, and patients were not pre-exposed to oxaliplatin in adjuvant setting contrary to our patient population, where 34% of patients received oxaliplatin in the experimental arm (vs 9% in the control arm).
Regarding the study population, 54.2% of patients had KRAS mutations; however, KRAS status did not influence survival in either the experimental or control arm. As mentioned in the Methods section, only the KRAS exon 2 and 3 statuses were determined at the time of the study, as the prognostic and predictive value of the BRAF  V600E status had not yet been well established. Consequently, we cannot rule out its potential negative impact on the survival of our patients. In contrast, within the experimental arm, we observed a significant difference in OS rates based on tumor location, with patients having left-sided tumors experiencing better outcomes than those with right-sided tumor (11.6 months vs 4.6 months, P < .001). This result is consistent with previous findings in the literature.25 Only 7.2% patients were treated with anti-EGFR in first-line treatment. This can be explained by the guidelines on the use of these targeted therapies administered after failure of irinotecan-based chemotherapy at the time of the study.
Both the mFOLFOX6 and TOMOX regimens were well tolerated, with similar overall grade 3-4 toxicities, though some differences in the safety profiles were observed. As expected, patients receiving mFOLFOX6 experienced a higher incidence of neutropenia and leukopenia, while those treated with TOMOX showed a higher incidence of diarrhea and paresthesia, likely related to the oxaliplatin dose (130 mg/m2). However, these differences were not statistically significant. Unlike 5-FU-based chemotherapy, no cases of hand-foot syndrome were reported with raltitrexed-based chemotherapy. No treatment-related deaths were reported in the overall cohort.
At the time of the study, developing new strategies to promote ambulatory care over hospitalization in the oncology setting was a key objective, aiming to minimize patients’ hospital stays and reduce time and staff resource demands. The use of raltitrexed could support this goal by reducing hospital visits and the need for extended inpatient care. Currently, this is no longer a concern, as both regimens are administered on an outpatient basis.
Since TOMOX did not present any additional safety or tolerability concerns compared to mFOLFOX6, and considering the relative convenience of raltitrexed administration, TOMOX may serve as an interesting second-line treatment option for patients with advanced CRC. This could be particularly suitable for those who have difficulty adhering to the infusion schedule of 5-FU or cannot tolerate 5-FU-based regimens. Raltitrexed could also serve as a potential alternative treatment for patients with a contraindication to fluoropyrimidine, such as those with complete DPD deficiency or those with severe diarrhea, stomatitis, and/or neutropenia due to partial DPD deficiency. Indeed, since 2018, the Agence Nationale de Sécurité du Médicament et des produits de santé (French National Agency Authority for the Safety of Health Products) has recommended routine screening for DPD deficiency before starting 5-FU treatment, using plasma uracil concentration.26 Additionally, the TOMOX regimen could provide an alternative to fluoropyrimidine-based treatments for patients with severe cardiovascular comorbidities or those at risk for coronary vasospasm, a complication that affects 1.5% to 12% of patients treated with fluoropyrimidine.27
The main limitations of our study include the lack of extended RAS, BRAF  V600E, and dMMR/MSI status assessment, elements now required before initiating systemic treatment in the metastatic setting. In addition, the non-comparative design of this randomized study, missing data, the time elapsed since the study was conducted, and recruitment challenges limit our ability to draw definitive conclusions regarding the efficacy of the TOMOX plus bevacizumab combination.

Conclusion
The TOMOX-Bevacizumab regimen is feasible as second-line treatment in mCRC patients, with a favorable safety profile, and may represent an alternative option to the FOLFOX chemotherapy doublet in combination with bevacizumab after failure of irinotecan-based chemotherapy in patients with BRAF  V600E wt, RAS wt or mt, pMMR/MSS mCRC, especially for patients with contraindications or reluctance to fluoropyrimidines.