Bosutinib versus imatinib for newly diagnosed chronic phase chronic myeloid leukemia: final results from the BFORE trial

This analysis from the multicenter, open-label, phase 3 BFORE trial reports efficacy and safety of bosutinib in patients with newly diagnosed chronic phase (CP) chronic myeloid leukemia (CML) after five years’ follow-up. Patients were randomized to 400-mg once-daily bosutinib (n = 268) or imatinib (n = 268; three untreated). At study completion, 59.7% of bosutinib- and 58.1% of imatinib-treated patients remained on study treatment. Median duration of treatment and time on study was 55 months in both groups. Cumulative major molecular response (MMR) rate by 5 years was higher with bosutinib versus imatinib (73.9% vs. 64.6%; odds ratio, 1.57 [95% CI, 1.08–2.28]), as were cumulative MR4 (58.2% vs. 48.1%; 1.50 [1.07–2.12]) and MR4.5 (47.4% vs. 36.6%; 1.57 [1.11–2.22]) rates. Superior MR with bosutinib versus imatinib was consistent across Sokal risk groups, with greatest benefit seen in patients with high risk. Treatment-emergent adverse events (TEAEs) were consistent with 12-month data. After 5 years of follow-up there was an increase in the incidence of cardiac, effusion, renal, and vascular TEAEs in bosutinib- and imatinib-treated patients, but overall, no new safety signals were identified. These final results support 400-mg once-daily bosutinib as standard-of-care in patients with newly diagnosed CP CML. This trial was registered at www.clinicaltrials.gov as #NCT02130557.


INTRODUCTION
Bosutinib is approved for the treatment of patients with Philadelphia chromosome-positive (Ph+) chronic myeloid leukemia (CML) resistant/intolerant to prior therapy and patients with newly diagnosed chronic phase (CP) CML [1][2][3][4][5][6]. Approval of first-line bosutinib was based on primary results from the phase 3 BFORE trial, which showed superior efficacy of bosutinib versus imatinib in the modified intentto-treat (mITT) population (Ph+ patients with e13a2/e14a2 transcripts) after ≥12 months of follow-up [7]. We report the final efficacy and safety results from BFORE after five years of follow-up. BFORE (ClincalTrials.gov, NCT02130557) was an open-label, randomized, multicenter, phase 3 study; methods have been published [7,8]. Patients aged ≥18 years, with newly diagnosed BCR::ABL1-positive CP CML, were randomized 1:1 to receive (starting dose) bosutinib or imatinib 400 mg once daily. On-study treatment was continued for five years (240 weeks; end of study) or until treatment failure, unacceptable toxicity, death, or withdrawal of consent. Patients who discontinued treatment prior to completing five years were followed for survival until completion of five years on study, death, or withdrawal of consent. At the end of the planned five years, patients could continue with their ongoing treatment at the discretion of the investigator.

METHODS Study design and patients
The primary endpoint was major molecular response (MMR; BCR:: ABL1 ≤ 0.1% on the international scale [IS]) at 12 months (mITT population).
The study was conducted in accordance with the Declaration of Helsinki. Patients provided written informed consent, and the protocol was approved by study-site institutional review boards. This final analysis was based on a last patient/last visit of 17 April 2020 (12 June 2020 database lock), five years after the last patient enrolled.

Efficacy and safety assessments
The short-term secondary endpoint, MMR by month 18 (not previously reported), is included. Long-term secondary endpoints included duration of complete cytogenetic response (CCyR), duration of MMR, on-treatment event-free survival (EFS), and overall survival (OS). Exploratory endpoints included time to response (TTR), on-treatment transformation to accelerated phase (AP) or blast phase (BP) CML, and newly observed BCR::ABL1 mutations. Post-hoc analyses included cumulative response rates by five years, cumulative molecular response (MR) rate by Sokal risk group, duration of MR 4 , sustained 1-year MR 4 (≥3 years on treatment, and BCR:: ABL1 ≤ 0.01% IS in all consecutive assessments for ≥1 year), sustained twoyear MR 4 (≥4 years on treatment, and BCR::ABL1 ≤ 0.01% IS in all consecutive assessments for ≥2 years), and efficacy by BCR::ABL1 ≤ 10% IS at three months in evaluable patients with ≥3000 ABL1 copies at three months. Additional methods are provided in the supplementary material.

Statistical analysis
This analysis evaluated efficacy in the ITT population (all randomized patients), with the exception of cytogenetic endpoints, which were evaluated in the mITT population (Ph+ patients with e13a2 and/or e14a2 transcripts). Results for prespecified endpoints in the hierarchical testing strategy are displayed for the ITT population (results were consistent with those in the mITT population [data not shown]). Per protocol, CCyR was imputed on any date where MMR was achieved and no valid cytogenetic assessment was available. Definitions for TTR, duration of response, on-treatment EFS and OS, censoring for time-to-event endpoints, and imputation methods have been described [7].
All efficacy analyses were based on assessments up through 28 days after the last dose of study medication except for OS, which included posttreatment follow-up data. Response data after treatment discontinuation were not collected.
Confirmed loss of response was defined as two consecutive assessments at least 28 days apart, treatment discontinuation due to suboptimal response/treatment failure or progressive disease, or death due to progressive disease within 28 days of last dose. Confirmed loss of BCR:: ABL1 transcripts ≤1% IS was included as an additional EFS event for Philadelphia chromosome-negative/unknown Philadelphia chromosome status e13a2/e14a2 patients. Duration of response was measured from the first date of response until the first date of loss of response that was subsequently confirmed. Loss of CCyR was defined as ≥1 Ph+ metaphase from <100 metaphases analyzed. Loss of MMR and MR 4 was defined as BCR::ABL1 transcripts >0.1% and >0.01% IS, respectively, with ≥5-fold increase from the lowest recorded value.
Safety data were summarized descriptively and included all randomized patients who received ≥1 dose of study medication.
All hazard or odds ratios are bosutinib vs imatinib. Ratios <1 for duration of response, EFS, and OS, and ratios >1 for response and TTR were considered to favor bosutinib. For all endpoints, 95% confidence intervals (CIs) excluding 1 were considered predictive of the outcome of interest.

RESULTS
Disposition, demography, and baseline characteristics A total of 536 patients were randomized to bosutinib (n = 268) or imatinib (n = 268), of whom 268 and 265, respectively, received study treatment (Fig. 1). Patient baseline demographics and disease characteristics (ITT population) were well balanced across treatment arms (Supplementary Table 1   bosutinib-versus imatinib-treated patients were still receiving treatment; 86.6% versus 86.2% of randomized patients completed five years of follow-up. The most common primary reasons for permanent treatment discontinuation were adverse events (AEs) in the bosutinib arm (bosutinib, 25.0% vs. imatinib, 12.5%) and lack of efficacy (suboptimal response, treatment failure, or disease progression) in the imatinib arm (imatinib, 17.7% vs. bosutinib, 5.6%; Fig. 1).
The rate of patients achieving a sustained MR 4 was also assessed. In the bosutinib versus imatinib arms, 42.9% (95% CI, 37.0-48. 8 4 . In a subdistributional hazards model, BCR::ABL1 transcript level ≤10% at three months was predictive of time to a one-year sustained MR 4 , and Eastern Cooperative Oncology Group performance status 0 and BCR::ABL1 transcript level ≤10% at three months were predictive of time to a two-year sustained MR 4 (Supplementary Table 3).
On-treatment transformations to AP/BP CML occurred in six bosutinib-and seven imatinib-treated patients. Of these, six (three in each arm) met AP criteria within two weeks of randomization based solely on increased basophil count and did not appear to be true transformations, as their clinical course was not consistent with AP/BP. None of these six patients discontinued treatment due to progression to AP/BP or death. Of the remaining patients, all three bosutinib-treated patients progressed to BP; three . In both treatment arms, the cumulative incidence function of MMR (Fig. 3A), as well as MR 4 and MR 4.5 (data not shown), was higher in patients who had BCR:: ABL1 transcripts ≤10% at three months versus those who did not. Cumulative incidence function of on-treatment EFS events by BCR::ABL1 transcript level (≤10% vs. >10% IS) at three months is shown in Fig. 3B; a lower rate of EFS events was observed for patients with BCR::ABL1 transcript level ≤10% (vs. >10%) at three months in the imatinib arm.
Renal TEAEs were reported in 28 (10.4%) versus 26 (9.8%) patients treated with bosutinib versus imatinib; increased blood creatinine was the most common TEAE in both arms ( Table 2). Decreases from baseline in estimated glomerular filtration rate (eGFR) based on the Modification of Diet in Renal Disease method and increases in serum creatinine were observed over time in both treatment arms (Supplementary Fig. 2A, B). At 60 months, median decline from baseline eGFR was 14.

DISCUSSION
This final analysis of the BFORE trial demonstrated long-term efficacy and safety of bosutinib in patients with newly diagnosed CP CML. After 5 years of follow-up, superior MR was demonstrated with bosutinib versus imatinib. An improvement in MR in favor of bosutinib was identified across all Sokal risk groups, with the greatest improvement observed in patients with Sokal high-risk, which is an important factor if treatment-free remission is considered as a treatment goal [9]. Furthermore, the rate of early MR at 3 months was higher with bosutinib than with imatinib and, in both treatment arms, the cumulative incidence of MMR and deep molecular response (DMR; defined as MR 4 and MR 4.5 ) was higher in patients who had BCR::ABL1 transcripts ≤10% at 3 months versus those who did not. The 5-year follow-up of the second-generation tyrosine kinase inhibitors (TKIs) nilotinib (ENESTnd trial) and dasatinib (DASISION trial) in patients with newly diagnosed CP CML has been reported [10,11]. Although comparisons between trials should be considered with caution, MR rates with bosutinib align with the MMR and DMR rates observed with nilotinib and dasatinib. An improvement in cumulative MR rates by 60 months with bosutinib versus imatinib (difference [Δ] in response) was demonstrated for  Treatment-free remission is an emerging treatment goal of increasing importance, with several studies demonstrating that a substantial proportion of patients who achieve stable DMR maintain response after TKI discontinuation [12]. Few clinical trials have prospectively evaluated the incidence of patients achieving a sustained DMR. One study in de novo imatinib-treated patients reported a cumulative incidence of sustained (≥2 years) MR 4.5 of 36.5% after eight years of treatment [13]. In a retrospective analysis of patients treated with frontline TKIs, with a median follow-up of 103 months, 47% of patients achieved a sustained (≥2 years) MR 4.5 at any time [14]. In our study with a median follow-up of 55.2 months, a two-year sustained MR 4 was achieved by 32.5% of patients treated with bosutinib versus 26.5% with imatinib. Some patients with a confirmed loss of MR subsequently regained the respective response with continued treatment, reflecting the fluctuations in BCR::ABL1 often observed in patients before the achievement of a sustained DMR. This suggests that a follow-up ≤60 months may be insufficient to adequately assess the proportion of patients achieving sustained DMR. However, the rate observed with bosutinib by five years was similar to imatinib by eight years, suggesting (with acknowledgement of caution in comparison across studies) that treatment with second-generation TKIs may allow patients to achieve a sustained DMR faster, as would be expected based on the earlier achievement of DMR with second-generation TKIs [15]. This study also confirmed the achievement of BCR::ABL1 transcript level ≤10% at three months as predictive of sustained MR 4 , as previously suggested in studies with other TKIs [15].
Despite the increasing interest in treatment-free remission, the majority of patients with CML will still require lifelong TKI treatment, and therefore preserving or improving health-related quality of life (HRQoL) remains an important consideration for treatment selection [16]. A previous analysis of BFORE demonstrated that HRQoL was maintained or improved compared with baseline after 12 months of bosutinib or imatinib treatment [17]. In addition, a pooled analysis of the bosutinib and imatinib arms showed that a better molecular response with tyrosine kinase inhibitor treatment was generally associated with improved HRQoL [8].
Safety data were consistent with the known safety profiles of bosutinib and imatinib in newly diagnosed patients with CP CML, and with second-line or later bosutinib treatment, with no new safety signals identified [4,5,7,[18][19][20][21][22]. The onset of TEAEs occurred primarily during the first year of treatment and they were generally manageable, with few new TEAEs (eg, effusion events) occurring in later years. In general, permanent treatment discontinuations due to AEs occurred early during treatment, most during the first year, confirming the importance of closely monitoring patients following initiation of treatment, particularly since rechallenge after temporary discontinuation due to toxicity has often been shown to be successful if management recommendations are followed [23,24]. In patients receiving bosutinib or imatinib, there was a slight increase in the overall incidence of AEs of special interest; however, few patients in either arm discontinued treatment due to these AEs.
Liver function abnormalities were the most common AEs leading to treatment discontinuation of bosutinib. Although diarrhea was frequently reported in bosutinib-treated patients, few permanently discontinued treatment due to diarrhea, and the event rate was similar between treatment arms. Guidelines for the management of AEs occurring with bosutinib treatment have been published [23,25].
As opposed to cerebrovascular and peripheral vascular events, which did not differ between treatments, cardiovascular TEAEs, although they remained low (≤5%) in both arms, were higher in the bosutinib versus imatinib arm (Table 3).  Exposure-adjusted incidence rates of cardiac and vascular TEAEs were slightly higher with bosutinib 400 mg once daily versus those observed in the phase 3 BELA trial of bosutinib 500 mg/d for newly diagnosed CP CML; however, patients in BFORE had a higher cardiovascular comorbidity burden at baseline compared with patients in BELA (Supplementary Tables 2, 9) [18,19,26].
Hyperlipidemia and hyperglycemia are major cardiovascular risk factors [27]. In this study, the overall rate of metabolic TEAEs was similar in the bosutinib and imatinib arms, with hyperlipidemia and hyperglycemia reported in ≤5% of patients in the bosutinib arm. These rates appear to be lower with bosutinib than those previously reported with nilotinib [10,28].
Pleural effusions are more commonly associated with dasatinib; after five years of follow-up, 28% of patients receiving first-line dasatinib reported pleural effusions. Although their occurrence was higher with bosutinib than with imatinib, the incidence (6%) after five years appears to be lower compared with dasatinib. Importantly, pleural effusions can also first occur years after treatment (Table 3); however, they were generally manageable and rarely led to treatment discontinuation.
Renal dysfunction has been reported with imatinib and bosutinib and, to a lesser degree, with dasatinib [29,30]. In this study, there was a similar decline in eGFR over time with both treatments; however, few patients in either treatment arm had a decline to Kidney Disease Improving Global Outcomes grade ≥3b, and~50% of those patients had returned to grade ≤3a at their last assessment, suggesting a reversible mechanism.
Although the efficacy of bosutinib, nilotinib, and dasatinib is similar, bosutinib has a distinct safety profile, with a low incidence of some TEAEs compared with other TKIs (eg, vascular and effusion TEAEs) but higher incidence of other AEs (eg, diarrhea, liver) [6,11]. A number of factors, including patients' comorbidities and risk factors as well as the safety profile and schedule of administration of TKIs, should be considered when selecting the most appropriate TKI for the treatment of newly diagnosed patients with CP CML [23].
In conclusion, first-line bosutinib continued to show superior efficacy versus imatinib, with patients who received bosutinib achieving earlier and deeper MR. AEs were generally manageable, reversible, and consistent with the known safety profiles of both drugs. These results confirm the use of bosutinib as a standard of care in patients with newly diagnosed CP CML.
Information about this study in a plain language format is available in the supplementary materials.

DATA AVAILABILITY
Upon request, and subject to review, Pfizer will provide the data that support the findings of this study. Subject to certain criteria, conditions and exceptions, Pfizer may also provide access to the related individual de-identified participant data. See https://www.pfizer.com/science/clinical-trials/trial-data-and-results for more information.