A phase II study of FOLFOXIRI plus bevacizumab as initial chemotherapy for patients with untreated metastatic colorectal cancer: TRICC1414 (BeTRI)

Purpose FOLFOXIRI plus bevacizumab is regarded as a first-line therapeutic option for selected patients with metastatic colorectal cancer (mCRC). Our aim was to assess the efficacy and safety of induction treatment with FOLFOXIRI plus bevacizumab in patients with untreated mCRC harboring UGT1A1 wild (*1/*1), or single-hetero (*1/*6 or *1/*28) genotypes. Methods Twelve cycles of FOLFOXIRI plus bevacizumab were administered to patients with untreated mCRC. The primary endpoint was the overall response rate (ORR) assessed by central independent reviewers. Secondary endpoints included time to treatment failure (TTF), progression-free survival (PFS), overall survival (OS), relative dose intensity (RDI), R0 resection rate, and safety. The exploratory objectives were early tumor shrinkage (ETS) and depth of response (DoR). Results Of the 47 patients enrolled, 46 and 44 patients were eligible for the safety and efficacy analysis, respectively. The primary endpoint was met. The ORR was 63.6% (95% CI 47.8–77.6). At a median follow-up of 25.4 months, median TTF, PFS, and OS was 8.1, 15.5, and 34.4 months, respectively. The median RDI of 5-fluorouracil, irinotecan, oxaliplatin, and bevacizumab was 72, 69, 62, and 71%, respectively. R0 resection rate was 22.7%. Grade 3 or higher adverse events (≥ 10%) included neutropenia (65.2%), febrile neutropenia (26.1%), leukopenia (23.9%), anorexia (10.9%), nausea (10.9%), and diarrhoea (10.9%). No treatment-related deaths were observed. ETS and DoR were 70.5 and 45.4%, respectively. Conclusions FOLFOXIRI plus bevacizumab induction treatment of Japanese patients was shown to be beneficial and manageable, although caution is required since the treatment causes febrile neutropenia. Electronic supplementary material The online version of this article (10.1007/s10147-020-01811-w) contains supplementary material, which is available to authorized users.


Introduction
Colorectal cancer (CRC) is the third most common malignant neoplasm and the second major cause of cancer death in 2018 worldwide for both sexes [1]. Most patients with metastatic CRC (mCRC) have unresectable disease and cannot be cured. However, long-term survival or even cure is reported to have been attained in 20-50% of patients who underwent complete R0 resection of their metastases [2].
The major clinical practice guidelines, including the Japanese Society for Cancer of the Colon and Rectum (JSCCR) guideline 2019, all recommend FOLFOXIRI plus bevacizumab as a first-line therapeutic option for selected patients with mCRC [5][6][7][8]. Only two prospective, single-arm phase Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1014 7-020-01811 -w) contains supplementary material, which is available to authorized users. II trials of FOLFOXIRI plus bevacizumab have been conducted to assess the safety and efficacy in Japan. One is the QUATTRO study using the GONO-FOLFOXIRI regimen [9]. The other is the JACCRO-CC11 trial using a modified FOLFOXIRI regimen that consisted of oxaliplatin (85 mg/ m 2 ), irinotecan (150 mg/m 2 ), l-leucovorin (200 mg/m 2 ), and 5-fluorouracil (2400 mg/m 2 ) [10]. Both studies concluded that FOLFOXIRI plus bevacizumab was effective. The incidence of febrile neutropenia was as low as 5% in modified FOLFOXIRI compared to 21.7% in GONO-FOLFOXIRI. However, mature PFS and OS have not been documented for either of these studies. The recommended doses of FOL-FOXIRI also remain controversial.
Adverse events associated with the FOLFOXIRI regimen include higher risks of neutropenia and diarrhoea [11]. Irinotecan especially has significant adverse effects, including myelosuppression and diarrhoea. In Japanese patients, homozygosity for UGT 1A1*28 or UGT 1A1*6 and heterozygosity for both UGT 1A1*6 and UGT 1A1*28 are associated with severe irinotecan-related neutropenia [12,13].
Thus, we conducted a single-arm, multicenter, phase II study to assess the efficacy and safety of induction treatment with FOLFOXIRI plus bevacizumab in patients with untreated mCRC harboring UGT 1A1 wild (*1/*1) or singlehetero (*1/*6 or *1/*28) genotypes. In our study, sequential treatment with the remaining drugs was continued after termination of protocol therapy or discontinuation of oxaliplatin and/or irinotecan at the investigator's discretion.

Study design
The Bevacizumab plus Triplet (BeTRI) study was a multisite, open-label, single-arm, phase II clinical trial. This trial was registered on the UMIN Clinical Trials Registry (UMIN000017102), followed by Japan Registry of Clinical Trials (jRCTs061180021).

Patients
Patients who met the following criteria were enrolled in this study: aged between 20 and 70 years; histologically confirmed adenocarcinoma of the colon or rectum; unresectable or recurrent CRC patient; no prior chemotherapy, immunotherapy, or radiation therapy (but can be enrolled 6 months after the date of completion of adjuvant chemotherapy); one or more measurable lesions; ECOG PS of 0 or 1; patients harbored UGT 1A1 wild (*1/*1) or single-heterozygous (*1/*6 or *1/*28) genotypes. Patients were ineligible if they had severe, uncontrolled organ or metabolic dysfunction.

Endpoints
The primary endpoint was the objective response rate (ORR) assessed by central independent reviewers. Secondary endpoints were TTF and PFS, OS, R0 resection rate, relative dose intensity (RDI), and safety. TTF represented the time from the initial day of protocol therapy to the first day when we observed any of the following events: discontinuation of protocol therapy; initial progression; death with any cause. PFS was defined as the time from the initial day of protocol therapy to the initial progression day or death for any reason. The exploratory objectives were early tumor shrinkage (ETS) and depth of response (DoR).

Treatment and evaluation
We defined protocol therapy as 12 cycles of FOLFOXIRI plus bevacizumab induction treatment consisting of a 30-90-min infusion of bevacizumab at a dose of 5 mg/kg, a 90-min infusion of irinotecan at a dose of 165 mg/m 2 , a 120-min infusion of oxaliplatin at a dose of 85 mg/m 2 and a concomitant 120-min infusion of l-leucovorin at a dose of 200 mg/m 2 . These were followed by a 48-h continuous infusion of 5-fluorouracil to a total dose of 3200 mg/m 2 . Cycles were repeated every 14 days.
The chemotherapy was continued until disease progression, an unacceptable adverse event, tumor resection, a delay of more than 29 days due to an adverse event, reduction of 5-fluorouracil dose to less than 2000 mg/m 2 , or consent withdrawal. In cases of prespecified adverse events, treatment modification was permitted according to the study protocols (Supplementary Tables 1 and 2). If oxaliplatin and/or irinotecan was discontinued at the investigator's discretion, sequential treatment consisting of the remaining drugs was continued.
We analyzed safety with a safety analysis set (SAS) in which patients received at least one cycle of protocol therapy. Efficacy was analyzed with a full analysis set (FAS) consisting of eligible patients. Based on the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1, central independent reviewers assessed response and progression, referring to CT or MRI taken every 8 weeks. ETS was defined as a reduction of at least 20% in the sum of the longest diameters of target lesions at week 8 compared with baseline. DoR was defined as the relative change in the sum of the longest diameters of RECIST target lesions at the nadir, in the absence of new lesions or progression of non-target lesions, when compared with baseline. Adverse events were evaluated according to CTCAE version 4.0.

Sample size
We assumed the expected response rate to be 70% and the threshold response rate to be 45%, with a two-sided α error of 0.05 and a power of 90%. The expected response rate was grounded in the results of the GONO and TRIBE trials, of which the confirmed ORRs were 65 and 77% for FOLFOX-IRI plus bevacizumab, respectively [3,14]. We also referred to the threshold response rate reported by the NO16966 and TRIBE trials, of which the confirmed ORRs were 34 and 53% for doublet plus bevacizumab, respectively [3,15]. We estimated a minimum of 40 patients, but we planned to enrol 45 or more patients, taking account of dropouts and withdrawals.

Statistical analyses
We concluded that the treatment could be regarded as promising if the lower limit of the 95% CI exceeded 45%. For the evaluation of the ORR and ETS, we calculated the 95% CI of the rate using the Clopper-Pearson method with F distribution. We used the Kaplan-Meier method to estimate TTF, PFS, and OS. Statistical analyses were carried out using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and R × 64 version 3.5.2.

Patients
Between April 2015 and May 2017, a total of 47 patients were enrolled from 16 Japanese study sites. One patient was excluded from the study because of a deviation of the inclusion criteria. Two more patients were removed after the central reviewers determined there were no measurable lesions. The numbers in the SAS and FAS were, therefore, 46 and 44, respectively (Supplementary Fig. 1). Table 1 shows patient characteristics in the SAS. There were twenty-six male and twenty female patients with a median age of 58 years (range 29-68 years). ECOG PS scores of 0 and 1 were 76% and 24%, respectively. The site of the primary tumor was right colon in 10 (22%) patients, left colon in 16 (35%), and rectum in 20 (43%) patients. Resection rate of the primary tumor was 37%. Seventeen (37%), 24 (52%) and 5 (11%) patients had RAS wild-type, RAS mutated, and RAS unknown tumor, respectively.

Study treatment
All 46 enrolled patients received FOLFOXIRI plus bevacizumab. The commencement of second cycle treatment was delayed in 33 (72%) patients. In addition, the dose level was reduced in the second cycle in 23 (50%) patients. Meanwhile, twelve (26%) patients completed the second cycle treatment as scheduled without dose modification.
In the FAS, the median relative dose intensity of 5-fluorouracil, irinotecan, oxaliplatin, and bevacizumab was 72, 69, 62 and 71%, respectively. The median number of cycles administered per patient as FOLFOXIRI plus or minus bevacizumab was eight (range 2-12).

Subgroup analyses by primary tumor sidedness and RAS mutational status
We evaluated the activity of FOLFOXIRI plus bevacizumab in an ad hoc analysis on the basis of primary tumor sidedness or RAS status (Table 3). Following classification of the 44 patients in the FAS on the basis of primary tumor sidedness and RAS mutation status, left-sided colon tumors were associated with a higher frequency of mutated RAS tumors than right-sided colon tumors (47.7 vs. 6.8%) (Supplementary Table 3). In stratification by sidedness, the ORR was numerically higher in right-sided colon tumors compared with left-sided ones (70.0 vs. 61.8%, respectively). As for ETS and DoR, there was no difference between the two sides.
In stratification by RAS status, the ORR was numerically higher in RAS wild-type tumors compared with RAS mutated ones (73.3 vs. 62.5%). ETS and median DoR were 73.3 and 47.6% in RAS wild-type tumors and 70.8 and 45.2% in RAS mutated tumors, respectively.

Conversion surgery
Fourteen patients (31.8%) had conversion surgery performed with curative intent. Nine of these patients underwent surgery during the protocol therapy, and two patients did after the end of protocol therapy. The remaining three patients did during sequential treatment following the completion of protocol therapy. The median time to surgery from the initial day of protocol therapy was 182.5 days (range 87-316 days).
R0 resection was performed in 10 (22.7%) patients. Five of these patients were treated with 12 cycles of protocol therapy (two patients) or sequential treatment (three patients). When we examined the presence of ETS in ten patients with R0 resection, eight patients had ETS and two patients had no ETS. The surgical procedures (number of patients) performed on the four patients who could not achieve R0 resection included primary resection (3) and liver and lung resection (1).
In total, 14 serious adverse events occurred in 11 patients (23.9%). Five of these adverse events were independent of protocol therapy. The remaining adverse events included infections (4), a thromboembolic event (1), an anastomosis fistula (1), ileus (1), nausea (1), and fever (1). All patients recovered from the serious adverse events at the end of the study. No treatment-related deaths occurred in the present study.
Right-sided colon tumors are associated with a poorer prognosis compared with left-sided tumors [16,17]. Considering only patients with previously untreated mCRC receiving first-line doublet chemotherapy plus or minus bevacizumab, Loupakis et al. observed that ORR and PFS were statistically significantly higher in patients with leftsided tumors [16]. Interestingly, our ad hoc analysis showed that FOLFOXIRI plus bevacizumab treatment produced a numerically higher ORR, TTF, and PFS in right-sided colon tumor cases (compared with left-sided tumors). In rightsided colon tumor cases, ORR was the same, irrespective of RAS mutational status. This finding supports the post hoc analysis of the TRIBE trial which found that FOLFOXIRI plus bevacizumab may be able to efficiently counteract the intrinsic aggressiveness of right-sided colon tumors [18].
FOLFOXIRI plus bevacizumab with an expected ORR of around 70% is an attractive strategy for conversion chemotherapy. Interestingly, conversion surgery (31.8%) and R0 resection (22.7%) rate in the present study were consistent with previous studies [9,10,14], despite the decreased dose intensity. The median time to surgery from the initial day of protocol therapy was 182.5 days (range 87-316 days). Of the 14 patients who underwent surgery, the chemotherapy regimen immediately prior to surgery was FOLFOXIRI (n = 11), FOLFIRI (n = 1), and 5-fluorouracil and leucovorin (n = 2). Among the 10 patients who had undergone R0 resection, 50% of patients were treated with 12 or more cycles of protocol therapy. The ten patients who underwent R0 resection included two patients without ETS. This finding supports previously published meta-regression analysis data suggesting that a high number of chemotherapy cycles (close to 12) is associated with conversion surgery [19]. Therefore, a sufficient treatment time period may be required for R0 resection in unresectable or recurrent CRC.
Considering the safety profile of this drug regimen, 26.1% of Japanese patients developed febrile neutropenia in our BeTRI study, which was consistent with the results of the QUATTRO study. It should be noted that only two (16.7%) of 12 patients were treated with G-CSF as a primary treatment or secondary prophylaxis of febrile neutropenia. The second cycle of treatment was performed without modification of dose or delay in 12 (26%) patients. The median relative dose intensities of 5-fluorouracil, irinotecan, oxaliplatin, and bevacizumab were 72, 69, 62, and 71%, respectively. Compared with earlier studies [9,14], the frequency of G-CSF use was extremely low. This may be related to the lower dose intensities in comparison with previous studies. The observed incidence of grade 3 or more diarrhea was low (10.9%) compared with the two GONO-FOLFOXIRI plus bevacizumab studies (14-18.8%) [4,14], and was comparable with the QUAT-TRO study [9]. The major differences from the TRIBE trial was that target patients who harbored UGT 1A1*1/*1, *1/*6, or *1/*28 were included in our BeTRI and QUAT-TRO studies. This inclusion criterion may be the reason for the lower frequency of diarrhea. Our results indicate that the GONO-FOLFOXIRI regimen could be managed with appropriate dose reductions and treatment delays in Japanese patients harboring UGT 1A1 wild (*1/*1), or single-hetero (*1/*6 or *1/*28) genotypes.
In clinical trials, chemotherapy is historically administered until disease progression, unacceptable toxicities, or patients' refusal. Oxaliplatin often causes cumulative neurotoxicity before clinical progression [20]. NCCN and ASCO guidelines recommend discontinuation of oxaliplatin from FOLFOX or CAPEOX three-four months after initiation of treatment or sooner for unacceptable neurotoxicity, with other drugs in the regimen maintained until time of tumor progression [6,7]. In contrast, continuation of FOLFIRI induction treatment is recommended for at least as long as tumor shrinkage continues, or disease stabilization is maintained with tolerable toxicities [5]. In the TRIBE trial, 76% of patients received secondary treatment, of which 63% were on irinotecan-containing regimens [4]. In our BeTRI study, a sequential treatment following FOLFOXIRI plus bevacizumab was not intentionally defined as maintenance treatment with 5-fluorouracil and leucovorin plus bevacizumab. A total of 27 (59%) patients completed twelve cycles of protocol therapy. In fact, at the twelfth cycle, 18 (39.1%) patients received FOLFOXIRI plus bevacizumab, six (13.0%) patients received FOLFIRI plus bevacizumab, two (4.3%) patients received FOLFIRI, and one (2.2%) patient received infusional 5-fluorouracil and leucovorin plus bevacizumab. It should be noted that a FOLFIRI regimen was opted for eight (30.8%) patients. Moreover, R0 resection was performed in 10 (22.7%) patients including five patients treated with more than 12 cycles. These may have impacted on the encouraging lengths of 15.5 months for PFS and 37.1 months for OS. Taking the above into account, FOL-FIRI appears to be suitable as a sequential treatment after three to four months of FOLFOXIRI or sooner for unacceptable neurotoxicity.
The current study had several potential limitations. A major weakness was that our trial was a single-arm, small sample size study with no comparators. Second, data on BRAF mutational status were not available. Finally, the follow-up time was short for evaluation of OS. In the future, appropriate sequential treatment strategies following FOL-FOXIRI induction treatment should be considered.
In conclusion, our BeTRI study demonstrated that GONO-FOLFOXIRI plus bevacizumab were highly beneficial and manageable with toxicities in selected Japanese patients with good performance status who can tolerate intensive treatment.
Acknowledgments We thank the patients who participated in this study and their families, the physicians and the medical staffs of the 16 sites registered in this study, and Dr. Hisayuki Endo, Dr. Yasuo Iwamoto, Dr. Takao Hinoi, and Dr. Satoshi Teramukai for comprising the independent data monitoring committee. We also thank Dr. Kunihisa Miyakawa and Dr. Masaki Matsusako for the central review. We acknowledge and appreciate the data analyses work carried out by Ms. Yoko Nakagawa and Mr. Hiromi Tsukada. We also appreciate the work of Mr. Tasuku Inaji in his duty as study manager in the Translational Research Center for Medical Innovation (TRI). Funding This study was funded by Yakult Honsha Co., Ltd under contract. Yakult Honsha Co., Ltd played no role in the design, collection, analysis, or interpretation of the data, or writing of this manuscript.
Ethical approval This study was conducted by the Translational Research Center for Medical Innovation (TRI) in accordance with the Declaration of Helsinki and Clinical Trials Act in Japan. The study protocol was approved by the Institutional Review Board of each participating institution and Certified Review Board. All patients provided written informed consent. This study was registered with UMIN Clinical Trial Registry (Trial Identifier UMIN000017102), the Japan Registry of Clinical Trials (jRCTs061180021), and ClinicalTrials.gov (NCT02497157).
Informed consent Informed consent was obtained from all individual participants included in this study.

Availability of data and material
The datasets generated during the current study are not publicly available because the informed consent form signed by the participants did not address an individual data sharing statement.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.