Second Tyrosine Kinase Inhibitor Discontinuation Attempt in Patients With Chronic Myeloid Leukemia

Transcription

1 Second Tyrosine Kinase Inhibitor Discontinuation Attempt in Patients With Chronic Myeloid Leukemia Laurence Legros, MD, PhD 1,2 ; Franck E. Nicolini, MD, PhD 3,4 ; Gabriel Etienne, MD, PhD 5 ; Philippe Rousselot, MD, PhD 6 ; Delphine Rea, MD, PhD 7 ;Stephane Giraudier, MD, PhD 8 ; Agnès Guerci-Bresler, MD, PhD 9 ; Françoise Huguet, MD 10 ; Martine Gardembas, MD 11 ; Martine Escoffre, MD 12 ; Jean-Christophe Ianotto, MD, PhD 13 ; Marie-Pierre No el, MD 14 ; Bruno R. Varet, MD, PhD 15 ; Thomas Pagliardini 1 ; Irit Touitou, PhD 1 ;Stephane Morisset, MS 16 ; and Francois-Xavier Mahon, MD, PhD 17 ; on behalf of the French Intergroup for Chronic Myeloid Leukemias BACKGROUND: Several studies have demonstrated that approximately one-half of patients with chronic myeloid leukemia (CML) who receive treatment with tyrosine kinase inhibitors (TKIs) and achieve and maintain a deep molecular response (DMR) are able to successfully discontinue therapy. In patients who have a molecular relapse, a DMR is rapidly regained upon treatment re-initiation. METHODS: The authors report the results from RE-STIM, a French observational, multicenter study that evaluated treatment-free remission (TFR) in 70 patients who re-attempted TKI discontinuation after a first unsuccessful attempt. After the second TKI discontinuation attempt, the trigger for treatment re-introduction was the loss of a major molecular response in all patients. RESULTS: The median follow-up was 38.3 months (range, months), and 45 patients (64.3%) lost a major molecular response after a median time off therapy of 5.3 months (range, 2-42 months). TFR rates at 12, 24, and 36 months were 48% (95% confidence interval [CI], 37.6%-61.5%), 42% (95% CI, 31.5%-55.4%), and 35% (95% CI, 24.4%-49.4%), respectively. No progression toward advanced-phase CML occurred, and no efficacy issue was observed upon TKI re-introduction. In univariate analysis, the speed of molecular relapse after the first TKI discontinuation attempt was the only factor significantly associated with outcome. The TFR rate at 24 months was 72% (95% CI, 48.8%-100%) in patients who remained in DMR within the first 3 months after the first TKI discontinuation and 36% (95% CI, 25.8%-51.3%) for others. CONCLUSIONS: This study is the first to demonstrate that a second TKI discontinuation attempt is safe and that a first failed attempt at discontinuing TKI does not preclude a second successful attempt. Cancer 2017;123: VC 2017 American Cancer Society. KEYWORDS: chronic myeloid leukemia, hematology, myeloproliferative disorders, therapy discontinuation, tyrosine kinase inhibitor. INTRODUCTION Adenosine triphosphate-competitive tyrosine kinase inhibitors (TKIs) in patients with chronic myeloid leukemia (CML) represent a major therapeutic breakthrough, and optimal responders have a near-normal life expectancy. 1 Although the recommendation is to continue lifelong treatment in clinical practice, a pilot study performed more than 10 years ago in patients with so-called molecularly undetectable disease for prolonged periods challenged the notion that TKIs may never be stopped. 2 TKI discontinuation was then evaluated prospectively, and it was demonstrated that imatinib, 2-9 nilotinib, or dasatinib 10,11 could be safely stopped in patients who had long-lasting and deep molecular responses (DMRs). In parallel, important efforts were made to better standardize and score DMRs at different levels of quantitative reverse transcriptase-polymerase chain reaction (RT-qPCR) sensitivity; namely, molecular responses of MR4, MR4.5, and MR5. 12 Corresponding author: Laurence Legros, MD, PhD, Service d Hematologie Clinique, H^opital Archet 1, 151 Route de St. Antoine de Ginestière BP 3079, Nice Cedex 3, Universite C^ote d Azur, CNRS UMR7277, Inserm U1091, Centre Hospitalier Universitaire de Nice, Service d Hematologie Clinique, IBV, Nice, France; legros@unice.fr 1 Hematology Department, Nice University Hospital, Nice, France; 2 Valrose Institute of Biology, National Center for Scientific Research (CNRS) Unit 7277, National Institute of Health and Medical Research (INSERM) Unit 1091, Nice, France; 3 Hematology Department, Lyon University Hospital, Pierre Benite, France; 4 INSERM Unit 1052, Leon Berard Center, Lyon, France; 5 Haematology Department, Bergonie Institute, Bordeaux, France; 6 Haematology and Oncology Department, Hopital A Mignot, INSERM Unit 1173, Versailles, University of Versailles St.-Quentin-Yvelines, France; 7 Adult Hematology Department, INSERM Unit 1160, St. Louis Hospital, Paris, France; 8 Hematology Laboratory, Henri-Mondor Hospital, Creteil, France; 9 Hematology Department, Brabois University Hospital, Vandoeuvre-les-Nancy, France; 10 Hematology Department, Toulouse-Oncopole University Cancer Institute, Toulouse, France; 11 Blood Disorder Department, Angers University Hospital, Angers, France; 12 Adult Hematology Department, Rennes University Hospital, Rennes, France; 13 Hematology Department, Brest University Hospital, Brest, France; 14 Blood Disorders Department, Lille University Hospital, Lille, France; 15 Blood Disorders Department, Necker Hospital, Paris-Descartes University, Paris, France; 16 Hematology Department, Lyon University Hospital, Pierre Benite, France; 17 Laboratory of Mammary and Leukemic Oncogenesis, Genetic Diversity, and Resistance to Treatment, INSERM Unit 1218, Bordeaux, France. We thank the patients and the French Groupe des Biologistes Moleculaires des Hemopathies Malignes (GBMHM) network of molecular biologists, the association Anim Montbernier (Ruy-Montceau, France), and its president Mr. Armand Glasson for its support to the Lyon team. We also thank Mrs. Barbara White and Mr. Ravi Nookala for editorial assistance. DOI: /cncr.30885, Received: April 10, 2017; Revised: May 18, 2017; Accepted: May 25, 2017, Published online July 25, 2017 in Wiley Online Library (wileyonlinelibrary.com) Cancer November 15,

2 Initial attempts of TKI cessation involved imatinib given either upfront or after failure on or intolerance to interferon-a (IFNa), followed by evaluations of discontinuation of the second-generation TKIs dasatinib and nilotinib given as first or subsequent lines of therapy The pioneer Stopping Imatinib (STIM) and Trial of Withdrawing Imatinib in Stable Remission (TWISTER) imatinib-discontinuation trials were based on 2 major eligibility criteria: 1) 3 years of exposure to imatinib, and 2) 2 years of sustained MR4.5 with undetectable peripheral blood breakpoint cluster region-abelson 1 (BCR-ABL1) transcripts (herein referred to as umr4.5). 3,5 Then, the observational A-STIM trial (A Study for Tyrosine Kinase Inhibitors Discontinuation) demonstrated the feasibility of imatinib discontinuation in patients with CML who maintained MR4.5 for 2 years regardless of whether they had detectable BCR- ABL1 transcripts. 6 Recently, the ongoing European Stop Tyrosine Kinase Inhibitor (EURO-SKI) trial explored less deep MR levels as thresholds for discontinuing imatinib, dasatinib, or nilotinib, enrolling patients who maintained MR4 for at least 1 year only. 13 In the STIM and TWISTER trials, the definitions for molecular relapse after TKI discontinuation were quite stringent. 5,13 In the STIM study, a 1-log increase in BCR-ABL1 transcripts between 2 consecutive assessments defined a molecular relapse and triggered imatinib resumption. 13 In the TWISTER study, detectable BCR- ABL1 transcripts at any level in 2 consecutive assessments were considered as a molecular relapse and a trigger for restarting imatinib. 4,7-11,13 Then, the loss of a major molecular response (MMR) was proposed as a more reliable and safe molecular definition for relapse and treatment-free remission (TFR) was preferred over molecular relapse-free survival as the primary endpoint in most TKI-cessation studies. Currently, TFR rates in TKI discontinuation studies are approximately 50%, although variations may be observed, depending on the definition of molecular relapse and characteristics of the study populations. 2-11,14 The most common factor associated with a successful outcome across studies is a long duration of TKI treatment before cessation. 3,5 It has also been reported that the Sokal risk score and DMR duration are factors in success, and the role of different DMR levels also is under investigation in EURO-SKI. 13 In case of a molecular relapse, patients must rapidly resume TKI therapy, and no major efficacy issues have been reported. Currently, whether such relapsing patients may hope for re-attempting TFR is uncertain. In recent years, we observed that imatinib discontinuation after a first unsuccessful cessation attempt in a small series of 16 patients was feasible and did not automatically lead to a molecular relapse. 15 Thus we initiated a larger observational multicenter study named RE-STIM, to evaluate TFR after a second TKI discontinuation attempt. MATERIALS AND METHODS Patients Patients aged 18 years who met criteria for chronicphase CML at diagnosis, as defined by the European LeukemiaNet, 16 were eligible for enrolment in RE-STIM, provided that they failed a first TKI discontinuation attempt, regained umr4.5 after TKI resumption, and discontinued TKI again, regardless of the type of TKI in use before the first and second treatment-free phases and regardless of the reason for a second TKI discontinuation attempt (tolerability issues, proposal by treating physician, or patient willingness). Patients who previously were included in clinical trials like STIM, A-STIM, or EURO- SKI were eligible for enrollment; therefore, failure of a first TKI discontinuation attempt consisted in the loss of umr4.5 with a 1-log increase in BCR-ABL1 transcripts between 2 consecutive assessments or loss of an MMR leading to treatment resumption. Patients who had undergone autologous or allogeneic hematopoietic stem cell transplantation, those with nonmajor BCR-ABL1 transcripts, and those with a history of progression to accelerated-phase or blast-crisis CML were excluded. RE- STIM was approved by the French Advisory Committee on Information Processing in Material Research in the Field of Health (CCTIRS) (no ) and French data protection authority (CNIL). Informed consent was obtained from all patients in accordance with the Declaration of Helsinki. Definition of Molecular Responses and Molecular Relapse Molecular response assessments for BCR-ABL1 messenger RNA quantification by RT-qPCR and scoring of DMR followed European LeukemiaNet recommendations. 12,17,18 In brief, MR4.5 was defined as detectable BCR-ABL/ABL1 (internationally standardize [IS] ratio %) or undetectable BCR-ABL1 in samples with 40,000 ABL1 transcripts until the end of 2012 and 32,000 ABL1 transcripts thereafter. MR4 was defined as detectable BCR-ABL/ABL1 (IS ratio 0.01%) or undetectable BCR-ABL1 in samples with 10,000 ABL1 transcripts. During the second attempt to discontinue TKI, a molecular relapse was defined as loss of MMR (BCR-ABL/ ABL1 IS ratio 0.1%) on any single test. BCR-ABL Cancer November 15, 2017

3 Second TKI Discontinuation in CML/Legros et al Figure 1. Patients disposition is illustrated. TKI indicates tyrosine kinase inhibitor. transcripts were monitored in the peripheral blood by RT-qPCR at the same frequency used in the STIM study, ie, monthly during the first 12 months, every 2 to 3 months during the second year, and every 3 to 6 months for up to 5 years. Monitoring of molecular response was performed in laboratories belonging to the French network of quality assurance and accreditation of molecular biology laboratories for hematologic malignancies (the French Molecular Biology Group in Hematology) (GBMHM). 19 After each TKI discontinuation attempt, the time of reappearance of BCR-ABL1 transcripts was defined as the first assessment indicating detectable BCR- ABL1 transcripts. Endpoints and Statistical Methods The primary endpoint was TFR, as estimated by Kaplan- Meier method and defined as the time elapsed between TKI discontinuation and the date of restarting TKI or MMR loss, whichever came first. For all other patients, data were censored at the date of last molecular evaluation. Log-rank tests were performed if necessary for the graphic illustrations of survival analyses. Secondary endpoints included safety (lack of progression to advancedphase CML and efficacy of treatment resumption). We also searched for factors associated with TFR after the second TKI discontinuation attempt. Descriptive analyses comparing cohorts were performed using common statistical tests, such as the t test or the Mann-Whitney U test for quantitative continuous variables and the chi-square test or the Monte-Carlo test for qualitative variables. We also searched for factors associated with TFR after the second TKI discontinuation attempt through simple regressions using a Cox model. The level of significance was set at 5%, and estimated hazard ratios were described with 95% confidence intervals (CIs). All statistical analyses were performed using R software (version 3.2.3; R Foundation for Statistical Computing, Vienna, Austria). RESULTS Patient Characteristics Seventy patients were included in the RE-STIM study as of October 1, At the time of CML diagnosis, the median patient age was 51 years (range, years) and the Sokal risk score was low, intermediate, high, and unknown in 35 patients (50%), 25 patients (36%), 9 patients (13%), and 1 patient (1%), respectively. Imatinib was received as the first-line TKI in all patients, either alone or combined with IFN (n 5 5) or cytarabine (n 5 9) within the French SPIRIT (STI571 Prospective Randomized Trial) clinical trial. 20 Nineteen patients received IFN before imatinib initiation. At the first TKI discontinuation attempt, the median duration of TKI treatment was 59 months (range, months), and the median duration of umr4.5 was 32 months (range, months). TKI type at the first discontinuation attempt consisted of imatinib in 60 patients and secondline or subsequent line nilotinib or dasatinib in 10 patients (Fig. 1). The reason for switching to nilotinib or dasatinib was imatinib intolerance in 8 patients and lack Cancer November 15,

4 TABLE 1. Characteristics at Second Attempt of Discontinuation Parameter Second Discontinuation, mo TKI duration 32 (6-72) Total TKI duration 103 (55-173) umr4.5 duration 25 (4-68) Total umr4.5 duration 68 (34-113) Time to umr4.5 loss from 2 (1-37) TKI discontinuation TKI-free period 5.3 (2-42) a Time to umr4.5 after 6.5 (2-30) b TKI re-challenge Abbreviations: TKI, tyrosine kinase inhibitor; umr4.5, undetectable molecular response. a N 5 44 patients. b N 5 30 patients. of an MMR in 2 patients. In agreement with RE-STIM inclusion criteria, all patients experienced a molecular relapse. The median time to umr4.5 loss was 3 months (range, 1-22 months), and the median time to TKI resumption was 5 months (range, 1-36 months). The same TKI that was received before treatment cessation was restarted in 61 patients (Fig. 1). Eight patients received dasatinib (n 5 3) or nilotinib (n 5 5) instead of imatinib, and 1 patient was switched back from dasatinib to imatinib (Fig. 1). The median time to regain umr4.5 after TKI resumption was 4.5 months (range, 1-26 months). At the second TKI discontinuation attempt, the median patient age was 60 years (range, years). The median time from TKI resumption to the second TKI discontinuation attempt was 32 months (range, 6-72 months), and the median duration of second umr4.5 before second TKI discontinuation attempt was 25 months (range, 4-68 months) (Table 1). TKI type at the second discontinuation attempt consisted of imatinib in 50 patients, dasatinib in 7 patients, and nilotinib in 13 patients (Fig. 1). The median follow-up after the second TKI discontinuation attempt was 38.3 months (range, months). Molecular Relapses and TFR After the Second TKI Discontinuation Attempt At the time of this analysis, 45 patients (64%) lost an MMR after a median time off therapy of 5.3 months (range, 2-42 months). Most molecular relapses (54%) occurred during the first year after TKI discontinuation, with TFR probabilities at 6 and 12 months of 66% (95% CI, 55.5%-77.9%) and 48% (95% CI, 37.6%- 61.5%), respectively (Fig. 2). However, later onset molecular relapses occurred, as indicated by the steady decrease beyond 12 months: 42% (95% CI, 31.6%- 55.4%) at 24 months and 35% (95% CI, 24.4%- 49.4%) at 36 months (Fig. 2). All relapsing patients restarted TKI treatment after a median of 5.3 months (range, 2-42 months). It is noteworthy that no progression toward advanced-phase CML was observed. The same TKI that was received before the second cessation was restarted in 27 (60%) relapsing patients. Dasatinib (n 5 5) or nilotinib (n 5 4) rather than imatinib was chosen in 9 patients because of prior tolerance issues, and 8 patients switched from a second-generation TKI to another (n 5 5) or to imatinib (n 5 3). The median follow-up of patients who resumed therapy was 39 months (range, 5-90 months). At the last follow-up, 34 patients regained umr4.5 within a median timeframe of 6.5 months (range, 2-30 months), and 8 had an MMR Figure 2. Treatment-free remission (TFR) is illustrated after a second attempt to discontinue tyrosine kinase inhibitor (TKI) treatment. The solid curve represents the TFR probability after the second attempt at TKI discontinuation, and the dashed curves define the confidence interval. The TFR probability at 6, 12, 24, and 36 months after the second attempt to discontinue TKI was 66%, 48%, 42%, and 35%, respectively Cancer November 15, 2017

5 Second TKI Discontinuation in CML/Legros et al TABLE 2. Potential Predictive Factors of Treatment-Free Remission by Univariable Cox Regression Model Analysis Variable No. of Patients HR 95% CI P Age at second discontinuation: < 60 vs > 60 y Sokal risk score 69 Low and intermediate High Prior exposure to IFN 70 No Yes 19 TKI duration at first discontinuation: < 59 vs > 59 mo umr4.5 duration at first discontinuation: < 32 vs > 32 mo TKI type at first attempt 70 Imatinib Dasatinib and nilotinib First discontinuation molecular criteria 70 1yuMR4.5 < 2 y umr4.5 2 y Time to umr4.5 loss from first TKI discontinuation 70 <3 mo >3 mo Time to umr4.5 loss from first TKI discontinuation 70 <6 mo >6 mo Reason for first TKI re-challenge 70 umr4.5 loss MMR loss First TKI-free duration: < 5vs > 5 mo Switch from imatinib to 2G TKI after first discontinuation 60 No Yes 8 Second TKI duration at second discontinuation: < 32 vs > 30 mo Total TKI duration at second discontinuation: < 103 vs < 103 mo Second umr4.5 duration at second discontinuation: < 25 vs > 25 mo Total umr4.5 duration at second discontinuation: < 68 vs > 68 mo Abbreviations: 2G, second-generation; CI, confidence interval; HR, hazard ratio; IFN, interferon; MMR, major molecular response; TKI, tyrosine kinase inhibitor; umr4.5, undetectable molecular disease. response within a median of 4.6 months (range, 2-71 months). Two patients have not yet regained an MMR (with IS ratios of 0.16% and 0.24%), but follow-up is still short (<3 months). We note that 1 patient who lost an MMR at 6 months declined a TKI re-challenge. Consequently, BCR-ABL1 transcripts increased, and a complete hematologic response was lost at 24 months. Since the last blood test, the patient has been lost to follow-up. The median follow-up of the 25 patients (36%) who did not lose an MMR and remained treatment-free was 33.7 months (range, months). Thirteen of these patients maintained umr4.5 during the entire follow-up, whereas 12 had intermittently detectable BCR-ABL1 transcripts, but levels were always below the MMR threshold (results not shown). One patient died at age 93 years from a CML-unrelated cause (myocardial infarction 8 months after a second attempt to discontinue imatinib). Factors Associated With Molecular Relapse We tried to determine whether clinical or biologic variables were associated with outcome after a second TKI discontinuation attempt. In univariate analysis, we failed to identify any association between TFR and the following factors: age, sex, Sokal risk score, prior exposure to IFN, TKI in combination versus monotherapy, TKI type, TKI treatment duration, umr4.5 duration before the first and second discontinuation attempts, and type of molecular relapse after the first discontinuation attempt (umr4.5 vs MMR loss) (Table 2). However, patients who lost umr4.5 later than the median time to umr4.5 loss after the first TKI discontinuation attempt (>3 months) had significantly less molecular relapses after the second TKI discontinuation attempt than those who lost umr4.5 in less than 3 months (P 5.024) (Table 2). Therefore, we analyzed TFR after the second TKI discontinuation attempt according to BCR-ABL1 transcript levels soon after the first TKI discontinuation attempt and observed that TFR probabilities were significantly higher in patients who were still in MMR at 3 months than in those who had already lost MMR (P 5.002) (Fig. 3). Indeed, TFR probabilities at 36 months were 46% (95% CI, 30.6%-68.8%) Cancer November 15,

6 Figure 3. The probability of achieving treatment-free remission (TFR) is illustrated according to molecular status at 3 months after the first attempt to discontinue tyrosine kinase inhibitor (TKI) therapy. The blue curve represents patients who remained in persistent, undetectable molecular disease (umr4.5) (undetectable BCR-ABL1 in samples with 40,000 ABL1 transcripts until the end of 2012 and 32,000 ABL1 transcripts later on; see definition page 2 in definition of molecular responses and molecular relapse ) at 3 months (M3); the green curve represents those who lost umr4.5 but retained a major molecular response (MMR); and the red curve represents patients who lost an MMR within 3 months. TFR probabilities at 36 months were 46% (95% confidence interval [CI], 30.6%-68.8%) in the umr4.5 group, 31% (95% CI, 17.9%-54.8%) in the MR4.5 group with conserved MMR, and only 0% (95% CI, 100%-100%) in the group that lost an MMR. Figure 4. The probability of achieving treatment-free remission (TFR) is illustrated according to undetectable/detectable status at 3 months after the first attempt to discontinue tyrosine kinase inhibitor (TKI) therapy. The blue curve represents patients who remained in persistent, undetectable molecular disease (umr4.5) (undetectable BCR-ABL1 in samples with 40,000 ABL1 transcripts until the end of 2012 and 32,000 ABL1 transcripts later on; see definition page 2 in definition of molecular responses and molecular relapse ) at 3 months (M3), and the red curve represents patients who lost umr4.5 within 3 months. TFR probabilities at 36 months were 72% (95% confidence interval [CI], 48.8%-100%) in the persistent umr4.5 group and 27% (95% CI, 16.8%- 44%) in the group that lost umr4.5. in the group that maintained persistent umr4.5, 31% (95% CI, 17.9%-54.8%) in the group that lost umr4.5 but conserved MMR, and 0% (95% CI, 100%-100%) in the group that lost MMR (Fig. 3). Consequently, TFR probabilities were significantly higher for patients who were still in umr4.5 at 3 months than in those who had already lost umr4.5 (P 5.021) (Fig. 4). Indeed, TFR probabilities at 36 months were 72% (95% CI, 48.8%- 100%) in the persistent umr4.5 group and 27% (95% CI, 16.8%-44%) in the group that lost umr4.5. DISCUSSION Currently, patients with CML who achieve stable DMR on long-lasting TKI therapy have the opportunity to attempt treatment discontinuation, but only one-half of them successfully discontinue TKI therapy. 21 Relapsing patients must resume therapy; and, although they regain DMR upon TKI re-challenge, whether a second treatment discontinuation attempt is feasible is unknown. For the first time, we demonstrate that a second TKI cessation attempt is safe and potentially successful Cancer November 15, 2017

7 Second TKI Discontinuation in CML/Legros et al On the basis of the loss of MMR criterion as a trigger for treatment resumption, we report TFR rates of 42% at 24 months and 35% at 36 months. Currently, the only point of comparison is the first attempt TFR studies, such as STIM, A-STIM, the Imatinib Suspension and Validation (ISAV) Study, the Dutch-Belgian Cooperative Trial for Haemato-Oncology (HOVON), the Japan survey, and TWISTER. 3-8 Studies in which TKIs were discontinued under comparable circumstances (minimal treatment length and DMR level and with the same molecular relapse definition) reported estimated TFR rates at 24 months from 52% to 64%. Thus, a second attempt of TKI discontinuation appears to be less successful than a first attempt. We failed to identify the usual predictive factors associated with molecular relapse during a first attempt, such as the Sokal risk score and umr4.5 and TKI duration (Table 2). This difference could have been explained by the first umr4.5 duration, which was 2 years in TFR studies and a minimum of 1 year in the current RE-STIM study; however, molecular stringency (umr4.5 2 years) and second TKI duration 3 years do not appear to be the main factors for a successful or unsuccessful second attempt of TKI discontinuation (data not shown). It is noteworthy, in the patients who maintained a persistent umr4.5 at 3 months, TFR rates were similar to those from the first attempt cessation studies. Indeed, patients who remain in umr4.5 at 3 months during the first discontinuation have a better TFR at the second discontinuation, with a somewhat identical TFR probability at 24 and 36 months (72%; 95% CI, 48.8%- 100%) compared with 36% (95% CI, 25.8%-51.3%) and 27% (95% CI, 16.8%-44%) in the other patients. Another difference observed in our study compared with the majority of TFR studies is the molecular relapse pattern. Mathematical models of CML indicate that the average rate of exponential increase after stopping imatinib is per day, corresponding to a doubling time of 8 days. 22 This explains why, in most of the TKI cessation studies, the majority of molecular relapses occurred during the first 6 months after discontinuation, with only a few relapses observed at later time points. 3,5,6,11 Differing from molecular relapse profiles after the first attempt, here, we have demonstrated more delayed relapses spread over time, leading to a lower TFR rate of 35% (95% CI, 24.4%-49.4%) at 36 months and encouraging us to propose close molecular monitoring even after 24 months. Molecular relapses are related to the reactivation of clones entering the cell cycle 23,24 and may be related to various factors, including residual leukemic stem cell burden, but the lack of stringent criteria impacting on the success of second TKI discontinuation does not support this hypothesis. The delayed relapse pattern observed in the RE-STIM study might be better explained by an indirect mechanism as the late failure or exhaustion of the autologous immune system 25 or as long-term microenvironment alterations Patients who remain in umr4.5 at 3 months during the first discontinuation have a higher rate of TFR at the second discontinuation compared with the other patients. Therefore, the impact of delaying a molecular relapse after discontinuation on the success of this strategy could support the idea of controlling the reactivation of residual leukemic stem cells. In conclusion, a second attempt to discontinue TKI therapy in patients who failed a first discontinuation and subsequently regained umr4.5 after TKI re-challenge is safe. This strategy is more successful in patients who relapse later than 3 months during a first discontinuation attempt. Very close and prolonged molecular monitoring is required. To confirm the long-term safety of this strategy, patients should be registered or included in a prospective trial. FUNDING SUPPORT No specific funding was disclosed. CONFLICT OF INTEREST DISCLOSURES Laurence Legros reports personal fees from Bristol Myers-Squibb (BMS), Incyte, Novartis, and Pfizer and research support from Novartis, Pfizer, and Incyte outside the submitted work. Franck E. Nicolini reports grants from Novartis and personal fees from BMS during the study; grants, personal fees, and other support from Novartis outside the submitted work; and personal fees and other support from BMS outside the submitted work. Gabriel Etienne reports grants and personal fees from Novartis and BMS and personal fees from Ariad, Incyte, and Pfizer, all outside the submitted work. Philippe Rousselot reports research grants from Ariad, BMS, and Pfizer outside the submitted work. Delphine Rea reports personal fees from BMS, Incyte, Novartis, and Pfizer outside the submitted work. Stephane Giraudier reports grants and personal fees from BMS and Novartis outside the submitted work. Agnès Guerici-Bresler reports personal fees from Ariad, BMS, Novartis, and Pfizer outside the submitted work. Françoise Huguet reports personal fees from BMS, Incyte, Novartis, and Pfizer outside the submitted work. Martine Gardembas reports personal fees from Ariad, BMS, and Novartis outside the submitted work. Francios- Xavier Mahon reports personal fees from BMS, Incyte, Novartis and Pfizer outside the submitted work. The remaining authors made no disclosures. AUTHOR CONTRIBUTIONS Laurence Legros: Designed the study, collected data, interpreted the results, provided patient care, writing initial draft, writing review and editing, and approved the final version. Franck E. Cancer November 15,

8 Nicolini: Designed the study, interpreted the results, provided patient care, review and editing, and approved the final version. Gabriel Etienne, Philippe Rousselot, Delphine Rea, Stephane Giraudier, Agnès Guerci-Bresler, Françoise Huguet, Martine Gardembas, Martine Escoffre, Jean-Christophe Ianotto, Marie- Pierre No el, and Bruno R. Varet: Provided patient care, and approved the final version. Stephane Morisset: Analyzed data and approved the final version. Thomas Pagliardini and Irit Touitou: Collected data, and approved the final version. Francois-Xavier Mahon: Designed the study, interpreted the results, provided patient care, and approved the final version. REFERENCES 1. Bower H, Bjorkholm M, Dickman PW, Hoglund M, Lambert PC, Andersson TM. Life expectancy of patients with chronic myeloid leukemia approaches the life expectancy of the general population. J Clin Oncol. 2016;34: Rousselot P, Huguet F, Rea D, et al. Imatinib mesylate discontinuation in patients with chronic myelogenous leukemia in complete molecular remission for more than 2 years. Blood. 2007;109: Mahon FX, Rea D, Guilhot J, et al. Discontinuation of imatinib in patients with chronic myeloid leukaemia who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial. Lancet Oncol. 2010;11: Mori S, Vagge E, le Coutre P, et al. Age and dpcr can predict relapse in CML patients who discontinued imatinib: the ISAV study. Am J Hematol. 2015;90: Ross DM, Branford S, Seymour JF, et al. Safety and efficacy of imatinib cessation for CML patients with stable undetectable minimal residual disease: results from the TWISTER study. Blood. 2013;122: Rousselot P, Charbonnier A, Cony-Makhoul P, et al. Loss of major molecular response as a trigger for restarting tyrosine kinase inhibitor therapy in patients with chronic-phase chronic myelogenous leukemia who have stopped imatinib after durable undetectable disease. J Clin Oncol. 2014;32: Takahashi N, Kyo T, Maeda Y, et al. Discontinuation of imatinib in Japanese patients with chronic myeloid leukemia. Haematologica. 2012;97: Thielen N, van der Holt B, Cornelissen JJ, et al. Imatinib discontinuation in chronic phase myeloid leukaemia patients in sustained complete molecular response: a randomised trial of the Dutch- Belgian Cooperative Trial for Haemato-Oncology (HOVON). Eur J Cancer. 2013;49: Lee SE, Choi SY, Bang JH, et al. Predictive factors for successful imatinib cessation in chronic myeloid leukemia patients treated with imatinib. Am J Hematol. 2013;88: Imagawa J, Tanaka H, Okada M, et al. Discontinuation of dasatinib in patients with chronic myeloid leukaemia who have maintained deep molecular response for longer than 1 year (DADI trial): a multicentre phase 2 trial. Lancet Haematol. 2015;2:e528-e Rea D, Nicolini FE, Tulliez M, et al. Discontinuation of dasatinib or nilotinib in chronic myeloid leukemia: interim analysis of the STOP 2G-TKI study. Blood. 2017;129: Cross NC, White HE, Colomer D, et al. Laboratory recommendations for scoring deep molecular responses following treatment for chronic myeloid leukemia. Leukemia. 2015;29: Mahon FX, Richter J, Guilhot J, et al. Cessation of tyrosine kinase inhibitors treatment in chronic myeloid leukemia patients with deep molecular response: results of the Euro-Ski Trial [abstract]. Blood. 2016;128: Etienne G, Guilhot J, Rea D, et al. Long-term follow-up of the French Stop Imatinib (STIM1) study in patients with chronic myeloid leukemia. J Clin Oncol. 2017;35: Legros L, Rousselot P, Giraudier S, et al. Second attempt to discontinue imatinib in CP-CML patients with a second sustained complete molecular response. Blood. 2012;120: Baccarani M, Saglio G, Goldman J, et al. Evolving concepts in the management of chronic myeloid leukemia: recommendations from an expert panel on behalf of the European LeukemiaNet. Blood. 2006;108: Cross NC, White HE, Muller MC, Saglio G, Hochhaus A. Standardized definitions of molecular response in chronic myeloid leukemia. Leukemia. 2012;26: Muller MC, Erben P, Saglio G, et al. Harmonization of BCR-ABL mrna quantification using a uniform multifunctional control plasmid in 37 international laboratories. Leukemia. 2008;22: Flandrin-Gresta P, Cornillet P, Hayette S, et al. Recommendations for accreditation of laboratories in molecular biology of hematologic malignancies. Ann Biol Clin (Paris). 2015;73: Preudhomme C, Guilhot J, Nicolini FE, et al. Imatinib plus peginterferon alfa-2a in chronic myeloid leukemia. N Engl J Med. 2010; 363: Hughes TP, Ross DM. Moving treatment-free remission into mainstream clinical practice in CML. Blood. 2016;128: Michor F, Hughes TP, Iwasa Y, et al. Dynamics of chronic myeloid leukaemia. Nature. 2005;435: Prost S, Relouzat F, Spentchian M, et al. Erosion of the chronic myeloid leukaemia stem cell pool by PPARgamma agonists. Nature. 2015;525: Ross DM, Branford S, Seymour JF, et al. Patients with chronic myeloid leukemia who maintain a complete molecular response after stopping imatinib treatment have evidence of persistent leukemia by DNA PCR. Leukemia. 2010;24: Ilander M, Olsson-Stromberg U, Schlums H, et al. Increased proportion of mature NK cells is associated with successful imatinib discontinuation in chronic myeloid leukemia. Leukemia. 2017;31: Krause DS, Fulzele K, Catic A, et al. Differential regulation of myeloid leukemias by the bone marrow microenvironment. Nat Med. 2013;19: Lane SW, Scadden DT, Gilliland DG. The leukemic stem cell niche: current concepts and therapeutic opportunities. Blood. 2009; 114: MacLean AL, Filippi S, Stumpf MP. The ecology in the hematopoietic stem cell niche determines the clinical outcome in chronic myeloid leukemia. Proc Natl Acad Sci U S A. 2014;111: Cancer November 15, 2017