Durable treatment-free remission in patients with chronic myeloid leukemia in chronic phase following frontline nilotinib: 96-week update of the ENESTfreedom study

Purpose ENESTfreedom is evaluating treatment-free remission (TFR) following frontline nilotinib in patients with chronic myeloid leukemia (CML) in chronic phase. Following our primary analysis at 48 weeks, we here provide an updated 96-week analysis. Methods Attempting TFR required ≥ 3 years of nilotinib, a molecular response of MR4.5 [BCR-ABL1 ≤ 0.0032% on the International Scale (BCR-ABL1IS)], and sustained deep molecular response (DMR) during a 1-year consolidation phase. Patients restarted nilotinib following loss of major molecular response (MMR; BCR-ABL1IS ≤ 0.1%). Results Ninety-six weeks after stopping treatment (3.6-year median prior nilotinib duration), 93 of 190 patients (48.9%) remained in TFR. Of 88 patients who restarted nilotinib following loss of MMR, 87 regained MMR and 81 regained MR4.5 by the data cut-off. Ninety-six-week TFR rates were 61.3, 50.0, and 28.6% in patients with low, intermediate, and high Sokal risk scores at diagnosis, respectively. Patients consistently in MR4.5 during consolidation had higher TFR rates (50.6%) than patients with ≥ 1 assessment without MR4.5 during consolidation (35.0%). In a landmark analysis, 96-week TFR rates for patients with MR4.5, MR4 (BCR-ABL1IS ≤ 0.01%) but not MR4.5, and MMR but not MR4 at TFR week 12 were 82.6, 23.1, and 0%, respectively. There were no reports of disease progression or death due to CML; overall adverse event frequency decreased following TFR. Within the follow-up period, TFR did not adversely affect disease outcomes. Conclusions These results demonstrate the feasibility and durability of TFR following frontline nilotinib and emphasize the importance of sustained DMR for TFR.


Introduction
For patients with chronic myeloid leukemia in chronic phase (CML-CP) who have achieved a stable deep molecular response (DMR) using BCR-ABL1 tyrosine kinase inhibitors (TKIs), treatment-free remission (TFR) following TKI cessation is an emerging goal (Hughes and Ross 2016;Mahon et al. 2010;Etienne et al. 2017;Rousselot et al. 2014;Ross et al. 2013;Lee et al. 2016;Mori et al. 2015;Hochhaus et al. 2017a;Imagawa et al. 2015;Nakamae et al. 2017;Saussele et al. 2017;Rea et al. 2017;Boquimpani et al. 2014;Villemagne Sanchez et al. 2018). Potential reasons why patients or physicians may wish to attempt TFR include the possibility of reducing adverse events (AEs), avoiding long-term toxicities, reducing costs, improving quality of life, the convenience of having fewer medications to take, and the ability to plan pregnancies (Boquimpani et al. 2014;Villemagne Sanchez et al. 2018;Jiang et al. 2016). In the original Stop Imatinib (STIM1) trial, which has the longest TFR follow-up to date, patients attempted TFR following achievement of sustained undetectable minimal residual disease with imatinib (Etienne et al. 2017;Mahon et al. 2010). Sixty months after TKI cessation, the Kaplan-Meier-estimated molecular recurrence-free survival rate was 38%, and most instances of molecular recurrence occurred during the first 6 months. Other TFR studies have reported comparable results (Ross et al. 2013;Rousselot et al. 2014;Lee et al. 2016;Mori et al. 2015;Hochhaus et al. 2017a;Imagawa et al. 2015;Nakamae et al. 2017;Saussele et al. 2017;Rea et al. 2017;. In light of the success of TFR studies and the possible advantages that TFR might confer, the US National Comprehensive Cancer Network and the European Society for Medical Oncology have published guidelines that include criteria for attempting TFR outside of clinical trials (National Comprehensive Cancer Network 2017;Hochhaus et al. 2017b).
For patients with newly diagnosed CML-CP, frontline treatment with the second-generation TKI nilotinib can lead to a higher rate of stable DMR (and TFR eligibility) than frontline treatment with imatinib (Hochhaus et al. 2015). In the Evaluating Nilotinib Efficacy and Safety in Clinical Trials-Newly Diagnosed Patients (ENESTnd) study, nilotinib resulted in higher 5-year cumulative rates of MR 4.5 [BCR-ABL1 ≤ 0.0032% on the International Scale (BCR-ABL1 IS )] than imatinib (nilotinib 300 mg twice daily, 54%; nilotinib 400 mg twice daily, 52%; imatinib, 31%) , and an estimated ≥ 80% of patients achieving MR 4.5 with nilotinib maintained the response for ≥ 1 year (Hochhaus et al. 2015). The ENESTfreedom study is the first study specifically to assess the feasibility of TFR in patients with CML-CP achieving sustained DMR with frontline nilotinib (Hochhaus et al. 2017a). The primary analysis reported that 48 weeks after attempting TFR, 51.6% of patients remained off treatment without loss of major molecular response (MMR; BCR-ABL1 IS ≤ 0.1%). This is comparable to TFR rates following imatinib treatment, despite a shorter duration of prior TKI therapy (median of 3.6 years in ENESTfreedom vs ≈ 5-8 years in prior imatinib studies) (Hochhaus et al. 2017a;Etienne et al. 2017;Ross et al. 2013;Rousselot et al. 2014;Lee et al. 2016;Mori et al. 2015). Based on the results of the primary analysis from ENESTfreedom, as well as results from the ENESTop study of TFR following second-line nilotinib, nilotinib became the first TKI with TFR in its product label (Hochhaus et al. 2017a;Mahon et al. 2018;Novartis Pharmaceuticals Corporation 2017a, b). We now report an updated analysis of the ENESTfreedom study based on 96 weeks of follow-up in the TFR phase.

Study design
ENESTfreedom is an ongoing, single-arm, phase 2 study (NCT01784068) that has been previously described (Hochhaus et al. 2017a). Briefly, adult patients (aged ≥ 18 years) with Philadelphia chromosome-positive CML-CP, ≥ 2 years of frontline nilotinib therapy, and MR 4.5 were eligible to enroll. Patients having received prior interferon alfa therapy or > 4 weeks of any other BCR-ABL1 TKI were not eligible. Following enrollment, patients entered a 1-year consolidation phase ( Fig. 1) during which they continued nilotinib treatment and were monitored by real-time quantitative polymerase chain reaction (RQ-PCR) every 12 weeks. Patients who maintained DMR during the consolidation phase, meaning no assessment worse than MR 4 (BCR-ABL1 IS ≤ 0.01%), ≤ 2 assessments between MR 4 and MR 4.5 , and MR 4.5 in the last assessment, could enter the TFR phase. Patients in the TFR phase were monitored by RQ-PCR every 4 weeks during the first 48 weeks, every 6 weeks during the next 48 weeks, and every 12 weeks thereafter up to 264 weeks after the last patient entered the TFR phase. Nilotinib reinitiation was triggered by any single assessment showing BCR-ABL1 IS > 0.1% (loss of MMR). Patients in the reinitiation phase were monitored every 4 weeks for the first 24 weeks and every 12 weeks thereafter (or as clinically indicated for patients who had not regained MMR) to assess molecular response.

Updated and additional analyses
Here, we report updated results based on a data cut-off date of 31 October 2016, at which time all patients who entered the TFR phase had completed 96 weeks of TFR, transitioned to the reinitiation phase, or discontinued from the study. The TFR rate was calculated as a percentage with an exact 95% Clopper-Pearson confidence interval. Treatment-free survival (TFS) was defined as the time from TFR start to the loss of MMR, nilotinib reinitiation for any reason, progression to accelerated phase/blast crisis (AP/BC), or death from any cause and was estimated using the Kaplan-Meier method. For patients without any such event, TFS was censored at the date of last assessment. Molecular response rates in patients who entered retreatment due to loss of MMR in the TFR phase were calculated as cumulative incidences.
To investigate factors potentially associated with TFR, we stratified all patients attempting TFR by Sokal risk score at diagnosis and by MR 4.5 stability during the consolidation phase and calculated TFR rates at 48 and 96 weeks in each subgroup. Patients' Sokal risk scores at diagnosis were not collected at study entry but rather retrospectively following a protocol amendment; therefore, data on Sokal risk at diagnosis were not available for all patients. MR 4.5 stability during the consolidation phase was defined based on patients' responses during the 4 protocol-mandated RQ-PCR assessments collected during that phase. TFR rates were evaluated in patients with MR 4.5 in all consolidation phase assessments vs those with ≥ 1 assessment worse than MR 4.5 . To assess the association between BCR-ABL1 IS levels at week 12 of the TFR phase and future TFR maintenance, we stratified patients remaining in the TFR phase for ≥ 12 weeks by BCR-ABL1 IS level at week 12 (MR 4.5 , MR 4 but not MR 4.5 , or MMR but not MR 4 ) and calculated TFR rates at 48 and 96 weeks. Patients with missing BCR-ABL1 IS values at week 12 were excluded from this analysis.
Updated safety data included a characterization of AEs across the consolidation phase and the first and second 48 weeks of the TFR phase for patients remaining in the TFR phase for > 48 weeks. The musculoskeletal-pain AE grouping included events reported using the preferred terms of myalgia, arthralgia, bone pain, spinal pain, pain in extremity, and musculoskeletal pain. Time to first musculoskeletal-pain event among patients entering the TFR phase was estimated using the Kaplan-Meier method. AEs were assessed according to the Common Terminology Criteria for Adverse Events version 4.03.

Ethical approval
This study was designed and conducted in accordance with the ethical principles of the Declaration of Helsinki and the ICH Harmonized Tripartite Guidelines for Good Clinical Practice. An independent ethics committee or institutional review board for each study center reviewed the study protocol and its amendments. All patients provided written informed consent before any study procedures and in accordance with local laws and regulations.

Patients
Patients who entered the TFR phase (n = 190) had a median age at study entry of 55 years (range 21-86 years), median total nilotinib duration prior to entering the TFR phase of 43.5 months (range 32.9-88.7 months), and a median time from first MR 4.5 until entering the TFR phase of 30.4 months (range 12.3-83.0 months). The Sokal score at diagnosis was low in 62 patients (32.6%), intermediate in 50 patients (26.3%), high in 28 patients (14.7%), and unavailable for the remaining patients. The median duration of follow-up in the TFR phase at the 96-week data cut-off date was 75.9 weeks (range 8.4-133.0 weeks). At the data cut-off date, of the 190 patients who entered the TFR phase, 93 (48.9%) remained 1 3 in the TFR phase, 88 (46.3%) reinitiated nilotinib, and 9 (4.7%) discontinued from the study while in the TFR phase (Fig. 2). Of the 88 patients who entered the reinitiation phase by the data cut-off, 69 (78.4%) remained in this phase and 19 (21.6%) had discontinued from the study.  (Fig. 4). As previously described (Hochhaus et al. 2017a), the single patient who did not regain MMR discontinued from the study due to patient decision 7 weeks after reinitiating treatment. Of the 6 patients who regained MMR but not MR 4.5 , 1 remained on study and 5 discontinued from the study by the data cut-off date, between 5 and 25 weeks after nilotinib reinitiation (2 due to AEs, 1 due to lack of efficacy, and 2 due to individual decisions).  (Fig. 5). The TFR rates in patients with MR 4.5 , MR 4 but not MR 4.5 , and MMR but not MR 4 at 12 weeks were 86.2, 23.1, and 0% at 48 weeks and 82.6, 23.1, and 0% at 96 weeks, respectively. Among patients with MMR but not MR 4 at 12 weeks, the last patient to lose MMR did so at 48 weeks.

Safety
Eight deaths were reported on study by the data cut-off date, three of which occurred after the 48-week data cutoff (Table 2): one death from respiratory failure during the treatment reinitiation phase and two deaths from non-CML cancers > 30 days after study discontinuation. To date, no deaths due to CML or progressions to AP/BC have been reported in the TFR population. Among patients who remained in the TFR phase for > 48 weeks (n = 100), the overall frequency of AEs decreased from 85.0% during the consolidation phase to 76.0% during the first 48 weeks of the TFR phase and 62.0% during the second 48 weeks of the TFR phase. In this group of patients, the most frequent all-grade AEs reported during the second 48 weeks of TFR were nasopharyngitis (9.0%) and back pain (5.0%). Overall, among these 100 patients, cardiovascular events were reported in 3 during the consolidation phase, 2 during the first 48 weeks of TFR, and 1 during the second 48 weeks of TFR (Table 3); the frequency of AEs in the musculoskeletal-pain grouping increased upon starting TFR (34.0% during the first 48 weeks of TFR vs 17.0% during the Fig. 3 Kaplan-Meier-estimated TFS among all patients who entered the TFR phase. a MMR major molecular response (BCR-ABL1 ≤ 0.1% on the International Scale), TFR treatment-free remission, TFS treatment-free survival. a TFS was defined as the time from the start of TFR until the earliest of any of the following: loss of MMR, reinitiation of nilotinib for any reason, progression to accelerated phase/blast crisis, or death due to any cause. By the data cut-off date, one patient had lost MMR at week 120, at which time only eight patients were considered at risk, resulting in the artificial drop seen at the end of the curve     consolidation phase) but then decreased as TFR continued (9.0% during the second 48 weeks of TFR). At the time of the 96-week data cut-off, with < 50% of patients experiencing musculoskeletal pain-grouping AEs, a median time to the first occurrence of these AEs after treatment cessation could not be estimated.

Conclusions
Results of the 96-week analysis affirm findings from the 48-week analysis. The TFR rate was 48.9% at 96 weeks and 51.6% at 48 weeks (Hochhaus et al. 2017a), suggesting a very low risk of relapse in patients remaining in TFR for > 48 weeks. As already noted in the 48-week analysis (Hochhaus et al. 2017a), most TFS events occurred during the first 24 weeks of TFR, after which the Kaplan-Meier curve for TFS approached a plateau. Of the 98 patients in the TFR phase at 48 weeks, only 3 exited this phase due to loss of MMR during the second 48 weeks, and 2 others discontinued the study despite remaining in MMR. Patients reinitiating nilotinib following loss of MMR promptly regained molecular responses; 98.9% regained MMR and 92.0% regained MR 4.5 by the data cut-off, demonstrating that temporary treatment cessation is safe in patients who experience molecular relapse. This 96-week TFR rate is comparable to TFR rates reported in other studies (Etienne et al. 2017;Ross et al. 2013;Mahon et al. 2010;Nakamae et al. 2017;Mori et al. 2015;Rea et al. 2017;Saussele et al. 2017;Rousselot et al. 2014;Lee et al. 2016;Imagawa et al. 2015), including the ENESTop study (96-week TFR rate of 53.2%) (Mahon et al. 2018), and the high response rate following nilotinib reinitiation is also consistent with results from ENESTop (Mahon et al. 2018). The frequency of AEs, including musculoskeletal pain AEs, decreased during the second 48 weeks of the TFR phase. This pattern of musculoskeletal pain AEs occurring early during the TFR phase and decreasing later is similar to that reported following second-line nilotinib cessation in the ENESTop study (Mahon et al. 2018). A TKI withdrawal syndrome of musculoskeletal pain has also been reported following imatinib or dasatinib treatment cessation in other TFR studies, although it has not been rigorously characterized, and the underlying biological mechanisms are currently unknown (Lee et al. 2016;Richter et al. 2014;Saußele et al. 2016;Shah et al. 2017).
Importantly, TFR did not negatively affect patients' clinical outcomes. There were no reports of progression to AP/ BC or deaths due to CML, and AE frequencies generally decreased following nilotinib cessation. The number of cardiovascular events was lower in the second year of TFR than in the first year of TFR or during nilotinib consolidation, although much larger numbers of patients would be required to examine whether the vascular risk associated with nilotinib treatment is reversed during TFR. Together with the rapid restoration of DMR in patients who reinitiated  treatment and the decreasing frequency of musculoskeletal pain AEs over time during TFR, these results suggest that TFR is safe and does not adversely impact patients' clinical outcomes or lead to TKI resistance. We previously performed multivariate logistic regression analyses to evaluate whether patient baseline characteristics (including sex, age, and duration of nilotinib treatment or MR 4.5 prior to study entry) were predictive of successful TFR (Hochhaus et al. 2017a). However, none of the evaluated characteristics were found to be strong predictors of successful TFR. In this 96-week analysis, the subgroup of patients having low Sokal risk scores at diagnosis and the subgroup maintaining MR 4.5 consistently throughout the consolidation phase both had numerically higher TFR rates. In agreement with this, the STIM1 study reported that low and intermediate Sokal risk scores were associated with lower molecular recurrence rates (Etienne et al. 2017). However, the Korean Imatinib Discontinuation and STOP Second-Generation TKI studies reported finding no association between Sokal risk scores and TFR rates, highlighting that predictors of successful TFR are not completely defined (Lee et al. 2016;Rea et al. 2017). Other studies have reported that TFR rates may be higher in patients with longer durations of TKI therapy (Etienne et al. 2017;Lee et al. 2016;Saussele et al. 2017) and those with longer DMR duration prior to treatment cessation (Saussele et al. 2017). As more patients attempt TFR, especially with CML treatment guidelines now noting that TFR may be attempted outside of clinical trials (National Comprehensive Cancer Network 2017; Hochhaus et al. 2017b), the identification of predictors of successful TFR will be critical to enabling physicians to determine which patients would be most likely to benefit from treatment cessation.
We additionally found that a deeper molecular response at 12 weeks of TFR appeared to be predictive of maintaining TFR through 96 weeks. Thus, for patients with MMR but not MR 4.5 at 12 weeks, close monitoring is especially critical, since they may have an increased risk of molecular relapse. This is consistent with the fact that many TFR studies reported that most molecular relapses occurred early in TFR (Etienne et al. 2017;Mahon et al. 2010;Ross et al. 2013;Rousselot et al. 2014;Lee et al. 2016;Mori et al. 2015;Imagawa et al. 2015;Rea et al. 2017;Nakamae et al. 2017;Saussele et al. 2017) and highlights the importance of frequent monitoring early during TFR. However, although molecular relapses tend to be less frequent later in TFR, regular monitoring remains essential for all patients during long-term followup, because late molecular relapses do occur. Every patient who lost MR 4 during the first 12 weeks of the TFR phase went on to lose MMR within the first 48 weeks. We have demonstrated that loss of MMR is a safe and reproducible trigger for retreatment with nilotinib; for patients in whom BCR-ABL1 IS levels rise to above 0.01% within the first 12 weeks, diligent follow-up is critical to ensure that treatment is resumed as soon as loss of MMR is evident.
The updated results from ENESTfreedom presented here demonstrate the durability of TFR following frontline nilotinib and continue to demonstrate the safety of TFR. The results suggest that patients with consistent MR 4.5 throughout the year prior to nilotinib cessation or low Sokal risk scores at diagnosis have numerically higher TFR rates. Strong predictors of successful TFR after stopping nilotinib remain to be identified. Overall, these findings support the use of frontline nilotinib in patients with newly diagnosed CML-CP for whom TFR might be a future treatment goal.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.