U.S. Food and Drug Administration Benefit‐Risk Assessment of Nilotinib Treatment Discontinuation in Patients with Chronic Phase Chronic Myeloid Leukemia in a Sustained Molecular Remission

Considering the activity of nilotinib in treatment of chronic myeloid leukemia, two single‐arm trials, ENESTop and ENESTfreedom, were conducted to examine whether the use of nilotinib might lead to sustained remission off therapy. The results form the basis for recommended updates to the nilotinib prescribing information for treatment discontinuation in select patients.

Abstract

On December 22, 2017, the U.S. Food and Drug Administration (FDA) updated the product label for nilotinib to include information for providers on how to discontinue this drug in certain patients. With the updated dosing recommendations, select patients with chronic phase myeloid leukemia (CML) taking nilotinib for 3 years or more and whose leukemia has responded with sustained molecular remission (MR4.5, BCR‐ABL transcripts of ≤0.0032%) as determined by a FDA‐approved test may be eligible to discontinue nilotinib. The updated dosing regimen was based on the efficacy results from two trials that measured how long patients could stop taking nilotinib without the leukemia returning (treatment‐free remission). Trial results demonstrated that, among selected patients who received nilotinib as first‐line therapy or after transition from imatinib, approximately 50% continued to be in remission at 96 weeks after stopping therapy. Relapses continued to occur throughout the study, indicating that long‐term monitoring is needed for safety and disease monitoring. Discontinuation of treatment was associated with an increased risk of low grade musculoskeletal adverse events, some of which were prolonged. Overall, the results support the approval of updates to the dosing recommendations with regard to treatment discontinuation in selected patients who have received nilotinib for at least 3 years, are in a sustained molecular remission, and who can undergo appropriate monitoring.

Implications for Practice.

The updated dosing information provides eligibility criteria for treatment discontinuation, strict monitoring criteria after nilotinib discontinuation, and guidance for treatment reinitiation in eligible patients with chronic phase myeloid leukemia. About half of appropriately selected patients remained in remission 96 weeks after treatment discontinuation. Patients may experience musculoskeletal pain on withdrawal of treatment, incidence of which appears to decrease over time; however, some patients may have long lasting events. The decision to withdraw or continue treatment with nilotinib should be based on clinical condition and patient preferences.

Introduction

Chronic myeloid leukemia (CML) is a rare form of leukemia characterized by a chronic phase (CP) with slow growth and progression, which eventually leads to accelerated phase (AP) and blast crisis (BC), typically within months to years of diagnosis. The disease is characterized by the presence of the fusion protein BCR‐ABL, which drives the malignancy. Prior to 2001, CML had a poor prognosis, with only about 10% of patients surviving 10 years after diagnosis [1]. However, in 2001, the first BCR‐ABL‐selective tyrosine kinase inhibitor (TKI), imatinib, was approved for use in patients with CML [2]. This approval was followed in the early 2000s by further approvals of dasatinib and nilotinib [3]. Nilotinib is currently indicated for the following: treatment of newly diagnosed adult patients with Philadelphia chromosome‐positive chronic myelogenous leukemia (Ph+ CML) in CP and treatment of chronic phase and accelerated phase Ph+ CML in adult patients resistant to or intolerant to prior therapy that included imatinib. These drugs have resulted in astonishing improvements in the prognosis of CML patients, particularly in population‐level survival [4], [5]. However, a few patients can develop resistance to treatment because of emerging mutations in the BCR‐ABL protein. In addition, patients are prescribed life‐long treatment with TKIs, and adverse events associated with the use of TKIs may be burdensome to patients.

Patients with CP‐CML often experience a molecular remission, with BCR‐ABL levels decreasing to extremely low or undetectable levels. Despite this, discontinuation of treatment with TKIs has not, in the past, been recommended because of the risk of recurrence and uncertainty concerning whether intermittent treatment would increase the risk of development of resistant mutations. Nonetheless, case reports and series of patients who discontinued for a variety of reasons and experienced a sustained remission off therapy [6], [7] led to an interest in formal evaluation of the risks and benefits of treatment discontinuation. The Stop Imatinib (STIM) trial, which evaluated a discontinuation protocol in patients treated with imatinib and experienced a sustained molecular remission, found that 41% of patients remained in remission at 12 months follow‐up off therapy [8]. Although most relapses occurred in the first 24 weeks after therapy discontinuation, occasional late relapses continued for up to 5 years after therapy discontinuation [9].

Given the activity of nilotinib, two single‐arm trials (ENESTop and ENESTfreedom) were conducted to examine whether the use of nilotinib might lead sustained remission off therapy above prespecified thresholds. The 96‐week efficacy results from these two trials formed the basis for the recommendations to update information in the nilotinib (Tasigna; Novartis Pharmaceuticals, Basel, Switzerland) prescribing information for treatment discontinuation with nilotinib in select patients with chronic phase CML.

Clinical Trials

ENESTop (CAMN107A2408, NCT01698905, hereafter referred to as A2408) and ENESTfreedom (CAMN107I2201, NCT01784068, hereafter referred to as I2201) examined treatment‐free remission (TFR) by looking at several endpoints including treatment‐free‐survival and time to progression in selected patients.

Both studies enrolled similar patient populations: achievement of MR4.5 at screening and treated for a minimum of 2 years with nilotinib prior to study entry (Fig. 1). After enrolling in the trial, all patients received treatment for 1 year with standard dose of nilotinib. After the 1 year of treatment with nilotinib, a determination was made as to whether they met criteria for entering the TFR phase. Criteria for entering the TFR phase varied slightly between the trials. For A2408, patients who did not demonstrate a confirmed loss of MR4.5, defined as two consecutive measurements of BCR‐ABL1 of >0.0032% international scale (IS), were eligible to transition to the TFR phase. For I2201, those who were in MR4.5 around 52 weeks, had no measured BCR‐ABL levels greater than MR4 (BCR‐ABL <0.01% IS), and had no more than two assessments between MR4.0 and MR4.5 at any time during the consolidation phase, were eligible to transition to the TFR phase. Once patients entered the TFR phase, treatment with nilotinib was stopped. Patients were monitored frequently for relapse using a highly sensitive U.S. Food and Drug Administration (FDA)‐approved test for BCR‐ABL transcripts. Currently, only the MRDx BCR‐ABL test from Molecular Diagnostics is approved for monitoring of BCR‐ABL levels in context of treatment discontinuation. However, other diagnostics may become available in the future, and practitioners are encouraged to refer to the FDA web site on companion diagnostics (below) for confirmation of the suitability of a specific test. https://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/InVitroDiagnostics/ucm301431.htm.

Figure 1.

Trial schema. (A): A2408. (B): I2201. All study phases (TFR, NTRI, TFR‐2, NTRI‐2, NTCT‐P, post‐treatment follow‐up) will last up to 192 weeks after the last patients enters the TFR.

Abbreviations: I/E, inclusion/exclusion; IM, imatinib; MMR, major molecular response; MRD, minimal residual disease; NIL, nilotinib; NTCS, nilotinib consolidation phase; NTCT, nilotinib treatment continuation phase; NTRI, nilotinib treatment reinitiation; PCR, polymerase chain reaction; TFR, treatment‐free remission.

Patients who did not have a sustained MR4.5 during the nilotinib treatment consolidation phase entered a prolonged consolidation treatment period. If patients obtained a sustained MR4.5 after a further year's treatment, they could enter the treatment‐free remission‐2 phase (TFR‐2), which was conducted in a similar manner to the TFR phase. Patients not in MR4.5 after the second year of treatment were eligible for prolonged continued treatment using nilotinib. This review will concentrate on the TFR and nilotinib treatment reinitiation (NTRI) phases, as few patients entered prolonged treatment or the TFR‐2 phase.

NTRI

In patients with newly diagnosed CML who lost major molecular response (MMR), patients reinitiated treatment within 4 weeks at the dose level prior to discontinuation of therapy. Patients who reinitiated therapy underwent monitoring of their BCR‐ABL transcripts every 4 weeks during the first 24 weeks, then once every 12 weeks thereafter (or more frequently as clinically indicated in patients not regaining MMR). For patients with resistance or intolerance to prior therapy with imatinib with a confirmed loss of MR 4.0, defined as loss of MR 4.0 on two consecutive measurements, or loss of MMR reinitiated treatment within 4 weeks at the dose level prior to discontinuation. These patients had BCR‐ABL transcript levels monitored monthly until MMR or MR 4.0 was reestablished and then monitored every 12 weeks thereafter.

Patient Selection

Patients with newly diagnosed CP‐CML treated with nilotinib for at least 2 years who had achieved a MR4.5 (corresponding to BCR‐ABL/ABL1 ≤ 0.0032% IS) with the last assessment taken immediately prior to treatment discontinuation and had no atypical BCR‐ABL transcripts were eligible to participate in I2201. Patients with previously chronic phase CML who were treated initially with imatinib for 4 weeks to 1 year, then nilotinib for at least 2 years, achieved and sustained an MR4.5 on nilotinib for a minimum of 1 year prior to treatment discontinuation, and had no atypical BCR‐ABL transcripts were eligible to participate in A2408.

Endpoints

The primary endpoint in both studies was the percentage of patients who were in MMR at 48 weeks after starting the TFR phase. However, the FDA evaluated the efficacy based on the percentage of patients in MMR at the end of 96 weeks to fully evaluate the longer‐term risks and benefits of discontinuation. Secondary endpoints included the percentage of patients in MMR at later timepoints (96, 144, 192, and 264 weeks), treatment‐free‐survival, overall survival (OS), the percentage of patients who lost MMR who subsequently recovered MMR at 12 weeks after reinitiation of therapy, the percentage of patients who experienced progression to accelerated phase or blast crisis, the mutation rate, and the safety profile in each phase.

For the 96‐week efficacy results, patients retreated were considered to have lost remission. Patients discontinuing the study were also considered as loss of remission, regardless of whether the dropout was related to disease status. The Kaplan‐Meier plot in Figure 2 differs slightly from the primary endpoints, since dropout was considered censoring as opposed to failure. The dotted lines in Figure 2 represent a 95% confidence interval for the percentage remaining in remission.

Figure 2.

Treatment‐free survival curves. (A): Curve for A2408. (B): Curve for I2201.

Abbreviations: Cen, number of censored observations; Evt, number of events; Pat, number of patients; TFR, treatment‐free remission.

Efficacy Analysis

Baseline demographic characteristics of patients included in the trials are presented in Table 1. Baseline characteristics were similar for patients in each trial. Gender balance, median age, and median weight were similar for patients in both trials. The majority of patients entering either trial were white, and most had an Eastern Cooperative Oncology Group performance status of 0. Geographic site of enrollment varied, with 50% of patients in trial A2408 from sites in countries outside Europe or North America, whereas over 80% of patients in I2201 were from countries in Europe.

Table 1. Patient demographics for patients in the intent‐to‐treat populations of the included clinical trials.

Totals may not sum to 100% because of rounding.

^aUnavailable for two subjects on A2408.

^bUnavailable for one subject on A2408.

Abbreviations: EU, Europe; NA, North America; NTCS, nilotinib treatment consolidation phase; NTRI, nilotinib treatment reinitiation phase; ROW, rest of world; TFR, treatment‐free remission phase.

At 96 weeks after treatment discontinuation, 53.2% (95% confidence interval [CI] 44.1%–62.1%) of patients in study A2408 and 48.9% (95% CI 41.6%–56.3%) of patients in trial I2201 continued to be in remission off treatment (Fig. 2). Low Sokal score was weakly associated with maintaining remission off treatment, but no other demographic or disease factors could be identified that predicted maintenance of remission. Most relapses occurred in the first 24 weeks after treatment discontinuation, but rare relapses continued at later time points, including one patient who lost remission at 120 weeks. No deaths from CML or transition to blast crisis or accelerated phase were observed. One patient developed a new mutation of uncertain significance. Of patients who relapsed during TFR phase and entered the NTRI phase, 91.1% in A2408 and 92.0% in I2201 regained MR4.5 after reinitiating treatment. No patients lost hematologic response or progressed while on the NTRI phase.

Safety Assessment

Eleven deaths occurred in the trials (three in A2408, eight in I2201). None of the deaths were related to CML. Three deaths occurred in the NTCS phase, one in the TFR phase, three in the NTRI phase, and four in follow‐up. There were three deaths from malignancies other than CML; three from cardiopulmonary failure, cardiac arrest, or respiratory failure; two deaths of unknown cause; and one each from arterial hemorrhage, suicide, and acute myocardial infarction. No deaths were attributed to the intervention.

The incidence of adverse reactions in the TFR phase was compared with the incidence in the NTRI phase for patients in the trials. The incidence of most adverse events was roughly equal or more common in the NTRI for most events (Table 2). The primary exception was musculoskeletal adverse events (MAEs), which occurred more frequently in the TFR phase of both trials, with events occurring in 45% of patients in the TFR phase of A2408 and 28% of patients in I2201 compared with 34% and 22%, respectively, in the NTRI phases of the trials. The majority of these events occurred early in the TFR phase, with 62% of events in A2408 and 61% in I2201 occurring in the first 24 weeks after treatment discontinuation and 14% and 7%, respectively, occurring in the first 30 days after discontinuation. Most MAEs were grade 1 or 2, with a combined incidence of five grade 3 MAEs in both trials in the TFR phase and four grade 3 events in the NTRI phase, and no grade 4 events observed in either phase. Although the MAEs were generally low grade, approximately one third of patients had events that were not resolved by data cutoff date. There were no factors identified that clearly predicted increased risk for MAE, including age, sex, race, and history of arthritis.

Table 2. Common adverse events of any grade occurring in the TFR or NTRI phase of the examined clinical trials.

Musculoskeletal pain: arthralgia, pain in extremity, bone pain, back pain, myalgia, musculoskeletal pain, musculoskeletal chest pain, joint stiffness, neck pain, musculoskeletal stiffness, pain, noncardiac chest pain, chest discomfort, spinal pain.

Hypertension: hypertension, hypertensive crisis, blood pressure increased.

Pancreatitis: pancreatitis/increased lipase, increased lipase, amylase increased, hyperlipasemia.

Hepatotoxicity: alanine aminotransferase increased, aspartate aminotransferase increased, blood alkaline phosphatase increased, gamma‐glutamyltransferase GGT increased, cholestasis, hepatic function abnormal, hepatotoxicity, hyperbilirubinemia, cholelithiasis.

Hyperglycemia: hyperglycemia, increased glucose, impaired glucose tolerance, diabetes mellitus, glycosylated hemoglobin increased.

Hyperlipidemia: hypercholesterolemia, blood cholesterol increased, blood triglycerides increased, low density lipoprotein increased, hypertriglyceridemia, hyperlipidemia, dyslipidemia.

URI: upper respiratory tract infection, viral upper respiratory tract, pharyngitis, sinusitis, laryngitis, tracheitis, bronchitis, rhinitis, tonsillitis, influenza, influenza‐like illness.

Abbreviations: NTRI, nilotinib treatment reinitiation; TFR, transmission‐free survival.

Patient‐Reported Outcomes

The EQ‐5D scale was used to measure patient‐reported outcomes for patients involved in the two trials. In a descriptive, exploratory analysis, most patients reported little or no change in scores over the course of the study, with at least 60% of patients reporting no change from baseline (start of TFR phase) at 96 weeks in any category.

Discussion

Discontinuation of therapy after at least 3 years of treatment with nilotinib can be safely performed in selected patients with a sustained MR4.5 and no history of aberrant BCR‐ABL transcript or accelerated phase/blast crisis. Patients eligible for discontinuation require close monitoring for recurrence using an FDA‐approved test of high sensitivity to detection of BCR‐ABL transcripts. About half of the patients maintained a continuous remission without further treatment for at least 2 years in the trials reviewed. The results of the trials provided support for an updated dosing regimen for nilotinib discontinuation in the context of TFR for select patients who have a sustained molecular remission (Table 3).

Table 3. U.S. Food and Drug Administration benefit‐risk assessment.

The current National Comprehensive Cancer Network (NCCN) guidelines for CML [10] include a section on TKI discontinuation in patients with CML based on a number of clinical trials examining discontinuation after achievement of a major molecular response. Although there is substantial agreement between the NCCN guidelines and those provided in the nilotinib label, there are modest differences. Most notably, the NCCN guidelines do not restrict discontinuation to patients previously treated with a specific TKI. At this time, nilotinib is the only TKI for treatment of CML with an FDA‐approved dosing regimen for discontinuation.

Despite the positive results of the trials and the growing acceptance of TKI discontinuation as a general strategy in patients with CML, there are a few unanswered questions and caveats that physicians and patients should consider when deciding whether discontinuing treatment is the right choice for an individual patient.

First, there are currently no definitive markers to predict which patients will be able to successfully discontinue treatment and remain in sustained remission and which patients will relapse upon treatment discontinuation. In addition, although the majority of relapses occur in the first 24 weeks after discontinuation, there is no time yet identified after which no further relapses occur. Thus, it is not possible to discontinue monitoring at any specific time point nor can any patients be considered “cured,” even with long‐term remissions; long‐term monitoring is needed for patient safety. As part of an FDA postmarketing requirement, the two TFR studies will be continued for 10 years to collect additional data.

It has been established that quiescent CML cells persist in patients, despite no clinical signs of disease, following optimal treatment with TKIs. These cells may become active when treatment is stopped. However, the specific triggers that lead to reactivation are currently unknown. Further research to identify and, ideally, prevent reactivation of these cells might lead to increased ability to maintain remission.

Second, patients who discontinue nilotinib are at risk of developing MAEs in the first 24 weeks after discontinuation. The mechanism by which this occurs is unclear. Although most patients report the pain to be grade 1 or 2, the pain may be long lasting, with about one third of patients who experienced pain having pain that was not resolved or resolving at the time of data cutoff. No demographic factors clearly predicted the risk of occurrence of MAEs in any individual patient. Thus, patients should be cautioned as to the risk of this adverse event as part of the discussion of whether to discontinue treatment.

Third, it is unclear whether the long‐term risks associated with use of nilotinib (i.e. pancreatitis, hyperglycemia, arterial vascular disease) are decreased with discontinuation, especially if the discontinuation lasts less than one year. Indeed, there was a numerically higher rate of some adverse events in the NTRI phase compared with the consolidation phase, reflecting the known safety profile of nilotinib in patients starting medication anew in the NTRI phase compared with patients on nilotinib for a longer duration of time in consolidation phase. In contrast, there were instances in the two trials in which absence of treatment in the period following discontinuation provided an opportunity for women with childbearing potential to pursue pregnancies, with uneventful outcomes in both cases. However, these were unplanned post hoc observations and should be taken as hypothesis generating only.

Taken together, the evidence suggests that the decision to continue or discontinue treatment with Tasigna should be individualized. Approval of the new dosing regimen for discontinuation provides guidance to physicians on how to safely discontinue treatment in qualified patients who have achieved sustained MR4.5 and are interested in treatment discontinuation.

Author Contributions

Conception/design: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Provision of study material or patients: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Collection and/or assembly of data: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Data analysis and interpretation: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Manuscript writing: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Final approval of manuscript: E. Dianne Pulte, Tanya Wroblewski, Erik Bloomquist, Shenghui Tang, Ann Farrell, Albert Deisseroth, Amy E. McKee, Richard Pazdur

Disclosures

The authors indicated no financial relationships.