CADTH Reimbursement Review

Azacitidine (Onureg)

Sponsor: Celgene Inc., a Bristol Myers Squibb Company

Therapeutic area: Acute myeloid leukemia

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

Acute myeloid leukemia (AML) is the most common form of acute leukemia in adults. It is an aggressive hematological malignancy that affects the production of normal blood cells (i.e., clonal expansion of myeloid blasts in the peripheral blood, bone marrow, and/or other tissue), causing weakness, infection, bleeding, and other symptoms and complications.^3,4 According to the most recent incidence statistic available for AML in Canada, 1,090 new cases were reported in 2016; recent mortality statistics indicate that 1,184 Canadians died from AML in 2017.⁵

Patients with AML who can tolerate aggressive treatment receive initial induction therapy followed by consolidation therapy, and approximately 50% of patients aged 60 years or younger are cured in this fashion.^4,6,7 Allogeneic transplantation, which is potentially curative (the probability of long-term survival is 50%), is pursued as a consolidation strategy in patients in first remission with intermediate or poor cytogenetic risk who are eligible for transplantation.¹ However, the procedure is associated with a high risk of morbidity and mortality,⁸ and reasons for transplantation ineligibility vary, including old age, comorbidities, and decreased organ function.¹ For patients who are ineligible for transplantation, maintaining first remission to delay relapse and prolong survival is an important clinical goal.¹ The clinical experts consulted by CADTH noted that there are currently no standard funded maintenance regimens specific for patients with AML who are in first remission by achieving either complete remission (CR) or complete remission with incomplete blood count recovery (CRi) following induction therapy with or without consolidation treatment and who are not eligible for transplantation. In the absence of proven treatment options for patients in this setting, a watch-and-wait approach or best supportive care (BSC) is recommended.¹ Survival of patients with relapsed disease after induction or consolidation treatment is poor; if first remission lasts for less than 1 year, the expected median overall survival (OS) is 4 to 6 months.¹ The clinical experts consulted by CADTH agreed that there is a need for effective treatments that prolong life, delay relapse, maintain or improve patient quality of life, have an acceptable safety profile, and have a convenient oral route of administration.

Oral azacitidine is a cytidine nucleoside analogue that incorporates into both DNA and RNA after phosphorylation to the nucleotide and produces anti-leukemic activity in cancers such as AML.¹ On January 4, 2021, oral azacitidine was approved by Health Canada as maintenance therapy in adult patients with AML who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for hematopoietic stem cell transplantation (HSCT).² The requested reimbursement criteria for oral azacitidine are as per the Health Canada–approved indication. Oral azacitidine has no other Health Canada–approved indication and has not previously been reviewed by CADTH. The recommended starting dosage is 300 mg administered orally once daily on day 1 through day 14 of continuous 28-day cycles. The product monograph states that if the absolute neutrophil count (ANC) is less than 500 µL on day 1 of a cycle, oral azacitidine should not be administered and the start of the cycle should be delayed until the ANC is 500 µL or more.²

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups who responded to CADTH’s call for patient input and from clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Input

One patient group, the Leukemia and Lymphoma Society of Canada (LLSC), provided input for this review. The input was based on an online survey created by the LLSC to which a total of 29 patients with AML responded.

Respondents to the LLSC survey indicated a varying range of AML symptoms affecting their daily lives, including disruption to their social lives and relationships, causing detrimental effects on their health. Patient respondents reported being easily fatigued, unable to exercise or work, losing vision in 1 eye, nausea, bruising, numbness or body aches, and being immunocompromised. Many of these symptoms reportedly led to feelings of isolation and fear of relapse, with negative impacts on their psychological well-being.

According to the patient input received, patients expect the following key outcomes to be improved with any new drug or treatment: maintaining remission, fewer side effects, lower cost, and being accessible in their neighbourhoods. Additionally, the following factors were noted in choosing a new cancer treatment: physician recommendation, impact on disease, quality of life, closeness of home, and outpatient treatment. No survey respondents indicated that they had experience taking oral azacitidine.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The following input was provided by 3 clinical specialists with expertise in the diagnosis and management of AML.

The clinical experts consulted by CADTH indicated that there are currently no standard funded maintenance regimens, and watch-and-wait or BSC are recommended for patients in the present target setting. The clinical experts identified an unmet need for effective therapies with acceptable toxicity profiles that delay relapse, prolong survival, maintain health-related quality of life (HRQoL), and potentially lead to a cure in patients with AML who are in their first remission and are not eligible for transplantation. They emphasized a need for a convenient oral route of administration to achieve high adherence and reduce hospital-based resource utilization. It was agreed that, while oral azacitidine would likely shift the current treatment paradigm, more experience with maintenance oral azacitidine therapy will be necessary to determine the potential impacts on the current treatment paradigm (e.g., potentially fewer patients requiring transplant, a reduction of cycles of consolidation chemotherapy, and refinement of the target population).

The clinical experts agreed that patients as selected per the inclusion and exclusion criteria of the QUAZAR AML-001 trial should be eligible for maintenance oral azacitidine therapy. While the clinical experts agreed that there is currently insufficient evidence to guide a recommendation on which patient subgroups would be best suited for (most likely to show a response to) oral azacitidine, the experts identified the following potential patient subgroups as having the highest risk of relapse and therefore being most in need of maintenance therapy: patients aged 65 years or older, patients with minimal residual disease (MRD)-positive status, patients who have not received consolidation chemotherapy, and patients with poor-risk karyotypes. In the opinion of the clinical experts, patient subgroups that would potentially benefit the least from oral azacitidine may include patients with an MDR-negative status, patients with low-risk features for relapse, patients who develop unacceptable toxicities, and patients lacking the social and/or medical support necessary to be safely treated with oral azacitidine.

The clinical experts agreed that patients would be identified as possible candidates for oral azacitidine treatment if they are considered ineligible for transplantation. The transplantation ineligibility status may be known in patients at the time of diagnosis or may develop over the course of their treatment. Clinical assessments to evaluate the response to treatment with oral azacitidine would include regular monitoring of blood counts (every 1 to 2 weeks) to determine if a patient maintains CR or CRi. If changes in blood count signal a potential relapse, a bone marrow examination may be required to determine if a patient is still in CR or CRi or has relapsed AML. The clinical experts indicated that the most clinically meaningful responses to treatment include prolonged OS and relapse-free survival (RFS) while maintaining or improving HRQoL and reducing symptom burdens. No increase in drug-related toxicities, such as infections, neutropenia, and thrombocytopenia, and a reduced risk of relapse were also noted as clinically meaningful outcomes.

In the opinion of the clinical experts, treatment with maintenance therapy of oral azacitidine should be discontinued if a patient experiences relapsed AML, has a markedly impaired performance status, or is intolerant to or experiences unacceptable toxicity from oral azacitidine. If AML recurs with at least 5% of blasts in the peripheral blood or bone marrow, oral azacitidine should be discontinued. There is currently insufficient evidence from the QUAZAR AML-001 trial to recommend dose escalation when AML recurs with blast levels between 5% and 15% in either the peripheral blood or bone marrow.

Clinician Group Input

Two clinician group inputs were provided, 1 from the Ontario Health–Cancer Care Ontario (OH-CCO) Hematological Cancer Drug Advisory Committee (H-DAC) and 1 from the Alberta Tumour Board Myeloid Physician Group (ATB-MPG). Overall, the views of the clinician groups were consistent with those of the clinical experts consulted by CADTH; all agreed that the most important treatment goals are prolongation of life and remission as reflected by the pivotal trial’s primary and key secondary end points of OS and RFS, respectively, as well as improvement in quality of life. Similar to the clinical experts consulted by CADTH, the clinicians from the OH-CCO H-DAC reported that they would generalize the QUAZAR AML-001 trial results to patients younger than 55 years of age. In addition, the clinician group also stated that pharmacokinetic data suggest that patients with myelodysplastic syndrome (MDS) who have progressed on subcutaneous azacitidine and subsequently received induction chemotherapy with or without consolidation and have achieved CR may also benefit from oral azacitidine if they are not eligible for transplantation. There was consensus among the clinical experts consulted by CADTH and the clinicians from both clinician groups that oral azacitidine would be a standard-of-care maintenance therapy for patients with AML who are in first CR and ineligible for HSCT.

Drug Program Input

Input from the drug programs identified factors pertaining to relevant comparators, prior therapies required for eligibility, therapy continuation criteria, dosing considerations, and generalizability that may affect their ability to implement a recommendation. The clinical experts consulted by CADTH weighed evidence from the QUAZAR AML-001 trial and other clinical considerations to provide responses to questions from the drug programs related to these factors, which can be found in Table 4.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Study

The QUAZAR AML-001 trial is an ongoing, international, multicentre, double-blind, placebo-controlled randomized phase III trial that compared the efficacy and safety of maintenance therapy with oral azacitidine plus BSC with placebo plus BSC in patients with AML in first CR. A total of 472 patients were randomized in a 1:1 ratio to receive maintenance oral azacitidine (300 mg tablets once daily for the first 14 days of each 28-day cycle) plus BSC or oral placebo (matching placebo tablets once daily for the first 14 days of each 28-day cycle) plus BSC. Randomization was stratified by age at time of induction therapy (55 to 64 years and ≥ 65 years), prior history of MDS or chronic myelomonocytic leukemia (CMML) (yes/no), cytogenetic risk category at time of induction therapy (intermediate risk/poor risk), and receipt of consolidation therapy following induction (yes/no). No crossover between the treatment groups was permitted. The primary outcome was OS, and the key secondary outcome was RFS. Other secondary end points included time to relapse and time to discontinuation from treatment. Two HRQoL measures, the EuroQol 5-Dimensions 3-Levels questionnaire (EQ-5D-3L) and the Physical Impairment Numeric Rating (PINR) scale, were included as secondary and exploratory outcomes, respectively. A symptom severity measure, the Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-F) scale was included as a secondary outcome.⁹

The QUAZAR AML-001 trial enrolled adults 55 years of age and older diagnosed with AML or AML secondary to prior MDS or CMML with intermediate- or poor-risk cytogenetics and who had achieved first remission (CR or CRi) following induction with or without consolidation chemotherapy. Patients were excluded if they were eligible for HSCT. At screening patients had to have an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0 to 3 and adequate marrow function.¹⁰ The mean age of patients in the trial was 68.0 (standard deviation [SD] = 5.66) years of age. The oral azacitidine group had a lower proportion (versus placebo) of male patients (49.6% versus 54.3%) and patients enrolled in trial sites in Asia (2.5% versus 7.3%) and a higher proportion of White patients (90.8% versus 84.2%) and patients enrolled in trial sites in Europe (70.2% versus 62.8%). The majority of patients had an ECOG PS of 0 (48.7% versus 47.4%) or 1 (42.4% versus 45.3%) and intermediate cytogenetic risk (85.3% versus 86.8%) compared with patients given placebo. Most of the patients received 1 cycle (46.2% versus 43.6%) or 2 cycles (29.4% versus 32.9%) of consolidation therapy. The most common reason for transplantation ineligibility was age (64.7% versus 65.0%), followed by comorbidities (21.8% versus 21.4), and no available donor (15.5% versus 15.0%).⁹ The oral azacitidine group had a lower proportion of patients with a negative MRD status at randomization (55.9% versus 47.4%) and a lower proportion of patients who achieved CR after induction therapy with or without consolidation compared with placebo (78.6% versus 84.2%).⁹ The clinical experts consulted by CADTH noted that the differences observed in patient characteristics at baseline between the oral azacitidine and the placebo groups are unlikely to have had an impact on the treatment differences observed between the treatment groups.

A post-relapse dose escalation (300 mg once daily from 14 days to 21 days of the 28-day cycle) was planned for patients with disease relapse with blast levels of between 5% and 15% in either the peripheral blood or bone marrow.⁹ This dose escalation was used in the context of re-induction of remission, which is not consistent with the Health Canada indication and the requested reimbursement criteria, which are for oral azacitidine as maintenance therapy.¹¹ The QUAZAR AML-001 trial was not designed to assess if dose escalation of oral azacitidine produces benefits for patients. Dose escalation has not been authorized by Health Canada as per the Health Canada product monograph.¹¹

The primary analysis, which coincided with the final data cut-off date (July 15, 2019) was conducted once the total number of events (n = 330 deaths) occurred.^9,10 The key secondary and other secondary end points were also planned to be assessed at the time of the final data cut-off date. The trial is ongoing, with an estimated completion date of December 31, 2021. Data from an updated data cut-off date (September 2020), which was not planned a priori in the statistical analysis plan, were not provided to CADTH by the sponsor at the time of this review.^12,13

Efficacy Results

The key efficacy results from the QUAZAR AML-001 trial are summarized in Table 2. At the final data cut-off date (July 15, 2019) the median duration of follow-up for OS was ||| months in the oral azacitidine group and ||| months in the placebo group. Median OS was 24.7 (95% confidence interval [CI], 18.7 to 30.5) months in the oral azacitidine group compared with 14.8 (95% CI, 11.7 to 17.6) months in the placebo group, with a stratified hazard ratio (HR) of 0.69 (95% CI, 0.55 to 0.86; P = 0.0009) in favour of the oral azacitidine group. Testing of treatment-by-time interactions in the overall Cox proportional hazards model suggested deviation from the proportional hazards assumption. The results of sensitivity analyses for OS that were planned to be conducted if the proportional hazards assumption was violated suggested robustness and consistency of the primary OS result. The OS results for the subgroups of interest, as prespecified a priori in the protocol for this CADTH review, suggested that the treatment effect on OS for the subgroups was generally consistent with the primary analysis. The subgroup analysis by cycles of consolidation therapy suggested possible heterogeneity of treatment effect; however, a number of methodological issues limit the ability to interpret these results.⁹

The median RFS was 10.2 months (95% CI, 7.9 to 12.9) in the oral azacitidine group and 4.8 months (95% CI, 4.6 to 6.4) in the placebo group, with a stratified HR of 0.65 (95% CI, 0.52 to 0.81; P = 0.0001) in favour of the oral azacitidine group.⁹

As of the final analysis, the percentage of patients who had relapsed was 64.7% in the oral azacitidine group compared with 76.5% in placebo group. Median time to relapse was 10.2 months (95% CI, 8.3 to 13.4) in the oral azacitidine group and 4.9 months (95% CI, 4.6 to 6.4) in the placebo group. The relapse rate at 1 year was 0.53 (95% CI, 0.46 to 0.59) in the oral azacitidine group and 0.71 (95% CI, 0.65 to 0.77) in the placebo group.⁹

Most patients in both treatment groups had discontinued study treatment (81.1% in the oral azacitidine group and 88.9% in the placebo group) at the time of the final analysis. The median time to treatment discontinuation was 11.4 months (95% CI, 9.8 to 13.6) in the oral azacitidine group and 6.1 months (95% CI, 5.1 to 7.4) in the placebo group. The 1-year treatment discontinuation rate was |||||||||||||||||||||||||||||||| in the oral azacitidine group and |||||||||||||||||||||||||||||||| in the placebo group.⁹

Overall, there were no statistically significant or clinically meaningful differences between the oral azacitidine group and the placebo group in the observed mean changes from baseline at any post-baseline assessment for the EQ-5D-3L health utility index and EuroQol Visual Analogue Scale (EQ VAS), PINR scale, and FACIT-F scale. Results using analysis of covariance (ANCOVA) models and mixed-effect model for repeated measures (MMRM) analyses showed similar results overall, with no statistically significant or clinically meaningful differences in least squares (LS) mean changes evident between treatment groups across assessments for any of the 3 instruments (EQ-5D-3L, PINR scale, and FACIT-F scale). Furthermore, no statistically significant differences in the proportion of patients with clinically meaningful deterioration were evident between the oral azacitidine and placebo groups for the FACIT-F scale and the EQ-5D-3L health utility index and EQ VAS. The time to definitive deterioration was similar between the oral azacitidine and the placebo groups regardless of the instrument used (i.e., FACIT-F, EQ-5D-3L health utility index, or EQ VAS). Clinically meaningful deterioration and time to definitive deterioration were not reported for the PINR scale.¹⁴ All analyses performed on the HRQoL outcomes and symptom severity were noninferential.

Harms Results

Key harms reported in the QUAZAR AML-001 trial as of the final analysis (July 15, 2019, data cut-off date) are summarized in Table 2. Nearly all patients in both study groups (97.9% of patients in the oral azacitidine group and 96.6% in the placebo group) experienced at least 1 treatment-emergent adverse event (TEAE). The most commonly reported TEAEs in the oral azacitidine and the placebo groups were nausea (64.8% and 23.6%, respectively), vomiting (59.7% and 9.9%, respectively), diarrhea (50.4% and 21.5%, respectively), and neutropenia (44.5% and 26.2%, respectively). TEAEs led to discontinuation of study treatment in 13.1% of patients in the oral azacitidine group and 4.3% of patients in the placebo group. The percentage of patients experiencing serious TEAEs was 33.5% in the oral azacitidine group compared to 25.3% in the placebo group. The most common serious TEAEs reported in the oral azacitidine and the placebo groups were febrile neutropenia (6.8% and 3.9%, respectively), pneumonia (3.8% and 3.0%, respectively), and pyrexia (2.1% and 0.4%, respectively). As of the final analysis, 66.4% of patients in the oral azacitidine group and 73.1% of patients in the placebo group had died, with most deaths attributed to AML.⁹

The percentage of patients experiencing gastrointestinal TEAEs was 91.1% in the oral azacitidine group compared to 66.5% in the placebo group (Table 2). The most commonly reported gastrointestinal TEAEs in the oral azacitidine and the placebo groups were nausea (64.8% and 23.6%, respectively), vomiting (59.7% and 9.9%, respectively), diarrhea (50.4% and 21.5%, respectively), and constipation (38.6% and 24.0%, respectively).⁹

The percentage of patients experiencing myelosuppression TEAEs was |||| in the oral azacitidine group compared to |||| in the placebo group (Table 2). The most commonly reported myelosuppression event was neutropenia, which was reported in 44.5% of patients in the oral azacitidine group compared to 26.2% of patients in the placebo group.⁹

The percentage of patients experiencing infections was 62.3% in the oral azacitidine group and 52.8% in the placebo group (Table 2). The most commonly reported infections in the oral azacitidine group included upper respiratory tract infections (13.1% and 13.7% in the oral azacitidine and placebo groups, respectively), |||||||||| (|||||||||||||| respectively), and |||||||||||||||||| (|||||||||||||||||| respectively).⁹

The percentage of patients experiencing fatigue was 29.7% in the oral azacitidine group and 19.3% in the placebo group.⁹

The types and frequency of hemorrhagic events were similar across study groups. The most commonly reported events (oral azacitidine versus placebo) included |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||.⁹

Critical Appraisal

Testing of treatment-by-time interactions in the overall Cox proportional hazards model suggested deviation from the proportional hazards assumption. Sensitivity analyses were performed to adjust for nonproportional hazards. Results of these analyses were consistent with the results of the proportional hazards analysis, which supported the HR obtained in the primary analysis demonstrating that oral azacitidine improved patients’ OS relative to placebo.

A higher proportion of patients in the placebo group compared with the oral azacitidine group received subsequent systemic therapy and subsequent transplantation after discontinuing study treatment. Results from planned sensitivity analyses, including censoring for the use of subsequent therapy or subsequent transplantation, were consistent with the primary OS results. Given the consistent results obtained in the sensitivity analyses and the fact that the clinical experts consulted by CADTH anticipated limited impact of post-relapse therapy on survival, the potential of confounding effects on survival outcomes from subsequent treatment is expected to be low.

Only the primary end point, OS, and the key secondary end point, RFS, were included in the statistical hierarchy and adjusted for multiplicity. No adjustment for multiplicity was made for any other analyses of other outcomes in the trial, such as HRQoL and symptom severity.

Methodological issues limited the ability to interpret the results from subgroup analyses. The subgroup analyses were noninferential, wide CIs reflected uncertainty in the effect estimates, and small sample sizes limited the generalizability to a broader population.

The ability to interpret results from the EQ-5D-3L and FACIT-F scale is limited and results from the PINR scale are inconclusive, due to the noninferential analyses, the significant decline in patients available to provide assessments over time, and the use of an unestablished method to derive a minimally important difference (MID).

The standard dose of oral azacitidine in the trial would be the dose used in Canadian clinical practice and is in line with the Health Canada–approved dosing. Post-relapse dose escalation, which was attempted in the QUAZAR AML-001 trial to reinitiate remission, is not generalizable to clinical practice in Canada according to the clinical experts consulted by CADTH and has not been approved by Health Canada. The clinical experts noted that it is unlikely that dose escalation would significantly influence the clinical outcomes observed in the trial due to the small numbers of patients who received dose escalation relative to the study sample size.

Indirect Comparisons

No indirect treatment comparisons were included in the sponsor’s submission to CADTH or identified in the literature search.

Other Relevant Evidence

No long-term extension studies or other relevant studies were included in the sponsor’s submission to CADTH or identified in the literature search.

Conclusions

One phase III, double-blind, placebo-controlled randomized controlled trial (RCT) (QUAZAR AML-001) provided evidence regarding the efficacy and safety of maintenance therapy with oral azacitidine plus BSC compared with placebo plus BSC in patients with AML in first CR. Compared with patients treated with placebo, patients who were treated with oral azacitidine showed statistically significant improvements in OS, the primary end point, and RFS, the key secondary outcome. The differences in median OS and RFS between treatment groups (approximately 10 months and 5 months, respectively) were considered clinically meaningful by the clinical experts consulted by CADTH. Other secondary efficacy outcomes, time to relapse from CR or CRi and time to discontinuation from treatment, were also supportive of the observed OS and RFS gains. The level of fatigue and HRQoL over the assessment period was similar between the 2 study groups and oral azacitidine did not appear to significantly delay time to deterioration compared with placebo. However, results for HRQoL and symptom severity remain uncertain due to a number of important limitations. The toxicity profile of oral azacitidine was considered manageable by the clinical experts consulted by CADTH.

Introduction

Disease Background

Acute myeloid leukemia is the most common form of acute leukemia in adults. It is an aggressive hematological malignancy that affects the production of normal blood cells (i.e., clonal expansion of myeloid blasts in the peripheral blood, bone marrow, and/or other tissue), causing weakness, infection, bleeding, and other symptoms and complications.^3,4,15 According to the most recent incidence statistics available for AML in Canada, 1,090 new cases of AML were reported in 2016.⁵ The most recent mortality statistics indicate that 1,184 Canadians died from AML in 2017.⁵ The median age of diagnosis has been estimated to be 65 years, with approximately one-third of diagnoses occurring in those aged 75 years or older.^4,16

In Canada, the 5-year and 10-year estimated net survival rates for AML are 21% and 18%, respectively.¹⁷ Median survival of untreated AML has been estimated to range between 11 and 20 weeks.¹⁶ The prognosis of patients with AML is influenced by several factors, including age at diagnosis, cytogenetics, and MRD. Older people have worse prognoses; those diagnosed after the age of 85 have an estimated 5-year survival rate of 0.4%.¹ Patients with AML are stratified into those with favourable, intermediate, and adverse risk, primarily mediated by the molecular genetic profile of the AML.⁶

AML is classified into 4 main disease subtypes by the WHO classification system: AML with recurrent genetic abnormalities (11% of cases), AML with myelodysplasia-related features (6% of cases), therapy-related AML (2% of cases), and AML not otherwise specified (81% of cases).¹⁶ Commonly associated mutations in AML include mutations in the FMS-like tyrosine kinase 3 (FLT3) gene and mutations in nucleophosmin 1 (NPM1) both of which are found in approximately 30% of patients with AML.⁶ Generally, 20% blasts in the marrow or blood is required for a diagnosis of AML; however, a diagnosis of AML may be made with a myeloid blast level of 20% with the following cytogenetic abnormalities: t(15;17), t(8;21), t(16;16), or inv(16).¹⁶ A diagnosis of AML is confirmed by bone marrow histology and ancillary tests such as cytogenetics and molecular testing.

Standards of Therapy

Although there are no overarching national Canadian guidelines for the management of AML, several international guidelines harmonize with practice in Canada.^4,6,7 Different therapeutic approaches are pursued for patients who can tolerate aggressive therapy and those who cannot. For patients who are not candidates for intensive chemotherapy because of advanced age or frailty, treatment with low-intensity therapies without curative intent is a reasonable option.⁴

In patients who can tolerate aggressive treatment, the first step of treatment is induction therapy, with the goal of achieving a first CR. Standard induction treatment involves combination chemotherapy (7 days of cytarabine and 3 days of anthracycline therapy [7 + 3]). Targeted therapies are added to this backbone as indicated. There is evidence to support that the combination of gemtuzumab ozogamicin with 7 + 3 prolongs progression-free survival and OS in patients with AML,¹⁸ particularly those with a favourable or intermediate cytogenetic risk. Gemtuzumab ozogamicin was recently reviewed by CADTH and is recommended for reimbursement for adults with previously untreated, de novo CD33-positive AML, with the exception of those with acute promyelocytic leukemia (APL), and who have good performance status and favourable, intermediate, or unknown cytogenetics (ELN [European LeukemiaNet] 2017).¹⁹ For patients with an FLT3 mutation, combining midostaurin with standard remission induction (7 + 3) and consolidation chemotherapy is associated with an OS benefit. Midostaurin has been reviewed by CADTH and is funded in most jurisdictions in Canada for this indication.²⁰ In patients with a higher risk of disease, a more aggressive regimen than the 7 + 3 regimen, such as FLAG (fludarabine plus high-dose cytarabine plus granulocyte colony-stimulating factor) with or without idarubicin may be considered. As was indicated by the clinical experts consulted by CADTH, if first remission is not achieved after 1 cycle of induction therapy, a second cycle of induction therapy is given. This may include a repeat of 7 + 3 chemotherapy or alternatively FLAG-Ida (FLAG plus idarubicin).¹⁶

The goal of consolidation therapy is to maintain first CR.¹ Post-remission consolidation therapy involves up to 4 cycles of high-dose cytarabine.¹⁶ Approximately 50% of patients aged 60 years or younger are cured in this fashion.^4,6,7 Allogeneic transplantation, which is potentially curative, is pursued as a consolidation strategy in first CR patients with intermediate or poor cytogenetic risk who are eligible for transplantation.¹ The clinical experts consulted by CADTH noted that patients with favourable cytogenetic risk are not generally considered for transplantation given their positive outcomes and likelihood of being cured after consolidation therapy with standard chemotherapy treatment alone. Allogeneic transplantation for AML in first remission is associated with a long-term survival probability of 50%; however, the procedure is associated with a high risk of morbidity and mortality.⁸ Reasons for transplantation ineligibility include old age, comorbidities, and decreased organ function.¹ The clinical experts consulted by CADTH noted that the stem cell donor option or a patient’s preference also determines transplant eligibility. Furthermore, according to the clinical experts, at times the optimal number of cycles of consolidation chemotherapy cannot be administered for similar reasons, as noted for transplantation ineligibility. The percentage of patients proceeding to transplantation is approximately 30% in those younger than 70 years of age and lower than 5% in those older than 70 years.¹

For patients who are ineligible for transplantation, maintaining first CR to delay relapse and prolong survival is an important clinical goal, particularly in patients with intermediate or poor cytogenetic risk who have a higher risk of relapse.¹ The clinical experts consulted by CADTH noted that there are currently no standard funded maintenance regimens specific for patients with AML who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for transplantation. In the absence of proven treatment options in this setting, a watch-and-wait approach or BSC is recommended. Treatment options of relapsed AML are limited and commonly involve palliative chemotherapy.¹⁶ Survival of patients with relapsed disease after induction and/or consolidation treatment is poor (if first CR lasts 1 to 2 years, median OS is estimated at 1 year or more; if first CR lasts less than 1 year, expected median OS is 4 to 6 months).¹ The clinical experts agreed that there is a need for effective treatments that prolong life, delay relapse, maintain or improve patient’s quality of life with an acceptable safety profile, and have a convenient oral route of administration.

The clinical experts noted that injectable azacitidine may be considered as maintenance therapy in patients who are ineligible for standard consolidation therapy. The use of injectable azacitidine for this particular clinical situation is considered off-label. Funding criteria for injectable azacitidine as maintenance therapy vary by jurisdiction and there is currently no standardized dose protocol. The product monograph for oral azacitidine discourages substituting oral azacitidine for IV or subcutaneous azacitidine.² Consequently, injectable azacitidine as maintenance therapy is given to a small number of patients. Due to its infrequent use, injectable azacitidine was not included as relevant comparator in the protocol for this CADTH systematic literature review.

The clinical experts also noted that MRD monitoring is currently being studied in clinical practice of AML management, as it may help identify patients at high risk of an impending relapse of AML, providing an opportunity to intervene with therapeutics before full-blown relapse. However, MRD monitoring in AML is not widely available and is not yet part of formal standard of care.

Drug

Oral azacitidine is a formulation of the hypomethylating drug azacitidine, which is administered orally.¹ It is a cytidine nucleoside analogue that incorporates into both DNA and RNA after phosphorylation to the nucleotide and produces anti-leukemic activity in cancers such as AML.¹

On January 4, 2021, oral azacitidine was approved by Health Canada as maintenance therapy in adult patients with AML who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT.² The sponsor’s requested reimbursement criteria for oral azacitidine are as per the Health Canada–approved indication. Oral azacitidine received a priority review by Health Canada. Oral azacitidine has no other Health Canada–approved indication and has not previously been reviewed by CADTH.

Oral azacitidine is available as 200 mg and 300 mg tablets. The recommended starting dosage is 300 mg administered orally once daily on day 1 through day 14 of continuous 28-day cycles. The product monograph states that if the ANC is less than 500/µL on day 1 of a cycle, oral azacitidine should not be administered and the start of the cycle should be delayed until the ANC is 500/µL or more.²

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

About the Patient Group(s) and Information Gathered

One patient group, the LLSC, provided input for this review. The LLSC has the following mission statement: Cure leukemia, lymphoma, Hodgkin’s disease and myeloma, and improve the quality of life of patients and their families. This is done through research, patient support, and advocacy.

The patient input was collected through an online survey created by the LLSC, available in both English and French, during the period from December 7, 2020, to January 24, 2021. A total of 29 patients with AML responded: 18 females, 10 males, and 1 patient who did not identify their gender. Most patients were from Ontario (n = 13); the remaining patients were from Quebec and British Columbia (n = 6 each) and Alberta (n = 4). The majority of patients were aged between the ages of 55 years and 74 years (n = 18); 9 patients were younger than 55 years of age, while 2 were older than 75 years. Most patients were diagnosed with AML between 2015 and 2020, with the exception of 2 patients who were diagnosed in 2014 (n = 1) and 2021 (n = 1).

Disease Experience

Respondents in the LLSC survey indicated that a varying range of AML symptoms affected their daily lives, including their social lives and relationships, causing a number of detrimental effects on their health. Patient respondents reported being easily fatigued, unable to exercise or work, losing vision in an eye, nausea, bruising, numbness or body aches, and being immunocompromised. Many of these symptoms left patients feeling physically and/or socially isolated and in fear of relapse, leading to negative impacts on their psychological well-being. Some patient respondents indicated that physical activity helped alleviate their fatigue.

The following quotes illustrate patients’ experiences of living with AML:

• “The initial year was absolutely overwhelming, in terms of treatment and expenses.”

• “Everything in my life stopped cold turkey-employment, social life, relationships, etc. I made a complete personal 360 degree pivot to focus on my healing and living.”

• “COVID and my compromised immune system has caused me to be very socially isolated. I haven’t seen some very important people in my life for almost 2 years at this point.”

The effects of AML are experienced by both patients and their caregivers. According to the survey, patients reported that caregivers or family members might feel stress, worry, sadness, insecurity, and fear of dying. Caregivers were also reported to express more tolerance toward the patients. One individual responded, “It affects those in your immediate household since they will most probably be your primary caregivers. However, anyone that is close to you will be affected as well.”

Experiences With Currently Available Treatments

According to the survey, the majority of patients received chemotherapy and stem cell or bone marrow transplantation after diagnosis, while a small number of patients received drug or radiation therapy and 1 patient received chimeric antigen receptor T-cell therapy. The following treatment-associated side effects had a large impact on patients’ quality of life (in order of most frequently reported): hair loss; weakness; extreme fatigue; diarrhea; infections; anemia; mouth sores; nausea and vomiting; fever; low blood cell counts; tingling sensations; constipation; graft-versus-host disease; lung, heart, kidney or nerve problems, cough; rashes; shortness of breath; and psychological distress. Chemotherapy and transplantation reportedly had 1 of the largest negative impacts on quality of life, due to changes in physical activity, anxiety, problems in mental health and overall happiness, eating challenges, and social and educational development. Feelings of isolation during hospitalization for stem cell transplantations, was another common issue reported among respondents. Opportunistic infection could occur due to their immunocompromised state.

Comments from patients regarding their experiences with AML treatments included:

• “Your whole world changes when you are diagnosed with AML. Suddenly, you confront your mortality. You feel extremely weak, you have to go into hospital for months, and you don't realize you MUST go into remission to have a stem cell transplant.”

• “The Chemo, transplant, radiation and drug therapies are all hard on the body and short term, long term effects are tough as well. The worst part is that these therapies can cause other illnesses.”

Patient respondents reported a mix of both positive and negative experiences accessing treatments. Almost half of the patients reported generally positive experiences and support from medical staff. Six patients reported negative experiences, resulting from a long referral period with their physician, lack of treatment information, difficulty getting a room for consolidation therapy, having to receive treatment in another city or province, being away from their support systems as a result of moving, and a general lack of psychological support.

Improved Outcomes

Patients reported that they expect any new drug or treatment to produce the following key outcomes: maintaining remission, fewer side effects, lower cost, and being accessible in their neighbourhoods. Additionally, the following factors were noted in choosing a new cancer treatment: physician recommendation, impact on disease, quality of life, closeness to home, and outpatient treatment. The availability of supportive options, such as meditation, hypnosis, neuro-linguistic programming support, and awareness support (thoughts, emotions, and behaviours), were also mentioned.

Experience With Drug Under Review

No survey respondents indicated that they had experience taking oral azacitidine.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process, providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 3 clinical specialists with expertise in the diagnosis and management of AML.

Unmet Needs

The clinical experts consulted by CADTH noted that there are currently no standard funded maintenance regimens specific to patients with AML who have achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT. In the absence of proven treatment options in this setting, watch-and-wait or BSC is recommended. There was consensus among the clinical experts that there is an unmet need for effective therapies with an acceptable toxicity profile that delay relapse, prolong survival, maintain HRQoL and potentially lead to a cure in patients with AML who are in their first remission and not eligible for transplantation. They emphasized the need for a convenient oral route of administration to achieve high adherence and reduce hospital-based resource utilization. One clinical expert suggested that the benefit-risk profile of new therapies may be more beneficial for patients with high-risk features for relapse (e.g., MRD-positive or intermediate- or poor-cytogenic risk [ELN 2017]).

Place in Therapy

Oral azacitidine was intended to be used by patients with AML who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT as per the QUAZAR AML-001 trial. While the clinical experts consulted by CADTH agreed that oral azacitidine would likely shift the current treatment paradigm, they also agreed that more experience with maintenance oral azacitidine therapy will be necessary to determine the potential impacts on the current treatment paradigm (e.g., potentially fewer patients requiring transplantation, a reduction in the number of cycles of consolidation chemotherapy, and refinement of the target population). The experts added that, given the lack of comparative data, the relative efficacy of hypomethylating drugs, such as oral azacitidine, compared with targeted therapy in subtypes of AML with certain mutations (e.g., midostaurin in the FLT3-ITD-mutated AML subtype) is currently unknown.

Patient Population

Overall, the clinical experts agreed that patients selected according to the inclusion and exclusion criteria of the QUAZAR AML-001 trial should be eligible for maintenance oral azacitidine therapy. While it was agreed that there is currently insufficient evidence to recommend which patient subgroups would be best suited for or most likely show a response to oral azacitidine, the clinical experts identified the following potential subgroups as having the highest risk of relapse and therefore being most in need of maintenance therapy: patients 65 years or older, patients with an MRD-positive status, patients who have not received consolidation chemotherapy, and patients with poor-risk karyotypes. Subgroups who may benefit the least from oral azacitidine include patients with an MDR-negative status, patients with low-risk features for relapse, patients who develop unacceptable toxicities, and patients lacking the social and/or medical support necessary to be safely treated with oral azacitidine (e.g., those with no family or friends, living in a remote community without medical support services, or an inability to understand the treatment risks and benefits). The clinical experts agreed that patients would be identified as possible candidates for oral azacitidine treatment if they are considered ineligible for transplantation. The transplantation ineligibility status may be known in patients at the time of diagnosis or may develop over the course of their treatment.

Assessing Response to Treatment

In the opinion of the clinical experts consulted by CADTH, clinical assessments to evaluate the response to treatment include regular monitoring of blood counts (every 1 to 2 weeks) to determine if a patient maintains CR or CRi. It was noted that 28-day treatment cycles would likely lead to drugs being dispensed 1 month at a time due to drug costs. If changes in a blood count signal a potential relapse (e.g., blood counts drop significantly below the normal range, particularly if they are trending downward) a bone marrow examination may be required to determine if a patient is still in CR or CRi or has relapsed AML. Bone marrow exams are usually not done unless a relapse is suspected. Some clinicians may routinely monitor MRD status using peripheral blood every 3 to 6 months; however, this is not currently part of standard of care.

The clinical experts indicated that the most clinically meaningful responses to treatment include prolonged OS and RFS while maintaining or improving HRQoL and reducing symptom burden. No increase of drug-related toxicities, such as infections, neutropenia, and thrombocytopenia, was also noted as a clinically meaningful outcome.

Discontinuing Treatment

In the opinion of the clinical experts consulted by CADTH, maintenance therapy with oral azacitidine should be discontinued if a patient experiences relapsed AML, has a markedly impaired performance status, or is intolerant to or experiences unacceptable toxicity from oral azacitidine. If AML recurs with blast counts of 5% or greater in the peripheral blood or bone marrow, oral azacitidine should be discontinued. The clinical experts agreed that there is currently insufficient evidence to recommend dose escalation as attempted in the QUAZAR AML-001 trial when AML recurs with blast levels reaching between 5% and 15% in either the peripheral blood or bone marrow. It was stated that the aim of dose escalation is to reinitiate remission, which is beyond the scope of the requested reimbursement criteria, which are for oral azacitidine maintenance therapy.

Prescribing Conditions

In the opinion of the clinical experts consulted by CADTH, oral azacitidine is a drug that is self-administered in a patient’s home. Monitoring of blood work and evaluation for toxicity from treatment, which may warrant dose modification, is required. The clinical experts agreed that patients should have access to a regional cancer clinic to ensure that treatment tolerance is confirmed, and that the disease has not relapsed.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups.

Input was received from 2 clinician groups, the OH-CCO H-DAC and the ATB-MPG, on the reimbursement review of oral azacitidine (Onureg) for the maintenance therapy in adult patients with AML who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT.

The clinicians from the OH-CCO H-DAC noted that they provide evidence-based clinical and health system guidance on drug-related issues in support of the OH-CCO mandate, which includes the Provincial Drug Reimbursement Programs and the Systemic Treatment Program. Information for this review was jointly discussed at an H-DAC meeting.

The ATB-MPG is a group of physicians who treat myeloid malignancies and acute leukemias within Alberta and functions within the Alberta Hematology Tumour Group. Members meet weekly as Edmonton and Calgary groups to discuss patient cases and upcoming and open clinical trials. They also meet quarterly to discuss clinical trials, the streamlining of patient care across the province, and improvements to the diagnosis, management, and follow-up of patients with myeloid diseases. The group meets annually to update Alberta provincial treatment guidelines. Information for this review was gathered through a review of publications and group discussions.

Unmet Needs

The ATB-MPG noted that a majority of new AML cases are diagnosed in patients who are elderly and/or comorbid. The group added that approximately 80% of patients are diagnosed with intermediate or high-risk AML. Although these patients will often achieve remission with induction chemotherapy, they will invariably relapse without further chemotherapy to consolidate and maintain remission. Allogeneic stem cell transplantation, which offers a potential cure, is standard-of-care post-remission therapy in patients who are fit and younger than 70 years old. The ATB-MPG noted that, because of the serious and prohibitive morbidity and mortality associated with transplantation in older or comorbid patients, many of these patients are not eligible for transplantation.

Both clinician groups agreed that there are currently no standard funded maintenance regimens specific for patients with AML who are in first remission following induction therapy with or without consolidation treatment and who are not eligible for transplantation. The clinicians from the OH-CCO H-DAC added that, in the absence of proven treatment options, patients are currently being followed by surveillance. Similarly, clinicians from the ATB-MPG noted that current options for care involve a palliative approach focused on transfusion support, hydroxyurea, treatment for infections, and general symptom-based care. The clinicians from ATB-MPG described the current treatment landscape for these patients as bleak and added that treatment options are needed to offer improved survival and quality of life for these patients, who generally only have months to live once they relapse.

The ATB-MPG added that, based on the results reported in the QUAZAR AML-001 trial, oral azacitidine appeared to be a promising option, potentially addressing a large unmet need in the AML disease space. The ATB-MPG noted that oral azacitidine was associated with a significant prolongation of remission and a significant improvement in OS compared with placebo, as well as being generally well tolerated. The group emphasized that oral azacitidine is an oral medication that offers important quality-of-life and feasibility advantages by avoiding the need for frequent visits to treatment centres for therapy. The ATB-MPG noted oral azacitidine is accessible in Alberta via compassionate access from the drug manufacturer and access is expected to be temporary and cumbersome to navigate and may not be available to all candidate patients. The ATB-MPG added that, given the non-curative nature of nontransplantation approaches and the elderly and/or comorbid nature of this target population, treatments that are easy and convenient to administer and well tolerated are urgently needed to prolong remission and survival without adversely affecting quality of life.

According to the clinicians from the OH-CCO H-DAC, the most important goals that an ideal treatment would address are to prolong life as reflected by the end point of OS and improve quality of life. The clinician group commented that maintenance oral azacitidine demonstrated an OS benefit in a patient population that has not seen the benefit of new therapies since 1973, when 7 + 3 became the standard of care. Similarly, the ATB-MPG noted that prolonging life and remission, as well as improving HRQoL, are the most important treatment goals for the target population. It was noted that improving quality of life includes reducing transfusion needs, hospital admissions, severity of symptoms, adverse events (AEs), and decreasing hospital visits to receive therapy. The latter is particularly important for patients who live far from treatment centres and/or are not independent (i.e., caregivers may struggle to manage the demands of intensive therapies).

Both clinician groups agreed that AML remains a highly “lethal” cancer with poor outcomes in patients who are not eligible for transplantation (including those who do not have an available donor). Therefore, these patients have the greatest unmet need for an intervention. The ATB-MPG noted that, given that the large majority of patients who are newly diagnosed with AML have intermediate- or high-risk disease and that a considerable proportion of these patients are not eligible for transplantation, these represent a substantial cohort of patients with AML who will invariably relapse with a prognosis that is measured in months. The group noted that oral azacitidine would fill a large therapeutic hole in the treatment landscape and has the potential to address a substantial unmet need.

Place in Therapy

With respect to how oral azacitidine would fit into the current treatment paradigm, both clinician groups noted that the treatment would be a standard-of-care maintenance therapy for patients with AML who are in first CR and ineligible for HSCT.

The ATB-MPG added that, because oral azacitidine offers the prospect of prolonged remission and survival, decreased disease-related complications, an acceptable safety profile, and a convenient oral route of administration with the potential for improved quality of life, it could shift the current treatment paradigm for patients for whom few other therapeutic options exist. The group noted that the use of oral azacitidine could result in a substantial reduction in the health care burden to patients, their caregivers, and the health care system as a whole.

The OH-CCO H-DAC clinicians commented that oral azacitidine would not affect the sequencing of available therapies for AML. The ATB-MPG added that no other treatment options are available to recommend for this select group of patients before initiating oral azacitidine. However, the ATB-MPG clinicians noted that patients may have received additional cycles of chemotherapy to consolidate their remission before starting oral azacitidine, as per the eligibility criteria of the QUAZAR AML-001 trial. The ATB-MPG also noted that, after receiving oral azacitidine, patients could be enrolled in a clinical trial if available or be given palliative supportive therapy (transfusion, hydroxy urea, or symptom-driven care) as second-line treatment.

Both clinician groups agreed that the patients best suited for treatment with oral azacitidine would be those who meet the QUAZAR AML-001 trial inclusion criteria; this would include patients who do not have an available donor. The OH-CCO H-DAC clinicians added that, although not included in the pivotal trial, maintenance oral azacitidine would be appropriate for patients younger than 55 years of age. Based on pharmacokinetic data, patients with MDS who have progressed on subcutaneous azacitidine, subsequently received induction chemotherapy with or without consolidation, achieve CR, and are transplantation ineligible, may also benefit from oral azacitidine.

According to the clinicians from OH-CCO’s H-DAC, the patients best suited for treatment with oral azacitidine would be identified through a diagnosis of AML, having received treatment with induction therapy with or without consolidation, having achieved CR, and being ineligible for transplantation. The clinicians added that all transplantation-ineligible patients in first CR would be equally suitable for oral azacitidine. The ATB-MPG clinicians noted that patients would be identified through clinician examination and assessment. Testing is widely available as this is an objectively diagnosed disease. The group added non-candidacy for allogeneic stem cell transplantation would involve formal consultation and assessment to confirm transplantation ineligibility. The ATB-MPG group noted that patients who were eligible for transplantation would be the least suitable for treatment with oral azacitidine. Additionally, treatment of patients who are too elderly or comorbid to be included in the clinical trial for oral azacitidine would be subject to a multifaceted clinical judgment by the treating physician. The group commented that patients who choose not to receive oral azacitidine therapy for personal or other reasons would also not be suitable.

Both clinician groups agreed that it would not be possible to identify patients who are most likely to exhibit a response to treatment with oral azacitidine at this time and there were no subgroups of patients who did not benefit from oral azacitidine in the pivotal trial. The ATB-MPG noted that potential identification of particular subpopulations that may experience greater benefit would require further long-term analyses supported by additional patient and clinical experience with oral azacitidine use.

Assessing Response to Treatment

The OH-CCO H-DAC clinicians noted that outcomes for these patients should be assessed according to usual clinical practice, and that oral azacitidine should be continued until overt clinical relapse. The clinicians added that bone marrow sampling is not necessarily required unless there is suspicion of relapse. According to the clinicians, a clinically meaningful response to treatment would be the absence of relapse. The clinicians commented that treatment response should be assessed through a complete blood count and bone marrow surveillance as per clinician judgment.

The ATB-MPG noted that the following outcomes would be used to determine patients’ response to treatment: bone marrow biopsy to assess disease response, improvement in cytopenia and transfusion needs, decrease in blast counts in peripheral blood, improvement in disease-related symptoms (infections, bleeding, and fatigue), improvement in quality of life (measured subjectively by patient-described experiences and objectively through quality-of-life scores), and improvement in functional status (due to improvements in cytopenias, decreased disease-related complications, and decreased need for hospital admission and outpatient visits for therapy administration, transfusion, or other support).

The clinicians added that these outcomes are aligned with those used in clinical trials, although bone marrow biopsies to assess relapse are more frequently performed in a clinical trial setting.

The ATB-MPG clinicians noted that patients who are undergoing initial therapy will have laboratory work performed weekly to assess transfusion needs and drug side effects. However, the group noted that this monitoring will decrease in frequency to monthly depending on individual patients’ transfusion needs, cytopenias, drug tolerability, and stability of treatment response. The group commented that improvement in cytopenias and transfusion requirements will be monitored as patients proceed through each cycle of treatment. The clinicians added that a marrow biopsy would be performed after 4 to 6 cycles and would act as a definitive indicator of disease response. According to the clinicians, once remission or maximal response is obtained, a repeat bone marrow biopsy would be indicated if there is clinical deterioration, significant cytopenias, or other clinical findings requiring reassessment of disease status.

Discontinuing Treatment

The OH-CCO H-DAC clinicians noted that disease relapse or treatment intolerance should be considered when deciding whether to discontinue treatment with oral azacitidine. The ATB-MPG added that assessment of disease progression as measured by peripheral blood indices and transfusion needs, and definitively by a bone marrow biopsy, is required to determine treatment discontinuation. Additionally, the ATB-MPG noted that treatment intolerance, which could include cytopenias complicated by recurrent severe infections, transfusion needs that do not improve after 4 to 6 cycles of therapy, intolerable adverse effects (i.e., gastrointestinal intolerance), or patient preference are factors to be considered by the treating physician for treatment discontinuation.

Prescribing Conditions

Both clinician groups agreed that outpatient cancer clinics in the community setting where regular laboratory and transfusion support are available are most appropriate for treatment with oral azacitidine as the drug can be taken at home.

Additional Considerations

No additional considerations were identified.

Drug Program Input

Drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation.

The drug programs indicated that the standard approach for patients who have achieved CR or CRi after potential consolidation therapy is to “watch and wait” without any disease-targeting therapies. It was noted that the Health Canada product monograph² for oral azacitidine indicates that oral azacitidine is not interchangeable with injectable azacitidine, which is available in generic form. As indicated in the product monograph, oral azacitidine is available as 200 mg and 300 mg tablets in a blister pack containing 7 tablets. The list price per tablet was described as extremely high. Furthermore, oral azacitidine would be supplied in 7-day blister packs and dose adjustments or extending days of treatment (i.e., from days 1 through 14 to days 1 through 21) would have a significant effect on treatment costs or risk of wastage. It was suggested that jurisdictions may wish to limit the quantity dispensed if part fills are mandated (e.g., in 7-day increments versus a full cycle) and that dispensed quantities should align with the timing of clinical assessments and blood work. The drug programs emphasized that, because the most commonly reported serious AE of oral azacitidine is febrile neutropenia, patients may require granulocyte colony-stimulating factor, which will add to the cost of patient treatment. It was noted that several drugs in the AML treatment space have been either reviewed or are currently under review by CADTH; midostaurin, gemtuzumab ozogamicin, and gilteritinib were previously reviewed, and azacitidine (injectable) with venetoclax, low-dose cytarabine with venetoclax, and daunorubicin and cytarabine are currently under review.

The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence

The clinical evidence included in the review of oral azacitidine is presented in this section. The systematic review includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. No indirect or other evidence was identified.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of azacitidine tablets as maintenance therapy for the treatment of AML in adult patients who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT.

Methods

Studies selected for inclusion in the systematic review included the pivotal study provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the Peer Review of Electronic Search Strategies checklist (https://www.cadth.ca/resources/finding-evidence/press).²³

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974–) via Ovid. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were Onureg, oral azacitidine, and acute myeloid leukemia. Clinical trials registries searched included the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

Search filters were applied to limit retrieval to randomized controlled trials or controlled clinical trials. Retrieval was not limited by publication date or language. Conference abstracts were excluded from the search results. Appendix 1 provides detailed search strategies.

The initial search was completed on March 25, 2021. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on August 11, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist (https://www.cadth.ca/grey-matters).²⁴ Included in this search were the websites of regulatory agencies (US FDA and European Medicines Agency [EMA]). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the sponsor of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

A focused literature search for network meta-analyses dealing with acute myeloid leukemia was run in MEDLINE All (1946–) on March 26, 2021. No language or date limits were applied.

Findings From the Literature

Six reports presenting data from 1 unique study were identified from the literature for inclusion in the systematic review (Figure 1). The included study is summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

The QUAZAR AML-001 trial is an ongoing, international, multi-centre, double-blind, placebo-controlled, randomized phase III trial that compared the efficacy and safety of maintenance therapy with oral azacitidine plus BSC compared with placebo plus BSC in patients with AML in first CR. The primary objective of the trial was to compare OS between maintenance therapy with oral azacitidine and placebo in patients with AML, aged 55 years or older who had achieved first CR or CRi after induction with intensive chemotherapy with or without consolidation chemotherapy.¹⁰ Patients in this international trial were randomized at 147 sites across 23 countries, which are listed in Table 6. The majority of sites were in North America and Europe, including 5 sites in Canada.⁹ According to the sponsor, 　|　 Canadian patients were randomized in the QUAZAR AML-001 trial.²⁷

A total of 472 patients were randomized (using an interactive voice response system) in a 1:1 ratio to receive maintenance oral azacitidine plus BSC or oral placebo plus BSC. Enrolment began on May 10, 2013; the enrolment end date has not been determined.¹² Best supportive care used in combination with either oral azacitidine or placebo was administered as needed. The BSC products were recorded as concomitant medications, therapies, or procedures. A description of BSC is provided in Table 6.¹⁰ Randomization was stratified by age at the time of induction therapy (55 to 64 years and ≥ 65 years), prior history of MDS or CMML (yes/no), cytogenetic risk category at time of induction therapy (intermediate/poor risk), and receipt of consolidation therapy following induction (yes/no). The study design is depicted in Figure 6.

The study consisted of 3 phases: pre-randomization (screening), treatment, and follow-up. No crossover between treatment groups was allowed at any point during the study and blinding was upheld until study closure and database lock.¹⁰ During the follow-up phase, patients were monitored for survival every month for the first year and then every 3 months. An extension phase was added to the trial as protocol amendment 2 to ensure drug access for patients who continued to benefit from azacitidine as well as for the collection of survival data after study closure.¹⁰ Data from the extension phase are not included in this review.

Response assessment, according to AML response criteria defined by the International Working Group (IWG) for maintaining CR or CRi was planned to occur every 3 cycles starting at cycle 3 until cycle 24, at cycles 30 and 36, and at the treatment discontinuation visit. After cycle 36, response assessments occurred at the investigator’s discretion.⁹ Relapse was defined as the appearance of a blast level of at least 5% in peripheral blood or bone marrow.¹¹ A post-relapse dose escalation (300 mg once daily from 14 days to 21 days of the 28-day cycle) was planned for patients with disease relapse with a blast level of between 5% and 15% in either the peripheral blood or bone marrow.⁹ This dose escalation was used in the context of re-induction of remission, which is not consistent with the Health Canada indication and the requested reimbursement criteria, which are for oral azacitidine as maintenance therapy.¹¹ The QUAZAR AML-001 trial was not designed to assess if dose escalation of oral azacitidine produces benefits for patients. Dose escalation has not been authorized by Health Canada according to the Health Canada product monograph.¹¹

The primary analysis, which coincided with the final data cut-off date (July 15, 2019) was conducted once the total number of events (n = 330 deaths) occurred.^9,10 The key secondary and other secondary end points were also planned to be assessed at the final data cut-off date. An updated analysis, which was not planned a priori in the statistical analysis plan, occurred based on a September 2020 data cut-off date. Data from this updated data cut-off date have not been provided by the sponsor to CADTH at the time of this review.¹² The trial is still ongoing with an estimated completion date of December 31, 2021.¹³ The QUAZAR AML-001 trial was sponsored by Celgene (a wholly owned subsidiary of Bristol Myers Squibb).¹⁰

Figure 2: QUAZAR AML-001 Trial Study Design

Source: Clinical Study Report.⁹

Populations

Inclusion and Exclusion Criteria

The key inclusion and exclusion criteria used in the QUAZAR AML-001 trial are described in Table 6. Briefly, the trial enrolled adults, 55 years of age and older diagnosed with AML or AML secondary to prior MDS or CMML, with intermediate- or poor-risk cytogenetics, and who had achieved first remission (CR or CRi) following induction with or without consolidation chemotherapy. Patients were excluded if they were considered eligible for HSCT. Transplantation eligibility criteria were not predefined in the trial protocol. At screening patients had to have an ECOG PS of 0 to 3 and adequate marrow function (based on area under the curve for ANC ≥ 0.5 × 10⁹/L and a platelet count ≥ 20 × 10⁹/L).¹⁰

Baseline Characteristics

The baseline characteristics of patients who comprised the intention-to-treat (ITT) population of the QUAZAR AML-001 trial are summarized in Table 7. In the ITT population the mean age was 68.0 (SD = 5.66). The oral azacitidine group had a lower proportion (versus placebo) of male patients (49.6% versus 54.3%) and patients enrolled in trial sites in Asia (2.5% versus 7.3); and a higher proportion of White patients (90.8% versus 84.2%) and patients enrolled in trial sites in Europe (70.2% versus 62.8%). A total of 90.9% of patients had de novo AML, and the most common initial AML classification was AML not otherwise specified (62.2% versus 62.0%). The majority of patients had an ECOG PS of 0 (48.7% versus 47.4%) or 1 (42.4% versus 45.3%) and intermediate cytogenetic risk (85.3% versus 86.8%). Subsequent to induction therapy 80.1% of patients received consolidation therapy, with most of these patients receiving 1 cycle (46.2% versus 43.6%) or 2 cycles (29.4% versus 32.9%). The most common reason for transplantation ineligibility was age (64.7% versus 65.0%), followed by comorbidities (21.8% versus 21.4%), and no available donor (15.5% versus 15.0%). The oral azacitidine group had a lower proportion of patients with a negative MRD status at randomization compared with the placebo group (55.9% versus 47.4%) and of patients who achieved CR after induction therapy with or without consolidation (78.6% versus 84.2%).⁹ The clinical experts consulted by CADTH noted that the differences observed in patient characteristics at baseline between the oral azacitidine and the placebo groups are unlikely to have had an impact on the treatment differences observed between the treatment groups.

All patients in the ITT population received at least 1 induction therapy for AML before enrolment into the trial, and the proportions and types of therapies were balanced across groups as summarized in Table 8. The most commonly received induction therapy (oral azacitidine versus placebo) was cytarabine (99.6% versus 99.1%) followed by idarubicin (54.2% versus 55.6% placebo), and daunorubicin (33.2% versus 32.9%).⁹

In the ITT population, most of the patients received at least 1 consolidation therapy for AML before enrolment into the trial, and the proportions and types of therapies were balanced across groups as summarized in Table 9. The most commonly received consolidation therapy was cytarabine (78.2% oral azacitidine; 81.6% placebo) followed by idarubicin (21.4% oral azacitidine; 18.8% placebo), and daunorubicin (9.2% oral azacitidine; 6.4% placebo). Most of the patients received 1 cycle (46.2% oral azacitidine; 43.6% placebo) or 2 cycles (29.4% oral azacitidine; 32.9% placebo) (Table 9).⁹

Interventions

Patients randomized in the QUAZAR AML-001 trial were allocated to either the oral azacitidine plus BSC group or the placebo plus BSC group. Treatments and doses are described in Table 10.

Dose Escalation in Patients With Disease Relapse

A dose escalation (300 mg once daily from 14 days to 21 days of the 28-day cycle) was planned for all patients with disease relapse with blasts between 5% and 15% either in the peripheral blood or bone marrow at the discretion of the investigator and under the following criteria¹⁰:

• For patients currently on blinded treatment with oral azacitidine 300 mg or placebo once daily for 14 days: Escalate schedule to oral azacitidine 300 mg or placebo once daily for 21 days

• For patients currently on blinded treatment of oral azacitidine 200 mg or placebo once daily for 14 days in the event that the dosage had previously been reduced: Escalate schedule as a first step to 21 days once daily and then escalate dosage to oral azacitidine or placebo to 300 mg once daily for 21 days

• For patients currently on blinded treatment with oral azacitidine 200 mg or placebo once daily for 7 days: Escalate schedule as a first step, to oral azacitidine 200 mg or placebo once daily for 14 days and then escalate dosage to oral azacitidine 300 mg or placebo once daily for 14 days.

Dose Modifications

Dose interruptions, delays, or modifications were permitted per guidelines of version 4.0 of the National Cancer Institute Common Terminology Criteria for Adverse Events for the grading of severity.¹⁰ A maximum of 1 dose reduction (to a daily dose of 200 mg) was permitted. One treatment schedule modification was allowed from 14 to 7 days (21 days to 14 days for patient on dose escalation). Patients could not receive less than 200 mg or be on a treatment schedule for less than 7 days. Once a dose or schedule modification occurred, titration back up to the original dose or schedule was permitted.¹⁰

Concomitant Medication

Best supportive care was documented as concomitant medication in the trial. The following BSC medications were allowed (but not limited to): red blood cell and platelet transfusions, use of an erythropoiesis-stimulating drug, antibiotic, antiviral, and antifungal therapy, nutritional support, and granulocyte colony-stimulating factor for patients experiencing neutropenic infections.¹⁰

Outcomes

A list of the efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 11 and summarized in the following section. A detailed discussion and critical appraisal of the HRQoL and symptom severity outcomes is provided in Appendix 4.

Overall Survival

Overall survival was the primary outcome of the QUAZAR AML-001 trial and was defined as the time from the date of randomization until the date of death due to any cause.¹⁰

Relapse-Free Survival

Relapse-free survival was designated the key secondary outcome and defined for patients who achieved CR or CRi and measured as the time from the date of randomization until the date of documented relapsed disease or death by any cause, whichever occurred first.¹⁰ Disease relapse for the analysis of RFS was determined according to the IWG AML response criteria: relapse after CR or CRi was defined as reappearance of leukemic blasts in the peripheral blood of 5% or more in the bone marrow not attributable to any other cause (e.g., bone marrow regeneration after consolidation therapy). In the setting of recent treatment, if there were no circulating blasts and the bone marrow contained 5% to 20% blasts, a repeat bone marrow sample taken at least 1 week later was necessary to distinguish relapse from regeneration. A separate central review (blinded to patients’ treatment assignment) was used to confirm all clinical laboratory value assessments.¹⁰ A sensitivity analysis was conducted to assess RFS based on investigator-assessed response.¹⁰

Time to Relapse From CR or CRi

Time to relapse from CR or CRi was designated an additional secondary outcome and was defined as the interval from the date of randomization to the date of documented relapse after CR or CRi, as determined according to the IWG’s AML response criteria.¹⁰

Time to Discontinuation From Treatment

Time to discontinuation from treatment was designated an additional secondary outcome and was defined as the interval from the date of randomization to the date of discontinuation from investigational product.¹⁰

Health-Related Quality of Life

The HRQoL outcomes measured in the trial included the EQ-5D-3L instrument and the PINR scale. The EQ-5D-3L instrument was an additional secondary outcome, whereas the PINR scale was an exploratory outcome.¹⁰ The primary analysis of the EQ-5D-3L assessed the treatment-effect difference between oral azacitidine and placebo on the mean change from baseline. The secondary analyses of the EQ-5D-3L questionnaire evaluated the proportion of patients with a clinically meaningful deterioration at a given post-baseline assessment visit, as well as the time to a definitive deterioration (defined as a clinically meaningful deterioration for at least 2 consecutive assessment visits).¹⁴ The sponsor defined a clinically meaningful deterioration as a decline of at least 0.10 and 11 points on the EQ-5D-3L health utility index and the EQ VAS, respectively.¹⁴ The EQ-5D-3L was assessed on day 1 of each 28-day cycle until discontinuation of study treatment and at the end-of-treatment visit.¹⁴ A detailed discussion and critical appraisal of the HRQoL measures are provided in Appendix 4.

The EQ-5D-3L is a standardized, patient self-administered questionnaire that consists of 2 parts:

• Health utility index: A descriptive section comprising 5 dimensions. Each of the 5 dimensions (anxiety/depression, pain/discomfort, mobility, self-care, usual activities) has 3 response levels reflecting health and well-being (no problems, moderate problems, and extreme problems).¹⁴ The health utility index generates a population preference–weighted score in that it reflects societal preferences for the health states.

• EQ VAS: A visual analogue scale that provides a quantitative measure by recording a respondent’s self-rated health on a 100-point scale between 100 (“best imaginable health state”) and 0 (“worst imaginable health state”).¹⁴ The higher the score the better the imaginable health state. The EQ VAS captures the individual’s value or judgment of his or her present health state.

The validity of the EQ-5D-3L questionnaire was assessed in 168 patients with AML in China.²⁸ Known-group and discriminative validities were assessed and both demonstrated an ability to distinguish groups based on various baseline factors. Construct validity was assessed through the convergent approach, indicating overall good validity. Test-retest reliability was assessed using the intraclass correlation coefficient (ICC), which showed good reproducibility. No relevant studies assessing responsiveness were found. An MID was not identified in the literature for patients with AML. The sponsor referenced studies in the published literature that reported MIDs ranging from 0.06 to 0.12 for the EQ-5D-3L health index score and from 6 to 11 for the EQ VAS in cancer patients.^29,30 Other reported MIDs for the health index score range from 0.033 to 0.074 in the general population.³¹ The sponsor’s submission noted that, although the terms MID and RD have gradually differentiated over time, these terms are still being used interchangeably in the literature.¹⁴ The sponsor therefore used the same cut-off values for the RD as for the MID. In addition, the sponsor presented the entire distribution of responses (cumulative distribution frequency curves of patients experiencing different degrees of change at cycles 3, 6, 12, 18, and 24).³²

The analysis of the PINR scale assessed the treatment-effect difference between oral azacitidine and placebo on the mean change from baseline.³² The PINR scale was assessed on day 1 of each 28-day cycle until discontinuation of the study treatment and at the end-of-treatment visit.¹⁴ The PINR scale is a self-administered patient questionnaire for evaluating the level of physical impairment, difficulty completing outdoor physical tasks, and difficulty completing indoor tasks over the past 24 hours.¹⁴ The PINR scale consists of 3 questions, each reflecting the level of impairment and difficulty (0 = “no impairment or not difficulty” and 10 = “worse impairment or extremely difficult”). The higher the score, the greater the difficulty or impairment respondents experienced. Items were scored separately (i.e., each item was scored and not combined into a single overall score). The PINR scale was assessed on day 1 of each 28-day cycle and at the end-of-treatment visit.¹⁴

No study was found that assessed the psychometric properties related to validity or reliability for the PINR scale in patients with AML or leukemia. No studies of the validity or reliability of the PINR scale were provided in the sponsor’s submission to CADTH. An MID for the PINR scale was not identified in the literature for AML or leukemia patients. The sponsor’s submission noted that neither the MID nor the responder definition (RD) for the PINR scale were available in the literature.¹⁴ The sponsor’s submission reported that half of the SD at baseline from all ITT patients was used as an approximation for the MID.

Symptom Severity

Symptom severity was assessed in the trial using version 4 of the FACIT-F scale and was designated an additional secondary outcome.¹⁰ The primary analysis of the FACIT-F scale assessed the treatment-effect difference between oral azacitidine and placebo in the mean change from baseline; the secondary analyses of the FACIT-F scale evaluated the proportion of patients with a clinically meaningful deterioration at a given post-baseline assessment visit and the time to a definitive deterioration.¹⁴ Assessments of the FACIT-F scale occurred on day 1 of each 28-day cycle until discontinuation of the study treatment and at the end-of-treatment visit.¹⁴ A detailed discussion and critical appraisal of the FACIT-F scale is provided in Appendix 4.

The FACIT-F scale is a standardized, self-administered patient questionnaire for evaluating the level of fatigue in an individual during usual daily activities over the past week.¹⁴ The FACIT-F scale consists of a 5-point Likert scale (0 = not at all fatigued to 4 = very much fatigued) and a summary score is calculated based on the FACIT-F scoring guidelines. A higher score corresponds to a lower level of fatigue (scores range from 0 to 52).¹⁴

No study assessing the psychometric properties related to the validity or reliability of the FACIT-F scale in patients with AML was found. One study, which assessed the content validity of the FACIT-F scale in 40 patients in the first line or relapsed or refractory chronic lymphocytic leukemia (CLL) setting in the US, found that interviewed patients understood the terminology and agreed with response options.³³ The same study investigated structural validity, construct validity, known-groups validity, and internal consistency using data from a phase III RCT (ASCEND)³⁴ in 263 patients with relapsed or refractory CLL. An assessment of structural validity using confirmatory factor analysis (CFA) supported the unidimensionality of the FACIT-F scale.³³ Construct validity, as assessed through the convergent approach, indicated overall good validity. Known-groups validity was supported by differences between groups based on various baseline factors. Internal consistency reliability was assessed using Cronbach alpha. The FACIT-F total, symptom, and impact subscales demonstrated good internal consistency.³³ No relevant studies were found to assess responsiveness. An MID was not identified in the literature in patients with AML. The sponsor referenced studies in the published literature that reported MIDs ranging from 3 to 4 points in cancer patients.^35,36 It was noted in the sponsor’s submission that, although the terms MID and responder definition have gradually differentiated over time, these terms are still being used interchangeably in the literature.¹⁴ In consequence, the sponsor used the same cut-off values for the responder definition as for the MID for the FACIT-F scale. In addition, the sponsor presented the entire distribution of responses (cumulative distribution frequency curves of patient experiencing different degrees of change at cycles 3, 6, 12, 18, and 24).¹⁴

Measures of Health Care Resource Utilization

Health care resource utilization was designated an additional secondary outcome. Health care resource use was recorded in specifically designed case report forms adapted from the universal billing (UB) 90 billing form and/or through routine study activities. Resource utilization data collected included, but was not limited to, hospitalizations, medications, clinic visits, medical and diagnostic procedures, and treatment for AEs. Health care resource utilization data were collected until the last dose of study treatment or the date of last study visit, whichever occurred later.¹⁰

Transfusion Independence

This outcome was not assessed in the QUAZAR AML-001 trial.

Safety

The QUAZAR AML-001 trial included assessment of the following safety parameters: TEAEs (summarized by worst severity grade, system organ class, and preferred term), TEAEs leading to death or to discontinuation from treatment, TEAEs leading to dose modification, AEs classified as grade 3 or grade 4 by the National Cancer Institute Common Terminology Criteria for Adverse Events (Version 4.0), AEs related to investigational product, serious AEs, time to discontinuation from the study due to AE, and deaths.¹⁰

All AEs were documented by the investigator from the date a patient signed the informed consent document to 28 days after the last dose of investigational drug or until the date of the last study visit, whichever occurred later.¹⁰

The following parameters were monitored: a patient’s clinical signs and symptoms, laboratory, pathological, radiological or surgical findings, physical examination findings; and results of any other appropriate test or procedure.¹⁰

Statistical Analysis

Details related to statical analyses of efficacy end points are summarized in Table 12.

Sample Size Determination

The trial sample size was determined based on the comparison of the primary end point of OS between the 2 treatment groups, with approximately 460 patients planned for randomization. The data cut-off for the primary analysis of OS was planned for the time point at which 330 deaths were observed across both treatment groups. If at the time of the primary analysis the average OS HR was 0.70, the study would have 90% power to demonstrate a statistically significant difference in OS at a 1-sided alpha of 0.025. No rationale was provided in the sponsor’s submission for assuming an OS HR of 0.70. An assumption of exponential OS distribution with a constant hazard rate was made with median OS values of 22.9 months and 16 months for the oral azacitidine and placebo groups, respectively.¹⁰ No rationale was provided in the sponsor’s submission for assuming a median OS of 22.9 months. However, with regards to the assumption of a median OS of 16 months, the sponsor’s submission referenced 2 studies: Baer et al. (2011)³⁷ reported outcomes for 610 patients aged 60 years or older with untreated de novo and secondary AML who received induction intensive chemotherapy, and Baer et al. (2010)³⁸ provided a historical comparison of patients aged 60 or older with de novo AML who either received 1 course or multiple courses of post-remission consolidation chemotherapy.

Interim Analysis

An interim analysis for futility after 30% of the events (99 deaths) had occurred was planned. The pre-specified futility boundary of a z value of less than −1.9796, using gamma(−10) to correspond to a conditional power level of approximately 18%, was calculated using a beta-spending function.⁹ At the data cut-off date of December 31, 2015, at which point 104 deaths had occurred, the interim analysis was conducted. The data monitoring committee reviewed unblinded results²⁷ and determined that the HR of oral azacitidine versus placebo remained within the pre-specified futility boundary (the z score was 3.4883 and the P value was 0.0538 based on stratified log-rank test)²⁷ and that no safety concerns were observed.¹⁰ Therefore, the data monitoring committee recommended that the study could proceed as planned.¹⁰ As the purpose of the interim analysis was to stop the trial for futility, no alpha adjustment was applied in the final analysis.²⁷

Primary Outcome

The difference in OS between the 2 study groups was compared using a log-rank test to assess superiority of oral azacitidine, at a 5% (2-sided) significance level and stratified by age at the time of induction therapy, prior history of MDS, cytogenetic risk category, and receipt of consolidation therapy following induction therapy. If the minimum number of patients in a stratum was less than 16, then only the 3 stratification variables that resulted in the largest minimum stratum size were used for analysis purposes. The HRs and 95% CIs were estimated from a stratified Cox proportional hazards regression model. The proportional hazards assumption was assessed by a time-dependent Cox model with interaction of treatment and time. A Kaplan–Meier (KM) plot of OS was presented by treatment group, unadjusted for stratification variables. Medians and 25th and 75th percentiles with corresponding 95% CIs as well as the numerical difference (and 95% CI difference) between groups were presented. KM estimates of probabilities and corresponding 95% CIs at several time points (i.e., 1 year and 2 year) were also reported. Reasons for censoring included losses to follow-up, withdrawal of consent, or alive at the data cut-off date.¹⁰

Several sensitivity analyses for OS were planned as follows:

• Investigation of potentially confounding effects of subsequent therapies¹⁰:

  • censoring for the use of any subsequent therapy (including post-treatment transplantation) for AML

  • censoring for the use of disease-modifying subsequent AML therapy, defined as any subsequent AML therapy that is not hydroxycarbamide

  • censoring for post-treatment transplantation

  • a Cox proportional hazard model with adjustment for covariates that included, but was not limited to, subsequent AML therapy as a time-varying covariate and treatment-by-subsequent AML therapy interaction, and baseline characteristics (such as age, ECOG score, cytogenetic risk status, and CR or CRi status at randomization)

  • inverse probability of censoring weighted (IPCW) method

  • regression-based imputation analysis method that allows for inferences about the treatment effect in the presence of confounding due to additional therapy received subsequent to the study treatment.

• Additional analyses were planned if the proportional hazards assumption was violated. As the proportional hazard assumption did not hold, the following analyses were performed:

  • restricted mean survival time

  • a piecewise Cox proportional hazard model with interval time points of 3 or less, 3 to 6, 6 to 12, and greater than 12 months

  • a generalized Wilcoxon test in the presence of nonproportional and early OS separation

• Not censoring for withdrawal of consent

• Primary analysis on a modified intention-to-treat population (mITT) population (defined in the section that follows).

Secondary Outcomes

Relapse-free survival, the key secondary outcome in the QUAZAR AML-001 trial, was planned to be analyzed in the same manner as OS. If superiority of oral azacitidine was demonstrated for OS at the 2-sided 0.05 significance level, then RFS was formally tested for statistical significance in the primary analysis. The P value from the stratified log-rank test was the confirmatory P value. This hierarchical testing scheme ensured that the overall alpha level at 0.05 across the OS and RFS end points was preserved.¹⁰ Reasons for censoring included no documented relapse and still alive, lost to follow-up, or withdrawal of consent without documented relapse. A date of documented relapse was censored if subsequent therapy for AML was initiated before the date of relapse or if the interval between the date of relapse and the previous bone marrow assessment date or the randomizing date (if no post-baseline bone marrow assessment), was more than 200 days.

Several sensitivity analyses for RFS were planned as follows¹⁰:

• Primary analysis was performed on the mITT population (defined in the section that follows)

• The RFS definition was replaced with the documented relapse based on investigator-assessed response

• Censoring was based on EMA guidance, with the following censoring rules²⁶:

  • for patients who experienced documented relapse or death, define outcome as event and event date as the earliest of documented relapse or death

  • for patients without documented relapse or death, define outcome as censoring and event date as the date of last bone marrow assessment or date of randomization if no post-baseline bone marrow assessment.

Time to relapse was a descriptive end point that was calculated using a competing risk analysis whereby death without documented relapse was considered a competing risk for relapse. The median time to relapse was calculated using an unstratified KM analysis. A cumulative incidence function, as described by Kalbfleisch and Prentice (1980),³⁹ was calculated and displayed graphically for each treatment group. Censoring rules similar to those used for RFS were applied.¹⁰

Time to discontinuation from treatment was a descriptive end point that was calculated using a competing risk model with the following competing events¹⁰:

• disease relapse

• adverse event(s)

• eligibility for bone marrow or stem cell transplantation

• withdrawal of consent, lost to follow-up, or protocol violation or other event

• death.

Cumulative incidence functions for each specific reason for discontinuation from treatment were calculated and displayed graphically for each study group.¹⁰ The median time to discontinuation was calculated based on an unstratified KM analysis. The CI for the difference in median times between treatment groups was obtained using the Kosorok method. Treatment discontinuation rates were estimated using KM methods. The CIs for the difference in treatment discontinuation rates were obtained using a Greenwood variance estimate.⁹ The reason for censoring is ongoing in treatment at the time of study closure.¹⁰

All analyses performed on the HRQoL outcomes (EQ-5D-3L and PINR scale) and symptom severity outcome (FACIT-F scale) were based on the July 15, 2019, data cut-off, considered descriptive (i.e., noninferential) in nature, and unadjusted for multiplicity.^9,14 The HRQoL-evaluable population (defined as all ITT patients who had a valid [i.e., non-missing] assessment at baseline and at least 1 valid post-baseline assessment) was used to analyze all HRQoL and symptom severity outcomes, with the exception of the calculation of adherence and completion rates, which were based on the ITT population. Statistical tests were conducted at a 2-sided 0.05 level of significance and noninferential P values were presented.¹⁴ Point estimates of treatment difference and associated 95% CIs were presented for all end points.¹⁴ Missing data at a given assessment visit were not imputed. Missing items for a measure at a given assessment visit were handled according to the scoring guidance of the respective measure.¹⁴

In the HRQoL-evaluable population, descriptive statistics (mean SD, median, first quartile, third quartile, minimum, and maximum) were used to summarize observed scores at each assessment visit and the change from baseline.¹⁴ For the FACIT-F scale, the EQ-5D-3L instrument (health utility index and EQ VAS) and each of 3 PINR questions, within-group changes from baseline at each post-baseline assessment visit were tested using paired 1-sample t-tests, and between-group differences in changes from baseline were tested using independent 2-sample t-tests.¹⁴ Additionally, ANCOVA models were used to assess change from baseline to each post-baseline assessment visit, including covariates (treatment, baseline score, and trial randomization stratification factors). The treatment differences between treatment groups were evaluated using differences in LS mean changes (95% CIs and P values). Plots were generated for each treatment group showing the mean changes (and standard errors) from baseline across all post-baseline assessment visits that had 25 patients or more in each treatment group.¹⁴ Furthermore, an MMRM analysis was used to analyze the FACIT-F scale, the EQ-5D-3L instrument (health utility index and EQ VAS), and each of 3 PINR questions using the HRQoL-evaluable population but excluding the end-of-treatment visit. The MMRM analysis controlled for the baseline HRQoL score and other pre-selected covariates of interest. The MMRM analysis was used to assess the overall LS means for change from baseline at each post-baseline assessment visit as well as to calculate the treatment differences between treatment groups using differences in LS mean change (95% CI and P value) at each post-baseline assessment visit. Mixed-effect models with a random intercept or slope were used to assess treatment differences in change scores over time between treatment groups.

To determine if the between-group difference was considered clinically meaningful in the MMRM analysis, the sponsor defined the following MIDs⁹:

• FACIT-F scale: change from baseline of 3 points

• EQ-5D-3L health utility index: change from baseline of 0.10 points

• EQ VAS: change from baseline of 11 points

• PINR scale, physical impairment: change from baseline of 1.12

• PINR scale, difficulty completing outdoor physical tasks: change from baseline of 1.59

• PINR scale, difficulty completing indoor tasks: change from baseline of 1.28.

As the MID for the PINR scale has not been described in the literature, the sponsor used half of a SD (0.5 × SD) at baseline from all ITT patients as an approximation for the MID.¹⁴

To compare the proportions of patients with clinically meaningful change between treatment groups, Cochran-Mantel-Haenszel tests were performed (an odds ratio and a 95% CI with an associated P value were provided). This analysis was performed for the EQ-5D-3L and FACTI-F scale but not for the PINR scale. The sponsor defined a change of 3 points or greater as a clinically meaningful change at the individual level for the FACIT-F scale. For the EQ-5D-3L, a change of 0.10 points or greater from baseline was defined as a clinically meaningful improvement, whereas a change of 0.08 points or greater was used to define clinically meaningful worsening. A change of 11 points or more from baseline was used to define clinically meaningful change in the EQ VAS.¹⁴ Specifically, each patient was categorized into 3 response levels at each post-baseline visit: improvement (i.e., change from baseline ≥ RD), no change (i.e., −RD < change from baseline < RD), and deterioration (i.e., change from baseline ≤ −RD).

Time to definitive deterioration (defined as having 2 consecutive clinically meaningful deteriorations) was assessed using the KM product-limited method to estimate survival distribution functions. For the median time to deterioration, a 2-sided 95% CI was estimated. Treatment groups were compared using a non-stratified log-rank test. The Cox proportional hazards regression analysis (adjusting for treatment group [oral azacitidine versus placebo], baseline score, and stratification factors) was used to calculate HRs, 95% CIs and P values. Patients who never experienced a clinically meaningful deterioration were censored at the time of the last assessment visit. Sensitivity analysis was performed with death as an event.¹⁴

Measures of Health Care Resource Utilization

Summaries of hospitalizations were specified a priori in the statistical analysis plan and included number of events, total number of days hospitalized, rate of events, days of hospitalized per person-year of exposure, and associated relative risk of hospitalization with the corresponding 95% CI.¹⁰ How missing data were handled was not reported.

Subgroup Analyses

For each subgroup, the HR and 95% CI were calculated from an unstratified Cox proportional hazards model.¹⁰ Subgroup analyses were planned a priori in the statistical analysis plan for the following groups of patients¹⁰:

• age at induction therapy (< 65, ≥ 65, ≥ 75 years)

• sex (male, female)

• race (White, Asian, Black, or other)

• CR or CRi status at randomization (CR, CRi)

• CR or CRi status at first achieving response (CR, CRi)

• CR or CRi status at randomization and use of consolidation (CR with consolidation, CR without consolidation, CRi with consolidation, and CRi without consolidation)

• prior history of MDS or CMML (yes, no)

• cytogenetic risk category at induction therapy (intermediate, poor)

• MRD status at randomization (positive, negative)

• CR or CRi status at randomization and MRD status at randomization (CR with MRD-positive, CR with MRD-negative, CRi with MRD-positive, and CRi with MRD-negative)

• consolidation therapy following induction (yes, no)

• consolidation therapy following induction (1 or 2 cycles, 3 or 4 cycles)

• geographic region (North America, Europe, Asia, Australia)

• ECOG PS (0 or 1, 2 or 3)

• WHO AML classification (AML with myelodysplasia-related changes, AML with recurrent genetic abnormalities, AML not otherwise specified)

• types of first-line subsequent therapy:

  • high-intensity and low-intensity chemotherapy

  • hypomethylating drug monotherapy and non-hypomethylating subsequent therapy

  • injectable azacitidine monotherapy and other subsequent therapy (excluding decitabine monotherapy).

In addition, OS and RFS were analyzed for the subset of patients whose doses were escalated to 300 mg × 21 days per 28-day cycle to explore the efficacy of oral azacitidine in this patient group. This subgroup analysis was also planned a priori in the statistical analysis plan.

The following subgroups, planned a priori in the statistical analyses plan, aligned with the subgroups prespecified in the protocol for this CADTH review: age, ECOG PS, prior history of MDS or CMML, cytogenetic risk, consolidation therapy following induction (yes, no), consolidation therapy following induction (1 or 2 cycles, 3 or 4 cycles), CR or CRi status at randomization (CR, CRi), MRD status at randomization. Only those subgroups identified in the CADTH review protocol are reported in the efficacy section below.

Multiplicity

Apart from the prespecified hierarchical testing of OS and RFS, no adjustments for multiplicity were performed to control the type I error rate.¹⁰

Dose Escalation

A post-relapse dose escalation (300 mg once daily from 14 days to 21 days of the 28-day cycle) was planned for all patients with disease relapse with blasts of between 5% and 15%, in either the peripheral blood or bone marrow.¹⁰ Blinding was upheld until study closure; therefore, all patients, irrespective of the study group they belonged to, could be selected for dose escalation. Because dose escalation was only implemented post-relapse (relapse was defined in the trial as the appearance of relapse was defined in the trial as the appearance of blast counts of ≥ 5% in the peripheral blood or bone marrow) the key secondary outcomes of RFS and the secondary outcome of time to relapse were not affected by dose escalation. However, the primary analysis of OS included all patients in the ITT population, whether they had received dose escalation or not. The OS outcome and the secondary outcome of time to discontinuation from treatment were not adjusted for the impact of patients who received a dose escalation. Subgroup analyses for OS and RFS were planned a priori according to the statistical analysis plan for the subset of patients who had experienced dose escalation in each treatment group. In addition, overall response and best response (using the IWG’s response criterion for AML) were planned a priori to be assessed in each treatment group in patients who experienced dose escalation.¹⁰ Overall response was defined as CR plus CRi plus partial response, and best response categories included CR, CRi, and partial response.⁹ Safety in the subset of patients who received a dose escalation was also assessed.^9,10

Amendments

The study protocol of the QUAZAR AML-001 trial was amended twice (Amendment 1: December 29, 2015; Amendment 2 on November 8, 2018).¹⁰ Amendment 1 included changes to inclusion criteria and the schedule of examinations. Amendment 2 implemented an extension phase in which patients who were receiving oral azacitidine and demonstrated clinical benefit after the end of study were allowed to continue receiving oral azacitidine until they met the study discontinuation criteria or oral azacitidine became commercially available. As well, patients who were discontinued from study treatment (irrespective of randomization group) and continued in the follow-up phase were followed up for survival for at least another 12 months, until death, withdrawal of consent, or study closure, or they were lost to follow-up.⁹ Data from the extension phase are not included in this review.

Analysis Populations

The primary and secondary efficacy end points were analyzed using the ITT population as defined in Table 13.¹⁰ The EQ-5D-3L instrument, the PINR scale, and the FACIT-F scale were analyzed using the HRQoL-evaluable population as defined in Table 13. In sensitivity analyses, OS and RFS were analyzed using the mITT population as defined in Table 13.¹⁴ Analyses of safety were performed using the safety population (Table 13).¹⁰

Results

Patient Disposition

Details of patient disposition in the QUAZAR AML-001 trial are summarized in Table 14. A total of 555 patients were screened and of those 85.0% (N = 472) were randomized to receive oral azacitidine and BSC (N = 238) or placebo and BSC (N = 234). All patients who were not randomized failed to meet the trial eligibility criteria. A total of 469 randomized patients (84.5%) were treated: 236 of 238 (99.2%) in the oral azacitidine group and 233 of 234 (99.6%) in the placebo group. In the oral azacitidine group, 193 patients (81.1%) had discontinued treatment as of the final analysis data cut-off date. In the placebo group, 208 patients (88.9%) had discontinued placebo as of the final analysis.⁹

As of the final analysis, 347 patients (74%) had terminated the study: 167 patients (70.2%) in the azacitidine group and 180 patients (76.9%) in the placebo group. The main reason for termination was death, and the percentages of patients who died were 66.0% in the oral azacitidine group and 73.1% in the placebo group. In the oral azacitidine group, 10 patients (4.2%) discontinued the study due to withdrawal of consent and no patient discontinued due to being lost to follow-up. In the placebo group, 8 patients (3.4%) discontinued due to withdrawal of consent and 1 (0.4%) was lost to follow-up. At the time of the final data cut-off date, 18.9% of patients in the oral azacitidine group and 11.1% of patients in the placebo group were ongoing on study treatment, respectively.⁹

Protocol violations were reported for |||||||||| patients randomized to the oral azacitidine group and |||||||||| patients in the placebo group. Patients may have had more than 1 violation. The most commonly reported violations (oral azacitidine versus placebo) included |||||||||||||||||||||||||||||||||| |||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||| (|||| |||||| ||||), |||||||||||||||||||||||||||||||||||||||||||| 　|　|||||||||||||||||||||| No violations that substantially affected the internal validity of the study were reported.

Exposure to Study Treatments

Exposure to oral azacitidine and placebo as of the final analysis is summarized in Table 15. The median duration of treatment with oral azacitidine of 11.6 months (range = 0.5 to 74.3) was close to twice that of the treatment duration with placebo of 5.7 months (range = 0.7 to 68.5). The median dose intensity and median relative dose intensity of treatments were similar between groups, indicating high treatment adherence in both study groups. The median dose intensity (Table 15) for azacitidine was |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||, compared to |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| with placebo. The median relative dose intensities (Table 15) for oral azacitidine and placebo were |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| respectively. The median overall adherence (Table 15) was |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||| in both study groups.^9,10

Dose Modifications (Interruption, Reduction, Escalation)

As of the final analysis (July 15, 2019, data cut-off date), 102 patients (43.2%) in the oral azacitidine group and 40 patients (17.2%) in the placebo group of the safety population required a study treatment interruption due to a TEAE. The most common TEAE leading to dose interruption was neutropenia.¹⁰ More patients in the oral azacitidine group compared with the placebo group required dose reductions. One or 2 dose reductions, respectively, were experienced by |||||||||||||||||| in the oral azacitidine group and |||||||||||||||| in the placebo group. A higher proportion of patients in the oral azacitidine group compared with the placebo group experienced a dose reduction due to a TEAE (n = 37 [15.7%] and n = 6 [2.6%] in the oral azacitidine and placebo groups, respectively). The most common TEAE leading to dose reduction was neutropenia.¹⁰ Information regarding dose delays was not reported and not available at this time as confirmed with the sponsor.¹²

More dose escalations were reported for the oral azacitidine group compared to the placebo group in the ITT population (N = 51 [21.4%] and N = 40 [17.1%] in the oral azacitidine and placebo groups, respectively). The most common reason for dose escalation was |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||⁹

Concomitant Medication

Concomitant medications, as of the July 15, 2021, data cut-off date, were generally similarly administered across treatment groups and reported for |||||||||||||||||||||||||||||||| in the ITT population. Medications for |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| were those most commonly used. The most frequently reported concomitant medications included |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||⁹

Subsequent Treatments

The proportion of patients receiving subsequent AML therapies was lower in the oral azacitidine group compared with the placebo group (Table 16). At least 1 subsequent therapy was received by 137 patients (57.6%) in the oral azacitidine group compared to 170 patients (72.6%) in the placebo group. The percentage of patients who received subsequent intensive chemotherapy was 29.0% in the oral azacitidine group compared to 37.6% in the placebo group. The most common subsequent systemic treatments (reported in > 10% in either study group; oral azacitidine versus placebo) received by patients included |||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||⁹

The proportion of patients receiving subsequent transplantation was 6.3% in the oral azacitidine group compared with 13.7% the placebo group (Table 16).⁹

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported below. Refer to Appendix 3 for detailed efficacy data.

Overall Survival

The OS results of the QUAZAR AML-001 trial for the oral azacitidine and the placebo groups are summarized in Table 17.

As of the final analysis (July 15, 2019, data cut-off date), 329 death events occurred across both study groups. The median duration of follow-up for OS for all patients was 41.2 months, ||| months for patients in the oral azacitidine group, and ||| months for those in the placebo group. The median OS was 24.7 months (95% CI, 18.7 to 30.5) in the oral azacitidine group compared to 14.8 months (95% CI, 11.7 to 17.6) in the placebo group, with a stratified HR of 0.69 (95% CI, 0.55 to 0.86; P = 0.0009) in favour of the oral azacitidine group. The KM curves are depicted in Figure 3. The survival curves converged at the end of the follow-up period. The survival probabilities of patients surviving to 12 and 24 months were 0.73 (95% CI, 0.66 to 0.78) and 0.51 (95% CI, 0.44 to 0.57), respectively, in the oral azacitidine group, and 0.56 (95% CI, 0.49 to 0.62) and 0.37 (95% CI, 0.31 to 0.43), respectively, in the placebo group. The sensitivity analyses of OS using the mITT and ITT populations produced consistent results. The sensitivity analysis performed to assess the impact of censoring patients who withdrew consent from survival follow-up was also consistent with the OS results for the primary analysis.⁹

Graphical assessment and testing of treatment-by-time interactions in the overall Cox proportional hazards model suggested deviation from the proportional hazard assumption. The results of pre-specified sensitivity analyses for OS (using a piecewise Cox model, general Wilcoxon test, and restricted mean survival time) that were planned to be conducted if the proportional hazards assumption was violated suggested the robustness and consistency of the primary OS results (Table 25, Table 26, Table 27, and Table 28 in Appendix 3).

Figure 3: [Figure Redacted]

Note: This figure has been redacted as per sponsor’s request.

Planned sensitivity analyses, which censored for the use of subsequent therapy or subsequent transplantation, were conducted to assess potential confounding effects on the survival analyses from subsequent treatments (Table 18). The results of these sensitivity analyses were consistent with the primary OS results. Given that censoring for patients in each treatment group was greater than 80% in the analyses censoring for subsequent therapy, the ability to interpret these sensitive analyses remains limited. Additional sensitivity analyses for OS (a Cox proportional hazard model with covariates adjustment, IPCW method, and regression-based imputation analysis) were performed to further assess potential biases from subsequent treatments. The results of these sensitivity analyses also suggested consistency with the OS results for the ITT population (Table 29, Table 30, Table 31, Table 32, and Table 33 in Appendix 3).

The OS results by subgroups of interest, as specified a priori in the protocol for this CADTH review, are summarized in Figure 4.

The treatment effect on OS was consistent with the primary analysis across patient subgroups, except for the subgroup of 3 or 4 cycles of consolidation therapy. The sample size of this subgroup was small (6 patients in the oral azacitidine group and 13 patients in the placebo group). Several other subgroups (≥ 75, prior history of MDS or CMML [yes], poor cytogenetic risk status at induction, and CRi at randomization) had relatively small sample sizes (< 50 patients in either group).⁹ The wide CIs in subgroups reflected uncertainty in the effect estimates.

Figure 4: Forest Plot of OS by Prespecified Subgroups — ITT Population (July 15, 2019, Data Cut-Off Date)

CC-486 = oral azacitidine; CI = confidence interval; CMML = chronic myelomonocytic leukemia; CR = complete remission; CRi = complete remission with incomplete blood count recovery; ECOG = Eastern Cooperative Oncology Group; HR = hazard ratio; ITT = intention-to-treat; MDS = myelodysplastic syndromes; MRD = minimal residual disease; OS = overall survival.

^a Number of events per number of patients.

^b Median OS in months.

Source: Clinical Study Report.⁹

Exploratory OS results in the subset of patients who had dose escalation are summarized in Table 35 in Appendix 3.

Relapse-Free Survival

The RFS results of the QUAZAR AML-001 trial for the oral azacitidine and the placebo groups are summarized in Table 17. Median RFS was longer in the oral azacitidine group compared with the placebo group. As of the final analysis (July 15, 2019, data cut-off date), median RFS was 10.2 (95% CI, 7.9 to 12.9) months in the oral azacitidine group and 4.8 (95% CI, 4.6 to 6.4) months in the placebo group, with a stratified HR of 0.65 (95% CI, 0.52 to 0.81; P = 0.0001) in favour of the oral azacitidine group. The KM curves are depicted in Figure 5. The survival curves converged at the end of the follow-up period. The probabilities of RFS at 12 and 24 months were 0.53 (95% CI, 0.46 to 0.59) and |||||||||||||||||||||||||||||||||| respectively, in the oral azacitidine group, and 0.72 (95% CI, 0.65 to 0.77) and 0.81 (95% CI, 0.76 to 0.86), respectively, in the placebo group. The results of the RFS sensitivity analysis using the mITT population were consistent with the RFS results for the ITT population.⁹

Figure 5: [Figure Redacted]

||||||||||||||||||||||||||||||||||||||||||||||||||　|　⁹

Note: This figure has been redacted at the sponsor’s request.

Planned sensitivity analyses, using censoring rules based on EMA guidance, which accounts for all events regardless of when they occurred, were consistent with the primary analysis of RFS. The median RFS values were 10.2 months for the oral azacitidine group and 4.8 months for the placebo group, with a stratified HR of 0.66 (95% CI, 0.54 to 0.81; P = 0.0001). For this analysis, |||| of patients in the oral azacitidine group and |||| of patients in the placebo group were censored.⁹

In addition, a sensitivity analysis was conducted using investigator-assessed response according to the IWG’s AML criteria. The results for this analysis were consistent with the primary analysis of RFS. Median RFS was |||||||||||||||||||||||||||||||| months in the oral azacitidine group and |||||||||||||||||||||||||||| months in the placebo group, with a stratified HR of |||||||||||||||||||||||||||||||||||||| |||||| in favour of the oral azacitidine group.⁹

Exploratory RFS results in the subset of patients who had dose escalation are summarized in Table 35 in Appendix 3.

Time to Relapse From CR or CRi

The results for time to relapse for oral azacitidine and the placebo groups are summarized in Table 17.

As of the final analysis (July 15, 2019, data cut-off date), the percentage of patients who had relapsed was |||| in the oral azacitidine group compared with |||| in the placebo group. Median time to relapse was |||||||||||||||||||||||||||||| months in the oral azacitidine group and |||||||||||||||||||||||||||| months in the placebo group. The 1-year and 2-year relapse rates were |||||||||||||| |||||||||||||| and |||||||||||||||||||||||||||||||||| respectively, in the oral azacitidine group and |||||||||||||||||||||||||||||||| and |||||||||| |||||||||||||||||| respectively, in the placebo group. These reported relapse rates were based on the cumulative incidence function from a competing risk analysis with death as a competing risk for relapse from CR or CRi. Figure 11 in Appendix 3 provides the cumulative incidence distribution of relapse for each treatment group.⁹

Time to Discontinuation From Treatment

The results for time to discontinuation from treatment for the oral azacitidine and the placebo groups are summarized in Table 17.

As of the final analysis (July 15, 2019, data cut-off date), most patients in both treatment groups had discontinued study treatment (81.1% in the oral azacitidine group and 88.9% in the placebo group). The median time to treatment discontinuation was 11.4 (95% CI, 9.8 to 13.6) months in the oral azacitidine group and 6.1 (95% CI, 5.1 to 7.4) months in the placebo group. The 1-year treatment discontinuation rate was |||||||||||||||||||||||||||||||| in the oral azacitidine group and |||||||||||||||||||||||||||||||| in the placebo group.⁹

Results for time to discontinuation from treatment were also reported using competing risks. As of the final analysis (July 15, 2019, data cut-off date) 143 patients (60.1%) in the oral azacitidine group and 180 patients (76.9%) in the placebo group had discontinued treatment due to disease relapse. Additional reasons for treatment discontinuation, representing the competing risks, are described in Table 34 in Appendix 3. The median time to treatment discontinuation due to disease relapse was 14.6 months (95% CI, 11.3 to 20.1) in the oral azacitidine group and 6.9 months (95% CI, 5.3 to 7.9) in the placebo group (Appendix 3).⁹

Health-Related Quality of Life

EQ-5D-3L

Adherence rates for the EQ-5D-3L were highest at baseline (> 95%) and similar between the oral azacitidine and the placebo groups over all post-baseline assessments (> 85%), except for the end-of-treatment visit (|||||||||||||||||||| in the oral azacitidine group and placebo groups, respectively). However, the completion rates declined over time in both study groups. After cycle 12 less than 50% of patients were still available for completion in the oral azacitidine group while the number of patients available for completion declined more rapidly in the placebo group, with less than 50% of patients available beyond cycle 6.¹⁴

The baseline characteristics based on the HRQoL-evaluable population were balanced between study groups (see Table 36 in Appendix 3 for baseline characteristics). Baseline scores were balanced between treatment groups (Table 19).⁹

The observed mean changes from baseline for the EQ-5D-3L health utility index and the EQ VAS are summarized in Figure 6 and Figure 7, respectively.

Analyses performed on the HRQoL outcomes were noninferential. Overall, no statistically significant or clinically meaningful differences were observed between the oral azacitidine group and the placebo groups based on the mean changes from baseline at any post-baseline assessment for the EQ-5D-3L health utility index and EQ VAS. The results generated by the ANCOVA models were consistent with the summary statistics of observed scores. The MMRM analyses showed results similar to those from the ANCOVA models, with no statistically significant or clinically meaningful differences observed in LS mean changes between treatment groups at any assessment time point for the EQ-5D-3L health utility index and EQ VAS.

The proportions of patients experiencing clinically meaningful deterioration over time by treatment group are summarized in Figure 9.

Overall, no statistically significant differences were observed in the proportion of patients with clinically meaningful deterioration between the oral azacitidine and placebo groups for the EQ-5D-3L health utility index and EQ VAS across all post-baseline visits, except for 1 visit (cycle 19, day 1) for which statistical significance was detected. However, the sponsor’s submission noted that, given the lack of adjusting for multiple testing, the statistical significance may be due to chance.¹⁴

Times to definitive deterioration for the EQ-5D-3L health utility index and EQ VAS are summarized in Figure 10. The time to definitive deterioration was similar between the oral azacitidine and the placebo groups. For the EQ-5D-3L health utility index and EQ VAS, respectively, the median times to definitive deterioration were |||| weeks in the oral azacitidine group versus |||| weeks for the placebo group (||||||||||||||||||||||||||||||||||||||||||||||||||||||) and “not yet reached” for the oral azacitidine group versus |||| weeks for the placebo group (||||||||||||||||||||||||||||||||||||||||||||||||||||). Sensitivity analyses in which deaths were censored showed similar results. Given that both treatment groups had greater than 65% and greater than 70% censored patients in the time to definitive deterioration analysis and its sensitivity analysis, respectively, the ability to interpret the time to definitive deterioration analyses remains limited.

PINR

Adherence and completion rates for the PINR scale were not reported.¹⁴ Overall, baseline characteristics based on the HRQoL-evaluable population were balanced between study groups (Table 36 in Appendix 3 provides baseline characteristics). Baseline scores were balanced between treatment groups (Table 19).

The summary statistics of observed mean changes from baseline for the FACIT-F scale are summarized in Figure 8. Overall, no statistically significant or clinically meaningful differences were observed between the oral azacitidine group and the placebo group based on the mean changes from baseline at any post-baseline assessment for the PINR scale (i.e., level of physical impairment, level of difficulty completing outdoor physical tasks, and level of difficulty completing indoor tasks). The results produced by the ANCOVA models were consistent with the summary statistics of observed scores. The MMRM analyses showed results similar to those of the ANCOVA models, with overall no statistically significant or clinically meaningful differences observed in LS mean changes between treatment groups at any assessment time point for the PINR scale.

The proportion of patients experiencing clinically meaningful deterioration and time to definitive deterioration were not reported for the PINR scale.¹⁴

Symptom Severity

FACIT-F Scale

Adherence rates for the FACIT-F scale were highest at baseline (> 95%) and similar between the oral azacitidine and placebo groups over all post-baseline assessments (> 85%), except for the end-of-treatment visit (69.6% and 63.3% in the oral azacitidine group and placebo groups, respectively). However, the completion rates declined over time in both study groups. After cycle 12 less than 50% of patients were still available for completion in the oral azacitidine group while the number of patients available for completion declined more rapidly in the placebo group, with less than 50% of patients available beyond cycle 6.

Overall, baseline characteristics based on the HRQoL-evaluable population were balanced between study groups (see Table 36 in Appendix 3 for baseline characteristics). Baseline scores were balanced between treatment groups (Table 19).

The summary statistics of observed mean changes from baseline for the FACIT-F scale are summarized in Figure 6. Analyses performed were noninferential. Overall, no statistically significant or clinically meaningful differences were observed between the oral azacitidine group and the placebo group based on the mean changes from baseline at any post-baseline assessment for the FACIT-F scale. The results using the ANCOVA models were consistent with the summary statistics of observed scores. The MMRM analyses showed results similar to those of the ANCOVA models; overall, no statistically significant or clinically meaningful differences were observed in LS mean changes between treatment groups at any assessment time point for the FACIT-F scale.

The proportions of patients experiencing clinically meaningful deterioration over time by treatment group are summarized in Figure 9. Overall, no statistically significant differences were observed in the proportion of patients with clinically meaningful deterioration between the oral azacitidine and placebo groups for the FACIT-F scale across all post-baseline visits, except for 1 visit (cycle 29, day 1) for which statistical significance was detected. However, the sponsor’s submission noted that, given the lack of adjusting for multiple testing, the statistical significance found may be due to chance.¹⁴

Time to definitive deterioration for the FACIT-F scale is summarized in Figure 10. The time to definitive deterioration was similar between the oral azacitidine and the placebo groups. The median time to definitive deterioration was ||| weeks for the oral azacitidine group and ||| weeks for the placebo group (||||||||||||||||||||||||||||||||||||||||||||||||||||).¹⁴ Sensitivity analyses in which deaths were censored showed similar results. Given that in both treatment groups there were more than 47% and more than 48% of censored patients in the time to definitive deterioration analysis and its sensitivity analysis, respectively, the ability to interpret the time to definitive deterioration analyses remains limited.

Figure 6: Results of HRQoL Analyses: Observed Mean Change From Baseline by Visit and Treatment Group — HRQoL-Evaluable Population^a

CC-486 = oral azacitidine; EQ-5D-3L = EuroQol 5-Dimensions 3-Levels questionnaire; FACIT = Functional Assessment of Chronic Illness Therapy; HRQoL = health-related quality of life; MID = minimally important difference; SE = standard error.

Note: Data are reported up to the cycle with an n of at least 25 in both treatment groups. A positive change indicates an improvement from baseline for FACIT-Fatigue and EQ-5D-3L.

^a Intention-to-treat patients with a valid (i.e., non-missing) HRQoL assessment at baseline and at least 1 valid post-baseline assessment.

* Indicates a statistically significant difference at P < 0.05 in mean change from baseline between treatment groups at a specific post-baseline assessment visit based on a 2-sample, 2-sided t-test.

Source: Clinical Study Report.⁹

Figure 7: Observed Mean Change From Baseline in EQ VAS by Visit and Treatment Group — HRQoL-Evaluable Population (July 15, 2019, Data Cut-Off Date)

C = cycle; CC-486 = oral azacitidine; D = day; EQ VAS = EuroQol Visual Analogue Scale; HRQoL = health-related quality of life; SE = standard error.

Note: HRQoL-evaluable population were ITT patients with a non-missing assessment at baseline and at least 1 post-baseline visit based on the EQ VAS. Data are reported up to the cycle with an n or least 25 in both treatment groups. A positive change indicated an improvement from baseline. Dashed lines indicate clinically meaningful thresholds of improvement and worsening at the group level for the EQ VAS.

Source: Clinical Study Report.¹⁴

Figure 8: Observed Mean Change From Baseline in PINR Levels by Visit and Treatment Group — HRQoL-Evaluable Population (July 15, 2019, Data Cut-Off Date)

C = cycle; CC-486 = oral azacitidine; D = day; HRQoL = health-related quality of life; ITT = intention-to-treat; MID = minimally important difference; PINR = Physical Impairment Numeric Rating; SE = standard error.

Note: HRQoL-evaluable population were ITT patients with a non-missing assessment at baseline and at least 1 post-baseline visit based on the PINR item. Data were reported up to the cycle with an n of at least 25 in both treatment groups. A positive change indicated a deterioration from baseline. Dashed lines indicate clinically meaningful thresholds of improvement and worsening at the group level for the PINR item, defined as 0.5 × standard deviation.

* Indicates a statistically significant difference at P < 0.05 in mean change from baseline between treatment groups at a specific post-baseline assessment visit based on a 2-sample, 2-sided t-test.

Source: Clinical Study Report.¹⁴

Figure 9: Proportion of Patients Experiencing Clinically Meaningful Deterioration Over Time by Treatment Group — HRQoL-Evaluable Population (July 15, 2019, Data Cut-Off Date)

C = cycle; CC-486 = oral azacitidine; D = day; EQ-5D-3L = EuroQol 5-Dimensions 3-Levels; EQ-5D VAS = EuroQol 5-Dimensions Visual Analogue Scale. FACIT = Functional Assessment of Chronic Illness Therapy; HRQoL = health-related quality of life; OR = odds ratio.

Note: HRQoL-evaluable population comprised ITT patients with a non-missing assessment at baseline and at least 1 post-baseline visit based on the HRQoL measure. Data were reported up to the cycle with an n of at least 25 in both treatment groups. A clinically meaningful deterioration was defined as at least 3, 0.10, and 11 points worsening for the FACIT-Fatigue, EQ-5D-3L health utility index, and EQ-5D VAS, respectively. The denominator (N) at a given visit was based on the number of the HRQoL-evaluable patients who had valid non-missing scores at that visit. The ORs (and P value) were estimated using Cochran-Mantel-Haenszel tests, stratified by randomization stratification factors, to compare the odds of experiencing clinically meaningful deterioration between CC-486 versus placebo.

* Indicates a statistically significant difference at P < 0.05 in the proportion of patients with a deterioration between treatment groups (CC-486 versus placebo) at a specific post-baseline assessment visit based on a Cochran-Mantel-Haenszel test, stratified by randomization stratification factors

Source: Clinical Study Report.¹⁴

Figure 10: Kaplan-Meier Plot of Time to Definitive Deterioration in HRQoL Measures When Treating Death as an Event — HRQoL-Evaluable Population (Data Cut-Off Date: July 15, 2019)

CI = confidence interval; CC-486 = oral azacitidine; EQ-5D-3L = EuroQol 5-Dimensions 3-Levels questionnaire; FACIT = Functional Assessment of Chronic Illness Therapy; HRQoL = health-related quality of life; HR = hazard ratio; EQ-5D VAS = EuroQol 5-Dimensions Visual Analogue Scale.

Note: HRQoL-evaluable population was intention-to-treat patients with a non-missing assessment at baseline and at least 1 post-baseline visit based on the HRQoL measure. Time to definitive HRQoL deterioration was defined as time to a clinically meaningful HRQoL deterioration sustained for at least 2 consecutive assessment visits. The HR was estimated from a Cox regression analysis with treatment group (CC-486 versus placebo), baseline domain score, and stratification factors (age group, prior history of myelodysplastic syndrome, cytogenetic risk category at time of induction therapy, and consolidation therapy following induction therapy) as covariates in the model.

Source: Clinical Study Report.¹⁴

Transfusion Independence

The outcome transfusion independence was specified in the CADTH review protocol but was not evaluated in the trial.

Health Care Resource Utilization

The total numbers of hospitalizations were 173 in the oral azacitidine group and 151 in the placebo group; hospitalization events after adjusting for duration of study drug exposure were 0.48 per person-year for the oral azacitidine group compared with 0.64 per person-year for the placebo group (relative risk = 0.740; 95% CI, 0.595 to 0.920; P = 0.0068) (Table 20).⁹ Analyses were noninferential.

A total of 2,872 hospitalization days were reported for the oral azacitidine group and 3,139 hospitalization days were reported for the placebo group. The number of days of hospitalization per person-year (exposure-adjusted rate) was lower in the oral azacitidine group compared with the placebo group: 7.89 days per person-year in the oral azacitidine group versus 13.36 days per person-year in the placebo group (relative risk = 0.591; 95% CI, 0.562 to 0.621; P < 0.0001) (Table 20).⁹ Analyses were noninferential.

Harms

Only those harms identified in the CADTH review protocol are reported below. Detailed harms data are presented in Table 21. Safety in the subset of patients who had dose escalation was also assessed (Table 35 in Appendix 3).

Adverse Events

Nearly all patients in both study groups experienced at least 1 TEAE (97.9% of patients in the oral azacitidine group and 96.6% in the placebo group). The most commonly reported TEAEs in the oral azacitidine and the placebo groups, respectively, were nausea (64.8% and 23.6%), vomiting (59.7% and 9.9%), diarrhea (50.4% and 21.5%), neutropenia (44.5% and 26.2%), constipation (38.6% and 24.0%), thrombocytopenia (33.5% and 27.0%), and fatigue (29.7% and 19.3%).

Grade 3 or 4 TEAEs occurred in 71.6% of patients in the oral azacitidine group and 63.1% of patients in the placebo group (Table 21). The most commonly reported grade 3 or 4 TEAE in both groups was neutropenia. The percentage of patients experiencing neutropenia was 41.1% in the oral azacitidine group and 23.6% in the placebo group. Other grade 3 or 4 TEAEs occurring across both treatment groups included thrombocytopenia (22.5% and 21.5% in the oral azacitidine and placebo groups, respectively), anemia (14.0% and 12.9% in the oral azacitidine and placebo groups, respectively), febrile neutropenia (11.4% and 7.7% in the oral azacitidine and placebo groups, respectively), and leukopenia (7.6% and 6.0% in the oral azacitidine and placebo groups, respectively).⁹

Serious Adverse Events

The percentage of patients experiencing serious TEAEs was 33.5% in the oral azacitidine group compared to 25.3% in the placebo group (Table 21). The most common serious TEAEs were febrile neutropenia, which occurred in 6.8% of patients in the oral azacitidine group and 3.9% of patients in the placebo group; pneumonia, in 3.8% of patients in the oral azacitidine group and 3.0% of patients in the placebo group; and pyrexia, in 2.1% of patients in oral azacitidine group and 0.4% of patient in the placebo group.⁹

Withdrawals Due to Adverse Events

Withdrawals specifically due to AEs were not reported in the QUAZAR AML-001 trial. Adverse events led to discontinuation of study treatment in 13.1% patients in the oral azacitidine group and 4.3% patients in the placebo group (Table 21). The most commonly occurring TEAE leading to treatment discontinuation (reported in more than 1 patient in either group) included nausea (n = 5 [2.1%] and n = 0 [0%] in the oral azacitidine and placebo groups, respectively), diarrhea (n = 4 [1.7%] and n = 0 [0%] in the oral azacitidine and placebo groups, respectively), vomiting (n = 3 [1.3%] and n = 0 [0%] in the oral azacitidine and placebo groups, respectively), and thrombocytopenia (n = 1 [0.4%] and n = 2 [0.9%] in the oral azacitidine and placebo groups, respectively).⁹

Mortality

On-treatment deaths were reported for |||||||||| patients in the oral azacitidine group and |||||||||| patients in the placebo group. On-treatment death included those occurring on or after the first dose date and within 28 days after the date of last study treatment. Most on-treatment deaths were due to the |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Post-treatment deaths were reported among |||||||||||| patients in the oral azacitidine group and |||||||||||| patients in the placebo group. Post-treatment deaths included those occurring beyond 28 days after the date of last study treatment. Most post-treatment deaths were due to the |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||).⁹

Overall, TEAEs leading to death occurred relatively rarely in both treatment groups (n = 9 [3.8%] and n = 4 [1.7%] in the oral azacitidine and placebo groups, respectively). The most commonly occurring TEAEs leading to death (reported in more than 1 patient in either group) included cerebral hemorrhage (n = 2 [0.8%] and n = [0.4%] in the oral azacitidine and placebo groups, respectively) and multiple organ dysfunction syndrome (n = 1 [0.4%] and n = 2 [0.9%] in the oral azacitidine and placebo groups, respectively). None of the TEAEs leading to death were considered treatment-related.⁹

Notable Harms

Notable harms specified in the protocol included gastrointestinal toxicities, hematologic toxicities, infections, fatigue, and bleeding.

Gastrointestinal Toxicities

The percentages of patients experiencing gastrointestinal TEAEs were 91.1% in the oral azacitidine group and 66.5% in the placebo group (Table 21). The most commonly reported gastrointestinal TEAEs in the oral azacitidine and placebo groups, respectively, included nausea (64.8% and 23.6%), vomiting (59.7% and 9.9), diarrhea (50.4% and 21.5%), and constipation (38.6% and 24.0). The percentage of patients experiencing gastrointestinal TEAEs of grade 3 or 4 was 16.9% in the oral azacitidine group compared with 6.4% in the placebo group. The most commonly reported grade 3 or 4 gastrointestinal TEAEs in the oral azacitidine group (versus placebo) were diarrhea (5.1% versus 1.3%), vomiting (3.0% versus 0.0%), nausea (2.5% versus 0.4), and constipation (1.3% versus 0%). Serious gastrointestinal TEAEs most commonly reported in the oral azacitidine group (versus placebo) included diarrhea (1.3% versus 0.0%), vomiting (0.8% versus 0.0%) gastroenteritis (0.8% versus 0.0%), and gastritis (0.8% versus 0.0%).⁹

Hematologic Toxicities

The percentage of patients experiencing myelosuppression TEAEs was |||| in the oral azacitidine group and |||| in the placebo group (Table 21). The most commonly reported myelosuppression event was neutropenia (44.5% and 26.2% in the oral azacitidine and placebo groups, respectively). Other myelosuppression TEAEs included thrombocytopenia (33.5% and 27.0% in the oral azacitidine and placebo groups, respectively), anemia (20.3 and 18.0% in the oral azacitidine and placebo groups, respectively), and febrile neutropenia (11.9% and 7.7% in the oral azacitidine and placebo groups, respectively). The percentage of patients experiencing grade 3 or 4 myelosuppression events was |||| in the oral azacitidine group and |||| in the placebo group; those most commonly reported (oral azacitidine versus placebo) included neutropenia (41.1% versus 23.6%), thrombocytopenia (22.5% versus 21.5%), and anemia (14.0% versus 12.9%). The percentage of patients experiencing serious myelosuppression events was |||| in the oral azacitidine group and |||| in the placebo group; those most commonly reported included |||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||⁹

Infections

The most commonly reported infections in the oral azacitidine group included upper respiratory tract infections (13.1% and 13.7% in the oral azacitidine and placebo groups, respectively), |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||). The percentage of patients experiencing grade 3 or 4 infections and serious infections was |||||||||||||||||||||||| 　||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. Grade 3 or 4 infections and serious infections most commonly reported in the oral azacitidine group (oral azacitidine versus placebo) included: |||||||||||| |||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||⁹

Fatigue

The percentage of patients experiencing fatigue TEAEs was 29.7% in the oral azacitidine group and 19.3% in the placebo group (Table 21). Fatigue of grade 3 or 4 or serious fatigue occurred rarely in patients in the trial. Fatigue of grade 3 or 4 was experienced by 3.0% and 0.9% of patients in the oral azacitidine group and placebo groups, respectively, and |||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||.⁹

Bleeding

The types and frequency of hemorrhagic events was similar across study groups (Table 21). The most commonly reported events in both groups included |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Grade 3 or 4 hemorrhagic events and serious hemorrhagic events occurred rarely in both groups (|||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||⁹

Critical Appraisal

Internal Validity

Baseline characteristics: A stratified randomization procedure was used based on known prognostic factors to minimize potential imbalances between the study groups that could lead to biased results. While imbalances were noted for a number of baseline characteristics, they were unlikely to influence clinical outcomes, according to the clinical experts consulted by CADTH.

Double-blind design: The QUAZAR AML-001 trial had a double-blind design. However, it is possible that some side effects of interest, such as gastrointestinal or myelosuppression events, allowed for the possible detection of the intervention being received for some participants. If trial investigators and/or patients were aware of the intervention assignment, this may have affected behaviour (such as initiation of subsequent treatment or adherence to treatment) or perceived HRQoL and symptom severity. The fact that a separate central review (blinded to patients’ treatment assignment) was used to confirm all clinical laboratory value assessments may have reduced the potential for bias arising from knowledge of treatments. Overall, the magnitude and direction of this bias remain unclear.

Statistical hierarchy and adjustment for multiplicity: Apart from the prespecified sequential testing of OS and RFS, no adjustments for multiplicity were performed to control the type I error rate; therefore, the analyses of the other secondary outcomes, such as HRQoL and symptom severity, were at risk of type I error.

Subgroup analysis: Methodological issues limited the ability to interpret the results from subgroup analyses. The subgroup analyses were noninferential, wide CIs reflected uncertainty in the effect estimates, and small sample sizes limited the generalizability to a broader population.

Proportional hazards: The analyses of OS and RFS were based on a stratified Cox proportional hazards model by which the HR between treatments is assumed to be unchanged over the duration of the comparison. The KM survival curves for OS initially diverged then converged at the end of the follow-up period, with a downward trend in survival. The clinical experts consulted by CADTH noted that the curves appeared to reflect their expectations of patients’ survival in clinical practice, i.e., patients with intermediate or poor cytogenic risk who are not eligible for transplantation (transplantation being the only potential cure in this setting) will relapse after some time. A pre-specified survival model that incorporated a treatment-by-time interaction suggested that the proportional hazards assumption did not hold. This affects the validity of the Cox regression models used to analyze OS in the trial. Prespecified sensitivity analyses were performed to adjust for nonproportional hazards such as restricted mean survival time, piecewise Cox regression, and generalized Wilcoxon tests. Results of these analyses were consistent with the results of the proportional hazards analysis, which supported the HRs obtained in the primary analysis demonstrating that oral azacitidine improved survival for patients relative to placebo. The graphical representation of the KM curves for RFS initially diverged and then converged, also suggesting a nonproportional hazard. However, no sensitivity analyses were pre-specified or conducted to evaluate the proportional hazards assumption.

Subsequent treatment: A higher proportion of patients in the placebo group compared with the oral azacitidine group received subsequent systemic therapy after discontinuing study treatment (57.6% in the oral azacitidine group versus 72.6% in the placebo group). Intensive chemotherapy and low-intensity therapy, respectively, were received by 29.0% and 39.5% of patients in the oral azacitidine group and by 37.6% and 47.0% of patients in the placebo group. Subsequent transplantation (implemented after discontinuation of study treatment)²⁷ was received by a small number of patients, 15 (6.3%) patients in the oral azacitidine group and 32 (13.7%) patients in the placebo group. Subsequent treatments could confound the assessment of OS if patients experienced additional clinical benefit. Planned sensitivity analyses, which censored for the use of subsequent therapy or subsequent transplantation, were conducted to assess potential confounding effects on the survival analyses from subsequent treatments or transplantation. The results of these sensitivity analyses were consistent with the primary OS results. However, given that greater than 80% of patients were censored in each treatment group in the analyses censoring for subsequent therapy, the ability to interpret these sensitive analyses remains limited. Additional sensitivity analyses for OS (a Cox proportional hazards model with covariates adjustment, the IPCW method, and regression-based imputation analysis) were performed to further assess potential biases from subsequent therapies. The results of these sensitivity analyses also suggested consistency with the primary OS results. The clinical experts consulted by CADTH noted that post-relapse salvage therapies in this setting are unlikely to have a significant impact on survival outcomes. It was further noted that the number of patients who would receive subsequent transplantation in clinical practice would be small and that there was no clinical rationale as to why the proportion of patients who went on to receive subsequent transplantation. Given the consistent results obtained in the sensitivity analyses and the anticipated limited impact of post-relapse therapy on survival, the potential of confounding effects on survival outcomes from subsequent treatment is expected to be low.

Censoring of RFS: The study provided 2 definitions of RFS based on different censoring rules. In the analysis of RFS using the primary definition, patients who received subsequent anticancer therapy before the date of relapse were censored in the analysis (overall,|| patients in the trial were censored for having a relapse event after initiating follow-up therapy: 　|　 patients in the oral azacitidine group and 　|　 patients in the placebo group).⁹ This may lead to biased results through informative censoring. The study conducted a sensitivity analysis of RFS based on EMA guidance that did not consider the receipt of subsequent therapy. The RFS results based on the sensitivity analysis were similar and supported the results of the primary definition. Given the consistent results obtained in the analyses using EMA guidance and the relatively small proportion of patients censored for having a relapse after initiating follow-up therapy in either study group, the potential for informative censoring bias affecting RFS is low.

Exposure to study: The median duration of 11.6 months (range = 0.5 to 74.3) of treatment with oral azacitidine was approximately twice as long as the treatment duration of 5.7 months (range = 0.7 to 68.5) with placebo. Given that most patients in the placebo group discontinued treatment beyond cycle 5 or 6, a safety comparison between the study groups across all cycles may be biased against the oral azacitidine group. The oral azacitidine group had a higher proportion of patients remaining on study treatment as well as longer exposure to study treatment compared with the placebo group; therefore, patients in the oral azacitidine group were more likely to report AEs relative to the placebo group during the treatment phase.

HRQoL and symptom severity assessments: The interpretation of results for the EQ-5D-3L and FACIT-F scale (i.e., the ability to assess trends over time and to make comparisons across treatment groups) at later cycles is limited by the significant decline in patients available to provide assessment over time (less than 50% of patients were available in the placebo group beyond cycle 6 and less than 50% of patients were available in the oral azacitidine group beyond cycle 13). In addition, selection bias over time should be considered when interpreting results, as the long-term survivors tended to be the healthier patients. Given that the analyses of HRQoL and severity of symptom outcomes were based on the HRQoL-evaluable population and not on the ITT population, they do not capture the experience of all patients randomized in each treatment group. Furthermore, multiple statistical testing was conducted with no adjustments for multiplicity to control the type I error rate.

The EQ-5D-3L has been validated in patients with AML in China, achieving overall good validity and reliability (responsiveness was not assessed), whereas ranges for MIDs have been reported for cancer patients but not specially for patients with AML. The sponsor used a conservative approach to defining MIDs (i.e., chose relative high cut-off values from the MID ranges reported in the literature) for the EQ-5D-3L health utility index and EQ VAS scale (i.e., relatively high cut-off values compared with the MID ranges reported in the literature). The PINR scale has not been validated nor have MIDs been reported in the literature. The sponsor’s submission stated that half of the standard deviation at baseline from all ITT patients was used as an approximation for MID. However, a distribution-based approach to estimate the MID internally from the trial data is not an established method; rather a triangulation with anchor-based approach would be required. It is therefore unclear if the MID used in the trial is appropriate and reflective of a clinically meaningful change in outcomes in patients with AML. The FACIT-F scale has been validated in patients with CCL and showed overall good validity and reliability (responsiveness was not assessed), and MIDs have been reported in the literature for cancer patients but not specially for patients with AML. The sponsor defined the MID by choosing a cut-off value within the range of reported values in the literature. Overall, these methodological issues limit the ability to interpret the results from the EQ-5D-3L and the FACIT-F scale and render results from the PINR scale inconclusive.

External Validity

The QUAZAR AML-001 trial was an international, multicentre trial. Although the majority of patients in the trial were enrolled in trial sites in Europe, according to the clinical experts consulted by CADTH, the enrolled population was consistent with the population expected to be treated for AML in first remission in Canadian clinical practice. The clinical experts consulted by CADTH agreed that no different treatment effect would be expected based on different disease-management practices across countries. They noted that few patients in the trial had received prior types of therapy that are not standard of care in Canada. While a small number of patients in both study groups received consolidation therapies that are not offered in Canada, the clinical experts indicated that this is unlikely to have affected the study outcomes. The clinical experts felt that the QUAZAR AML-001 trial results would be generalizable to patients who have received gemtuzumab ozogamicin as induction therapy. Gemtuzumab ozogamicin was recently reviewed by CADTH and is recommended for reimbursement for adults with previously untreated, de novo CD33-positive AML, with the exception of patients with APL, who have good performance status and favourable, intermediate, or unknown cytogenetics (ELN 2017).¹⁹ The most common reason for transplantation ineligibility reported in the trial was age (for 64.8% of patients). The clinical experts pointed out that, since the inception of the QUAZAR AML-001 trial, practice has shifted away from using a certain age cut-off as a single criterion for transplantation ineligibility. However, given that older age, more comorbidities, and worse ECOG PS correlate, overall, it was felt that the patients selected for transplantation ineligibility were reflective of Canadian clinical practice. In the trial most patients received 1 cycle (44.9%) followed by 2 cycles (31.1%) of prior consolidation therapy. The experts noted that fewer patients in Canadian clinical practice would be expected to only receive 1 cycle and more patients would receive 2 cycles of prior consolidation therapy.

Relevance of trial efficacy outcomes: The primary outcome in the QUAZAR AML-001 trial was OS and the key secondary outcome was RFS. According to the clinical experts consulted by CADTH, both are clinically meaningful end points for patients with AML in their first remission and both are used to guide treatment selection in clinical practice. Upon AML relapse, treatment options are limited and administered primarily with palliative intent. Extending the period patients remain in remission is therefore important, as AML relapse is associated with a higher burden of symptoms, a decrease in quality of life, and shorter time to death. The primary and key secondary outcomes were supported by secondary outcomes that included time to relapse from CR or CRi and time to discontinuation from treatment.

Excluded patient subgroups: The trial excluded patients aged less than 55 and patients with less-than-adequate organ function. The clinical experts consulted by CADTH felt that it would be reasonable to generalize the QUAZAR AML-001 trial results to patients younger than 55 years, given the acceptable safety profile of oral azacitidine and the lack of a biological rationale to assume that outcomes of oral azacitidine would be different in younger patients with AML who otherwise meet the trial’s inclusion criteria. As well, given the acceptable safety profile of oral azacitidine the clinical experts agreed that it would be reasonable to leave it up to the discretion of the treating physician to apply some flexibility in the terms of using oral azacitidine in patients with slightly lower laboratory parameters than those outlined in the trial.

Dose escalation: One design feature of the QUAZAR AML-001 trial was to allow post-relapse dose escalation for patients who relapsed with 5% to 15% of blasts in the bone marrow. The dosing schedule was extended from 14 to 21 days out of a period of 28 days at the investigator’s discretion. The dose escalation allowed for an exploration of whether oral azacitidine could be used to reinitiate remission, which is not consistent with the Health Canada indication or the reimbursement request for azacitidine as maintenance therapy. The percentage of patients that had relapsed was |||| in the oral azacitidine group compared with |||| in the placebo group (with a relapse defined as a blast count ≥ 5% in the peripheral blood or bone marrow). Dose escalation was implemented in 51 patients in the oral azacitidine group and 40 patients in the placebo group. The Health Canada product monograph does not recommend dose escalation as Health Canada concluded that there is insufficient evidence at this time to guide a recommendation (i.e., the trial was not designed to assess efficacy in patients who received dose escalation).¹¹ The clinical experts consulted by CADTH agreed that they would not attempt dose escalation in clinical practice as there is currently insufficient evidence to generalize the QUAZAR AML-001 trial results to oral azacitidine used to reinitiate remission. They further noted that it is unlikely that dose escalation would significantly influence OS outcomes observed in the trial due to the small numbers of patients who received dose escalation relative to the study sample size. Relapse-free survival was not affected by dose escalation as dose escalation was initiated after disease relapse, which was defined according to the IWG’s ALM response criteria: reappearance in the peripheral blood of 5% or greater blasts in the bone marrow.

Treatment discontinuation threshold: Treatment was continued until disease relapse (> 15% blasts in the bone marrow or peripheral blood), or unacceptable toxicity.¹⁰ The clinical experts consulted by CADTH noted that a 15% threshold is not a clinically established definition for relapse but appears to be an arbitrary choice in the trial to explore dose escalation in patients with bone marrow blast levels between 5% and 15%. The CADTH experts agreed that treatment with oral azacitidine would be discontinued in clinical practice at the time of relapse, as defined as reappearance of leukemic blasts in the peripheral blood or 5% or greater blasts in the bone marrow. An initially planned sensitivity analysis of RFS based on relapse using a 16% bone marrow blast threshold was not performed as it was determined that the interpretation of such analysis would be inconclusive given that not all patients with bone marrow blast levels between 5% and 15% received a dose escalation, and not all patients with dose escalation had marrow blast levels below 15%.⁹

Concomitant medications received by patients in the trial appeared reflective of the medications patients would receive in Canadian clinical practice, according to the clinical experts consulted by CADTH.

Placebo plus BSC is an appropriate comparator in Canadian clinical practice. The clinical experts by CADTH agreed that there is currently no active alternative treatment used in Canada in the target setting.

Indirect Evidence

No indirect treatment comparisons were included in the sponsor’s submission to CADTH or identified in the literature search.

Other Relevant Evidence

No long-term extension studies or other relevant studies were included in the sponsor’s submission to CADTH or identified from the literature.

Discussion

Summary of Available Evidence

The CADTH systematic review included 1 phase III RCT of oral azacitidine as maintenance therapy for the treatment of AML in adult patients who achieved CR or CRi following induction therapy with or without consolidation treatment and who are not eligible for HSCT. The QUAZAR AML-001 trial (N = 472) is an ongoing, international, multicentre, double-blind, placebo-controlled RCT that compares the efficacy and safety of maintenance therapy with oral azacitidine plus BSC with placebo plus BSC in patients with AML in first CR. Randomization was centrally performed in a 1:1 ratio and stratified by age at time of induction therapy, prior history of MDS or CMML, cytogenetic risk category at time of induction therapy, and receipt of consolidation therapy following induction. No crossover between the treatment groups was permitted. The primary outcome was OS, and the key secondary outcome was RFS. Other secondary end points included time to relapse, time to discontinuation from treatment, HRQoL, and symptoms.⁹

The QUAZAR AML-001 trial enrolled adults diagnosed with AML or AML secondary to prior MDS or CMML and intermediate- or poor-risk cytogenetics and who had achieved first remission (CR or CRi) following induction with or without consolidation chemotherapy. The majority of patients had an ECOG PS of 0 or 1, an intermediate cytogenetic risk, and had received 1 or 2 cycles of consolidation therapy. The most common reason for transplantation ineligibility was age, followed by comorbidities.⁹

No indirect treatment comparisons or other evidence were included in the sponsor’s submission to CADTH or identified in the literature search.

Interpretation of Results

Efficacy

The QUAZAR AML-001 trial met its primary end point, demonstrating statistically significant improvement in OS in favour of the oral azacitidine group compared with the placebo group. The results of prespecified sensitivity analyses for OS, which were conducted to adjust for nonproportional hazards and to assess potential biases from subsequent treatments, suggested the robustness and consistency of the primary OS results, indicating that oral azacitidine resulted in improved survival for patients. Results for subgroups of interest, as specified in the protocol for this CADTH systematic literature review, suggested that the OS benefit was consistently favourable across pre-specified subgroups of patients (except for the subcategory “3 or 4 cycles of prior consolidation therapy”). However, given that the trial was not designed to detect differences in treatment effects across subgroups, no conclusions can be drawn from the subgroup results. Results for the key secondary outcome, RFS, were supportive of the OS results and demonstrated a statistically significant improvement in favour of oral azacitidine. Other secondary efficacy outcomes, time to relapse from CR or CRi and time to discontinuation from treatment, were also supportive of the observed OS gain. According to the clinical experts CADTH consulted for this review, OS and RFS are both clinically meaningful end points for patients with AML in their first remission and are used to guide treatment selection in clinical practice. Upon AML relapse, treatment options are limited and administered primarily with palliative intent. Extending the period patients remain in remission is important as AML relapse is associated with a higher burden of symptoms, a decrease in quality of life, and shorter time to death. This view was echoed by the input provided by the patient advocacy group and the registered clinician input, both of which highlighted prolongation of life and maintenance of remission as important treatment goals for patients. The clinical experts further noted that the improvements in OS and RFS of the magnitude observed in the QUAZAR AML-001 trial (absolute differences for OS and RFS were approximately 10 months and 5 months, respectively) are of clinical importance in a patient population for which there is currently no standard maintenance treatment. Given that most patients in this setting relapse after a few months, the benefits observed with oral azacitidine over placebo were considered clinically meaningful.

Input received by the patient advocacy group, the registered clinicians, and the clinical experts consulted by CADTH highlighted HRQoL as an important outcome and treatment goal for patients. The level of fatigue and HRQoL over the assessment period was similar between the 2 study groups and oral azacitidine did not appear to significantly delay time to deterioration compared with placebo. The clinical experts consulted by CADTH considered this to be reasonable in this setting, where patients’ HRQoL is expected to be relatively high and stable. However, given several important limitations, including the noninferential analyses and the significant decline in patients available to provide assessment over time, the interpretation of results for the EQ-5D-3L and FACIT-F scale is limited.

Although patients in the QUAZAR AML-001 trial were considered representative of patients in Canadian clinical practice, the clinical experts consulted by CADTH noted that it would be reasonable to generalize the QUAZAR AML-001 trial results to patients younger than 55 years, who were excluded from the trial. The Health Canada indication does not have an age restriction. As well, given the acceptable safety profile of oral azacitidine, the clinical experts felt that it would be reasonable to leave it to the discretion of the treating physician to apply flexibility in the terms of using oral azacitidine in patients with slightly lower laboratory parameters than those outlined in the trial. Gemtuzumab ozogamicin was recently reviewed by CADTH and is recommended for reimbursement for adults with previously untreated, de novo CD33-positive AML (except APL) who have good performance status and favourable, intermediate, or unknown cytogenetics (ELN 2017).¹⁹ The clinical experts agreed that the trial results would be generalizable to patients who have received gemtuzumab ozogamicin as induction therapy.

The standard dose of oral azacitidine in the trial would be the dose used in Canadian clinical practice and is in line with the Health Canada–approved dosing. Dose escalation, which was attempted in the QUAZAR AML-001 trial to reinitiate remission, is not generalizable to clinical practice in Canada according to the clinical experts consulted by CADTH and has not been approved by Health Canada. The clinical experts further noted that it is unlikely that dose escalation would significantly influence the clinical outcomes observed in the trial due to the small numbers of patients who received dose escalation relative to the study sample size.

Harms

The median duration of treatment with oral azacitidine was about twice that of the treatment duration with placebo, which should be considered when reviewing the incidence of TEAEs. Almost all patients in both study groups experienced at least 1 TEAE. Differences in the proportion of patients experiencing grade 3 and 4 TEAEs and serious TEAEs were driven mainly by gastrointestinal toxicities and myelosuppression events. The most commonly reported TEAEs in the oral azacitidine and the placebo groups were nausea, vomiting, diarrhea, and neutropenia. The most commonly reported grade 3 or 4 and serious TEAEs, respectively, were neutropenia and febrile neutropenia. The clinical experts consulted by CADTH noted that most TEAEs associated with oral azacitidine could be managed with dose modifications and BSC, and, while treatment discontinuation due to TEAEs was relatively rare, it occurred more frequently in the oral azacitidine group. From the review of notable harms, it appeared that toxicities from oral azacitidine were seen mostly in the gastrointestinal system and as myelosuppression events. Few deaths were due to TEAEs overall, and no TEAE leading to death was considered treatment-related. The clinical experts consulted by CADTH agreed that the TEAEs observed with oral azacitidine were acceptable overall and could be adequately managed in clinical practice.

Conclusions

One phase III, double-blind, placebo-controlled RCT (QUAZAR AML-001) provided evidence regarding the efficacy and safety of maintenance therapy with oral azacitidine plus BSC compared with placebo plus BSC in patients with AML in first CR. Compared to placebo, patients treated with oral azacitidine showed statistically significant improvements in OS, the primary end point, and RFS, the key secondary outcome. The differences in median OS and RFS between treatment groups (approximately 10 months and 5 months, respectively) were considered clinically meaningful by the clinical experts consulted by CADTH. Other secondary efficacy outcomes, time to relapse from CR or CRi, and time to discontinuation from treatment were also supportive of the observed OS and RFS gains. The level of fatigue and HRQoL over the assessment period was similar between the 2 study groups, and oral azacitidine did not appear to significantly delay time to deterioration compared with placebo. However, these results remain uncertain due to a number of important limitations. The toxicity profile of oral azacitidine was considered manageable by the clinical experts consulted by CADTH.⁹

References

1.Drug Reimbursement Review sponsor submission: Onureg (azacitidine): 200 mg, 300 mg tablets [internal sponsor's package]. St. Laurent (QC): Celgene; 2021 Mar 1.

2.PrOnureg® (azacitidine): 200 mg, 300 mg tablets [product monograph]. St. Laurent (QC): Celgene; 2021 Jan 4.

3.Cogle CR, Scott BL, Garcia-Manero G. Oral azacitidine (Cc-486) for the treatment ofmyelodysplastic syndromes and acutemyeloid leukemia. Oncologist. 2015;20(12):1404-1412. PubMed

4.Acute myeloid leukemia (Version 3.2021). NCCN Clinical practice guidelines in oncology. Plymouth Meeting (PA): National Comprehensive Cancer Network; 2021.

5.Acute myelogenous leukemia statistics. Toronto (ON): Canadian Cancer Society; 2021: https://www.cancer.ca/en/cancer-information/cancer-type/leukemia-acute-myelogenous-aml/statistics/?region=on. Accessed 2021 May 12.

6.Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447. PubMed

7.Fey MF, Buske C. Acute myeloblastic leukaemias in adult patients: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2013;24(Suppl 6):vi138-143. PubMed

8.Koreth J, Schlenk R, Kopecky KJ, et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. JAMA. 2009;301(22):2349-2361. PubMed

9.Clinical study report CC-486-AML-001. A phase 3, randomized, double-blind, placebo controlled study to compare efficiency and safety of oral azacitidine plus best supportive care versus best supportive care as maintenance therapy in subjects with acute myeloid leukemia in complete remission [internal sponsor's report]. St. Laurent (QC): Celgene; 2020 Jan 8.

10.Wei AH, Dohner H, Pocock C, et al. Oral azacitidine maintenance therapy for acute myeloid leukemia in first remission. N Engl J Med. 2020;383(26):2526-2537. PubMed

11.Health Canada reviewer's report: Onureg (azacitidine) [internal sponsor's report]. St. Laurent (QC): Celgene; 2020 Dec 31.

12.Celgene response to May 18, 2021 request for additional information regarding Onureg review [internal additional sponsor's information]. St. Laurent (QC): Celgene; 2021.

13.Celgene. NCT01757535: Efficacy of oral azacitidine plus best supportive care as maintenance therapy in subjects with acute myeloid leukemia in complete remission (QUAZAR AML-001). ClinicalTrials.gov. Bethesda (MD): U.S. National Library of Medicine; 2012: https://clinicaltrials.gov/ct2/show/NCT01757535. Accessed 2021 Apr 22.

14.Study CC-486-AML-001. Patient-reported outcome/ health-realted quality of life report: oral azacitidine (CC-486) [internal sponsor's report]. St. Laurent (QC): Celgene; 2019 Dec 16.

15.Kolitz JE. Overview of acute myeloid leukemia in adults. In: Post TW, ed. UpToDate. Waltham (MA): UpToDate; 2021: https://www.uptodate.com. Accessed 2021 May 11.

16.Acute Myeloid Leukemia. Clinical practice guideline LYHE-006. Version 6. Edmonton (AB): Alberta Health Services; 2019: https://www.albertahealthservices.ca/assets/info/hp/cancer/if-hp-cancer-guide-lyhe006-aml.pdf Accessed 2021 May 12.

17.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics 2019. Toronto (ON): Canadian Cancer Society; 2019: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2020-statistics/canadian-cancer-statistics/res-cancerstatistics-canadiancancerstatistics-2019-en.pdf?rev=82dc3652fe3648988b9174ad4b397a24&hash=6D3186DF3AC76787C58EE95D1712033C. Accessed 2021 May 12.

18.Hills RK, Castaigne S, Appelbaum FR, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy in adult patients with acute myeloid leukaemia: a meta-analysis of individual patient data from randomised controlled trials. Lancet Oncol. 2014;15(9):986-996. PubMed

19.Gemtuzumab ozogamicin (Mylotarg) for acute myeloid leukemia. Drug reimbursement review. Ottawa (ON): CADTH; 2020: https://cadth.ca/gemtuzumab-ozogamicin-mylotarg-acute-myeloid-leukemia-details. Accessed 2021 May 12.

20.Provincial funding summary: Rydapt (midostaurin) for acute myeloid leukemia (pCODR 10108). Ottawa (ON): CADTH; 2020: https://www.cadth.ca/sites/default/files/pcodr/pcodr_profund_midostaurin_rydapt_aml.pdf. Accessed 2021 May 7.

21.Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447. PubMed

22.Hansen DK, Kim J, Thompson Z, et al. ELN 2017 genetic risk stratification predicts survival of acute myeloid leukemia patients receiving allogeneic hematopoietic stem cell transplantation. Transplant Cell Ther. 2021;27(3):256.e1-256.e7. PubMed

23.McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40-46. PubMed

24.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2021 Mar 16.

25.Celgene response to April 7, 2021 request for additional information regarding Onureg review [internal additional sponsor's information]. St. Laurent (QC): Celgene; 2021.

26.Center for Drug Evaluation Research. Multidiscipline review. Onureg (azacitidine): 200 mg, 300 mg tablets. Company: Celgene. Application No.: NDA214120. Approval date: Sept 1, 2020. Silver Spring (MD): U.S. Food and Drug Administration (FDA); 2020 Sep 1: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2020/214120Orig1s000MultidisciplineR.pdf. Accessed 2021 Apr 28.

27.Celgene response to May 28, 2021 request for additional information regarding Onureg review [internal additional sponsor's information]. St. Laurent (QC): Celgene; 2021.

28.Yu H, Zeng X, Sui M, et al. A head-to-head comparison of measurement properties of the EQ-5D-3L and EQ-5D-5L in acute myeloid leukemia patients. Qual Life Res. 2021;30(3):855-866. PubMed

29.Pickard AS, Neary MP, Cella D. Estimation of minimally important differences in EQ-5D utility and VAS scores in cancer. Health Qual Life Outcomes. 2007;5:70-70. PubMed

30.Kvam AK, Fayers PM, Wisloff F. Responsiveness and minimal important score differences in quality-of-life questionnaires: a comparison of the EORTC QLQ-C30 cancer-specific questionnaire to the generic utility questionnaires EQ-5D and 15D in patients with multiple myeloma. Eur J Haematol. 2011;87(4):330-337. PubMed

31.Sinnott PL, Joyce VR, Barnett PG. Preference measurement in economic analysis. Guidebook. Menlo Park (CA): Health Economics Research Center; 2007: https://www.herc.research.va.gov/files/BOOK_419.pdf. Accessed 2021 Apr 30.

32.Protocol CC-486-AML-001. Patient-reported outcomes/quality of life: statistical analysis plan [internal sponsor's report]. St. Laurent (QC): Celgene; 2019 Jul 26.

33.Eek D, Ivanescu C, Corredoira L, Meyers O, Cella D. Content validity and psychometric evaluation of the Functional Assessment of Chronic Illness Therapy-Fatigue scale in patients with chronic lymphocytic leukemia. J Patient Rep Outcomes. 2021;5(1):27. PubMed

34.Ghia P, Pluta A, Wach M, et al. ASCEND: Phase III, randomized trial of acalabrutinib versus idelalisib plus rituximab or bendamustine plus rituximab in relapsed or refractory chronic lymphocytic leukemia. J Clin Oncol. 2020;38(25):2849-2861. PubMed

35.Cella D, Eton DT, Lai JS, Peterman AH, Merkel DE. Combining anchor and distribution-based methods to derive minimal clinically important differences on the Functional Assessment of Cancer Therapy (FACT) anemia and fatigue scales. J Pain Symptom Manage. 2002;24(6):547-561. PubMed

36.Patrick DL, Gagnon DD, Zagari MJ, Mathijs R, Sweetenham J. Assessing the clinical significance of health-related quality of life (HrQOL) improvements in anaemic cancer patients receiving epoetin alfa. Eur J Cancer. 2003;39(3):335-345. PubMed

37.Baer MR, George SL, Sanford BL, et al. Escalation of daunorubicin and addition of etoposide in the ADE regimen in acute myeloid leukemia patients aged 60 years and older: Cancer and Leukemia Group B Study 9720. Leukemia. 2011;25(5):800-807. PubMed

38.Baer MR, George SL, Sanford BL, et al. Treatment of older patients with de novo acute myeloid leukemia (AML) with one or more postremission chemotherapy courses: analysis of four CALGB studies. J Clin Oncol. 2010;28(15 Suppl):6531-6531.

39.Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York (NY): John Wiley & Sons; 1980.

40.Luo X, Li M, Wu C, et al. A proposed approach for analyzing post-study therapy effect in survival analysis. J Biopharm Stat. 2016;26(4):790-800. PubMed

41.Pappa V, Girkas K, Sambani C, et al. Azacitidine in a patient with newly diagnosed acute myeloid leukemia and poor risk cytogenetics: outcomes of prolonged therapy. Leuk Res. 2012;36(4):e78-80. PubMed

42.Wehmeyer J, Zaiss M, Losem C, et al. Impact of performance status and transfusion dependency on outcome of patients with myelodysplastic syndrome, acute myeloid leukemia and chronic myelomonocytic leukemia treated with azacitidine (PIAZA study). Eur J Haematol. 2018;101(6):766-773. PubMed

43.Huls G, Chitu DA, Havelange V, et al. Azacitidine maintenance after intensive chemotherapy improves DFS in older AML patients. Blood. 2019;133(13):1457-1464. PubMed

44.Jehn U, Knuppel W, Wilmanns W. Intensive maintenance treatment in acute myelogenous leukemia (AML): single institution experience of a multicenter randomized trial. Onkologie. 1988;11(1):13-17. PubMed

45.Jehn U, Zittoun R. AML-6 study of the value of cyclic alternating chemotherapy during remission in acute myelocytic leukemia. Onkologie. 1985;8(2):94-95. PubMed

46.Rees JK, Gray R, Hayhoe FG. The ninth British Medical Research Council trial for the treatment of acute myeloid leukaemia. Haematol Blood Transfus. 1987;30:35-37. PubMed

47.Rees JK, Gray RG. Remission induction and postremission therapy in acute myelogenous leukemia: British MRC Study. Haematol Blood Transfus. 1990;33:243-248. PubMed

48.Dumas PY, Bertoli S, Berard E, et al. Azacitidine or intensive chemotherapy for older patients with secondary or therapy-related acute myeloid leukemia. Oncotarget. 2017;8(45):79126-79136. PubMed

49.Seymour JF, Dohner H, Butrym A, et al. Azacitidine improves clinical outcomes in older patients with acute myeloid leukaemia with myelodysplasia-related changes compared with conventional care regimens. BMC Cancer. 2017;17(1):852. PubMed

50.Seymour JF, Dohner H, Minden MD, et al. Incidence rates of treatment-emergent adverse events and related hospitalization are reduced with azacitidine compared with conventional care regimens in older patients with acute myeloid leukemia. Leuk Lymphoma. 2017;58(6):1412-1423. PubMed

51.Roboz GJ, Montesinos P, Selleslag D, et al. Design of the randomized, phase III, QUAZAR AML maintenance trial of CC-486 (oral azacitidine) maintenance therapy in acute myeloid leukemia. Fut Oncol. 2016;12(3):293-302. PubMed

52.Jehn U, Zittoun R, Lowenberg B. Study design of the AML 6 and 7 trial for treatment of acute myelogenous leukemia: cyclic alternating chemotherapy during remission and remission duration and survival in elderly patients. [German]. Onkologie. 1985;8(3):160-164. PubMed

53.Benefits of AML maintenance therapy extend to quality of life and hospitalization. Oncologist. 2021;26(Suppl 1):S11-S12. PubMed

54.Derissen EJ, Beijnen JH, Schellens JH. Concise drug review: azacitidine and decitabine. Oncologist. 2013;18(5):619-624. PubMed

55.Pan J, Altman D, Wilde L. Measurable residual disease-guided treatment to prevent relapse in acute myeloid leukemia and myelodysplastic syndrome. Front Oncol. 2020;10(576924). PubMed

56.Lowder JN, Taverna P, Issa JP. Will next-generation agents deliver on the promise of epigenetic hypomethylation therapy? Epigenomics. 2015;7(7):1083-1088. PubMed

57.Wei AH. Maintenance therapy for AML: are we there yet? Blood. 2019;133(13):1390-1392. PubMed

58.Sekeres MA, Watts J, Radinoff A, et al. Randomized phase 2 trial of pevonedistat plus azacitidine versus azacitidine for higher-risk MDS/CMML or low-blast AML. Leukemia. 2021;35(7):2119-2124. PubMed

59.Dinardo CD. Hypomethylating agents and venetoclax in acute myeloid leukemia. Clin Adv Hematol Oncol. 2021;19(2):82-83. PubMed

60.Protocol CC-486-AML-001. A phase 3, randomized, double-blind, placebo-controlled study to compare efficacy and safety of oral azacitidine plus best supportive care versus best supportive care as maintenance therapy in subjects with acute myeloid leukemia in complete remission [internal sponsor's report]. St. Laurent (QC): Celgene; 2018 Nov 8.

61.Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582-592. PubMed

62.Norman GR, Sloan JA, Wyrwich KW. The truly remarkable universality of half a standard deviation: confirmation through another look. Expert Rev Pharmacoecon Outcomes Res. 2004;4(5):581-585. PubMed

63.van Reenen M, Oppe M. EQ-5D-3L user guide: basic information on how to use the EQ-5D-3L instrument. Rotterdam (NL): EuroQol Research Foundation; 2015 Apr. .

64.FACIT-F: Functional Assessment of Chronic Illness Therapy – Fatigue. Munich (DE): FACIT.org: https://www.facit.org/measures/FACIT-F. Accessed 2021 Apr 26.

Appendix 1: Literature Search Strategy

Note that this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases:

• MEDLINE All (1946 to present)

• Embase (1974 to present)

• Note: Subject headings and search fields have been customized for each database. Duplicates between databases were removed in Ovid.

Date of search: March 25, 2021

Alerts: Bi-weekly search updates until project completion

Search filters applied: randomized controlled trials; controlled clinical trials

Limits: Conference abstracts: excluded

Table 22: Syntax Guide

Multi-Database Strategy

Search Strategy

1. Azacitidine/

2. (Onureg* or azacitidine* or azacytidin* or Vidaza* or Ladakamycin* or Mylosar* or 5-AZCR or 5AZCR or “BRN 0620461” or BRN0620461 or CCRIS 60 or CCRIS60 or EINECS 206-280-2 or EINECS 2062802 or EINECS2062802 or HSDB 6879 or HSDB6879 or NCI C01569 or NCIC01569 or NSC 102816 or NSC102816 or CC-486 or CC486 or U 18496 or U18496 or WR 183027 or WR183027 or M801H13NRU).ti,ab,ot,kf,hw,nm,rn.

3. 1 or 2

4. exp Leukemia, Myeloid, acute/

5. (acute adj3 (granulocytic or myeloblastic or myelocytic or myelogenous or myeloid or nonlymphoblastic or non-lymphoblastic or nonlymphocytic or non-lymphcytic or megakaryocytic or monocytic or myelomonocytic or basophilic or eosinophilic or erythroblastic or megakaryoblastic or promyelocytic) adj3 leuk?emia*).ti,ab,kf.

6. (erythroleuk?emia* or mast cell leuk?emia* or promyelocytic leuk?emia*).ti,ab,kf.

7. (AML or ANLL).ti,ab,kf.

8. or/4-7

9. 3 and 8

10. 9 use medall

11. *Azacitidine/

12. (Onureg* or azacitidin* or azacytidin* or Vidaza* or Ladakamycin* or Mylosar* or 5-AZCR or 5AZCR or “BRN 0620461” or BRN0620461 or CCRIS 60 or CCRIS60 or EINECS 206-280-2 or EINECS 2062802 or EINECS2062802 or HSDB 6879 or HSDB6879 or NCI C01569 or NCIC01569 or NSC 102816 or NSC102816 or CC-486 or CC486 or U 18496 or U18496 or WR 183027 or WR183027).ti,ab,kw,dq.

13. 11 or 12

14. exp Acute myeloid leukemia/

15. (acute adj3 (granulocytic or myeloblastic or myelocytic or myelogenous or myeloid or nonlymphoblastic or non-lymphoblastic or nonlymphocytic or non-lymphcytic or megakaryocytic or monocytic or myelomonocytic or basophilic or eosinophilic or erythroblastic or megakaryoblastic or promyelocytic) adj3 leuk?emia*).ti,ab,kw,dq.

16. (erythroleuk?emia* or mast cell leuk?emia* or promyelocytic leuk?emia*).ti,ab,kw,dq.

17. (AML or ANLL).ti,ab,kw,dq.

18. or/14-17

19. 13 and 18

20. 19 use oemezd

21. 20 not (conference abstract or conference review).pt.

22. 10 or 21

23. (Randomized Controlled Trial or Controlled Clinical Trial or Pragmatic Clinical Trial or Equivalence Trial or Clinical Trial, phase III).pt.

24. Randomized Controlled Trial/

25. exp Randomized Controlled Trials as Topic/

26. “Randomized Controlled Trial (topic)”/

27. Controlled Clinical Trial/

28. exp Controlled Clinical Trials as Topic/

29. “Controlled Clinical Trial (topic)”/

30. Randomization/

31. Random Allocation/

32. Double-Blind Method/

33. Double Blind Procedure/

34. Double-Blind Studies/

35. Single-Blind Method/

36. Single Blind Procedure/

37. Single-Blind Studies/

38. Placebos/

39. Placebo/

40. Control Groups/

41. Control Group/

42. (random* or sham or placebo*).ti,ab,hw,kf,kw.

43. ((singl* or doubl*) adj (blind* or dumm* or mask*)).ti,ab,hw,kf,kw.

44. ((tripl* or trebl*) adj (blind* or dumm* or mask*)).ti,ab,hw,kf,kw.

45. (control* adj3 (study or studies or trial* or group*)).ti,ab,kf,kw.

46. (Nonrandom* or non random* or non-random* or quasi-random* or quasirandom*).ti,ab,hw,kf,kw.

47. allocated.ti,ab,hw.

48. ((open label or open-label) adj5 (study or studies or trial*)).ti,ab,hw,kf,kw.

49. ((equivalence or superiority or non-inferiority or noninferiority) adj3 (study or studies or trial*)).ti,ab,hw,kf,kw.

50. (pragmatic study or pragmatic studies).ti,ab,hw,kf,kw.

51. ((pragmatic or practical) adj3 trial*).ti,ab,hw,kf,kw.

52. ((quasiexperimental or quasi-experimental) adj3 (study or studies or trial*)).ti,ab,hw,kf,kw.

53. (phase adj3 (III or “3”) adj3 (study or studies or trial*)).ti,hw,kf,kw.

54. or/23-53

55. 22 and 54

56. remove duplicates from 55

Clinical Trials Registries

ClinicalTrials.gov

Produced by the US National Library of Medicine. Targeted search used to capture registered clinical trials.

[Search – Onureg/oral azacitidine AND acute myeloid leukemia]

WHO ICTRP

International Clinical Trials Registry Platform, produced by the WHO. Targeted search used to capture registered clinical trials.

[Search terms – Onureg/oral azacitidine AND acute myeloid leukemia]

Health Canada’s Clinical Trials Database

Produced by Health Canada. Targeted search used to capture registered clinical trials.

[Search terms – Onureg/oral azacitidine AND acute myeloid leukemia]

EU Clinical Trials Register

European Union Clinical Trials Register, produced by the European Union. Targeted search used to capture registered clinical trials.

[Search terms – Onureg/oral azacitidine AND acute myeloid leukemia]

Grey Literature

Search dates: March 15 to 17, 2021

Keywords: Onureg/oral azacitidine AND acute myeloid leukemia

Limits: None

Updated: Search updated before the completion of stakeholder feedback period

Relevant websites from the following sections of the CADTH grey literature checklist Grey Matters: A Practical Tool For Searching Health-Related Grey Literature (https://www.cadth.ca/grey-matters) were searched:

• Health technology assessment agencies

• Health economics

• Clinical practice guidelines

• Drug and device regulatory approvals

• Advisories and warnings

• Drug class reviews

• Clinical trials registries

• Databases (free)