Q-HAM: a multicenter upfront randomized phase II trial of quizartinib and high-dose Ara-C plus mitoxantrone in relapsed/refractory AML with FLT3-ITD

In this multicenter, upfront randomized, open-label, phase II trial, all patients receive quizartinib combined with HAM during salvage therapy. During consolidation therapy (chemotherapy as well as allo-HCT), patients receive either prophylactic quizartinib therapy or MRD-triggered preemptive continuation therapy with quizartinib according to up-front randomization (see Fig. 1 and Additional file 1: SPIRIT figure). Efficacy is assessed by comparison to historical controls based on the matched threshold-crossing approach [10, 25] (Krisam J, Weber D, Schlenk RF, Kieser M: Enhancing single-arm phase II trials by inclusion of matched control patients, unpublished).

In Germany, patients with r/r-AML are usually referred to academic hospitals. Patients under treatment in the participating centers will be asked to participate in the study. Recruitment will take place in up to 20 academic centers registered in the Study Alliance Leukemia (SAL) group (http://​www.​sal-aml.​org/​). During the study conduct, participating centers (see Additional file 2) are contacted by study monitors and the medical coordinator monthly to promote patient recruitment. Furthermore, after protocol amendments or upon relevant updates during the study, a newsletter will be sent to all participating centers. Conferences with the participating centers of the SAL network are held regularly to share information regarding therapy responses and complications seen within the study. Expecting at least 5 potentially eligible patients per year and center, approximately 2 years are required to recruit the anticipated number of patients, when accounting for a consecutive study center initiation and non-participation.

Patients have to provide written informed consent and must meet all inclusion criteria before any protocol-specific procedures are performed.

Each patient must be registered in the electronic case report form (eCRF). During registration, a unique patient ID (PAT-ID) is assigned. Following registration, all eligible patients are upfront 1:1 randomized. The randomization step is triggered by the study sites using a centralized web-based tool (randomizer.at) configured by the responsible data management team. Randomization is performed stratified according to previous treatment with midostaurin. Forty patients are allocated to each arm. Twenty or more centers are selected to recruit the intended number of patients. The treatment arms are (a) the MRD-triggered arm and (b) the prophylactic arm.

In case a patient prematurely withdraws from the trial or is lost to follow-up before measurement of the primary endpoint, this patient’s response is set to missing. In case a patient dies due to any cause before CR/ or CR with incomplete hematological recovery (CRi) or complete remission with partial recovery of peripheral blood counts (CRh) measurement, the primary endpoint will be set to “no Composite Remission”.

A patient must be withdrawn from the trial (i) at any time at the patient’s own request, (ii) after salvage therapy if the patient fails to obtain CR/Cri/CRh, (iii) at any time if unacceptable toxicity necessitating cessation of treatment is observed, (iv) at any time if there are changes in the medical status of the patient that compromise the patient’s safety or if the investigator considers that the withdrawal is in the patient’s best interest, (v) in case of pregnancy, (vi) at any time a patient’s protocol compliance turns out insufficient, and (vii) if patient is lost to follow-up.

If any of the aforementioned criteria are met, and patients agree to follow-up, previous unresolved AEs will continuously be followed. If cases iii and iv are met we will continue with the follow-up of the patient until the end of the study. However, if the patient withdraws from the trial and also from consent for disclosure of future information (e.g., follow-up visits), further data collection is prohibited.

The primary endpoint of the study is the achievement of CR, Cri, or CRh after salvage therapy with HAM in combination with quizartinib. CR is defined as bone marrow blasts <5%, absence of circulating blasts and blasts with Auer rods, absence of extramedullary disease, neutrophil count ≥1000/μl, and platelet count ≥100,000/μl. CRi is defined as CR with residual neutropenia (neutrophils < 1000/μl) or thrombocytopenia (platelets < 100 000/μl) and CRh is defined as a CR with partial recovery of peripheral blood counts with residual neutropenia (neutrophils < 500/μl) or thrombocytopenia (platelets < 50,000/μl).

Refractory disease or treatment failure is defined as failure to achieve CR, Cri, or CRh, presence of Auer rods, or appearance of new or worsening extramedullary disease after one course of intensive treatment. Relapse after CR, CRi, or CRh is characterized by ≥5% blasts in the bone marrow aspirate and/or biopsy not attributable to any other cause, the reappearance of leukemic blasts in the peripheral blood, the new appearance of extramedullary leukemia, or presence of Auer rods. Platelet and neutrophil counts for the assessments of CR, CRi, and CRh are delineated based on the 2017 recommendations from the International Working Group commissioned by the European Leukemia Net [26].

Secondary survival endpoints are event-free survival (EFS), defined from randomization until one of the following events, whichever occurs first: (a) failure to obtain CR or CRi or CRh after Q-HAM therapy, (b) relapse from CR/Cri/CRh, or (c) death from any cause. Patients without an event are censored at the last follow-up. Overall survival (OS) is defined as the time from randomization until death from any cause and relapse-free survival (RFS) is measured from the first CR/Cri/CRh to the time of recurrence of the disease or death from any cause, whichever occurs first. Patients without event are censored at the last date of follow-up. Cumulative incidence of relapse (CIR) is defined as the time from the achievement of a CR/CRi/CRh after salvage therapy to the recurrence of the disease, whereby death from any cause is a competing event. Cumulative incidence of death (CID) is defined as the time from the achievement of a CR/CRi/CRh after salvage therapy to death from any cause whereby recurrence of the disease is a competing event.

Further secondary endpoints are patient-reported outcomes (PROs). PROs include assessments of (a) health-related quality of life (QoL), calculated as the EORTC QLQ-C30 Summary Score [27]; (b) the quality of sleep or sleep disorders, calculated with the “Sleep Quality Index” from the PSQI according to the corresponding scoring guidelines [28]; and (c) anxiety and depression, calculated from the PHQ-4 according to the corresponding scoring manual [29].

All adverse events (AEs) that occur after the screening visit (or as soon as the medical history of the patient has been examined) are documented. The period of observation ends with the last study visit. All patients who have AEs, whether considered associated with the use of the investigational medical products or not, are monitored for outcome determination. AEs are collected systematically based on patients’ answers to applicable questions by the investigator, but spontaneous reports are recorded as well if AE criteria apply. For analyses, AEs are centrally coded according to MedDRA. The Data Monitoring Committee (DMC) is composed of three independent experts, meets regularly to review all data relevant to safety, and provides the sponsor with recommendations regarding trial modification, continuation, or termination.

Audits are planned to be performed based on regular risk-based evaluation. Regulatory authorities and auditors authorized by the sponsor may request access to all source documents, the CRF, and other trial documentation. Investigators are contractually bound to enable direct access to these documents and to support audit activities.

Decisions regarding protocol amendments will be taken by the study core team encompassing the coordinating investigator, trial coordinator, trial statistician, medical coordinator, and data management. Meetings for reviewing all available findings and information are scheduled every 2 weeks.

The samples will be collected at the following time points: (i) at baseline, (ii) after salvage therapy, (iii) after consolidation therapy, (iv) after each cycle of consolidation therapy, (v) after each cycle of maintenance therapy, (vi) at the end of treatment, (vii) every 3 months during observational follow-up, and at end of study. Provided patients’ applicable informed consent is given, biological samples are stored to develop further knowledge and understanding of diagnostic, prognostic and predictive markers present in AML patients.

All data required as per the study protocol, including clinical and laboratory data, are documented by the investigator or an authorized member of the study team in the medical record of the patient and in the eCRF. Access to the eCRF is password protected and an audit trail is in place. Any entries are tracked and locked to prevent further editing. The investigator at the clinical site is responsible for ensuring that all sections of the eCRF are completed correctly. Entries are checked for plausibility and consistency via eCRF-inherent edit checks and visually by the monitors where necessary. Implausibility and missing entries are queried and to be clarified with the responsible investigator. All relevant documents and data collected within the study will be archived for at least 10 years after termination of the study.

For all patients in the prophylactic arm, treatment ends after the 12th cycle of maintenance therapy (EOT). For all patients randomized to the MRD-triggered arm, who did not cross-over in the prophylactic arm, treatment ends after the 2nd cycle of consolidation therapy (EOT). After EOT patients are routinely followed-up and treated as per the standard of care at the discretion of the treating physician. The period of observation (and the overall study) ends for all patients when the last patient being included and alive has been followed for at least 730 days (2 years) counted from this patient’s day 1 (EOS).

Study monitoring is done by the Heidelberg Clinical Studies Coordination Center (KKS). The first monitoring visit at each study center is scheduled to occur at the end of the second patient’s induction therapy. Further monitoring visits at each study site will depend on the (i) recruitment success of study participants, (ii) the monitor’s assessment of the trial site’s compliance with applicable stipulations (e.g., number and severity of protocol deviations or deficiencies detected during study visits), (iii) the deficiencies detected via Central Data Review, and (iv) the assessment of the coordinating team. The monitoring is carried out according to a monitoring manual giving comprehensive guidance on monitoring activities (Source Data Verification rules, corrective and preventive actions, documentation of protocol violations, escalation of findings, etc.).

All the procedures set out in the trial protocol are designed to ensure that all persons involved in the trial abide by Good Clinical Practice (GCP) and the ethical principles described in the current version of the Declaration of Helsinki. The trial is carried out in keeping with local legal and regulatory requirements. The Coordinating Investigator ensures that all trial personnel are adequately trained and maintains a list to whom he has delegated significant trial-related duties.

Before being admitted to the clinical trial, all patients must consent in writing to their participation after having understood the nature, scope, and possible consequences of the clinical trial. Patients are requested to consent to biobanking and to secondary use of their pseudonymized data for yet not determined further scientific analyses. The data obtained in the course of the trial are treated pursuant to the General Data Protection Regulation (EU-DSGVO, EU 2016/679) and the Data Protection Law of the Federal State (Landesdatenschutzgesetz), and the § 40 (2a) AMG. For protection of these data, organizational procedures are implemented to prevent the distribution of data to unauthorized persons.

All planned substantial changes to the study (protocol amendments) are to be submitted to the EC and the competent federal authority requesting their approval. Records of relevant communication with the EC and the regulatory authorities are kept by the coordinating investigator.

After the publication of the complete trial, access to selected raw data is intended. This must be done in accordance with the European data protection act and informed consent given by the patients.

The results from this trial will be presented at national (e.g., annual meeting of the German Society of Hematology/Oncology); meetings of the Competence Net “Acute and Chronic Leukemias”; and international meetings (e.g., meetings of the European Leukemia-Net; annual congresses of the European Hematology Association, the American Society of Hematology and the American Society of Clinical Oncology). The full results will be published in high-impact peer-reviewed medical journals in accordance with ICMJE guidelines [30]. It is planned to make trial data available for re- and meta-analyses, after the trial is completed and results are published.

To assess efficacy with respect to the primary endpoint CR/Cri/CRh of Q-HAM during salvage therapy, we will compare the patients in the study to historical controls from previous clinical trials based on the matched threshold-crossing approach [10, 25] (Krisam J, Weber D, Schlenk RF, Kieser M: Enhancing single-arm phase II trials by inclusion of matched control patients, unpublished). Matched controls are drawn from datasets available from individual participant data (IPD) meta-analyses as previously published [6, 7]. The matched threshold-crossing approach is a novel approach to enhance the classic single-arm trial design by including matched historical control patients [31]. This design overcomes common disadvantages of single-armed and small randomized studies [32], since the expected outcome of the observed study population can be adjusted based on the matched controls with a comparable distribution of known prognostic and predictive factors. Furthermore, balanced treatment groups lead to stable statistical models. The proposed adaptive design encompasses two stages with an interim analysis after the first stage [31, 32]. At the interim analysis, historical control patients are matched to the enrolled intervention patients with regard to known prognostic and predictive factors. The treatment effect of the intervention as compared to the control group is then estimated based on the enrolled intervention patients and the selected historical control patients. Results of the interim analysis will be presented to the Data Monitoring Committee (DMC) that will advise the Steering Committee of the trial either to terminate or to continue the trial. In case the estimated treatment effect is below the pre-specified threshold of OR_stop=1.3, the trial will be stopped for futility. However, in case the estimated treatment effect is above the pre-specified threshold, the sample size for the second stage of the trial will be recalculated to obtain a sufficiently high conditional power based on the results of the interim analysis and the trial continues to the second stage with the recalculated sample size.

In the proposed design, matched historical control patients from two large IPD meta-analyses datasets of r/r-AML patients will be matched by age, and high-risk cytogenetics in refractory patients. Additionally, first CR duration shorter than 18 months will be used as matching factor in relapsed patients. After the enrolment of 20 patients, the a priori unknown matching rate and the number of matched controls per patient in the intervention group are determined by the interim analysis using an iterative procedure (Krisam J, Weber D, Schlenk RF, Kieser M: Enhancing single-arm phase II trials by inclusion of matched control patients, unpublished). An (non-binding) interim stop for futility is done in case the odds ratio, estimated using a logistic regression model adjusting for the matching parameters, does not exceed a threshold of 1.3. Fitting a logistic regression model on both the trial patients and matching historical controls allows to estimate a treatment effect for the novel intervention as compared to standard treatment [32].

If the trial continues to a second stage, a sample size recalculation is performed using a conditional power argument taking the number of matched partners and observed treatment effect into account. In the final analysis, historical control patients are also matched to the patients from the second stage, and the p-values from the two stages are combined using the inverse normal approach and tested at a one-sided significance level of α=0.05.

The sample size required for a fixed design, assuming CR/CRi/CRh rates of 0.55 for the enrolled patients and 0.30 for the historical controls, would be a minimum of 60 assuming a logistic regression is performed at one-sided significance level α=0.05, the aspired power is 0.8, and number of matching partners per intervention patient is at least 1. Using this sample size as a starting point for several simulation studies, it was assessed that the maximal enrolment of 80 patients generated a desirable power of the used adaptive design for most scenarios. The trial sample size is recalculated based on a conditional power argument in an interim analysis after enrolling 20 patients and will hence range between 20 and 80. The assumed CR/CRi/CRh rate of 0.3 for historical controls was estimated based on two large IPD meta-analyses datasets of r/r-AML patients (“historical control”) [6, 7], while the assumed CR/Cri/CRh rate of 0.55 for the patients treated quizartinib was determined to reflect a treatment effect that can be considered as clinically relevant. The sample size for the fixed design was calculated using the software PASS v16, while the simulations we used to evaluate the adaptive design were done using R v3 (http://​r-project.​org). Randomized patients who do not receive the IMP quizartinib will be replaced. All patients receiving at least one dose of the IMP quizartinib must not be replaced. Due to the underlying disease, loss to follow-up is expected to be negligible.

For the analysis of the primary endpoint, a logistic regression model will be applied to assess the odds ratio of CR/CRi/CRh rate after salvage therapy for patients receiving Q-HAM versus matched controls not receiving Q-HAM. The null hypothesis H0: OR≤1 is tested against its alternative H1: OR>1 at a one-sided significance level of α=0.05. The logistic regression analysis will be adjusted for age, high-risk cytogenetics (yes/no), and CR1 duration <18 months (yes/no/not applicable due to refractory disease). Due to the adaptive nature of the trial, two separate logistic models will be fitted for the two trial stages and be combined using the inverse normal function approach. Determination of p-values, point, and interval estimates will take the adaptive nature of the trial into account. Concerning the secondary endpoints: EFS, OS, and RFS endpoints will be analyzed using a Cox regression model and Kaplan-Meier plots, CIR, and CID will be analyzed according to Gray 1988 [33] and via cumulative incidence plots, and quality of life will be analyzed descriptively according to the corresponding guidelines and EORTC recommendations. Safety endpoints will be analyzed descriptively. Analyses of the primary and secondary endpoints will be based on the full analysis population (all randomized patients with treatment groups assigned in accordance with the randomization, regardless of the treatment actually received (intention to treat (ITT)). For patients with incomplete follow-up, time to last follow-up date will be used as the censoring time in the analysis of time-to-event data. Otherwise, no imputation of missing data will be conducted. For the biometricians, patient treatment shall remain blinded from the time of randomization until the final database lock. The interim analysis and sample size recalculation will be conducted by an independent biometrician who is not involved in the conduct of the trial. Statistical analysis will be performed using SAS v9.4 or higher [34].