EUnetHTA Joint Action 3 WP4 · 1.1.1 International classification of diseases (Version 10 [ICD-10]...

Dec2015 ©EUnetHTA, 2015. Reproduction is authorised provided EUnetHTA is explicitly acknowledged 1

EUnetHTA Joint Action 3 WP4

Version v1.0, 30/07/2020

This Joint Assessment is part of the project / joint action ‘724130 / EUnetHTA JA3’ which has received funding from the European Union’s Health Programme (2014-2020)

Relative effectiveness assessment of pharmaceutical technologies

GLASDEGIB IN COMBINATION WITH LOW-DOSE CYTARABINE, FOR THE

TREATMENT OF NEWLY DIAGNOSED DE NOVO OR SECONDARY ACUTE

MYELOID LEUKAEMIA (AML) IN ADULT PATIENTS WHO ARE NOT

CANDIDATES FOR STANDARD INDUCTION CHEMOTHERAPY

Project ID: PTJA12

PTJA12 - Assessment Report Glasdegib for de novo or secondary acute AML

July 2020 EUnetHTA Joint Action 3 WP4 1

DOCUMENT HISTORY AND CONTRIBUTORS

Version Date Description

V0.1 09/06/2020 First draft

V0.2 10/07/2020 Input from dedicated reviewers has been processed

V0.3 28/07/2020 Input from medical editor and manufacturer(s) has been processed

V1.0 30/07/2020 Final Assessment Report

Disclaimer

The content of this Assessment Report represents a consolidated view based on the consensus within the Authoring Team; it cannot be considered to reflect the views of the European Network for Health Technology Assessment (EUnetHTA), EUnetHTA’s participating institutions, the European Commission and/or the Consumers, Health, Agriculture and Food Executive Agency or any other body of the European Union. The European Commission and the Agency do not accept any responsibility for use that may be made of the information it contains.

Assessment Team

Author(s) Federation of Social Insurances (DVSV), Austria

Co-Author(s) National Centre for Pharmacoeconomics (NCPE), Ireland

Dedicated Reviewer(s) French National Authority for Health (HAS), France Swiss Network for HTA (SNHTA), Switzerland National Authority of Medicines and Health Products (INFARMED), Portugal



Further contributors

Manufacturer [v0.2]

Pfizer INC Preparation of the Submission Dossier Factual accuracy check

Medical editor [v0.2]

National Centre for Pharmacoeconomics (NCPE), Ireland and Zorginstituut Nederland (ZIN), Netherlands

Medical editing was performed in house by: - NCPE performed grammar and spelling check - ZIN project managers performed layout updates

Patient(s) / patient organisation(s) / citizens

Leukaemia Care, United Kingdom Association of Cancer Patients in Finland, Finland Deutsche Leukämie- & Lymphom-Hilfe, Germany

Input in response to the Open Call for Patient Input published on 11th October, 2019

Project Management

Zorginstituut Nederland (ZIN), Netherlands

Coordination between involved parties throughout the assessment

Conflict of interest

All authors, co-authors, dedicated reviewers, observers, external experts (health care professionals, patients or patient representatives) involved in the production of this assessment have declared they have no conflicts of interest in relation to the technology and comparator(s) assessed according to the EUnetHTA declaration of interest (DOI) form. Conflict of Interest was evaluated following the EUnetHTA Procedure Guidance for handling DOI form (https://eunethta.eu/doi).

Copyright

EUnetHTA assessments are published under a “CC/BY/NC” Creative Commons Licence.

How to cite this assessment

Please cite this assessment as follows:

EUnetHTA PTJA12. Authoring Team. Glasdegib in combination with low-dose cytarabine, for the treatment of newly diagnosed de novo or secondary acute myeloid leukaemia (AML) in adult patients who are not candidates for standard induction chemotherapy. Joint Assessment. Diemen (The Netherlands): EUnetHTA; 2020. [date of citation]. 80 pages. Report No.: PTJA12. Available from: https //www.eunethta.eu

Contact the EUnetHTA Secretariat [email protected] with inquiries about this assessment.

https://eunethta.eu/wp-content/uploads/2019/11/EUnetHTA-Procedure-Guidelines-DOI.pdf

https://eunethta.eu/wp-content/uploads/2019/11/EUnetHTA-Procedure-Guidelines-DOI.pdf

https://eunethta.eu/doi

https://creativecommons.org/licenses/by-nc/2.0/es/deed.en

mailto:[email protected]



TABLE OF CONTENTS

DOCUMENT HISTORY AND CONTRIBUTORS .....................................................................................1

TABLE OF CONTENTS ...........................................................................................................................3

LIST OF TABLES AND FIGURES...........................................................................................................4

LIST OF ABBREVIATIONS .....................................................................................................................6

EXECUTIVE SUMMARY OF THE ASSESSMENT OF GLASDEGIB .....................................................9

INTRODUCTION .....................................................................................................................................9 OBJECTIVE AND SCOPE .........................................................................................................................9 METHODS .......................................................................................................................................... 10 RESULTS ........................................................................................................................................... 11 DISCUSSION ...................................................................................................................................... 12 CONCLUSION ..................................................................................................................................... 12

1 BACKGROUND .............................................................................................................................. 15

OVERVIEW OF THE DISEASE OR HEALTH CONDITION ....................................................................... 15 CURRENT CLINICAL PRACTICE ...................................................................................................... 17 FEATURES OF THE INTERVENTION ................................................................................................. 20

2 OBJECTIVE AND SCOPE .............................................................................................................. 23

3 METHODS ....................................................................................................................................... 25

INFORMATION RETRIEVAL ............................................................................................................. 25 DATA EXTRACTION ....................................................................................................................... 27 RISK OF BIAS ASSESSMENT .......................................................................................................... 27 CERTAINTY OF EVIDENCE ............................................................................................................. 27 RESULTS AND ANALYSES OF INCLUDED STUDIES ............................................................................ 27 PATIENT INVOLVEMENT ................................................................................................................ 28

4 RESULTS ........................................................................................................................................ 29

INFORMATION RETRIEVAL ............................................................................................................. 29 STUDIES INCLUDED IN THE ASSESSMENT ....................................................................................... 29 EXCLUDED STUDIES ..................................................................................................................... 29 CHARACTERISTICS OF BRIGHT AML 1003 .................................................................................. 30 STATISTICS ................................................................................................................................. 37 OUTCOMES INCLUDED .................................................................................................................. 38 PARTICIPANT FLOW ...................................................................................................................... 40 RISK OF BIAS ............................................................................................................................... 41 EXTERNAL VALIDITY ..................................................................................................................... 42

RESULTS ON CLINICAL EFFECTIVENESS AND SAFETY .............................................................. 43 INDIRECT EVIDENCE ............................................................................................................. 53

5 PATIENT INVOLVEMENT .............................................................................................................. 57

6 DISCUSSION .................................................................................................................................. 58

7 CONCLUSION ................................................................................................................................ 60

8 REFERENCES ................................................................................................................................ 61

APPENDIX 1: GUIDELINES FOR DIAGNOSIS AND MANAGEMENT ............................................... 64

APPENDIX 2: CERTAINTY OF EVIDENCE ......................................................................................... 66

APPENDIX 3: STUDY DETAILS FOR INDIRECT COMPARISONS ................................................... 73

APPENDIX 4: EVIDENCE GAPS ......................................................................................................... 79



LIST OF TABLES AND FIGURES

Tables

Table 0.1. Scope of the assessment ........................................................................................................9 Table 0.2. Summary of findings of glasdegib + LDAC vs. LDAC alone in AML patients according to

GRADE (6) ...................................................................................................................................... 14 Table 1.1. ELN Clinical definitions for outcomes in clinical trials .......................................................... 17 Table 1.2. Features of the intervention .................................................................................................. 20 Table 1.3. Administration and dosing of the technology ....................................................................... 22 Table 2.1. Scope of the assessment ..................................................................................................... 23 Table 3.1. Summary of information retrieval and study selection ......................................................... 26 Table 4.1. List of included studies ......................................................................................................... 29 Table 4.2. Cytogenetic risk conversion chart for AML patients ............................................................. 31 Table 4.3. Characteristics of the B1371003 trial / BRIGHT AML 1003 ................................................. 32 Table 4.4. Characterisation of the intervention and comparator of the B1371003 trial / BRIGHT AML

1003 ................................................................................................................................................. 33 Table 4.5. Information on the course of BRIGHT AML 1003 (including planned duration of follow-up)

......................................................................................................................................................... 34 Table 4.6. Baseline characteristics of BRIGHT AML 1003 ................................................................... 35 Table 4.7. Number (%) of AML patients meeting specified unfit criteria at baseline ............................. 36 Table 4.8. Matrix of outcomes in BRIGHT AML 1003 ........................................................................... 38 Table 4.9. Definition of hematologic responses to treatment for AML .................................................. 39 Table 4.10. Participant flow of AML subgroup ....................................................................................... 41 Table 4.11. OS results for BRIGHT AML 1003 ..................................................................................... 43 Table 4.12. Results summary for survival rates and survival probabilities for BRIGHT AML 1003 ...... 44 Table 4.13. Summary of OS for AML patients according to their cytogenetic risk by IVRS stratification

in BRIGHT AML 1003 ...................................................................................................................... 45 Table 4.14. Results summary for transfusion independence for BRIGHT AML 1003 ........................... 49 Table 4.15. Results for CR for BRIGHT AML 1003 ............................................................................... 49 Table 4.16. Results summary for objective response for BRIGHT AML 1003 ...................................... 50 Table 4.17. Results summary for total AEs for BRIGHT AML 1003 ...................................................... 52 Table A1. Overview of guidelines used for this assessment ................................................................. 64 Table A2. Overview of cytogenetic and molecular risk stratification tools used in B1371003, AZA-

AML-001 and DACO-016 ................................................................................................................ 65 Table A3. Risk of Bias 2 Assessment ................................................................................................... 66 Table A4. GRADE Evidence profile ....................................................................................................... 68 Table A5. Results summary for AEs (any CTCAE grade) occurring in ≥ 20 % of patients separately

by MedDRA SOC and PT for BRIGHT AML 1003 .......................................................................... 70 Table A6. Characteristics of the studies included - indirect comparison ............................................... 73 Table A7. Characteristics of the interventions and comparators – indirect comparison ....................... 74 Table A8. Information on the course of the studies (including planned duration of follow-up) – indirect

comparison ...................................................................................................................................... 75 Table A9. Baseline characteristics of the study populations (relevant subgroup or whole study

population) – indirect comparison ................................................................................................... 76 Table A10. Recommendations for research .......................................................................................... 79



Figures

Figure 1.1. Place of glasdegib in the treatment pathway (as per applicant’s submission).................... 19 Figure 4.1. Study design of B1371003 .................................................................................................. 30 Figure 4.2. AML subgroup (BRIGHT AML 1003) of B1371003 ............................................................. 30 Figure 4.3. Risk of bias for B1371003 trial ............................................................................................ 42 Figure 4.4. Kaplan-Meier plot of OS for BRIGHT AML 1003 ................................................................ 44 Figure 4.5. Forest plot of OS for BRIGHT AML 1003 ............................................................................ 46 Figure 4.6. Partitioned survival curve for glasdegib + LDAC................................................................. 47 Figure 4.7. Restricted mean duration of health states with TOX and REL downweighed by 0.5 ......... 48



LIST OF ABBREVIATIONS

AE Adverse event

AHD Antecedent haematologic disease

ALT Alanine aminotransferase

AML Acute myeloid leukaemia

ARR Absolute risk reduction

AST Aspartate aminotransferase

ATC Anatomical therapeutic chemical [classification system]

ATMP Advanced therapy medicinal product

BCL-2 B-cell lymphoma-2

BM Bone marrow

BSC Best supportive care

CENTRAL Cochrane Central Register of Controlled Trials

ChT Chemotherapy

CI Confidence interval

CID Cumulative incidence of death

CIR Cumulative incidence of relapse

CK Creatine kinase

CMML Chronic myelomonocytic leukaemia

CMH Cochran–Mantel–Haenszel test

CPG Clinical practice guideline

CR Complete remission

CRi CR with incomplete haematologic recovery

CRMRD- CR without minimal residual disease

CRF Case report form

CSR Clinical study report

CTCAE Common terminology criteria for adverse events

DOI Declaration of interest

ECA EUnetHTA confidentiality arrangement

ECG Electrocardiogram

ECOG PS Eastern Cooperative Oncology Group performance status

EFS Event-free survival

ELN European Leukaemia Net

EMA European Medicines Agency

Embase Excerpta Medica database

EMD Extramedullary disease

EORTC QLQ-C30 European Organisation for Research and Treatment of Cancer Core Quality of Life Questionnaire

EPAR European public Assessment Report

ESMO European Society for Medical Oncology

EU CTR European Union Clinical Trials Register

EUnetHTA European Network of Health Technology Assessment

FAB French-American-British

FAS Full analysis set

FDA Food and Drug Administration

G-CSF Granulocyte colony-stimulating factor

GRADE Grading of Recommendations Assessment, Development and Evaluation

HCT Hematopoietic cell transplantation

Hh Hedgehog

HMAs Hypomethylating agents



HR Hazard ratio

HRQoL Health-related quality of life

HSCT Haematopoietic stem cell transplantation

HTA Health Technology Assessment

HTAi Health Technology Assessment international

IC Induction chemotherapy

ICD International classification of diseases

ICTRP International Clinical Trials Registry Platform

IPSS International Prognostic Scoring System

IRS Interactive registration system

ITC Indirect treatment comparison

ITD Internal tandem duplication

ITT Intention-to-treat

IVRS Interactive voice response system

LDAC Low-dose cytarabine

MAH Market authorization holder

Max Maximum

MD Mean difference

MDS Myelodysplastic syndrome

MEDLINE Medical Literature Analysis and Retrieval System Online

MeSH Medical subject headings

MFC Multiparameter flow cytometry

Min Minimum

MLFS Morphologic leukaemia-free state

N Number of patients

NA Not applicable

NCCN National Comprehensive Cancer Network

NCI National Cancer Institute

n.e. Not estimable

NI No information

NIH US National Institute of Health

n.r. Not reported

OR Odds ratio

ORR Overall response rate

OS Overall survival

P2 FIT Phase 2 single arm component in fit patients

P2 unfit Phase 2 randomized component in unfit patients

PICO Patient, intervention, comparison, outcome

PN Probably no

PP Per protocol

PR Partial remission

PT Preferred term

Pts Patients

PY Probably yes

QoL Quality of Life

Q-TWiST Quality-adjusted time without symptoms of disease progression or toxicity

RAEB Refractory anemia with excess blasts

RCT Randomized controlled trial

RD Risk difference

REA Relative effectiveness assessment

REL Time after progression



RFS Relapse-free survival

RIC Reduced intensity conditioning

RMST Restricted mean survival time

RoB Risk of bias

RP2D Recommended Phase2 dose

RR Risk ratio

RT-qPCR Real-time quantitative polymerase chain reaction

SAE Serious adverse event

SAP Statistical analysis plan

SC Supportive care

SD Standard deviation

SEER Surveillance, Epidemiology and End Results

SLR Systematic literature search

SMD Standardized mean difference

SMO Smoothened

SmPC Summary of product characteristics

SOC System organ class

SOP Standard operating procedure

SMQ Standardised MedDRA query

SR Study report

STC Simulated treatment comparison

SWOG Southwest Oncology Group

TEAE Treatment-emergent adverse events

TOX Time with grade ≥ 3 toxicity

TWiST Time without symptoms of disease progression or toxicity

ULN Upper limit of normal

vs. versus

WBC White blood cell

WHO World Health Organisation

WP4 Work Package 4

Y Yes

Yrs Years



EXECUTIVE SUMMARY OF THE ASSESSMENT OF GLASDEGIB

Introduction

Acute myeloid leukaemia (AML) is a group of heterogeneous hematopoietic stem cell disorders which are characterised by incomplete maturation of blood cells and reduced production of other normal hematopoietic cells. Symptoms of AML include appetite loss, weight loss, fatigue, breathlessness, confusion, bruising, bleeding and frequent infections.

The incidence of AML increases with age, from ~1.3 per 100,000 population in patients less than 65 years old, to 12.2 cases per 100,000 population in those over 65 years (1). Although advances in the treatment of AML have led to significant improvements in outcomes for younger patients, prognosis in the elderly who account for the majority of new cases remains poor.

The backbone of therapy remains a combination of cytarabine- and anthracycline-based regimens with allogeneic stem cell transplantation for eligible candidates (1). Standard induction therapy with 3 days anthracycline (daunorubicin or idarubicin) and 7 days cytarabine (“7+3” regimens) is the first line therapy in those patients who can tolerate it. Elderly patients are often unable to tolerate such regimens, and carry a particularly poor prognosis. For patients not eligible for standard induction and consolidation chemotherapy, the European Society for Medical Oncology (ESMO) guidelines recommend enrolment in a clinical trial. However, if no clinical trial is available, decitabine or azacitidine are currently the first choice in newly diagnosed unfit (i.e. not eligible for standard induction and consolidation chemotherapy) AML patients. Further treatment options are low dose cytarabine (LDAC) or best supportive care (BSC) for patients who cannot tolerate any antileukaemic therapy, or who do not wish any therapy (2, 3). However, no commonly accepted definition exists for patients ineligible for intensive induction therapy.

Objective and scope

The objective of this assessment was to assess glasdegib in combination with LDAC, for the treatment of newly diagnosed de novo or secondary AML in adult patients who are not candidates for standard induction chemotherapy (approved indication) (see Table 0.1).

Table 0.1. Scope of the assessment

Description Assessment scope

Population Adult patients with newly diagnosed de novo or secondary acute myeloid leukaemia (AML) who are not eligible for standard induction chemotherapy ICD-10: 92.0 MeSH-terms: Leukemia, Myeloid, Acute Tree Number(s): C04.557.337.539.275 MeSH Unique ID: D015470

Intervention Glasdegib 100 mg orally once daily continuously (on days 1 to 28) in combination with low-dose cytarabine (LDAC) Synonyms for glasdegib: PF-04449913

Comparison Azacitidine;

Decitabine;

LDAC ;

Best Supportive Care (BSC: hydroxyurea, transfusion support)a.

Outcomes Effectivenessb:

Overall survival;

Health-related quality of life;

Transfusion independency;

Objective Response: o Overall Response Rate: Complete Remission (CR) + CR with incomplete blood count

recovery + morphologic leukaemia-free state. Safetyc:

Serious adverse events (AEs);

Grade ≥3 AEs;




Fatal AEs;

Overall AE;

AEs of special interest (febrile neutropenia, pneumonia, haemorrhage, QT Interval Prolongation);

Treatment discontinuation due to AE.

Study type Effectiveness:

Randomised controlled trials (RCT). Safety: If suitable evidence syntheses (SRs/HTA reports) are available:

Evidence syntheses (SRs/HTA reports); and

Primary studies (as described in next bullet) published after the last search date of the latest SR/HTA document.

If suitable evidence syntheses (SRs/HTA reports) are NOT available:

Randomised controlled trials;

Non-randomised controlled trials;

Observational studies. a All comparators were explicitly mentioned by patient organizations. b Including data at 6 and 12 months. c Including data at 90 days. Abbreviations: AE=adverse event; BSC=best supportive care; CR=complete remission; HTA=health technology assessment; LDAC=low-dose cytarabine; QT=Q-TWiST; RCT=randomized control trial; SR=study report.

Methods

The present assessment was based on the data and analyses included in the Submission Dossier prepared by the marketing authorisation holder (MAH). As part of the assessment, the completeness of the data and analyses in the Submission Dossier was verified. Additionally, methods for data analyses and syntheses as applied by the MAH [Pfizer INC] were checked against the requirements for the Submission Dossier and applicable EUnetHTA guidelines (https://www.eunethta.eu/methodology-guidelines/) and assessed with regard to scientific validity.

Literature search and assessment approach

The Submission Dossier submitted by the MAH included a systematic literature review (SLR) to identify randomised controlled trials (RCTs) in 3 bibliographic databases: Embase, MEDLINE and the Cochrane Central Register of Controlled Trials (CENTRAL). These were complemented by searches in three clinical trial registers. All searches were performed on 17th of January 2020. Search terms related to glasdegib were used, in Embase and MEDLINE, these were combined with validated study design search filters to identify RCTs. The MAH did not use any restrictions related to language or publication date.

Additional searches were developed for the comparators azacitidine and decitabine by combining search terms related to the comparator with those related to the target condition, using a validated search filter for RCTs where appropriate. Selection criteria were tuned to identify relevant RCTs evaluating glasdegib + LDAC versus the comparators identified in the scoping phase. Since no studies were available which directly compared glasdegib to relevant treatment options other than LDAC as defined in the project scope, selection criteria in the additional searches were tuned to identify RCTs evaluating decitabine versus LDAC or azacitidine versus LDAC, to allow for indirect comparisons of glasdegib to azacitidine and decitabine via the common comparator LDAC. Screening in all phases was performed in duplicate and independently, disagreements were resolved by involvement of a third reviewer.

The information specialist critically assessed the method and replicability of the information retrieval submitted, and verified completeness of the search in study registries. Completeness of the evidence base with regard to direct comparisons of glasdegib + LDAC versus a comparator was verified against the studies included in the regulatory Assessment Report.

https://www.eunethta.eu/methodology-guidelines/

https://www.eunethta.eu/methodology-guidelines/



The information specialist concluded that this SLR was well constructed and valid, so that the risk to miss relevant studies was deemed to be low. Details of the search can be found in the core Submission Dossier in section 5 “Clinical effectiveness and safety” and Appendix B (4). Since no flaws were identified in the MAH search strategy, no supplementary searches were performed.

Information used to assess the clinical effectiveness and safety was extracted from the core Submission Dossier and verified against the clinical study report (CSR) and other original documentation provided in the Submission Dossier.

The revised Cochrane risk-of-bias tool for randomized trials (RoB 2) quality rating tool (version of 22nd of August 2019) was used to assess the risk of bias in randomised trials for each relevant endpoint. The results of the risk of bias assessment were used in GRADE for rating the certainty of evidence on outcome level.

Three patient organisations provided input in response to the Open Call for Patient Input published on 11th of October 2019. Leukaemia Care (United Kingdom), Association of Cancer Patients in Finland (Finland) and Deutsche Leukämie- & Lymphom-Hilfe (Germany) provided their perspectives on the impact of AML, patient relevant outcomes and current therapy options. Details can be found in section 5 of this report.

Results

The SLR performed by the MAH identified a single relevant RCT evaluating glasdegib + LDAC versus LDAC alone - the B1371003 trial. Since this study included both AML and MDS patients, the MAH only included results from the AML subgroup i.e. BRIGHT AML 1003. In addition, two trials were identified by the MAH for indirect comparisons of decitabine or azacitidine via the common comparator LDAC.

Phase 2 results from a subgroup of 116 AML patients from the B1371003 trial, a single, randomised controlled phase 1b/2 study, which enrolled overall 132 patients (116 AML patients and 16 MDS patients) were submitted for this assessment. This open-label, multicentre trial compared glasdegib + LDAC to LDAC alone in adult patients with newly diagnosed, previously untreated AML or high-risk MDS considered not eligible for intensive chemotherapy. For being ineligible for intensive chemotherapy at least one of several criteria had to be fulfilled: age ≥ 75 years, serum creatinine >1.3 mg/dL, severe cardiac disease, or ECOG PS = 2. Patients were randomized in a 2:1 ratio to glasdegib + LDAC or to LDAC alone. The primary outcome was overall survival (OS) in the full analysis set (FAS). Other endpoints were disease response (e.g., complete remission (CR)) and adverse events (AEs). Further endpoints deemed relevant by the authoring team were transfusion independency and health-related quality of life (HrQoL).

However, the core Submission Dossier contains information only for the subgroup of 116 AML patients (BRIGHT AML 1003) (4). These analyses of the AML patient subgroup were not pre-planned and not controlled for multiple testing but contain the majority of randomised patients and directly correspond to the EMA licence. Baseline characteristics of these AML patients were overall well-balanced concerning age (median age of 77 years in the glasdegib + LDAC arm vs. 76 years in the LDAC alone arm), de novo AML (49 % vs. 47 %) and other characteristics such as median bone marrow blast and duration since histopathologic diagnosis. Descriptively, some imbalances existed concerning gender (male sex 76 % vs. 61 %), ECOG PS of 2 (53 % vs. 47 %), poor cytogenetic risk (according to IVRS 37 % vs. 45 %), serum creatinine > 1.3 mg/dL (19 % vs. 13 %), a history of severe cardiac disease (67 % vs. 53 %) and patients meeting two or more of the unfit criteria (71 % vs. 55 %) (4, 5).

No direct evidence is available comparing glasdegib with any other intervention included in the PICO apart from LDAC. An anchored indirect comparison was conducted by the MAH comparing glasdegib + LDAC (BRIGHT AML 1003) to azacitidine (AZA-AML-001 trial) and decitabine (DACO-016 trial) in terms of OS via the common comparator LDAC. Comparative effectiveness estimates were obtained by using hazard ratios and standard approaches for anchored ITCs (Bucher method), as well as simulated treatment comparisons as supportive analyses. However, this indirect comparison was not included in this assessment, mainly due to clear between-trial differences in patient populations and not fully compatible (common) comparator arms within available trials.



The results from the direct evidence (BRIGHT AML 1003) are as follows (see also Table 0.2):

For AML patients, glasdegib + LDAC showed superior OS (HR [95 % CI]: 0.46 [0.30, 0.72]; p = 0.0004) than LDAC alone (median OS: 8.3 months vs. 4.3 months). This corresponds to an estimated number needed to treat of 4 patients with glasdegib + LDAC instead of LDAC alone to prevent 1 additional death up to month 6. This result was consistent in sensitivity analyses for AML patients only (e.g. CRF based) as well as in the 132 AML + MDS patients initially randomised in the B1371003 trial (HR [95 % CI]: 0.51 [0.34, 0.77]; p = 0.0008). However, the certainty of evidence according to GRADE is considered low;

Exploratory results from a subgroup analysis according to IVRS based cytogenetic risk, descriptively showed a higher median OS improvement in AML patients with good/intermediate cytogenetic risk (median OS improvement of 6.7 months) compared to AML patients with poor cytogenetic risk (median OS improvement of 1.3 months);

Even though further efficacy endpoints (CR, ORR, transfusion independency, quality of survival measured by Q-TWiST) suggest improvements of glasdegib + LDAC in comparison to LDAC alone, these results were considered exploratory as supportive evidence for the primary endpoint. Reasons were that the pre-planned endpoint CR was not controlled for multiple testing and ORR, transfusion independence and quality of survival were not pre-planned endpoints. In addition, quality of life was not proactively collected from patients. Therefore, only retrospective results from a quality of survival analysis suggest that OS improvements for glasdegib + LDAC compared to LDAC alone also translate to quality-adjusted survival time improvements. Due to these limitations and due to the fact that mature OS data are available, the most important patient-relevant endpoint, no risk of bias or certainty of evidence assessment according to GRADE had been performed for these exploratory efficacy outcomes;

Over the entire study period with a median treatment duration of 83 days in the glasdegib + LDAC arm and 41 days in the LDAC alone arm, comparable rates for SAEs, fatal AEs and AEs leading to treatment discontinuation were observed. To adjust for differences in the treatment duration, results for the first 90 days of therapy were also reported: fewer fatal AEs compared to LDAC alone, and comparable rates for SAEs and treatment discontinuations due to AEs were observed with glasdegib + LDAC. However, the certainty of evidence according to GRADE is considered low for fatal AEs and very low for SAEs and AEs leading to treatment discontinuation.

Discussion

Evidence is available for one, phase 1b/2 study including 116 AML and 16 MDS patients, the B1371003 study. Since the EMA approved glasdegib for AML patients only, the core Submission Dossier presented phase 2 results also only for the subgroup of AML patients.

The available evidence has several shortcomings: data is only available for a subgroup of patients from phase 2 of a small, single phase 1b/2 trial. Thus, a substantial overestimation of the treatment effect can be considered likely. Furthermore, the study was not appropriately powered for the primary endpoint, patient reported outcomes were not proactively collected, no information on the random sequence generation exist, the design was open-label and hence the administration of concomitant therapies could have been influenced by knowledge of the assigned intervention.

No direct comparison of glasdegib + LDAC with azacitidine, decitabine, and BSC has been conducted and no valid conclusions can be drawn from an indirect comparison, mainly due to clear between-trial differences in patient populations and not fully compatible (common) comparator arms within available trials.

Conclusion

Phase 2 results from a single phase 1b/2 study, the B1371003 trial, showed a statistically significant improvement in median OS by 4.0 months for the combination of glasdegib + LDAC in comparison to LDAC alone in a subgroup of 116 newly diagnosed AML patients. Further supportive efficacy outcomes suggest an advantage of glasdegib + LDAC over LDAC alone but were considered exploratory and thus as supportive evidence for the primary endpoint. Patient reported outcomes were not proactively collected. AEs were not increased within the combination arm.



Grading of the evidence according to GRADE indicated that the certainty of the evidence is low for OS and fatal AEs and very low SAEs and AEs leading to treatment discontinuation. Whether specific subgroups of the licensed indication can benefit more from glasdegib than the overall AML population remains unclear.

No direct comparisons are available concerning other treatment options for patients not eligible for standard induction therapy. Especially the comparison to azacitidine and decitabine, both preferred over LDAC according to guidelines would be of interest. An indirect comparison submitted by the MAH was not considered to provide reliable results. Thus, no firm conclusions regarding the comparative effectiveness of glasdegib + LDAC vs. azacitine or decitabine or BSC can be drawn.



Table 0.2. Summary of findings of glasdegib + LDAC vs. LDAC alone in AML patients according to GRADE (6)

Outcomes Anticipated absolute effects* (95 % CI) Relative effect (95 % CI)

№ of participants (studies)

Certainty of the evidence (GRADE)**

Risk with LDAC Risk with glasdegib + LDAC

Overall survival (IVRS) 665 per 1.000 (risk of death up to month 6)

395 per 1.000

(280 to 545) (risk of death up to month 6)

HR 0.46

(0.30 to 0.72) 116

(1 RCT) ⨁⨁◯◯ LOW a,b

Health-related quality of life - - - - -

Transfusion independence - - - - -

Serious AEs (study period) 778 per 1.000 786 per 1.000

(638 to 972)

RR 1.01

(0.82 to 1.25) 111

(1 RCT) ⨁◯◯◯

VERY LOW c,d

Fatal AEs (study period) 444 per 1.000 293 per 1.000

(178 to 489)

RR 0.66

(0.40 to 1.10) 111

(1 RCT) ⨁⨁◯◯e,f

LOW

AEs leading to treatment discontinuation (study period)

472 per 1.000 307 per 1.000

(189 to 496)

RR 0.65

(0.40 to 1.05) 111

(1 RCT) ⨁◯◯◯

VERY LOW c,f

a Downgraded by 1 level: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) some concerns exist, and even though the B1371003 trial itself was not stopped early for benefit, no phase 3 trial was conducted due to the observed treatment effect in phase 2 of this small, single phase 1b/2 trial. A substantial overestimate of the treatment effect can be considered likely. b Downgraded by 1 level: optimal information size criterion not fulfilled. In phase 2 of the B1371003 trial, 94 OS events have been observed; however, to obtain a statistical power of at least 80 % to detect a true HR of 0.625 (as used within the SAP for sample size planning) at a conventional alpha level of 5 % (2-sided), 160 OS events would have been necessary (7, 8). c Downgraded by 2 levels: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) high risk of bias exists due to the open-label design and endpoint with subjective component. d Downgraded by 2 levels: very few events and CIs around both relative and absolute estimates of effect that include both appreciable benefit and appreciable harm. e Downgraded by 1 level: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) some concerns exist. f Downgraded by 1 level: few events and wide CI including 1.0 but also relative effects of potentially appreciable benefit. * The risk in the intervention group (and its 95 % CI) is based on the observed risk in the comparison group and the estimated relative effect of the intervention (and its 95 % CI). For OS the risk of death up to month 6 within the control group was estimated by the Kaplan-Meier method (i.e., considering censored observations). ** Definitions for ratings of the certainty of the evidence: (6)

High: This research provides a very good indication of the likely effect. The likelihood that the effect will be substantially different is low.

Moderate: This research provides a good indication of the likely effect. The likelihood that the effect will be substantially different is moderate.

Low: This research provides some indication of the likely effect. However, the likelihood that it will be substantially different (a large enough difference that it might have an effect on a decision) is high.

Very low: This research does not provide a reliable indication of the likely effect. The likelihood that the effect will be substantially different (a large enough difference that it might have an effect on a decision) is very high.

Abbreviations: AE=adverse event; CI=confidence interval; GRADE=Grading of Recommendations Assessment, Development and Evaluation; HR=hazard ratio; IVRS=interactive voice response system; LDAC=low-dose cytarabine; OS=overall survival; RCT=randomised controlled trial; RR=risk ratio.



1 BACKGROUND

Overview of the disease or health condition

1.1.1 International classification of diseases (Version 10 [ICD-10] Code: C92.0).

Acute myeloid leukaemia (AML) is a group of heterogeneous haematopoietic stem cell disorders which are characterised by incomplete maturation of blood cells and reduced production of other normal haematopoietic cells. While AML can arise in patients with underlying haematological disorders, or as a consequence of prior cytotoxic therapy, the majority of cases are de novo malignancies (9). Multiple risk factors are implicated including radiation exposure, certain blood disorders, cytotoxic therapy, smoking, alcohol use and a family history (10). While AML most commonly presents in older people, it can also affect children (11, 12). Most patients present with leukocytosis and signs of bone marrow failure such as thrombocytopenia, reflecting the accumulation of immature and non-functional myeloblasts within the bone marrow and peripheral blood (13). Myelodysplastic syndromes (MDS) are a group of heterogeneous haematologic disorders with risk of progression to AML. De novo, secondary AML (as a result of previous leukemic disorder such as MDS) and treatment related AML have been described as prognostically distinct subtypes. In a Danish population-based study the frequency of secondary AML was 19.8 % and was associated with a low likelihood of receiving intensive treatment (14).

Symptoms of AML include appetite loss, weight loss, fatigue, breathlessness, confusion, bruising, bleeding and frequent infections. Symptoms can present as vague and non-specific, with AML typically being considered as a potential diagnosis following the abnormal results of routine blood tests. AML is definitively diagnosed following an examination of peripheral blood and bone marrow specimens (aspiration and/or biopsy) indicating the presence of 20 % or more blasts (2).

Immunophenotyping, cytogenetics and molecular genetics are used to describe the features of the disease and to guide the therapy (2). Targeted testing for mutations in FLT3, NPM1, CEBPA, and KIT help determine the prognostic subgroup and have an important role in guiding therapy. Approximately 20-30 % of AML patients carry an internal tandem duplication (ITD) mutation in the FLT3 gene which is associated with particularly poor prognosis (10, 15). Cytogenetic changes constitute the single strongest prognostic factor for complete remission (CR) and overall survival (OS) (1).

1.1.2 Disease classification and risk stratification

The molecular structure of AML has been mapped recently and allowed the identification of distinct subgroups and cytogenetic abnormalities. How these cytogenetic abnormalities are associated with clinical presentations, therapeutic response, relapse rates and overall survival led to the development of the European Leukaemia Net (ELN) and the World Health Organization (WHO) molecular classification and risk stratification schemas (2, 16, 17).

Patients > 60 years of age are the most challenging to treat. The higher risk associated with older age is partly explained by both the aggregation of cytogenetic risk and gene mutations associated with adverse outcomes. Identifying those patients who are > 60 years without adverse genetics could improve treatment options for these patients (18).

Guidelines form the National Comprehensive Cancer Network (NCCN), ELN and Southwest Oncology Group (SWOG) use cytogenetic profile to stratify patients into three risk categories (favourable, intermediate and poor/adverse) to guide treatment decisions (2) (see Table A2 in the Appendix). They recognize six broad subtypes based on the presence of antecedent haematologic disorders, cytogenetic abnormalities, and morphological characteristics. In the WHO classification, AML is defined with a few exceptions (AML with recurrent genetic changes [t(15;17), t(8;21), inv(16) or t(16;16)], as well as some cases of erythrocyte leukaemia) via a blast percentage of ≥ 20 % in the bone marrow. In addition to morphological characteristics, cytogenetic and molecular genetic characteristics of leukemic cells are also considered.

1.1.3 Epidemiology, incidence and prevalence

AML accounts for approximately 80 % of acute leukaemias diagnosed in adults. Data from the Surveillance, Epidemiology and End Results (SEER) registry indicate that the age-adjusted incidence in the US is 4.3 per 100,000 population annually and that incidence increases with age (19).



Similar to the US, other European countries such as Sweden and Denmark have long-standing registry systems and have reported similar incidence rates (20, 21). A Danish registry-based analysis reported an overall age-adjusted AML incidence rate of 5.4 per 100,000 person-years with a slightly higher incidence in men vs. women (6.2 vs. 4.9 per 100,000) (21). Countries within Europe report varied AML incidence rates per 100,000 population: 2.7 in Serbia, 3.0 in Switzerland and the Netherlands, 4.7 in Italy, 5.1 in the United Kingdom, and 5.4 in Denmark. The median age at diagnosis is approximately 70 years for AML (3).

The incidence of AML increases with age, from ~1.3 per 100,000 population in patients less than 65 years old, to 12.2 cases per 100,000 population in those over 65 (1). Although advances in the treatment of AML have led to significant improvements in outcomes for younger patients, prognosis in the elderly who account for the majority of new cases remains poor (22). Even with current treatments, as much as 70 % of patients 65 years or older will die of their disease within 1 year of diagnosis (23). In a French registry study (N=1,582) from 2000 to 2015 the median survival was 420 days (95 % CI 349 to 491) vs. 157 days (95 % CI 118 to 196) in de novo versus secondary AML, respectively. The 5-year overall survival was 32 % compared to 5 % in de novo and secondary AML patients, respectively (24).

AML is designated an orphan disease by both the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) (25, 26).

1.1.4 Clinical outcomes

Most late-phase AML clinical trials are designed to evaluate OS, event-free survival (EFS), or CR as primary endpoints; disease-free relapse and relapse-free survival are less frequently used (27). CR is defined as the absence of extramedullary leukaemia, < 5 % blasts in the bone marrow, absence of circulating blasts and blasts with Auer rods and platelets ≥ 100 × 109/L and neutrophils ≥ 1.0 × 109/L (2, 28). A less complete response has been defined as CR with incomplete haematological recovery (CRi). Rates of CRi are likely higher in studies than CR rates and a number of studies have demonstrated that the CRi is an inferior response to CR (29). More recently CRMRD- has been used and is defined as complete response with no minimum residual disease. It is measured using laboratory tests such as real-time quantitative polymerase chain reaction (RT-qPCR) or MFC (multiparameter flow cytometry) It can be considered a surrogate outcome for EFS and OS (2) and further validation to ensure generalisability of this proxy outcome is required (30). In the recent ELN 2017 guidelines the panel aimed at harmonising clinical definitions for outcomes used in clinical trials (see Error! Reference source not found.). OS remains the gold standard for treatment of patients with AML. EFS while useful as an early indicator of benefit is not well correlated with OS as highlighted by (27). The relevance of newer outcomes in the context of overall survival requires more research.

https://www.sciencedirect.com/topics/medicine-and-dentistry/event-free-survival

https://www.sciencedirect.com/topics/medicine-and-dentistry/neutrophil



Table 1.1. ELN Clinical definitions for outcomes in clinical trials

ELN Clinical definitions for outcomes in clinical trials

Definition

Overall survival Defined for all patients of a trial; measured from the date of entry into a clinical trial or from the date of diagnosis (e.g., for correlative science studies) to the date of death from any cause; patients not known to have died at last follow-up are censored on the date they were last known to be alive

Relapse-free survival (RFS)*,†

Defined only for patients achieving CR, or CRi; measured from the date of achievement of a remission until the date of relapse or death from any cause; patients not known to have relapsed or died at last follow-up are censored on the date they were last examined

Event-free survival (EFS)†

Defined for all patients of a trial; measured from the date of entry into a study to the date of primary refractory disease, or relapse from CR, or CRi or death from any cause; patients not known to have any of these events are censored on the date they were last examined

Cumulative incidence of relapse (CIR)†,‡

Defined for all patients achieving CR, CRi; measured from the date of achievement of a remission until the date of relapse; patients not known to have relapsed are censored on the date they were last examined; patients who died without relapse are counted as a competing cause of failure

Source: (2) * RFS and disease-free survival have been used with the same definition. † In clinical trials in which the response criterion CRMRD− is used, consideration should be given to include molecular relapse as assessed by RT-qPCR or MFC as a criterion for relapse; similarly, for analysis of EFS, no achievement of CRMRD- may be regarded as an event. The definitions of RFS, EFS, and CIR must be clearly defined within each protocol. ‡ It is important to provide estimates of CID as well because just considering the results of CIR may be misleading if, for instance, CIR is lower for 1 group but CID is actually higher for that same group. Abbreviations: CID=cumulative incidence of death; CIR=cumulative incidence of relapse; CR=Complete Remission; CRi=Complete Remission with incomplete haematological recovery; CRMRD-=CR without minimal residual disease.

Other outcomes of interest include transfusion dependence and safety. Transfusion independence could be considered a surrogate outcome for improvement of symptoms and potentially quality of life.

Current Clinical Practice

The backbone of therapy remains a combination of cytarabine- and anthracycline-based regimens with allogeneic stem cell transplantation for eligible candidates. Standard induction therapy with 3 days anthracycline (daunorubicin or idarubicin) and 7 days cytarabine (“7+3” regimens) is the first line therapy in those patients who can tolerate it. CR is achieved in 60 % to 80 % of younger adults and 40 % to 60 % of older adults (> 60 years) (3).

Elderly patients are often unable to tolerate such regimens, and carry a particularly poor prognosis. Curative therapies, including intensive chemotherapy and allogenic stem cell transplantation are generally applicable to a minority of patients who are younger and fit, while most older individuals exhibit poor prognosis and survival. Differences in patient outcomes are influenced by disease characteristics, access to care including active therapies and supportive care, and other factors.

1.2.1 European clinical practice recommendations

An overview of guidelines used for this assessment is provided in Table A1 in the Appendix.

European Society for Medical Oncology (ESMO) clinical practice guidelines Disease classification and risk assessment (as described in Table A2 in the Appendix) are important in determining treatment pathways. The favourable-risk AML group comprises all patients in whom a relapse risk is predicted to be low if treated with induction and consolidation chemotherapy alone. The intermediate-risk AML group comprises patients with molecular or cytogenetic abnormalities not classified as favourable or adverse. The adverse-risk AML group includes patients with complex cytogenetics and other poor-risk genetic features. All patients failing to achieve CR after 2 induction cycles should also be considered as adverse-risk patients, regardless of cytogenetics and genetics.



The initial stage to managing patients with AML involves assessment of fitness for first-line treatment options i.e. induction and consolidation chemotherapy. Pre-existing co-morbidities including heart, kidney, lung or mental illness as well as ECOG scores of > 3 and age > 75 years are the strongest predictors of poor outcomes i.e. non relapse induction related mortality. For this reason, these predictors are considered exclusion criteria for standard induction and consolidation.

For the first-line treatment of AML patients not eligible for standard induction and consolidation chemotherapy (unfit patients), the ESMO clinical practice guidelines recommend enrolment in a clinical trial. However, if a clinical trial is not available, then treatment with the hypomethylating agents (HMAs) azacitidine and decitabine, low-dose cytarabine (LDAC) or best supportive care (BSC) with, for example, hydroxycarbamide are currently the first choice in newly diagnosed unfit AML patients (3).

A prospective randomised study comparing 5- and 10-day decitabine treatment in newly diagnosed AML patients found almost identical CR, early mortality rates, EFS and OS between the two arms (31). Thus, if decitabine is chosen, it is recommended to follow the 5-day schedule.

No predictive markers are known to recommend one HMA over the other. HMA treatment is usually continued until disease progression or intolerance but may be terminated after at least four consecutive cycles, if the patient has not responded or derived clinical benefit. Given the moderate effects of HMAs, LDAC remains an alternative to HMAs in the first-line treatment of AML patients who are ineligible for standard induction and consolidation chemotherapy, except in patients with adverse-risk cytogenetics,

where LDAC has very poor activity. First-line treatment with LDAC results in a median OS of ∼5 months (32).

Patients with MDS progressing to AML during treatment with azacitidine constitute a significant therapeutic challenge.

Current evidence shows that 21 % to 43 % AML patients pretreated with HMAs and who received HMA and venetoclax achieved a response, with a smaller proportion achieving CR (33). The evidence is however based on only 23 patients of whom 6 had relapsed post allogenic hematopoietic cell transplantation (HCT).

Patients should be treated for at least four cycles and, where clinical benefit is achieved, should continue until progression or intolerance. Patients responding to initial treatment should be re-evaluated regarding their ability to undergo allogenic HCT using reduced intensity conditioning (RIC), which may cure a proportion of these patients.

Venetoclax is a selective, orally bioavailable, small molecule, B-cell lymphoma-2 (BCL-2) inhibitor that restores programmed cell death (apoptosis) in cancer cells. BCL-2 over expression is a major contributor to the pathogenesis of some types of lymphoid malignancies. While venetoclax in combination with azacitidine, decitabine or LDAC is approved in the United States and Israel it is not licensed for this indication in Europe. Although venetoclax in combination with an HMA or LDAC is considered to be superior to currently available first-line treatments for AML patients ineligible for standard induction chemotherapy based on promising preliminary data, results of ongoing randomised trials are awaited before its use can be recommended with confidence. Further ventoclax is not currently recommended as a treatment option in Europe.

Glasdegib in combination with LDAC (Figure 1.1.Error! Reference source not found.) is approved in the United States for newly diagnosed AML patients aged ≥ 75 years or those with comorbidities that preclude use of intensive induction chemotherapy. It is not currently recommended by ESMO.

ELN recommendations The ELN published recommendations in 2017 on the diagnosis and management of patients with AML (2). For patients not eligible for standard induction therapy enrolment in a clinical trial is the preferred option, if possible. Alternatively, LDAC, or HMAs (decitabine or azacitidine) or BSC are recommended. LDAC is generally well-tolerated and produces CR rates in the order of 15 % to 25 % (32). An increase in median OS with decitabine vs. mostly LDAC (7.7 vs. 5.0 months) was observed, and azacitidine increased the median OS by 10.4 vs. 6.5 months vs. conventional care regimens (included standard induction chemotherapy, LDAC, or supportive care only). Azacitidine may be particularly advantageous in AML with adverse cytogenetics. Superiority of azacitidine over conventional care regimens was



previously shown in AML with 20 % to 30 % blasts. Up to 6 courses may be needed to observe maximal response with azacitidine or decitabine, although patients without response after 3 courses are unlikely to respond with further therapy.

BSC is an option for patients who choose not to have chemotherapy or who are unlikely to tolerate chemotherapy. BSC includes the use of anti-infectives, transfusion support with blood and blood products, hematopoietic growth factors and hydroxyurea.

Figure 1.1. Place of glasdegib in the treatment pathway (as per applicant’s submission)

Source: (4) Abbreviations: AML=acute myleoid leukaemia; HMAs=hypomethylating agents, LDAC=low-dose cytarabine.



Features of the intervention

Glasdegib is a potent hedgehog (Hh) pathway inhibitor that exerts its action by binding to and blocking Smoothened (SMO), which is a transmembrane protein involved in Hh signal transduction. Aberrant Hh signalling has been identified in many solid tumour types and in haematological malignancies. For features of the intervention see Table 1.2, for administration and dosing see

Table 1.3. Table 1.2. Features of the intervention

Non-proprietary name Glasdegib Azactidine Cytarabine Decitabine

Proprietary name Daurismo® Vidaza® Cytarabine Dacogen®

Registered EMA indication

In combination with LDAC, for the treatment of newly diagnosed de novo or secondary acute myeloid leukaemia (AML) in adult patients who are not candidates for standard induction chemotherapy.

Treatment of adult patients who are not eligible for haematopoietic stem cell transplantation (HSCT) with:

intermediate-2 and high-risk MDS according to the International Prognostic Scoring System (IPSS);

chronic myelomonocytic leukaemia (CMML) with 10-29 % marrow blasts without myeloproliferative disorder;

AML with 20-30 % blasts and multi-lineage dysplasia, according to WHO classification;

AML with >30% marrow blasts according to the WHO classification.

For induction of remission in AML in adults and for other acute leukaemias of adults and children.

Treatment of adult patients with newly diagnosed de novo or secondary acute myeloid leukaemia (AML), according to the World Health Organisation (WHO) classification, who are not candidates for standard induction chemotherapy.

Prospective Marketing authorisation holder

Pfizer Europe MA EEIG Celgene Ltd. Pfizer Ltd. Janssen-Cilag

Contra-indications Hypersensitivity to the active substance or to any of the excipients listed in SmPC section 6.1.

Hypersensitivity to the active substance or to any of the excipients listed in SmPC section 6.1. Advanced malignant hepatic tumours Breast-feeding

Therapy with cytarabine should not be considered in patients with pre-existing drug-induced bone marrow suppression, unless the clinician feels that such management offers the most hopeful alternative for the patient. Cytarabine should not be used in the management of non-

Hypersensitivity to decitabine or to any of the excipients, listed in section 6.1 of SmPC.

Breast-feeding.




malignant disease, except for immunosuppression. Hypersensitivity to the active substance or to any of the excipients listed in SmPC section 6.1.

Drug class Small-molecule hedgehog signalling inhibitor

Antineoplastic agent, pyrimidine analogue

Antineoplastic agent, pyrimidine analogue

Antineoplastic agents, antimetabolites, pyrimidine analogue

Active substance(s) PF-04449913-11 (1-((2R,4R)-2-(1H-benzo[d]imidazol-2-yl)-1-methylpiperidin-4-yl)-3-(4-cyanophenyl) urea maleate)

Azacitidine Cytarabine Decitabine

Pharmaceutical formulation(s)

25 mg and 100 mg tablets Powder for suspension for injection in 100mg vial (25mg/ml)

Solution for infusion or injection 20mg/ml.

Powder for concentrate for solution for infusion 50mg/10ml.

ATC code L01XX63 L01BC07 L01BC01 L01BC08

In vitro diagnostics required

- - - -

Monitoring required Complete blood counts, electrolytes, renal, and hepatic function should be assessed prior to the initiation and at least once weekly for the first month. Electrolytes and renal function should be monitored once monthly for the duration of therapy. Serum creatine kinase (CK) levels should be obtained prior to initiating and as indicated clinically thereafter (e.g., if muscle signs and symptoms are reported). Electrocardiograms (ECGs) should be monitored prior to initiation, approximately one week after initiation, and then once monthly for the next two months to assess for QT corrected for heart

Liver function tests, serum creatinine and serum bicarbonate should be determined prior to initiation of therapy and prior to each treatment cycle. Complete blood counts should be performed prior to initiation of therapy and as needed to monitor response and toxicity, but at a minimum, prior to each treatment cycle, cardiopulmonary assessment before and during the treatment should be considered. Patients should also be monitored for tumour lysis syndrome and necrotising fascitis.

Frequent platelet and leucocyte counts are mandatory. Periodic checks of bone marrow, liver and kidney functions should be performed Cardiopulmonary assessment before and during the treatment should be considered. Monitoring for neurological adverse reactions, tumour lysis syndrome and pancreatitis is also advised.

Baseline complete blood counts, platelet counts, hepatic function (liver function tests) and renal function (serum creatinine). Monitoring at each cycle of complete blood counts, liver and renal function. Bone marrow aspirate advised after 4 cycles and every 6 months thereafter. Patients especially those with cardiac disease history, should be monitored for signs and symptoms of heart failure.




rate (QTc) prolongation. ECG should be repeated if abnormal. Certain patients may require more frequent and ongoing ECG monitoring (see SmPC section 4.4). Abnormalities should be managed promptly.

Orphan Designation Yes No No No

ATMP No No No No

Source: (5, 34, 35) Abbreviations: AML=acute myeloid leukaemia; ATC=anatomical therapeutic chemical [classification system]; ATMP=advanced therapy medicinal product; CK=creatine kinase; CMML=chronic myelomonocytic leukaemia; ECG=electrocardiogram; HSCT=haematopoietic stem cell transplantation; IPSS=International Prognostic Scoring System; LDAC=low-dose cytarabine; SmPC=summary of product characteristics; WHO=World Health Organization.

Table 1.3. Administration and dosing of the technology

Glasdegib

Method of administration Tablets to be taken orally

Doses 25 mg tablets 100 mg tablets

Dosing frequency 100 mg orally once daily

Standard length of a course of treatment Continuously for 28-day cycles

Standard interval between courses of treatments

Continuously

Standard number of repeat courses of treatments

A minimum of 6 cycles

Dose adjustments Dose modifications may be required based on individual safety and tolerability. If dose reduction is necessary, then the dose should be reduced to 50 mg taken orally once daily.



2 OBJECTIVE AND SCOPE

The aim of this EUnetHTA Joint Relative Effectiveness Assessment is to compare the clinical effectiveness and safety of glasdegib in the target patient populations with relevant comparators. The target patient populations and relevant comparators (based on the requirements of EUnetHTA Partners) are defined in the project scope below (Table 2.1).

The assessment was based on the Submission Dossier submitted by the MAH [Pfizer INC].

The scope of the assessment deviates from the scope described in the Project Plan as follows:

According to the scope, results from non-randomised controlled trials (RCTs) should also be considered for safety endpoints, but the MAH conducted a systematic literature review (SLR) for RCTs only. However, the Committee’s for Medicinal Products for Human Use (CHMP) European Public Assessment Report (EPAR) assessment of clinical safety focused mainly on the overall population of the B1371003 study and reports that 101 patients (AML and MDS) were treated with the combination therapy assessed in this assessment. Of these, 89 were AML patients, with the majority (75 patients) from BRIGHT AML 1003 (5). Since safety outcomes for these patients are presented in this assessment, no additional literature search was conducted.

Table 2.1. Scope of the assessment


Population Adult patients with newly diagnosed de novo or secondary acute myeloid leukaemia (AML) who are not eligible for standard induction chemotherapy ICD-10: 92.0 MeSH-terms: Leukemia, Myeloid, Acute Tree Number(s): C04.557.337.539.275 MeSH Unique ID: D015470

Intervention Glasdegib 100 mg orally once daily continuously (on days 1 to 28) in combination with low-dose cytarabine (LDAC) Synonyms for glasdegib: PF-04449913

Comparison Azacitidine;

Decitabine;

LDAC ;

Best Supportive Care (BSC: hydroxyurea, transfusion support)a.

Outcomes Effectivenessb:

Overall survival;

Health-related quality of life;

Transfusion independency;

Objective Response: o Overall Response Rate: Complete Remission (CR) + CR with incomplete blood count

recovery + morphologic leukaemia-free state. Safetyc:

Serious adverse events (AEs);

Grade ≥3 AEs;

Fatal AEs;

Overall AE;

AEs of special interest (febrile neutropenia, pneumonia, haemorrhage, QT Interval Prolongation);

Treatment discontinuation due to AE.




Study type Effectiveness:

Randomised controlled trials (RCT). Safety: If suitable evidence syntheses (systematic reviews (SRs)/HTA reports) are available:

Evidence syntheses (SRs/HTA reports); and

Primary studies (as described in next bullet) published after the last search date of the latest SR/HTA document.

If suitable evidence syntheses (SRs/HTA reports) are NOT available:

RCTs;

Non-randomised controlled trials;

Observational studies. a All comparators were explicitly mentioned by patient organizations. b Including data at 6 and 12 months. c Including data at 90 days. Abbreviations: AE=adverse event; AML=acute myeloid leukaemia; BSC=best supportive care CR=complete remission; HTA=health technology assessment; LDAC=low-dose cytarabine; QT=Q-TWiST; RCT=randomized control trial; SR=study report.



3 METHODS

The assessment is based on the data and analyses included in the Submission Dossier prepared by the MAH. During the assessment, the completeness of data and analyses in the Submission Dossier was verified. Furthermore, the methods for data analysis and synthesis applied by the MAH were checked against the requirements of the Submission Dossier and applicable EUnetHTA Guidelines and assessed with regard to scientific validity.

Information retrieval

As can be seen in Table 3.1, the core Submission Dossier submitted by the MAH included a SLR to identify RCTs in 3 bibliographic databases: Embase, MEDLINE and the Cochrane Central Register of Controlled Trials (CENTRAL). We deemed the choice of databases complementary and sufficient with regard to the PICO of interest. The search strategy was well documented and reproducible. The search was adequately broad with respect to glasdegib, where relevant search terms related to the drug under evaluation were used in all three databases. In MEDLINE and Embase, the MAH combined the search filter for glasdegib with a validated study design search filter that optimizes both sensitivity and specificity to identify RCTs (36). The beforementioned search allowed the identification of RCTs providing direct comparative effectiveness and safety data for glasdegib + LDAC versus one of the comparators (i.e. LDAC) of interest to this report. The MAH performed additional searches to identify comparative data for the HMAs azacitidine and decitabine versus LDAC. These additional searches aimed at the identification of suitable trials using LDAC as common comparator for the indirect comparison of glasdegib + LDAC vs. azacitidine or decitabine (Appendix B of the core Submission Dossier) (4). Search terms related to the HMAs azacitidine and decitabine were well chosen and were combined with relevant search terms related to the target condition. Again, a validated search filter for RCTs was applied where appropriate (36).

The MAH complemented the searches in bibliographic databases by searches in three clinical trial registers:

US National Institutes of Health (NIH) Ongoing Trials Register (clinicaltrials.gov);

European Union Clinical Trials Register (EU CTR) (clinicaltrialsregister.eu/ctr-search/search);

International Clinical Trials Registry Platform (ICTRP) Search Portal (WHO, apps.who.int/trialsearch/AdvSearch.aspx).

In the first two trial registries, the search was exhaustive, only using search terms related to glasdegib. In the ICTRP search portal, such search terms were combined with relevant terms related to AML. Although the latter approach is more restrictive, it is deemed sufficiently sensitive to identify relevant entries. The three trial registries were also searched to identify comparative data on decitabine and azacitidine. Searches in trial registries could be reproduced with exception of searches in the ICTRP portal, as this portal was not functional at the time accessed (last search on 12th of June 2020) by the information specialist.

The MAH did not use any restrictions related to language or publication date. All searches were performed by the MAH on 17th of January 2020. Screening in both title and abstract phase and full text phase was conducted by two independent reviewers. Disagreements were resolved by involvement of a third reviewer.

There is no mentioning in the dossier of further search methods such as reference checking or search for studies identified in reports of other evaluation bodies. Details of the search can be found in the core Submission Dossier in section 5 “Clinical effectiveness and safety” and Appendix B.

With regard to selection criteria, these were tuned to identify relevant RCTs evaluating glasdegib + LDAC vs. the comparators identified in the scoping phase. Selection criteria in the additional searches were tuned to identify RCTs evaluating decitabine vs. LDAC or azacitidine vs. LDAC.



3.1.1 Literature search and assessment approach: Critical assessment of the method

The information specialist critically assessed the method of the information retrieval submitted. Data were also checked for completeness against a search in study registries and against the studies included in the regulatory Assessment Report. Completeness of data on glasdegib in combination with other drugs and/or in a different indication was not assessed.

The essential elements of the PICO are reflected in the search strategy. Search terms were relevant, involving the combination of text words with Medical Subject Headings (MeSH) or Emtree terms whenever available, applying Boolean operators adequately. The MAH chose a validated design type search filter and made a slight modification which will not impact on the sensitivity or specificity of the validated version, it merely ensures adequate use across different search platforms (details can be found in Appendix B of the core Submission Dossier). The information specialist concluded that the SLR was well constructed, adequately reported and reproducible, so that the risk to miss relevant studies was deemed to be low.

Screening methods used by the MAH to screen citations and full text reports were adequate. Overall, it can be concluded that search and screening methods were adequate. For this reason, the information specialist refrained from supplementary searches to check for possible incompleteness of the study pool.

Table 3.1. Summary of information retrieval and study selection

Elements Details

List of studies submitted by MAH

Glasdegib + LDAC vs. LDAC

B1371003 (NCT01546038; EudraCT: 2012-000684-24). Azacitidine or decitabine vs. LDAC;

AZA-AML-001 ((NCT01074047; EudraCT: 2009-012346-23);

DACO-016 (NCT00260832; EudraCT: 2005-004503-11).

Databases and trial registries searched

MEDLINE via OVID Embase via OVID Cochrane Central Register of Controlled Trials (CENTRAL) via OVID US National Institutes of Health (NIH) Ongoing Trials Register World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) European Union Clinical Trials Register (EU CTR)

Search date 17 January 2020

Keywords Details can be found in Appendix B of the core Submission Dossier

Inclusion criteria Study design RCT;

Meeting PICO elements. Details can be found in Table 10, Table 11 and Table 12 of the core Submission Dossier

Exclusion criteria Details can be found in Table 10, Table 11 and Table 12 of the core Submission Dossier

Date restrictions None

Other search limits or restrictions

None

Source: (4) Abbreviations: AML=acute myeloid leukemia; CENTRAL=Cochrane Central Register of Controlled Trials; Embase=Excerpta Medica database; EU CTR=European Union Clinical Trials Register; ICTRP=International Clinical Trials Registry Platform; LDAC=low-dose cytarabine; MAH=market authorization holder; MEDLINE=Medical Literature Analysis and Retrieval System Online; NIH=US National Institutes of Health; PICO=patient, intervention, comparison, outcome; RCT=randomised controlled trial; WHO=World Health Organization.

The study pool of the assessment was compiled on the basis of the following information:

Sources of the company in the Submission Dossier:

o Study list of MAH on glasdegib + LDAC, decitabine and azacitidine;

o Bibliographical databases (last search on 17th of January 2020);

o Trials registries (last search on 17th of January 2020).



Check of the completeness of the study pool:

o Trials registries (last search on 12th of June 2020).

No supplementary searches were conducted to check for possible incompleteness of the study pool.

Data extraction

Information used for the assessment of clinical effectiveness and safety were extracted from the core Submission Dossier and verified against the Clinical Study Reports (CSR) or other original documentation provided in the Submission Dossier.

Risk of bias assessment

The revised Cochrane risk-of-bias tool for randomized trials (RoB 2) quality rating tool (version of 22nd of August 2019) was used to assess the risk of bias in randomized trials (see Table A3 in the Appendix) (37). Risk of bias at outcome level was assessed for the following 5 different domains resulting in an overall risk of bias:

Risk of bias arising from the randomization process;

Risk of bias due to deviations from the intended interventions;

Missing outcome data;

Risk of bias in measurement of the outcome;

Risk of bias in selection of the reported result.

For each domain and for each endpoint, two independent assessors judged the risk of bias (‘low’, ‘high’ or ‘some concerns’) on the basis of the information retrieved from the full-text publications, the protocols and the Submission Dossier. Any disagreements were resolved by involving a third assessor.

Certainty of evidence

The quality of evidence for each outcome across all the studies (i.e. the body of evidence for an outcome) was rated according to the factors outlined in the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (see Table A4 in the Appendix), including the following five factors that may lead to rating down the quality of evidence of RCTs (6):

Study limitations (risk of bias);

Inconsistencies of results;

Indirectness of evidence;

Imprecision;

Publication bias.

The RoB 2 results were used for rating any study limitations, and two assessors rated the other factors independently. Any disagreements were resolved by involving a third assessor.

Results and analyses of included studies

The information in the Submission Dossier on the study design, study methods, populations, endpoints (patient relevance, validity, and operationalization) and study results were evaluated. The result of this evaluation is presented and was used for identification of relevant analyses and considered for the

conclusions of the Assessment Report.



3.5.1 Sensitivity analysis

To evaluate the robustness of results, sensitivity analyses with regard to methodological factors presented in the Submission Dossier and the corresponding methods applied were evaluated. These methodological factors arise from decisions made within the framework of the retrieval and assessment of information, for example, the specification of cut-offs for the time point of data collection or the choice of effect measure.

3.5.2 Subgroup analysis and other effect modifiers

During the assessment, the subgroup analyses examining potential effect modifiers presented in the Submission Dossier and the corresponding methods applied were evaluated. The evaluation also includes the justification for the choice of cut-offs, if quantitative characteristics were categorized.

3.5.3 Indirect comparisons

The methods of indirect comparisons applied, and, if applicable, the justification in the event of deviations from the required approaches were evaluated (38).

Patient involvement

For this Joint Assessment, an open call for patient input was published on the EUnetHTA website on the 11th of October 2019. The questions were based on the HTA international (HTAi) questionnaire template that was adapted for this project and covered the following topics:

The impact of AML on patients’ quality of life;

Impact of AML on carers/unpaid care-givers;

Experiences with currently available treatment options;

Expectations/requirements for a new medicine for AML.



4 RESULTS

Information retrieval

The MAH provided a flow chart showing that out of 1,928 citations identified through database searching, 415 concerned duplicates (454 records excluded) and 1,465 were excluded during title/abstract screening. During full text screening of 9 reports, another 6 were excluded as no outcomes of interest were addressed or because the target population was not relevant to this Assessment Report. A total of 3 studies met the selected PICO criteria and were included. Of these, one addressed glasdegib + LDAC vs. LDAC alone: the B1371003 trial of which only results from BRIGHT AML 1003 and thus only for the AML subgroup were mainly relevant for this assessment. No additional studies were identified that directly compared glasdegib + LDAC versus other PICO defined comparators i.e. azacitidine, decitabine or BSC. Two studies evaluated the effect of azacitidine and decitabine vs. LDAC and hence, were included for an indirect comparison via the common comparator LDAC.

Studies included in the assessment

The studies listed in the following Table 4.1 were included in the assessment. However, direct evidence for AML patients could be retrieved only from a subgroup (i.e. BRIGHT AML 1003) from one study, the B1371003 trial.

Table 4.1. List of included studies

Study reference/ID Study category

Study for marketing

authorization of the technology

under assessment

(yes/no)a

Sponsored or third-party studyb

Available documentationc

Direct comparison

B1371003 but only the AML subgroup i.e. BRIGHT AML 1003 (NCT01546038; EudraCT: 2012-000684-24)

Yes Sponsored Full text publication (39) Core Submission Dossier (4) Clinical study report (40) Final Protocol Amendment (41) Statistical Analysis Plan (42) EPAR (5)

Indirect comparison

AZA-AML-001 (NCT01074047; EudraCT: 2009-012346-23)

Yes Sponsored Full text publication (43) EPAR (35)

DACO-016 (NCT00260832; EudraCT: 2005-004503-11)

Yes Sponsored Full text publication (44) EPAR (34)

a If "yes," also indicate the respective reference of the data reference(s). b Study sponsored by the MAH or in which the MAH participated financially in some other way. c Include references of the study registry entries and, if available, the reports on study design and/or results listed in the study registries. Abbreviations: EPAR=European public Assessment Report.

Excluded studies

No studies included in the Submission Dossier were excluded.



Characteristics of BRIGHT AML 1003

A detailed description of the study characteristics can be found in Table 4.3 and Table 4.4.

BRIGHT AML 1003 contained the phase 2 results from the AML subgroup of patients initially enrolled into B1371003, a phase 1b/2, randomized, open-label, multicentre trial comparing glasdegib + LDAC to LDAC alone in adult patients with newly diagnosed, previously untreated AML or high-risk MDS considered not eligible for intensive chemotherapy (see Figure 4.1) (4, 39-42). For being ineligible for intensive chemotherapy at least one of several criteria had to be fulfilled, including age ≥ 75 years, serum creatinine > 1.3 mg/dL, severe cardiac disease or ECOG PS = 2 (see Table 4.3).

Figure 4.1. Study design of B1371003

Source: (40) Abbreviations: AML=acute myeloid leukaemia; LDAC=low-dose cytarabine; MDS=myelodysplastic syndrome; P2 FIT=Phase 2 single arm component in fit patients; P2 unfit=Phase 2 randomized component in unfit patients; pts=patients; RP2D=recommended Phase2 dose; vs.=versus.

As shown in Figure 4.2, within this assessment BRIGHT AML 1003 describes the AML subgroup only.

Figure 4.2. AML subgroup (BRIGHT AML 1003) of B1371003 Source: (4) Abbreviations: AML=acute myeloid leukaemia; LDAC=low-dose cytarabine; MDS=myelodysplastic syndrome.



AML patients included de novo AML, AML evolving from MDS or other antecedent haematologic diseases (AHD) and AML after previous cytotoxic therapy or radiation (secondary AML). Eligible patients with AML arising from an AHD or MDS, may have had one prior regimen with commercially available agent(s) (e.g., azacitidine or decitabine) for the treatment of their prior haematologic disease, but had to be untreated for their AML. The study was conducted in Europe and North America and started in June 2012; primary completion date was January 2017.

AML patients were grouped into one of four different risk groups based on their cytogenetic risk according to ELN 2010 guidelines (17). The factor(s) characterizing the patient risk were captured on the Case Report Forms (CRFs). For the randomization process itself, patients were stratified by two cytogenetic risk groups, whereby patients with favourable, intermediate-I, and intermediate-II risk groups were allocated to good/intermediate cytogenetic risk and patients in the adverse risk group were allocated to poor cytogenetic risk (see Table 4.2) (41). Patients with any one of the following cytogenetic features were classified as having poor risk disease: inv(3), t(6;9), 11q23, -5, -5q, -7, abnormal (17p),

or complex karyotype (3 clonal abnormalities). Patients with none of the features described above were classified as having good/intermediate risk disease. The patient risk was captured in an interactive registration system (IRS). Patient randomization was obtained by the investigator or the designee from an interactive voice response system (IVRS).

Table 4.2. Cytogenetic risk conversion chart for AML patients

Risk group per Case Report Form (CRF) Stratification factor for randomization (IVRS)

Favourable Good/intermediate

Intermediate-I Good/intermediate

Intermediate-II Good/intermediate

Adverse Poor

Source: (41) Abbreviations: CRF=Case Report Form; IVRS=interactive voice response system.

Overall, 132 patients were randomized 2:1 to glasdegib + LDAC (n = 88) or LDAC alone (n = 44) in the B1371003 trial. Of the 132 patients, 116 had AML of whom 78 received the combination therapy and 38 received LDAC alone. The 16 MDS patients initially allocated to either of the two groups were excluded from any analyses presented by the MAH.

As can be seen in Table 4.4, glasdegib 100 mg once daily was administered orally on a continuous basis and LDAC, either in combination with glasdegib or alone, was given at a dose of 20 mg administered subcutaneously twice daily for the first 10 days of a 28-day cycle. In the B1371003 trial, for unfit patients, study treatment with glasdegib + LDAC could continue for up to 1 year (12 cycles) from start of therapy or until disease progression or relapse, patient refusal or occurrence of unacceptable toxicity (whichever came first). Treatment with LDAC alone could continue for up to 1 year (12 cycles). Unfit patients who were treated for 1 year (12 cycles) with LDAC were considered to have completed treatment on trial (40). However, patients who completed the maximum number of cycles/months on study treatment, demonstrated clinical benefit with manageable toxicity, and who were willing to continue receiving assigned treatment could be given the opportunity to do so upon agreement between investigator, sponsor and pending study drug availability (5).



Table 4.3. Characteristics of the B1371003 trial / BRIGHT AML 1003

Study reference/ID

Study design Patient population Intervention (number of randomized patients)

Comparator (number of randomized patients)

Study duration and data cut off(s)

Primary outcome; patient-relevant secondary outcomes

B1371003 trial and BRIGHT AML 1003

Randomised, multi-centre, open-label, Phase 1b/2 Relevant population: Phase 2 Unfit/AML subgroup

Newly diagnosed de novo or secondary AML (≥20 % BM blasts) or refractory anemia with excess blasts (RAEB)-2 high-risk MDS who were newly diagnosed according to WHO 2008 Classification and previously untreated;

≥55 years;

Known cytogenetic profile at study entry and considered not suitable for intensive chemotherapy, defined by ≥1 of the following criteria:

o Age ≥75 years; o Serum creatinine >1.3 mg/dL; o Severe cardiac disease (e.g.,

left ventricular ejection fraction <45 %);

o ECOG PS = 2. Patients with ECOG PS = 0 or 1 who met ≥1 other inclusion criteria listed above were also eligible.

Glasdegib + LDAC (n = 88)

LDAC (n= 44)

Study duration: 4.6 years (27 June 2012 to 3 January 2017) Data cut-off: 3 Jan 2017 Planned follow-up period: 4 years from the first dose; median follow-up 21.7 months for glasdegib; 20.1 months for LDAC Type of analysis: Planned final analysis

Primary: OS (FAS);

Secondary: CR (FAS)*;

Adverse events (NCI CTCAE Version 4.0) (Safety analysis set).

Additional outcomes:

Quality of Survival evaluated by Q-TWiST analysis;

ORR;

Transfusion independence.

Relevant AML subgroup: (n = 78)

Relevant AML subgroup: (n = 38)

Source: (4, 40) * Assessment of response was made using response criteria for MDS and AML derived and defined by the disease specific International Working Groups and WHO Guidelines (Appendices 4 and 5 of the protocol [Section 16.1.1]) (40). Abbreviations: AML=acute myeloid leukaemia; BM=bone marrow; CR=complete remission; CTCAE=Common Terminology Criteria for Adverse events; ECOG PS=Eastern Cooperative Oncology Group performance status; FAS=Full Analysis Set, i.e. all randomized patients; LDAC=low-dose cytarabine; N=number of patients; NCI=National Cancer Institute; OS=overall survival.



Table 4.4. Characterisation of the intervention and comparator of the B1371003 trial / BRIGHT AML 1003

Study reference / ID Glasdegib + LDAC LDAC Treatment characteristics

B1371003 trial and BRIGHT AML 1003

Glasdegib 100 mg once daily orally in 28-day cycles on a continuous basis and LDAC 20 mg s.c. twice daily for 10 days every 28 days

LDAC 20 mg s.c. twice daily for 10 days every 28 days

Concomitant treatment considered necessary for the patient’s wellbeing could be given at the discretion of the treating physician. Treatment duration:

Glasdegib + LDAC: up to 1 year (12 cycles) from start of therapy or until disease progression or relapse, patient refusal or occurrence of unacceptable toxicity (whichever came first); patients who completed the maximum number of cycles/months on study treatment, demonstrated clinical benefit with manageable toxicity, and who were willing to continue receiving assigned treatment could be given the opportunity to do so upon agreement between investigator, sponsor and pending study drug availability;

LDAC alone: up to 1 year (12 cycles); or longer upon agreement between investigator, sponsor and pending study drug availability.

Source: (4, 40) Abbreviations: AML=acute myeloid leukaemia; LDAC=low-dose cytarabine; s.c.=subcutaneous.



Table 4.5 shows the [mean and median] treatment duration, and the planned duration of follow-up as well as the observation period for the individual outcomes in BRIGHT AML 1003.

The median (range) treatment duration was 83 (3 – 972) days in the glasdegib + LDAC arm and 41 (6 – 239) days in the LDAC alone arm. While about 37 % of patients received at least 6 cycles of treatment with glasdegib + LDAC, only less than 6 % of patients received at least 6 cycles of LDAC alone. Information concerning the median number of cycles administered is only available for the combined AML and MDS population. The median (range) number was 3 (1 – 35) cycles in the glasdegib + LDAC arm and 2 (1 – 9) cycles in the LDAC alone arm, respectively (39).

According to the study protocol of B1371003, it was planned that all patients will be followed for 4 years from the first dose (or randomization date for randomized patients who do not start treatment) and that the study will be considered complete once all patients have been followed for 4 years from first dose (or randomization date if applicable) (41).

The observation period for OS was planned until the date of death from any cause or the final data cut-off date (for a planned maximum observation time of 4 years from randomization), and the median follow-up for OS was reported at about 22 months in patients receiving glasdegib + LDAC vs. about 20 months in patients receiving LDAC alone (5). However, in the core Submission Dossier the MAH states a considerably shorter median follow-up for OS of approximately 7 months in the glasdegib + LDAC arm vs. less than 4 months in the LDAC alone arm. While the reason for this discrepancy remains unclear, it may have been caused by not censoring patients at the time of death.

Table 4.5. Information on the course of BRIGHT AML 1003 (including planned duration of follow-up)

BRIGHT AML 1003 Planned follow-up Intervention Control

ID Outcome category Glasdegib + LDAC N = 75

LDAC N = 36

Treatment duration [days]a

Median [Min; Max] 83.0 [3; 972] 40.5 [6; 239]

Mean (SD)

190.1 (n.r.) 61.6 (n.r.)

Observation period [months]

Overall survivalb

(Data cut-off 3rd Jan 2017)

Until the date of death from any cause (for 4 years from randomization)

Median [95% CI]c 21.7 [18.1; 26.5] 20.1 [n.e.; n.e.]

Mean (SD)

n.r. n.r.

Health-related quality of life Not pre-planned Not collected

Transfusion independence Not pre-planned Not collected

Complete remissionb During treatment n.r.

Adverse eventsd Through and including 28 calendar days after last administration of study medication

Median [Min; Max] n.r. n.r.

Mean (SD) n.r. n.r.

Source: (5, 45) a Patients receiving at least 6 cycles of treatment: Glasdegib + LDAC 28/75 (37.3%) vs. LDAC 2/36 (5.6%) b Glasdegib + LDAC: N = 78 vs. LDAC: N=38 c The MAH also states a shorter median follow-up for OS of 156.5 (range: 2 – 1005) days in total, thereof 226 (range: 5 – 1005) days in the glasdegib + LDAC arm and 115 (range: 2 – 613) days in the LDAC alone arm. The reason for this discrepancy is not clear from the Submission Dossier but might have occurred by not censoring patients on death. d SAEs that occurred to a subject after the active reporting period had ended were to be reported if the investigator had become aware of them; at a minimum, all SAEs that the investigator believed had at least a reasonable possibility of being related to investigational products were to be reported. Abbreviations: AML=acute myeloid leukaemia LDAC=low-dose cytarabine; max=maximum; min=minimum; N=number of patients; n.e.= not estimable; n.r.=not reported; SD=standard deviation; vs.=versus.



Table 4.6 shows the characteristics of the patients included in BRIGHT AML 1003.

Table 4.6. Baseline characteristics of BRIGHT AML 1003

BRIGHT AML 1003

Glasdegib + LDAC N = 78

LDAC N = 38

Demographic Characteristics

Sex, n (%)

Male 59 (75.6) 23 (60.5)

Female 19 (24.4) 15 (39.5)

Age (years)

Mean (SD) 76.4 (6.0) 74.8 (4.9)

Median (range) 77 (64-92) 76 (58-83)

Age categories, n (%)

< 75 years 30 (38.5) 15 (39.5)

≥ 75 years 48 (61.5) 23 (60.5)

Clinical characteristics

Type of AML, n (%)

De novo 38 (48.7) 18 (47.4)

Secondary 40 (51.3) 20 (52.6)

ECOG PS, n (%)

0-1 36 (46.2) 20 (52.6)

2 41 (52.6) 18 (47.4)

missing 1 (1.3) 0

History of severe cardiac disease, n (%)

no 26 (33.3) 18 (47.4)

yes 52 (66.7) 20 (52.6)

Serum creatinine, n (%)

≤ 1,3 mg/dL 62 (79.5) 32 (84.2)

> 1,3 mg/dL 15 (19.2) 5 (13.2)

Missing 1 (1.3) 1 (2.6)

ELN cytogenetic risk stratification (according to CRF, per ELN 2010 guidelines)

Favourable 5 (6.4) 3 (7.9)

Intermediate-I 27 (34.6) 11 (28.9)

Intermediate-II 21 (26.9) 8 (21.1)

Adverse 25 (32.1) 16 (42.1)

Cytogenetic risk, n (%)

According to CRF

(per ELN 2010 guidelinesa)

According to IVRS

According to CRF

(per ELN 2010 guidelinesa)

According to IVRS

Good/intermediate: including favourable, intermediate-I, and intermediate-II risk groups

53 (67.9) 49 (62.8) 22 (57.9) 21 (55.3)

Poor: including adverse risk group

25 (32.1) 29 (37.2) 16 (42.1) 17 (44.7)

Peripheral white blood cell count (103/mm3)

Median (range) 2.47 (0.6-64.0) 4.07 (1.1-45.2)

Bone marrow blasts (%)

Median (range) 41 (16-100) 46 (13-95)


< 20 % 3 (3.8) 1 (2.6)

≥ 20 - ≤ 30 % 21 (26.9) 9 (23.7)

> 30 - ≤ 50 % 20 (25.6) 8 (21.1)



BRIGHT AML 1003


LDAC N = 38

≤ 50 % n.r. n.r.

> 50 % 31 (39.7) 17 (44.7)

Not reported 3 (3.8) 3 (7.9)

Haemoglobin (g/dL)

Median (range) 9.1 (6.4-14.0) 9.3 (6.0-14.6)

Duration since histopathological diagnosis (months)

Median (range) 0.57 (0.0-3.5) 0.51 (0.1-3.8)

Source: Adapted from (39, 40, 46, 47) a According to ELN 2010 guidelines: Patients were classified as having poor-risk disease if they had one of the following cytogenetic features: inv(3), t(6;9), 11q23, –5, –5q, –7, abnormal (17p), or complex karyotype (≥3 clonal abnormalities). Patients with none of these features were classified as having good/intermediate-risk disease (17). Abbreviations: AML=acute myeloid leukaemia; CRF=case report form; ECOG PS=Eastern Cooperative Oncology Group performance status; ELN=European Leukaemia Net; IVRS=interactive voice response system; LDAC=low-dose cytarabine; N=number of patients; n.r.=not reported; SD=standard deviation.

Table 4.6 shows the characteristics of AML patients only. These data were verified with additional information provided by the MAH, since the CSR only provides information on the overall unfit population of the B1371003 trial and thus includes 16 MDS patients too.

The majority of AML patients were male, with a higher number of men in the glasdegib + LDAC arm than in the LDAC alone arm (glasdegib 76 % vs. LDAC 61 %). The median age at study entry was 77 years in the glasdegib + LDAC arm (62 % ≥ 75 years) and 76 years in the LDAC alone arm (61 % ≥ 75 years). Only one patient in each group was younger than 65 years of age (5).

De novo and secondary AML was nearly balanced in both treatment arms. 49 % in the glasdegib + LDAC arm and 47 % of the patients in the LDAC alone arm had de novo AML. Accordingly, 51 % in the glasdegib + LDAC arm and 53 % in the LDAC alone arm had secondary AML. Concerning the pre-specified criteria for not being suitable for intensive chemotherapy (i.e. unfit criteria), slight differences existed between the two groups in ECOG PS, since more patients in the glasdegib arm had ECOG PS 2 than in the LDAC only arm (53 % vs. 47 %). Also, more patients in the glasdegib arm had serum creatinine > 1.3 mg/dL (19 % vs. 13 %), a history of severe cardiac disease (67 % vs. 53 %) and met two or more of the unfit criteria (71 % vs. 55 %) (see Table 4.6 and Table 4.7) (5). Concerning cytogenetic risk, the majority had good/intermediate cytogenetic risk, but fewer patients in the glasdegib + LDAC arm had a poor cytogenetic risk than in the LDAC only arm (according to IVRS 37 % vs. 45 %; CRF 32 % vs. 42 %). The median peripheral blood white cell count was 2.5 103/mm3 and 4.1 103/mm3 in the glasdegib + LDAC or LDAC arm, respectively. The median bone marrow blasts were comparable between both treatment arms (41 % vs. 46 %). The duration since the histopathological diagnosis was nearly comparable between the treatment arms (median approximately two weeks).

Table 4.7. Number (%) of AML patients meeting specified unfit criteria at baseline

BRIGHT AML 1003 Glasdegib + LDAC N=78

LDAC N=38

Number of criteria met:

Exactly 1 23 (29.5) 17 (44.7)

Exactly 2 34 (43.6) 15 (39.5)

Exactly 3 19 (24.4) 5 (13.2)

Exactly 4 2 (2.6) 1 (2.6)

Source: Adapted from (5) Abbreviations: AML=acute myeloid leukaemia; LDAC=low-dose cytarabine; N=number of patients.



Statistics

The B1371003 trial included AML and MDS patients. The primary endpoint was OS defined as time from randomization to death by any cause. Patients not known to have died at the data cut-off date (3rd of January 2017) were right-censored at the last follow-up date they were known to be alive (4, 40). Initially planned subgroup analyses according to de novo/secondary/treatment-related AML, AML French-American-British (FAB) sub-types and whether patients had received prior HMAs were removed in Protocol Amendment 5 dated Feb 8, 2016 (41). Thus, the only pre-planned subgroup analyses mentioned in the statistical analysis plan (SAP) of the B1371003 trial was for cytogenetic risk groups (42).

The planned sample size was 132 patients (92 OS events), randomized 2:1, to accomplish 80 % power for testing the primary endpoint at a 10 % significance level (1-sided), assuming an expected hazard ratio (HR) of 0.625. Correspondingly, the assumed expected median OS was 5 months in the LDAC alone arm and 8 months in the glasdegib + LDAC arm. Hence, the final analysis was planned as soon as 92 OS events were observed (41). Of note, although within the trial 94 OS events were observed, 160 OS events would have been necessary to obtain a statistical power of at least 80 % to detect a true HR of 0.625 at a conventional alpha level of 5 % (2-sided) (8).

One interim analysis for futility was conducted at 46 OS events. The trial would have been stopped for futility at this interim analysis, if the observed HR had been larger than 0.92 (corresponding to a stopping probability of 61 % for an expected HR of 1 and 10 % for an expected HR of 0.625). Based on the results of the interim analysis, the trial continued as planned. The interim analysis was for futility only, therefore no alpha was spent. All efficacy and safety results presented in this assessment stem from the planned data cut-off for the final analysis (41, 42).

All efficacy analyses were based on all randomized patients, regardless of whether they have received the study drug (FAS). The primary endpoint (OS) was tested with a stratified log-rank test to compare the survival time distribution between both groups (significance level α = 10 %, 1-sided), stratified by cytogenetic risk (based on IVRS). The Kaplan-Meier method was utilized to estimate the survival distribution within both arms, the survival rates at month 6 and 12, as well as the median OS times with corresponding 80 % confidence intervals, respectively (42).

However, conclusions within this assessment (BRIGHT AML 1003) will be based on the more stringent significance level of 5 % (2-sided) and corresponding 95 % confidence intervals instead, in line with accepted standards for evaluating efficacy claims and as provided in the Submission Dossier by the MAH. For consistency reasons and if not explicitly stated otherwise, all reported p-values within this assessment are 2-sided (for potentially directional statistical hypothesis tests); if only 1-sided p-values were available these were converted to 2-sided p-values by multiplying by the factor two.

Kaplan-Meier estimates of OS time were also obtained for predefined subgroups according to good/intermediate or poor cytogenetic risk (based on IVRS). A Cox proportional-hazard regression model stratified by cytogenetic risk (based on IVRS) was used to obtain a HR estimate and corresponding 80 % confidence interval (42).

Sensitivity analyses of the primary endpoint regarding stratification (unstratified log-rank test and stratification by cytogenetic risk based on CRF instead of IVRS) and censoring rules (additional censoring of patients upon receiving transplant) were provided as well (42).

Predefined secondary endpoints were different forms of assessed clinical response: CR, CRi, and MFLS (for AML patients only). Secondary endpoints were summarized by response rates and 80 % confidence intervals. Differences between groups were tested with Cochran–Mantel–Haenszel tests stratified by cytogenetic risk (based on IVRS). Patients not known to have achieved a response endpoint were counted as non-responders (42). Secondary endpoints within the B1371003 trial were not adjusted for multiplicity and are therefore regarded as exploratory and as supportive evidence for the primary endpoint only.

Of note, the Submission Dossier exclusively focuses on results for the AML subgroup of the B1371003 trial. However, efficacy analyses within the AML subgroup were not specified in the study protocol, and were not controlled for multiple testing. From a statistical perspective, such analyses within a post-hoc



(potentially data-driven) selected subgroup are generally regarded as strictly exploratory. In the case of glasdegib, the following considerations indicate that a focus on AML subgroup results within this assessment might still be regarded reasonable. First, all statistically significant key efficacy results within the AML subgroup do also hold in all randomized patients (AML + MDS). Second, MDS patients make up only a small proportion (16/132 (12 %)) of all randomized patients and hence the majority of information stems from randomized AML patients. Third, based on the B1371003 trial, glasdegib was approved by the EMA for AML patients only. Last but not least, only the population of AML patients is stated in the assessment’s PICO.

Within the B1371003 trial, randomization was initially stratified by good/intermediate vs. poor cytogenetic risk (based on IVRS) to prevent imbalances between treatment arms regarding this potentially important prognostic factor. By focusing on AML patients only, stratification in the randomization process may not have been fully effective as suggested by baseline imbalances regarding IVRS-based cytogenetic risk between the glasdegib + LDAC and LDAC alone arm.

Within the core Submission Dossier and thus for AML patients only, the MAH provided additional results for quality-adjusted survival time (Q-TWiST analysis), transfusion independence, and overall response rate (ORR). ORR and transfusion independence were summarized by proportions and 95 % confidence intervals and group differences were tested with Cochran–Mantel–Haenszel tests stratified by cytogenetic risk (based on IVRS). For Q-TWiST, retrospective results were provided by partitioning the observed survival times of patients into time with toxicity (TOX; defined as days with AEs grade ≥ 3 prior to progression), time without symptoms of progression or toxicity (TWiST), and time post-progression (REL; progression defined as treatment discontinuation due to insufficient clinical response). Quality-adjusted survival times are obtained by down-weighting survival times with toxicity or post-progression. Average survival times in each state (TWiST, TOX, REL) were computed by restricted mean survival time (RMST) at month 20 for both arms. The base case assumed weights of 1 for survival times within the TWiST state and 0.5 for time spent in TOX and REL states, respectively (4). These additional post-hoc analyses of quality-adjusted survival time, transfusion independence, and ORR are regarded as exploratory evidence for the primary endpoint.

Safety analyses were based on all randomized AML patients, who received at least one dose of the study drug. Safety endpoints were summarized descriptively with absolute and relative frequencies (42).

Outcomes included

Table 4.8 shows for which outcomes data were available for BRIGHT AML 1003.

Table 4.8. Matrix of outcomes in BRIGHT AML 1003

Ov

era

ll s

urv

iva

l

Hea

lth

-rela

ted

Qo

L

Tra

ns

fus

ion

ind

ep

en

de

nce

Dis

ea

se r

es

po

ns

e

Se

rio

us

AE

s

Fa

tal

AE

s

Tre

atm

en

t

dis

co

nti

nu

ati

on

du

e t

o A

Es

BRIGHT AML 1003 Yes Noa Nob Yes Yes Yes Yes a Not pre-planned, but retrospective results from Q-TWiST available. b Not pre-planned, but retrospective results available. Abbreviations: AE=adverse event; AML=Acute myeloid leukaemia; QoL=quality of life.

Analyses for efficacy outcomes (i.e. OS, health-related quality of life, transfusion independence, disease response) were based on all randomized AML patients, regardless of whether they have received the study drugs (FAS). Safety analyses were based on all randomized AML patients who received at least one dose of any of the study medications (safety analysis set).

4.6.1 Overall survival

The primary outcome was defined as time from the date of randomization to the date of death from any cause (4). For life-threatening diseases, OS is considered the gold standard primary endpoint of clinical



trials by physicians and health regulatory agencies. OS is recognized as being unambiguous and not subject to investigator interpretation (48). Thus, OS will be included in the assessment.

4.6.2 Health-related quality of life

There was no pre-planned analysis of health-related quality of life and patient reported outcomes were not proactively collected. Instead, within the core Submission Dossier retrospective results from a Q-TWiST analysis were provided (4). However, this analysis was not pre-specified as efficacy endpoint within the B1371003 trial. It is therefore considered exploratory and will only be presented as supportive evidence for the primary endpoint.

4.6.3 Transfusion independence

Transfusion independence was not mentioned in the study protocol but according to the SAP, transfusion independence for unfit AML patients, was defined as free from platelets or packed red blood cells transfusion for at least one week. This was only of interest for AML patients who achieved CR on study (42).

However, transfusion independence was not measured at baseline. In the core Submission Dossier, transfusion independency was defined as a period of at least 56 consecutive days (≥ 8 weeks but also at 12, 16, 20 and 24 consecutive weeks during treatment phase) in which patients did not receive any type of blood transfusion - if patients achieved this during any time of their treatment, they were counted as responders for this endpoint (4).

Thus, information provided in the core Submission Dossier cannot be considered a pre-planned outcome as such. Results will be presented for information purposes as supportive evidence for the primary endpoint and have to be considered exploratory.

4.6.4 Disease response

Hematologic disease response was measured according to the revised recommendations of the International Working Group for diagnosis, standardization of response criteria, treatment outcomes, and reporting standards for therapeutic trials in AML (see Table 4.9) (5).

Disease response assessments were reported by investigators as well as derived by the sponsor. The investigator assessed morphologic disease response per patient. This was based on the locally analysed bone marrow aspirates or biopsies, blood samples, and other clinical assessments. Based on clinical assessment, the investigator entered a disease response for each patient in the CRF according to the criteria below (Table 4.9). Derived responses are the sponsor’s assessment of morphologic disease response determined programmatically. Algorithms based on AML and MDS response criteria used the same patient data in the CRF recorded by the investigator to derive each patient’s response (5).

Table 4.9. Definition of hematologic responses to treatment for AML

Response Criteria Peripheral Blood Bone Marrow Blasts (%)

Other

Neutrophils (µL)

Platelets (µL)

Morphologic Complete Response (CR) This is morphologic leukemia-free plus neutrophil & platelet response

≥1,000 ≥100,000 <5 with spicules present, no Auer rods

Transfusion independent, no EMD

Morphologic CR with incomplete blood count recovery (CRi)

<1,000 -or-

<100,000 <5 Either neutrophils or platelets not recovered, no EMD.

Morphologic leukemia-free state

<1,000 -and-

<100,000 <5 blasts in BM with spicules and no blasts with Auer rods

Neutrophils and platelets

not recovered, Flow cytometry negative. No EMD.



Response Criteria Peripheral Blood Bone Marrow Blasts (%)

Other

Neutrophils (µL)

Platelets (µL)

Partial remission (PR) ≥1,000 ≥100,000 Decrease to 5-25 & ≥50% decrease from start

Blasts ≤5% if Auer rod positive

Partial remission with incomplete blood count recovery (PRi)

<1,000 -or-

<100,000 Decrease to 5-25 & ≥50% decrease from start

Minor Response NA NA >25% decrease from start

Stable Disease NA NA Blasts stable ±25% from screening value

Source: (42) Abbreviations: BM=bone marrow; EMD=extramedullary disease; NA=not applicable. Comment to morphologic CR/column ”other”: Transfusion independent is defined as being free from platelets or packed red cell transfusions for at least one week (4).

According to the study protocol, a bone marrow evaluation (aspirate ± biopsy) was planned to determine disease response. Nonetheless, bone marrow evaluations were not performed for all patients (41). Therefore, these patients were non-evaluable and were counted as non-responders for disease response.

ORR, a composite endpoint (= CR + CRi + MLFS), was not a pre-planned endpoint (42), but was provided in the core Submission Dossier due to the request of the authors. However, this secondary endpoint was not controlled for multiple testing and is therefore regarded as exploratory, supportive evidence for the primary endpoint. In addition, since OS data are available, ORR results will be presented as supportive information but are not considered primarily relevant.

Similarly, even though CR is an outcome associated with improved survival and is by definition associated with partial regeneration of peripheral blood cell counts, so that transfusion of thrombocyte concentrates and hospitalization due to the high risk of infection is no longer required (49), CR results will be included in this assessment only as additional information since OS results are available.

4.6.5 Safety endpoints

Adverse events (AEs) were graded by the investigator according to the Common Terminology Criteria for Adverse events (CTCAE) version 4.0. In addition, an Internal Review Committee (IRC) was established. The Internal Oncology Business Unit-Safety Data Monitoring Committee (IOBU-SDMC) was a single, business unit-internal, project-independent advisory group established to enhance the early detection of potential safety signals in Oncology Business Unit-sponsored clinical trials (40).

Participant flow

As can be seen in Table 4.10, all patients in the LDAC alone arm and nearly all patients in the glasdegib + LDAC arm discontinued treatment. About 2.4-fold more patients discontinued the treatment because of death in the LDAC alone arm (31 %) than in the glasdegib + LDAC arm (13 %). However, the main reason for discontinuation was insufficient clinical response, with a higher proportion of patients in the combination arm (43 % vs. 33 %). At the cut-off date 15 patients (19 %) with glasdegib + LDAC and 1 patient (3 %) with LDAC alone were ongoing in the study. Of these, 4 patients with glasdegib + LDAC (5 %) remained on treatment (5).

Patients who discontinued study treatment could receive follow-up systemic therapies. Information is available only for the overall population of AML and MDS patients included in the B1371003 trial and not for the BRIGHT AML 1003 subgroup. 37 (44 %) of the 84 patients who had initially received glasdegib + LDAC received further therapies and 15 (37 %) of the 41 patients who were initially treated with LDAC alone (39). The vast majority (41 % vs. 34 %) received HMAs or palliative chemotherapy (i.e., azacitidine, decitabine, hydroxyurea, and hydroxycarbamide) (50).



Table 4.10. Participant flow of AML subgroup

BRIGHT AML 1003 Glasdegib + LDAC N (%)

LDAC N (%)

Assigned to study treatment 78 (100) 38 (100)

Analysed for efficacy 78 (100) 38 (100)

Discontinued from study 63 (80.8) 37 (97.4)

Patient died 59 (75.6) 35 (92.1)

Lost to follow-up 1 (1.3) 0

Patient refused further follow-up 3 (3.8) 2 (5.3)

Completed study at cut-off date 0 0

Ongoing in study at cut-off date 15 (19.2) 1 (2.6)

Treated 75 (100) 36 (100)

Treatment completed 0 0

Discontinued treatment 71 (94.7) 36 (100.0)

Insufficient clinical response 32 (42.7) 12 (33.3)

AE 18 (24.0) 11 (30.6)

Patient died 10 (13.3) 11 (30.6)

Global deterioration of health status

3 (4.0) 0

Other 3 (4.0) 0

Protocol violation 1 (1.3) 0

Patient refused continued treatment for other than AE

4 (5.3) 2 (5.6)

Treatment ongoing at cut-off date 4 (5.3) 0

Source: Adapted from (5, 45) Abbreviations: AE=adverse event; LDAC=low-dose cytarabine; N=number of patients.

Risk of bias

Risk of bias was assessed for each outcome according to the Cochrane RoB 2 tool (see Figure 4.3). For details concerning the risk of bias for each domain see Table A3 in the Appendix. Generally, the B1371003 trial is a Phase 1b/2, open-label study with few patients.

Limitations include lack of information concerning the random sequence generation. Even though the MAH states in the core Submission Dossier that the generation of the allocation sequence was adequate, this judgement cannot be verified, because the appendix of the CSR the MAH was referring to was not submitted. The only information available is that the investigator or designee registered eligible patients using an Interactive Registration System (IRS), patients were randomised (2:1) and were stratified in IRS based on prognosis (poor, good/intermediate). From the overall AML + MDS population, information exists that 6 (5 %) patients were randomized in error (no explanation provided) and 21 (16 %) patients were not stratified correctly at randomization in IVRS (40).

Even though the study initially included both AML and MDS patients, focusing on the AML subgroup was not considered to increase the risk of bias per se since random allocation within the AML subgroup was believed to be maintained: considering the small sample size, existing differences between arms in baseline characteristics within the AML subgroup could plausibly occur by chance alone, were not statistically significant (based on own calculation using information provided in Table 4.6 and Pearson chi-squared tests for categorical as well as t-tests for continuous baseline characteristics), and hence did not clearly indicate any systematic issues with the randomisation process. Also, the focus on the AML subgroup from BRIGHT AML 1003 did not raise concerns regarding a selection of reported outcomes because results from the B1371003 trial (including MDS and AML patients) were also available and widely consistent.



Due to the open-label design of the study, outcomes considered most relevant and which had a subjective component (SAE, discontinuation due to AE) were considered to be at risk of bias. In addition, the administration of concomitant therapies could have been influenced by knowledge of the assigned intervention.

Figure 4.3. Risk of bias for B1371003 trial

Abbreviations: AE=adverse event; OS=overall survival; SAE=serious AE.

External validity

Since phase 2 efficacy data for glasdegib + LDAC are available only from a subgroup of AML patients from one small phase 1b/2 study, the actual size of the treatment effect which can be expected in clinical practice remains uncertain.

Furthermore, since no direct comparisons have been performed with any therapy other than LDAC, the comparative effectiveness of glasdegib + LDAC foremost to HMAs but also to BSC cannot be assessed reliably. It also remains uncertain whether specific patient characteristics influence effectiveness and safety, e.g. good vs. adverse-risk cytogenetics, patients < 75 years vs. ≥ 75 years.

The optimal duration of therapy remains also unclear. In the B1371003 trial, it is mentioned that treatment with glasdegib + LDAC was considered complete after 12 cycles of therapy. Nonetheless, patients who completed the maximum number of cycles/months on study treatment could continue the therapy. In the US, the FDA label states that patients without unacceptable toxicity should be treated for a minimum of 6 cycles to allow time for clinical response. Although it seems reasonable to continue treatment with glasdegib + LDAC the criteria for prolonging therapy remain unclear.



Results on clinical effectiveness and safety

4.10.1 Overall survival

In BRIGHT AML 1003, the median follow-up for OS1 was 21.7 [95 % CI: 18.1, 26.5] months in glasdegib + LDAC arm vs. 20.1 [95 % CI: not estimable] months in the LDAC alone arm (5).

As can be seen in Table 4.11 and Figure 4.4, glasdegib + LDAC showed superior OS (HR [95 % CI]: 0.46 [0.30, 0.72]; p = 0.0004) than LDAC alone (median OS: 8.3 months vs. 4.3 months). Sensitivity analyses showed almost identical results. These results were also consistent for the 132 AML + MDS patients randomised in the B1371003 trial (HR [95 % CI]: 0.51 [0.34, 0.77]; p = 0.0008) (5).

Table 4.11. OS results for BRIGHT AML 1003

Outcome Glasdegib + LDAC LDAC Glasdegib + LDAC vs. LDAC

N Median time to event in

months [95 % CI]a

Patients with

event n (%)

N Median time to event in months

[95 % CI]a

Patients with

event n (%)

HR [95 % CI]b p-valuec

OS (IVRS based)

78 8.3 [4.7, 12.2]

59 (75.6) 38 4.3 [1.9, 5.7]

35 (92.1) Stratified analysisd: 0.46 [0.30, 0.72] 0.0004 Unstratified analysis: 0.45 [0.29, 0.69] 0.0002

Sensitivity analyses

OS (censoring patients upon receiving transplant)

78 8.3 [4.7, 12.2]

58 (74.4) 38 4.3 [1.9, 5.7]

35 (92.1) Stratified analysisd: 0.46 [0.30, 0.71] 0.0004

OS (CRF based)

78 8.3 [4.7, 12.2]

59 (75.6) 38 4.3 [1.9, 5.7]

35 (92.1) Stratified analysise: 0.47 [0.30, 0.72] 0.0005

Source: Adapted from (4, 45) a Based on the Brookmeyer und Cowley method. b Based on the Cox Proportional hazards model. c 2-sided p-value from the (stratified) log-rank test. d Stratification factor: IVRS based cytogenetic risk (good/intermediate vs. poor). e Stratification factor: CRF based cytogenetic risk (good/intermediate vs. poor). Abbreviations: CI=confidence interval; CRF=case-report form, HR=hazard ratio, IVRS=interactive voice response system, LDAC=low-dose cytarabine; n=number of patients with (at least one) event; N=number of analysed patients; OS=overall survival, RR=risk ratio; vs.=versus.

1 In the core submission dossier the MAH states a median follow-up for OS of 156.5 (range: 2 – 1005) days in total, thereof 226 (range: 5 – 1005) days in the glasdegib + LDAC arm and 115 (range: 2 – 613) days in the LDAC alone arm.



Figure 4.4. Kaplan-Meier plot of OS for BRIGHT AML 1003

Source: (4) Abbreviations: CI=confidence interval; HR=hazard ratio; LDAC=low-dose cytarabine; OS=overall survival. Comment: The reported p-value of 0.0002 is 1-sided and corresponds to a 2-sided p-value from the (stratified) log-rank test of 0.0004.

Survival rates and survival probabilities at month 6, and month 12 are presented in Table 4.12. At each time point considered in the analyses, the survival probabilities in the glasdegib + LDAC arm were higher than in the LDAC arm.

Table 4.12. Results summary for survival rates and survival probabilities for BRIGHT AML 1003


LDAC N = 38

Survival rates Survival probabilities (%) [95 % CI]

Survival rates Survival probabilities (%) [95 % CI]

patients died

patients ongoing

patients died

patients ongoing

At month 6a 30 44 59.8 [47.7, 69.9] 24 12 33.5 [18.9, 48.7]

At month 12a 45 29 39.4 [28.3, 50.3] 33 3 8.4 [2.2, 20.1]

Source: (4) a Four patients in the glasdegib + LDAC arm vs. two patients in the LDAC alone arm were lost to follow-up or refused further follow-up. Abbreviations: CI=confidence interval; LDAC=low-dose cytarabine; N=number of patients.



4.10.2 Overall survival - Subgroup analyses

A subgroup analysis was prespecified for cytogenetic risk (good/intermediate vs. poor) groups in the B1371003 trial. Additionally, subgroups by age, gender, race, region, baseline ECOG, baseline white blood cells count, baseline bone marrow blast count, prognostic risk factors and disease history were presented.

Subgroup analyses by cytogenetic risk As can be seen in Table 4.13, among the 116 AML patients, glasdegib + LDAC showed superior OS compared to LDAC alone (median OS improvement of 4 months) in BRIGHT AML 1003 (4). This result was also consistent for the 132 AML + MDS patients randomised in the B1371003 trial (HR [95 % CI]: 0.51 [0.34, 0.77]; p = 0.0008; median OS improvement of 3.9 months) (5).

In a subgroup analysis, stratified by IVRS based cytogenetic risk, 70 AML patients with good/intermediate cytogenetic risk showed an improvement in OS in the glasdegib + LDAC arm compared to LDAC alone (median OS improvement of 6.7 months). This result was also consistent within the prespecified subgroup analysis including 77 AML + MDS patients with good/intermediate cytogenetic risk (HR [95 % CI]: 0.43 [0.25, 0.73]; p = 0.0016; median OS improvement of 7.4 months) (5).

On the contrary, descriptively a smaller median OS improvement was observed in 46 AML patients with poor cytogenetic risk in favour of glasdegib + LDAC in comparison to LDAC alone (median OS improvement of 1.3 months), as well as for the 55 AML + MDS patients with poor cytogenetic risk within the prespecified subgroup analysis (HR [95 % CI]: 0.63 [0.35, 1.15]; p = 0.128; median OS reduction of 0.2 months) (5). However, although subgroup analyses by cytogenetic risk were pre-specified, the B1371003 trial was not designed to detect efficacy differences between cytogenetic risk groups and hence these results should be regarded as exploratory only.

Table 4.13. Summary of OS for AML patients according to their cytogenetic risk by IVRS stratification in BRIGHT AML 1003



months [95 % CI]a

Patients with

event n (%)


months [95 % CI]a

Patients with

event n (%)

HR [95 % CI] p-valuec

OS (IVRS based) 78 8.3 [4.7, 12.2]a

59 (75.6) 38 4.3 [1.9, 5.7]a

35 (92.1) 0.46 [0.30, 0.72]b 0.0004

Subgroup analysis

OS in patients with good/intermediate cytogenetic risk (by IVRS)

49 11.1 [7.1, 14.9]

35 (71.4) 21 4.4 [1.8, 8.7]

19 (90.5) 0.42 [0.23, 0.74] 0.0022

OS in patients with poor cytogenetic risk (by IVRS)

29 4.4 [3.4, 9.1]

24 (82.8) 17 3.1 [1.1, 6.4]

16 (94.1) 0.53 [0.27, 1.02] 0.0538

Source: (4, 5) a Based on the Brookmeyer und Cowley Method. b Based on the Cox Proportional hazards model. c 2-sided p-value from the (stratified) log-rank test. Abbreviations: CI=confidence interval; HR=hazard ratio; IVRS=interactive voice response system, LDAC=low-dose cytarabine; n=number of patients with (at least one) event; N=number of analysed patients; OS=overall survival.



Additional subgroup analyses Additional subgroup analyses are presented in Figure 4.5. The direction of subgroup treatment effects were consistent with the overall effect. Descriptively, differences in the treatment effect of potentially relevant magnitude were observed between subgroups according to age (smaller effects for patients aged 75 or older), ECOG subgroups (smaller effects for ECOG ≤ 1), subgroups according to baseline bone marrow blast count (smaller effects for ≥ 30 %), prognostic risk factor subgroups (smaller effects for patients with intermediate-II risk), and subgroups according to disease history (smaller effects for de-novo AML).

Figure 4.5. Forest plot of OS for BRIGHT AML 1003

Source: (4) Abbreviations: AML=acute myeloid leukaemia; CI=confidence interval; ECOG=Eastern Cooperative Oncology Group; HR=hazard ratio; LDAC=low-dose cytarabine; N=number of subjects; Yrs=years. Comment: The sample sizes for the subgroups of baseline age < 65 years old, race other than white and the prognostic risk factor as favourable were too small (n ≤ 10) for analysis. The HR and p-value presented were based on the unstratified analysis for all subgroups except for AML population. Prognostic risk factor (good/intermediate vs. poor) from IVRS was used as a stratification factor in the stratified analysis for AML population.

4.10.3 Quality of Survival – Q-TWiST

Patient reported outcomes (PRO) were not proactively collected because Study B1371003 was a phase 1b/2 study initially designed as a dose finding and proof of concept study.

The MAH provided a retrospective Q-TWiST analysis to partition the observed OS time into three different states: Survival time with toxicity (TOX state), survival time after disease progression (REL state), and survival time without toxicity or disease progression (TWiST). Assuming different utility values for the three different states, quality adjusted OS times were computed and compared between the two treatment arms.

Figure 4.6 depicts the Kaplan-Meier Survival curve of OS (magenta) within the glasdegib + LDAC arm. The area under each curve up to month 20 corresponds to the RMST in each included state. OS time is partitioned into time spent within TOX (area under the green curve), TWiST (area between the orange and green curve), and REL (area between the magenta and orange curve) health states.



Figure 4.6. Partitioned survival curve for glasdegib + LDAC

Source: (4) Abbreviations: AE=adverse event; LDAC=low-dose cytarabine; OS=overall survival; REL=time after progression/relapse; TOX=time with grade ≥ 3 toxicity; TWiST=time without symptoms of disease progression or toxicity (i.e., “good” health).

Further results of the initial Q-TWiST analysis submitted by the MAH within the core Submission Dossier contained major unexplained inconsistencies. In particular, the (unweighted) sum of RMST reported within the TOX, TWiST, and REL states did not produce the overall RMST up to month 20. This cannot be the case, if survival times in the different states are computed correctly (and survival time is wrongfully counted within multiple states). During the fact check procedure the MAH adapted the results.

Adapted results were as follows: Descriptively, the estimated RMST [95 % CI] at month 20 was 9.8 [8.2, 11.5] months in the glasdegib + LDAC group and 5.3 [3.9, 7.1] months in the LDAC alone group (RMST improvement of 4.5 [2.0, 6.9] months in favour of glasdegib + LDAC). If time spent in the TOX and REL states are both downweighed with utility values of 0.5, the quality adjusted survival time (RMST at month 20) was reported as 7.8 [6.5, 9.3] months in the glasdegib + LDAC group and 3.8 [2.8, 4.9] months in the LDAC alone group (quality adjusted RMST improvement of 4 [2.1, 5.8] months in favour of glasdegib + LDAC, comprising 3.4, 0.4, and 0.2 months in the TWiST, TOX, and REL state, respectively; Figure 4.7). For further details and sensitivity analyses see (4).



Figure 4.7. Restricted mean duration of health states with TOX and REL downweighed by 0.5

Source: (4) Abbreviations: CI=Confidence interval; LDAC=low-dose cytarabine; REL=time after progression/relapse; TOX=time with grade ≥ 3 toxicity; Q-TWiST=quality-adjusted time without symptoms of disease progression or toxicity.

4.10.4 Transfusion independence

Transfusion independence data for ≥ 8, 12, 16, 20 and 24 weeks as well as exposure-adjusted transfusion rates per month of treatment and transfusion days per month of treatment are presented in Table 4.14. In BRIGHT AML 1003, descriptively more patients receiving glasdegib + LDAC than patients receiving LDAC alone were transfusion independent for ≥ 8, ≥ 12, ≥ 16, ≥ 20, ≥ 24 weeks (e.g. transfusion independence ≥ 8 weeks: 28.2 % vs. 5.3 %; p=0.0056). The mean exposure-adjusted transfusion rate per month of treatment was 4.8 in the glasdegib + LDAC arm vs. 8.1 in the LDAC alone arm. Patients with glasdegib + LDAC received transfusions on average over 3.9 days per month of treatment, whereas patients with LDAC alone received transfusions on average over 5.9 days per month of treatment.



Table 4.14. Results summary for transfusion independence for BRIGHT AML 1003

Duration of transfusion independencea


LDAC N = 38

RR [95 % CI] p-value from CMHb

≥ 8 weeks 22 (28.2) 2 (5.3) 5.17 [1.25, 21.41] 0.0056

≥ 12 weeks 18 (23.1) 2 (5.3) 4.27 [1.02, 17.86] 0.0214

≥ 16 weeks 16 (20.5) 2 (5.3) 3.82 [0.91, 16.08] 0.0396

≥ 20 weeks 16 (20.5) 2 (5.3) 3.82 [0.91, 16.08] 0.0396

≥ 24 weeks 14 (17.9) 0 (0.0) n.e. 0.0067

Source: (4) a No transfusions for ≥ 8, 12, 16, 20 and 24 consecutive weeks during treatment phase. b Stratified Cochran–Mantel–Haenszel test and risk ratio estimate; Stratification factor: IVRS based cytogenetic risk (good/intermediate vs. poor). Abbreviations: CI=confidence interval; CMH=Cochran–Mantel–Haenszel test; IVRS=interactive voice response system; LDAC=low-dose cytarabine; n.e.=not estimable; RR=risk ratio.

4.10.5 Disease response

Complete Remission The analysis of CR (investigator-reported, stratified by IVRS based cytogenetic risk) is summarized in Table 4.15. More patients in the glasdegib + LDAC arm (n = 14, 17.9 %) achieved CR compared with LDAC alone (n = 1, 2.6 %) (RR [95 % CI]: 7.10 [0.89, 56.83]; p = 0.0235). Sensitivity analyses showed similar results. Consistent results were also found in the AML + MDS overall population in the B1371003 trial (RR [95 % CI]: 7.60 [1.01, 57.29]; p = 0.0152) (5).

Table 4.15. Results for CR for BRIGHT AML 1003


N n (%) N n (%) RR [95 % CI] p-value

Investigator-reported CR (IVRS)

78 14 (17.9)

38 1 (2.6)

Stratified analysisa: 7.10 [0.89, 56.83] 0.0235 Unstratified analysis: n.r. 0.0210b

Sensitivity analyses

Investigator-reported CR (CRF)

78 14 (17.9)

38 1 (2.6)

Stratified analysisc: 7.22 [0.86, 60.52] 0.0249

Derived responsed 78

14 (17.9)

38 1 (2.6) Stratified analysisa: 7.10 [0.89, 56.83] 0.0235 Stratified analysisc: 7.21 [0.86, 60.52] 0.0249

Source: (4) a Stratified Cochran–Mantel–Haenszel test and risk ratio estimate; Stratification factor: IVRS based cytogenetic risk (good/intermediate vs. poor). b Pearson Chi-Square Test. c Stratified Cochran–Mantel–Haenszel test and risk ratio estimate; Stratification factor: CRF based cytogenetic risk (good/intermediate vs. poor). d Derived by the sponsor. Abbreviations: CI=confidence interval; CR=complete remission; CRF=case-report form; IVRS=interactive voice response system, LDAC=low-dose cytarabine; n=number of patients with (at least one) event; N=number of analysed patients; RR=risk ratio; vs.=versus.



Objective response rate Table 4.15 and Table 4.16 summarize the ORR in terms of CR, CRi, MLFS and the composite endpoint ORR (CR + CRi + MLFS). Glasdegib + LDAC indicated better objective response assessed as CR + CRi + MLFS when compared to LDAC alone (RR [95 % CI]: 4.94 [1.19, 20.52], p = 0.0080).

Even though response should be assessed by a bone marrow evaluation, there was a relatively high number of patients that were not assessed: 24 (31 %) patients with glasdegib + LDAC and 16 (42 %) patients with LDAC alone. Main reasons for patients not being evaluated for response include AEs leading to study termination or death prior to on-study bone marrow sampling and withdrawal of consent.

Of the remaining 33 (42 %) patients in the glasdegib + LDAC arm, 23 (29 %) patients had either different forms of partial response or stable disease and 10 (13 %) patients had a resistant disease or a morphologic relapse (5). Of the remaining 20 (53 %) patients in the LDAC alone arm, 14 (37 %) had different forms of partial response or stable disease or their status could not be determined and 6 (16 %) patients showed resistant disease (5).

Table 4.16. Results summary for objective response for BRIGHT AML 1003


LDAC N = 38

n (% [95 % CI]a) n (% [95 % CI]a)

ORRb 21 (26.9 [17.5, 38.2]) 2 (5.3 [0.6, 17.7])

- Morphologic CR 14 (17.9 [10.2, 28.3]) 1 (2.6 [0.1, 13.8])

- Morphologic CRi 5 (6.4 [2.1, 14.3]) 1 (2.6 [0.1, 13.8])

- MLFS 2 (2.6 [0.3, 9.0]) 0 (0 [0.0, 9.3])

Not evaluablec 24 (30.8 [20.8, 42.2]) 16 (42.1 [26.3, 59.2])

Source: (4) a Using exact method based on binomial distribution. b Response is Investigator-reported. c Not evaluable is defined as patients that were not assessed for response. Abbreviations: CI=confidence interval; CR=complete remission; CRi=CR with incomplete blood count recovery; IVRS=interactive voice response system; LDAC=low-dose cytarabine; MLFS=morphologic leukaemia-free state; n=number of patients with (at least one) event; N=number of analysed patients; ORR=overall response rate.

4.10.6 Adverse events

In the glasdegib + LDAC arm, the median treatment duration was 83 (3 - 972) days vs. 41 (6 – 239) days for the LDAC alone arm. 28/75 (37 %) patients in the glasdegib + LDAC arm and 2/36 (6 %) patients in the LDAC alone arm received at least 6 cycles of treatment.

Table 4.17 summarizes the total AE endpoints for BRIGHT AML 1003, for the entire study period and for the first 90 days of therapy (to account for imbalances in treatment duration between the glasdegib + LDAC and LDAC alone arm).

Considering the entire study period, glasdegib + LDAC showed comparable rates to LDAC alone for AEs of any CTCAE grade (100 % in both groups). The most frequent AEs of any grade occurring in ≥ 20 % of patients were gastrointestinal disorders (77 % vs. 67 %), general disorders and administration site conditions (76 % vs. 67 %), and blood and lymphatic system disorders (71 % vs. 64 %). In addition, there was a higher frequency of specific AEs (nausea (36 % vs. 11 %), decreased appetite (32 % vs. 11 %), dysgeusia (24 % vs. 3 %)) in the glasdegib + LDAC arm compared to LDAC alone. Concerning

AEs of system organ classes in ≥ 20 % of patients, musculoskeletal and connective tissue disorders

(56 % vs. 31 %; e.g., muscle spasms), and nervous system disorders (59 % vs. 22 %; e.g., dizziness, dysgeusia) were more frequently reported in the glasdegib + LDAC arm than in the LDAC alone arm (see Table A5 in the Appendix).

No relevant differences between glasdegib + LDAC and LDAC alone occurred neither during the entire study period regarding the AEs commonly expected for antileukaemic treatments such as febrile neutropenia (35 % vs. 25 %), haemorrhage (48 % vs. 50 %), QT prolongation (20 % vs. 11 %), and infections including pneumonia (61 % vs. 56 %), nor for the first 90 days of therapy (see Table 4.17).



Comparable rates between the groups were also found concerning SAEs (79 % vs. 78 %), severe AEs grade 3 - 5 (92 % vs. 97 %), fatal AEs (29 % vs. 44 %) and AEs leading to permanent discontinuation of all study drugs (31 % vs. 47 %) during the entire study period. Considering the first 90 days of therapy, glasdegib + LDAC suggests a nominal advantage over LDAC alone regarding fatal AEs (16 % vs. 36 %), but comparable rates for SAEs (65 % vs. 72 %) and treatment discontinuations due to AEs (20 % vs. 33 %) (Table 4.17).

Considering the entire study period, the most frequently reported SAEs that occurred in ≥ 2 % of patients were pneumonia (21 % vs. 19 %), sepsis (4 % vs. 14 %), febrile neutropenia (28 % vs. 17 %), anaemia (7 % vs. 0), pancytopenia (0 % vs. 6 %) and disease progression (9 % vs. 11 %). During the same period, the most common AEs leading to treatment discontinuation in ≥ 2 % of patients comprised pneumonia (5 % vs. 3 %), sepsis (1 % vs. 6 %), and febrile neutropenia (3 % vs. 6 %). No information was provided concerning reasons for fatal AE.



Table 4.17. Results summary for total AEs for BRIGHT AML 1003

AE Outcome Study period Within the first 90 days of therapy

Glasdegib + LDAC N = 75 n (%)

LDAC N = 36 n (%)

RR [95 % CI]; p-value

RD [95 % CI]; p-value


LDAC N = 36 n (%)



Total patients with at least one TEAEa/TEAE90b

AEs (any CTCAE grade)

75 (100.0) 36 (100.0) n. e. n. e. 74 (98.7) 36 (100.0) 0.99 [0.96, 1.01]; 0.3173

-0.01 [-0.04, 0.01]; 0.3141

AEs commonly expected for antileukaemic treatments

- Febrile neutropenia (PT)

26 (34.7) 9 (25.0) 1.39 [0.73, 2.64]; 0.3209

0.10 [-0.08; 0.27]; 0.2866

23 (30.7) 8 (22.2) 1.38 [0.69, 2.78]; 0.3668

0.08 [-0.09, 0.26]; 0.3339

- Haemorrhage (SMQ)

36 (48.0) 18 (50.0) 0.96 [0.64, 1.44]; 0.8425

-0.02 [-0.22, 0.18]; 0.8436

27 (36.0) 17 (47.2) 0.76 [0.48, 1.21]; 0.2462

-0.11 [-0.31, 0.08]; 0.2616

- QT prolongation (SMQ)

15 (20.0) 4 (11.1) 1.80 [0.64, 5.04]; 0.2628

0.09 [-0.05, 0.23]; 0.2031

10 (13.3) 4 (11.1) 1.20 [0.40, 3.57]; 0.7429

0.02 [-0.11, 0.15]; 0.7342

- Infections (SOC)

46 (61.3) 20 (55.6) 1.10 [0.78, 1.56]; 0.5718

0.06 [-0.14; 0.24]; 0.5638

39 (52.0) 19 (52.8) 0.99 [0.68, 1.44]; 0.9386

-0.01 [-0.21, 0.19]; 0.9388

- Pneumonia (PT)

21 (28.0) 10 (27.8) 1.01 [0.53, 1.91]; 0.9805

0.00 [-0.18; 0.18]; 0.9805

14 (18.7) 9 (25.0) 0.75 [0.36, 1.56]; 0.4373

-0.06 [-0.23, 0.10]; 0.4565

SAEs

59 (78.7) 28 (77.8) 1.01 [0.82, 1.25]; 0.9158

0.01 [-0.16, 0.17]; 0.9156

49 (65.3) 26 (72.2) 0.90 [0.70, 1.17]; 0.4519

-0.07 [-0.25, 0.11]; 0.4574

Severe AEs (grade 3-5)

69 (92.0) 35 (97.2) 0.95 [0.87, 1.03]; 0.2116

-0.05 [-0.13, 0.03]; 0.2095

65 (86.7) 33 (91.7) 0.95 [0.83, 1.08]; 0.4070

-0.05 [-0.17, 0.07]; 0.4087

Fatal AEs

22 (29.3) 16 (44.4) 0.66 [0.40, 1.10]; 0.1080

-0.15 [-0.34, 0.04]; 0.1234

12 (16.0) 13 (36.1) 0.44 [0.23, 0.87]; 0.0184

-0.20 [-0.38, -0.02]; 0.0264

Treatment discontinuations due to AE

23 (30.7) 17 (47.2) 0.65 [0.40, 1.05]; 0.0810

-0.17 [-0.36, 0.03]; 0.0937

15 (20.0) 12 (33.3) 0.60 [0.31, 1.15]; 0.1216

-0.13 [-0.31, 0.05]; 0.1435

Source: (4) aTEAE are those events with onset date up to 28 days after last dose of study drug. bTEAE90 are those events with onset date up to 90 days after first dose of study drug. Abbreviations: AE=adverse event; CI=Confidence interval; CTCA=Common Terminology criteria for adverse events, LDAC=Low-dose cytarabine; n.e.=not estimable; PT=preferred term; RD=risk difference; RR=risk ratio; SAE=serious AE; SMQ=standardised MedDRA query; SOC=system organ class; TEAE=treatment-emergent adverse events.



Indirect evidence

4.11.1 Characteristics of AZA-AML-001 and DACO-016

AZA-AML-001 (Dombret 2015) This multicentre, randomized, open-label, phase 3 trial compared azacitidine vs. conventional care regimens in patients aged ≥ 65 years with newly diagnosed de novo or secondary AML presenting with > 30 % bone marrow blasts. Patients were not considered eligible for hematopoietic stem cell transplantation and had intermediate- or poor-risk cytogenetics (NCCN 2009 criteria; see Table A2 in the Appendix) and ECOG PS ≤2 (4). The study was conducted between October 2010 and January 2014 in 18 countries.

Before randomization, investigators determined which one of three protocol-designated conventional care regimens (BSC, LDAC, or induction chemotherapy) was most appropriate for each patient on the basis of age, ECOG PS, comorbidities, and regional guidelines and/or institutional practice. Within each of the three different conventional care regimens (BSC, LDAC, or induction chemotherapy), patients were then randomly assigned 1:1 to receive azacitidine or the pre-selected conventional care regimen by a central, stratified, and permuted block randomization method with an IVRS. Randomization was further stratified by ECOG PS (0-1 or 2), and cytogenetic risk (intermediate or poor). Patients assigned to a conventional care regimen received their preselected treatment. The study was conducted between October 2010 and January 2014; the database was locked in March 2014.

Overall, 488 patients participated in the study of whom 89 were preselected for BSC, 312 for LDAC and 87 to intensive chemotherapy. This led to overall 241 individuals randomised to azacitidine, 45 to BSC only, 44 to induction chemotherapy and 158 to LDAC. For use in the indirect treatment comparison (ITC), the MAH only used the data of the 312 patients preselected for LDAC of whom 154 were randomised to azacitidine and 158 to LDAC (see Table A6 in the Appendix) (4).

Azacitidine 75 mg/m2 per day was administered subcutaneously for 7 consecutive days per 28-day treatment cycle for at least 6 cycles. The conventional care regimens were as follows: BSC only consisting of blood product transfusions and antibiotics, with granulocyte colony-stimulating factor (G-CSF) for neutropenic infection; subcutaneous LDAC (20 mg twice per day for 10 days per 28-day treatment cycle for at least 4 cycles); or intensive chemotherapy for 1 cycle with intravenous cytarabine (for 7 days), followed by either daunorubicin or idarubicin (for 3 days), followed by up to 2 consolidation cycles with the same compound used at induction for those achieving CR or partial response. Reinduction was not allowed. All study participants could receive BSC, including transient use of hydroxyurea (see Table A7 in the Appendix) (35). For information on the course of the study see Table A8 in the Appendix.

DACO-016 (Kantarjian 2012) This randomized, open-label, phase 3 study compared decitabine with treatment choice in patients ≥65 years with newly diagnosed de novo or secondary AML (≥20 % blasts) and poor- or intermediate-risk cytogenetics (SWOG categorization; see Table A2 in the Appendix), ECOG PS of 0-2, and adequate organ function determined by laboratory evaluation (44, 51). The study was conducted between January 2006 and April 2009 in 15 countries.

Patients indicated, advised by their physician, their preferred treatment choice, either supportive care or LDAC. Patients were then randomly assigned 1:1 to receive decitabine or their preferred treatment choice by using a stratified permuted block method. Random assignment was stratified by age, cytogenetic risk, and ECOG PS (4).

Of overall 485 patients, 242 were assigned to decitabine, 215 to LDAC and 28 to supportive care. Decitabine was administered as 1-hour IV infusions of decitabine 20 mg/m2 one time per day for five consecutive days every 4 weeks, and LDAC 20 mg/m2 one time per day subcutaneously for 10 consecutive days every 4 weeks (see Table A6 and Table A7 in the Appendix). This dose schedule was considered by the MAH to have comparable drug exposition over time (area under the curve) including any associated cytotoxic effects (52) to twice per day dosing. Supportive care included, but was not limited to antibiotics, packed red blood cells or whole blood transfusions, fresh frozen plasma, platelet transfusions. Subjects were allowed to receive erythropoietin and darbepoetin during this study; however, use of G-CSF and Granulocyte Macrophage-CSF was restricted to the treatment of severe



infection (34). Treatment continued until relapse or progressive disease, death, unacceptable toxicity, lack of clinical benefit, intercurrent illness preventing treatment, or patient/physician request. Patients were allowed hydroxyurea until cycle 1 on day 15. For information on the course of the study see Table A8 in the Appendix

4.11.2 Baseline characteristics

AZA-AML-001 For the AZA-AML-001 trial, baseline characteristics for all patients randomised to azacitidine (n = 241) irrespective of their preselected conventional care regimen, and for patients randomised to LDAC in the conventional care arm (n = 158) are presented (see Table A9 in the Appendix). Thus, baseline characteristics are not expected to be entirely comparable between these treatment arms. No baseline characteristics of patients randomised to azacitidine who were preselected for LDAC (n = 154) prior to randomization were available, although only these patients from the azacitidine arm were included in the MAH’s ITC.

The majority of patients were male (azacitidine: 58 %; LDAC: 60 %). The median age at study entry was 75 years in both arms, with some differences in the patients aged ≥75 years (57 % vs. 53 %) The majority of patients showed good or intermediate cytogenetic risk, as defined by NCCN 2009 criteria: 64 % in the azacitidine arm and 66 % in the LDAC arm. The majority of patients enrolled in the study had de novo AML, but slightly more patients in the LDAC arm: 80 % in the azacitidine arm vs. 85 % in the LDAC arm. Regarding ECOG PS, the majority of patients in the study had ECOG PS 0 or 1 (77 % in the azacitidine arm vs. 78 % in the LDAC arm). The median peripheral blood white cell count was 3.1 103/mm3 and 2.3 103/mm3 in the azacitidine and LDAC arm, respectively. The median bone marrow blasts were nearly comparable between both treatment arms (70 % vs. 74 %).

DACO-016 As can be seen in Table A9 in the Appendix, overall, the majority of patients were male (decitabine: 57 %; treatment choice: 62 %). The median age at study entry was 73 years in both arms. The majority of patients showed good or intermediate cytogenetic risk, as defined by South West Oncology Classification: 63 % in the decitabine arm and 64 % in the treatment choice arm. The majority of patients enrolled in the study had de novo AML: 64 % in the decitabine arm vs. 65 % in the treatment choice arm. Regarding ECOG PS, the majority of patients in the study had ECOG PS 0 or 1 (76 % in the decitabine arm vs. 75 % in the treatment choice arm). The median peripheral blood white cell count was 3.1 103/mm3 and 3.7 103/mm3 in the decitabine or treatment choice arm, respectively. The median bone marrow blasts were nearly comparable between both arms (47 % vs. 45 %).

BRIGHT AML 1003 vs. AZA-AML-001 and DACO-016 Across all three studies (see Table A9 in the Appendix), the majority of patients were male, but the proportion of male patients was slightly higher in BRIGHT AML 1003 compared to AZA-AML-001 and DACO-016 (76 % or 61 % compared to 58 % or 60 % and 57 % or 62 %, respectively). The median age at study entry across treatment arms ranged from 73 to 77 years, with age ranges from 58 to 92 years.

Due to the different systems for assigning cytogenetic risk (see Table A2 in the Appendix), a comparison between trials is difficult. However, the majority of patients showed good or intermediate cytogenetic risk (between 55 % and 66 %). De novo and secondary AML was nearly balanced in both treatment arms of BRIGHT AML 1003, whereas in the AZA-AML-001 study and in the DACO-016 study the majority of patients had de novo AML (80 % and 85 % or 64 % and 65 % of patients, respectively). In BRIGHT AML 1003 the proportion of patients with ECOG PS 0-1 and ECOG PS 2 distributed almost equally, in AZA-AML-001 and DACO-016 most patients showed ECOG PS 0-1 (between 75 % and 78 % of patients). The median peripheral white blood cell count in the BRIGHT AML 1003 study was 2.5 103/mm3 and 4.1 103/mm3 in the glasdegib + LDAC or LDAC arm, respectively. In AZA-AML-001 and DACO-016 it ranged from 2.3 103/mm3 to 3.7 103/mm3. The bone marrow blasts were comparable between the treatment arms but differed between the studies: In BRIGHT AML 1003 and DACO-016 fewer patients had bone marrow blasts > 50 % (40 % or 45 % and 44 % or 42 %, respectively), whereas in AZA-AML-001 the majority of patients had bone marrow blasts > 50 % (72 % and 81 % respectively). The median haemoglobin level ranged from 9.1 g/dL to 9.5 g/dL. The duration since the histopathological diagnosis was comparable between treatment arms and across studies.



4.11.3 Critical Appraisal

No direct evidence is available comparing glasdegib with any other intervention included in the PICO apart from LDAC. An anchored indirect comparison was conducted by the MAH comparing glasdegib + LDAC (BRIGHT AML 1003) to azacitidine (AZA-AML-001 trial) and decitabine (DACO-016 trial) in terms of OS via the common comparator LDAC. Comparative effectiveness estimates were obtained by using hazard ratios and standard approaches for anchored ITCs (Bucher method). As additional supportive analyses simulated treatment comparisons (STCs) were conducted to ideally account for potential between trial differences in (known) patient characteristics using individual patient data from BRIGHT AML 1003 and aggregate data from the AZA-AML-001 and DACO-016 trial, respectively.

After carefully reviewing the submitted indirect evidence, the following limitations have been identified which question the validity of the estimated comparative effects from the indirect treatment comparisons.

There are considerable differences between trials in potentially effect modifying patient characteristics (e.g., type of AML, ECOG PS, in particular due to different trial-specific inclusion criteria (see Table A6 in the Appendix). Hence, the central similarity assumption for valid indirect treatment comparisons seems likely violated;

For inclusion into the B1371003 trial, patients had to meet at least one of the following criteria: age ≥ 75 years and/or ECOG PS 2 and/or serum creatinine > 1.3 mg/dL, and/or severe cardiac disease. These inclusion criteria presumably lead to evidently more patients having ECOG PS 2 in the BRIGHT AML 1003 compared to the AZA-AML-001 and DACO-016 trial included in the ITC. Also, because of these specific eligibility criteria in the B1371003 trial it seems reasonable to assume that systematically more patients in BRIGHT AML 1003 had a history of severe cardiac disease and/or serum creatinine > 1.3 mg/dL. However, history of severe cardiac disease and serum creatinine levels are unknown for patients in the AZA-AML-001 and DACO-016 trial included in the ITC. Subgroup analyses from BRIGHT AML 1003 descriptively show larger effects in patients with ECOG PS 2 and secondary AML, supporting the effect modifying potential of these factors. The effect modifier status of severe cardiac disease and serum creatinine > 1.3 mg/dL is unknown; no subgroup analyses regarding these factors were provided;

Although STCs were conducted to ideally control for between-trial imbalances, potentially relevant effect modifiers could not be controlled for because of unavailable information from trials included in the ITC (e.g., history of severe cardiac disease, serum creatinine > 1.3 mg/dL);

From the inclusion criteria of the AZA-AML-001 and DACO-016 trial it remains unclear, whether the patient population exclusively included individuals not eligible for standard induction chemotherapy, the population of interest for this assessment. Consistently to the questionable similarity assumption, the two studies possibly do not include the population of interest, making the results of the ITCs non-informative for decision-making. Due to differences between trials regarding patient populations and potential effect modifiers, biased indirect effect estimates seem likely. The magnitude and direction of this bias cannot be reliably established;

Specific to the AZA-AML-001 trial, the proportional hazards assumption might be violated. The Kaplan-Meier curves for OS of both arms cross at month 20 (4) and the small and insignificant HR (HR [95 % CI]: 0.90 [0.70, 1.16]; p = 0.43) is not consistent with the substantial difference in median survival time of 4.8 months (AZA: 11.2 [95 % CI: 8.8, 13.4] months; LDAC: 6.4 [95 % CI 4.8, 9.1] months). Therefore, basing the ITC on the HR from the AZA-AML-001 trial might not produce meaningful results;

In the AZA-AML-001 trial, treatment of patients within the control group was preselected by the investigator (LDAC or BSC or induction therapy). For the indirect treatment comparison, OS results only based on patients preselected for LDAC (the common comparator in the ITC) were used. Due to the non-transparent nature of treatment pre-selection criteria, the inclusion and exclusion criteria for the LDAC subgroup of the AZA-AML-001 trial are unknown and thus cannot be compared with those of BRIGHT AML 1003;

In the DACO-016 trial, the control group consisted of patients treated with LDAC (88.5 %) and BSC (11.5 %). Therefore, the control arm in the DACO-016 trial does not fully correspond to the common comparator treatment (LDAC) for the ITC. The biasing effect of the BSC patients in the control group on the indirect treatment comparison is unclear.



Due to the high uncertainty regarding the validity of the indirect comparison, no conclusions regarding comparative effectiveness based on these results will be drawn within this assessment. For completeness, the main results of the sponsor-provided ITC are descriptively and briefly stated below, but should be regarded as highly unsure.

4.11.4 Results

The sponsor-provided standard ITC (Bucher method) using HR for OS and the common comparator LDAC produced a HR [95 % CI] of 0.51 [0.31, 0.85] for the indirect comparison of glasdegib + LDAC with azacitidine and a HR [95 % CI] of 0.57 [0.35, 0.91] for the indirect comparison of glasdegib + LDAC with decitabine.

Supportive analyses (STCs) statistically adjusting for some between-trial differences in available patient characteristics (covariates considered for the STCs were age, sex, disease history, proportion of bone marrow blasts > 50 %, ECOG PS, cytogenetic risk, and haemoglobin level) show consistent results to the results from standard (Bucher) ITCs above (for details see (4)).



5 PATIENT INVOLVEMENT

Three patient organisations provided input in response to the Open Call for Patient Input published on 11th of October 2019:

Leukaemia Care, United Kingdom;

Association of Cancer Patients in Finland, Finland;

Deutsche Leukämie- & Lymphom-Hilfe, Germany.

Patient organisations stressed the fact that AML is an extremely serious and life-threatening illness with a huge impact on patients themselves but also on their families and care-givers. Current treatment options for patients not eligible for standard induction therapy are limited to azacitidine, decitabine, LDAC and BSC. Since these treatment options have only limited efficacy, low response rates and high relapse rates, further treatment options are highly welcomed. The potential of glasdegib + LDAC to extend life was mentioned as the most relevant benefit. Even though quality of life should ideally be enhanced, or at least not impaired, it was also mentioned that patients might be willing to tolerate additional side-effects when they can expect more time for them and their families. In any case, the decision not to have treatment in order to avoid side effects should be a personal one. The oral administration without the burden of intravenous treatment with azacitidine and decitabine was mentioned positively.

In contrast to recent therapies for AML which are focused on patients with higher cytogenetic risk, the trend towards a greater OS benefit of glasdegib + LDAC in patients with favourable or intermediate risk genetics than in those with unfavourable risk was raised. Due to the fact that mainly older patients and patients with co-morbidities tend to be the ones unable to tolerate chemotherapy or newer targeted agents, they are also the ones who have the fewest treatment options. New therapeutic alternatives for patients not suitable for these regimens should therefore be available as soon as possible, to not leave these patients behind.



6 DISCUSSION

Glasdegib is indicated, in combination with LDAC, for the treatment of newly diagnosed de novo or secondary AML in adult patients who are not candidates for standard induction chemotherapy (approved indication).

Current treatment options for patients not eligible for intensive regimens comprise the HMAs azacitidine and decitabine, LDAC and BSC. No trials directly comparing these treatment options exists, but generally treatment options for unfit patients are considered unsatisfactory (2). New modes of administration such as oral intake in the case of glasdegib would also address patient preferences.

No commonly accepted definition exists for patients ineligible for intensive induction therapy. Age > 65 years is one but not the only criterion to guide treatment decisions (2). According to the ELN recommendations, poor performance status and significant comorbidities and, in the case of conventional regimens such as a seven-day continuous infusion of cytarabine along with anthracycline treatment on days 1 to 3, adverse ELN cytogenetics/molecular genetics are the only factors which should be used for determining eligibility.

The search strategy and selection methods as described in the core Submission Dossier were deemed adequate. The SLR identified a single trial evaluating glasdegib + LDAC vs. LDAC alone. No other trials evaluating glasdegib with further comparators as defined in the PICO were identified.

Therefore, the assessment is based on a single, phase 1b/2 study, the B1371003 trial, which enrolled 132 AML and MDS patients. However, the core Submission Dossier only contains phase 2 results on the subgroup of 116 AML patients (BRIGHT AML 1003). Patients were defined as ineligible for standard induction therapy if they fulfilled at least one of several criteria: age ≥ 75 years, serum creatinine > 1.3 mg/dL, severe cardiac disease or ECOG PS 2.

The results are as follow:

For AML patients, median OS was 4.0 months longer for patients treated with glasdegib + LDAC than in patients treated with LDAC alone. This result was consistent in sensitivity analyses (e.g. CRF stratified). However, the certainty of evidence according to GRADE is considered low;

Exploratory results from a subgroup analysis according to IVRS based cytogenetic risk, descriptively showed a higher median OS improvement in AML patients with good/intermediate cytogenetic risk (median OS improvement of 6.7 months) compared to AML patients with poor cytogenetic risk (median OS improvement of 1.3 months);

All further efficacy endpoints (CR, ORR, transfusion independency, health-related quality of life

approximated by a Q-TWiST analysis) suggest improvements in the combination arm but are

considered exploratory and only supportive evidence for the primary outcome. The risk of bias was not assessed due to the availability of mature OS data, the most patient-relevant endpoint;

Over the entire study period with a median treatment duration of 83 days in the glasdegib + LDAC arm and 41 days in the LDAC alone arm, comparable rates for SAEs, fatal AEs and AEs leading to treatment discontinuation were observed. To adjust for differences in the treatment duration, results for the first 90 days of therapy were also reported: fewer fatal AEs compared to LDAC alone, and comparable rates for SAEs and treatment discontinuations due to AEs were observed with glasdegib + LDAC. However, the certainty of evidence according to GRADE is considered low for fatal AEs and very low for SAEs and AEs leading to treatment discontinuation.

The evidence for glasdegib has several limitations:

Evidence is available only from one, small open-label, phase 1b/2 study, the B1371003 trial. However, phase 2 of the study was underpowered for the primary endpoint because the final analysis was planned as soon as 92 OS events were observed and was conducted at 94 OS events but to obtain a statistical power of at least 80 % to detect a true HR of 0.625 at a conventional alpha level of 5 % (2-sided) 160 OS events would have been necessary;

https://www.uptodate.com/contents/cytarabine-drug-information?search=AML&topicRef=94786&source=see_link



Within the Submission Dossier phase 2 results are only presented for a subgroup of 116 AML patients, thus excluding MDS patients, BRIGHT AML 1003. These efficacy analyses within the AML subgroup were not specified in the original study protocol, and were not controlled for multiple testing. Strictly, a post-hoc selected subgroup is generally regarded as merely exploratory. However, these subgroup results might be meaningful because all statistically significant key efficacy results (OS) within the AML subgroup also held in all randomized patients (AML + MDS), only a small proportion of all randomized patients had MDS (16 of 132 randomized patients (12 %)) and hence the majority of information stems from randomized AML patients, and glasdegib was approved by the EMA only for AML patients;

Analyses of secondary endpoints (CR) were not controlled for multiple testing and have to be regarded as exploratory, providing supportive evidence for the primary endpoint only. Three additional endpoints (quality of survival, transfusion independence, ORR) were included in the core Submission Dossier but were not initially pre-specified as efficacy endpoints within the B1371003 study. These assessment specific endpoints are regarded as supportive evidence only;

No direct comparison of glasdegib + LDAC with azacitine, decitabine, or BSC has been conducted. Even though median OS was significantly improved when glasdegib + LDAC was directly compared to LDAC alone, indirect evidence suggests superiority of the HMAs azacitidine and decitabline over LDAC. However, as long as direct treatment comparisons are lacking and as long as valid indirect comparisons cannot be conducted, mainly due to clear between-trial differences in patient populations and not fully compatible (common) comparator arms within available trials, no firm conclusions regarding the comparative effectiveness of glasdegib + LDAC vs. azacitidine, decitabine or BSC can be drawn;

Patients were grouped and stratified according to IVRS based cytogenetic risk (good/ intermediate or poor). At baseline, fewer patients in the glasdegib + LDAC arm had poor cytogenetic risk than in the LDAC alone arm (37 % vs. 45 %). This imbalance is likely due to excluding MDS patients initially included in the stratified randomization process. Also, protocol violations led to the incorrect stratification at randomization according to the IVRS (n=17). However, provided sensitivity analyses indicated no large impact of these imbalances and protocol violations on efficacy outcomes;

Outcomes concerning disease response are compromised by the fact that in 31 % of patients with glasdegib + LDAC and in 42 % of patients with LDAC alone no bone marrow evaluation was performed. These non-evaluable patients were included in the denominator when calculating response rates and were counted as “non-responders”. Even though the decision to not perform a bone marrow evaluation was based on clinical judgement and is not unexpected in patients experiencing disease progression or complications of AML, this could have favoured the combination therapy, since fewer patients were not evaluable in this group;

In subgroup analyses, differences in the treatment effect of potentially relevant magnitude were shown descriptively for age (smaller effects for patients aged 75 or older), ECOG subgroups (smaller effects for ECOG ≤1), baseline bone marrow blast count (smaller effects for ≥ 30 %), prognostic risk factor subgroups (smaller effects for patients with intermediate-II risk), and subgroups according to disease history (smaller effects for de novo AML);

In general, the documents submitted by the MAH (i.e. Submission Dossier including core Submission Dossier, SAP, CSR, study protocol and additional tables) presented several inconsistencies. While some of the documents focus on the B1371003 trial, others present results for the AML subgroup (BRIGHT AML 1003) only. In addition, several key outcomes according to the PICO question (quality of survival, transfusion independence) were not pre-planned and hence were not included in the study protocol of B1371003. 1-sided and 2-sided p-values and 80 % and 95 % CI were presented inconsistently for different populations (AML only or AML + MDS patients) and for different outcomes, methodological descriptions for e.g. calculation of relative risk for disease response were lacking and inconsistencies in the initially submitted results of the Q-TWiST analysis were encountered. In addition, the MAH did not submit individual patient data. Hence, an independent evaluation of the reproducibility and robustness of reported results by the authoring team was not possible. Therefore, inconsistencies encountered in the Submission Dossier could not be resolved by the authoring team.



7 CONCLUSION

Phase 2 results from a single phase 1b/2 study, the B1371003 trial, showed a statistically significant improvement in median OS by 4.0 months for the combination of glasdegib + LDAC in comparison to LDAC alone in a subgroup of 116 newly diagnosed AML patients. Further efficacy outcomes suggest an advantage of glasdegib + LDAC over LDAC alone but were considered exploratory and thus as supportive evidence for the primary endpoint. Patient reported outcomes were not proactively collected. AEs were not increased within the combination arm.

Grading of the evidence according to GRADE indicated that the certainty of the evidence is low for OS and fatal AEs and very low for SAEs and AEs leading to treatment discontinuation. Whether specific subgroups of the licensed indication can benefit more from glasdegib than the overall AML population, remains unclear.

No direct comparisons are available concerning other treatment options for patients not eligible for standard induction therapy. Especially the comparison to azacitidine and decitabine, both preferred over LDAC according to guidelines would be of interest. An indirect comparison submitted by the MAH was not considered to provide reliable results. Thus, no firm conclusions regarding the comparative effectiveness of glasdegib + LDAC vs. azacitidine or decitabine or BSC can be drawn.



8 REFERENCES

1. De Kouchkovsky I, M A-H. Acute myeloid leukemia: a comprehensive review and 2016 update. Blood Cancer J. 2016;6(7):441-e.

2. Dohner H, Estey E, Grimwade D, Amadori S, Appelbaum FR, Buchner T, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-47.

3. Heuser M, Ofran Y, Boissel N, Mauri SB, Craddock C, Janssen J, et al. Acute myeloid leukaemia in adult patients: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Annals of oncology : official journal of the European Society for Medical Oncology. 2020.

4. Pfizer INC. Glasdegib Acute myeloid leukaemia - EUnetHTA pharmaceuticals evidence submission template - Short version (Amended July 2020). 2020.

5. European Medicines Agency. European Public Assessment Report. Daurismo. International non-proprietary name: glasdegib. 2020.

6. Grading of Recommendations Assessment Development and Evaluation Working Group. GRADE Handbook 2013 [Available from: https://gdt.gradepro.org/app/handbook/handbook.html.

7. Guyatt G, Oxman A, Kunz R, Brozek J, Alonso-Coello P, Rind D, et al. GRADE guidelines 6. Rating the quality of evidence—imprecision. . Journal of clinical epidemiology. 2011;64(12):1283-93.

8. Chow S, Wang H, Shao J. Sample size calculations in clinical research. Second Edition ed. CRC CH, editor2007.

9. Papaemmanuil E, Gerstung M, Bullinger L, Gaidzik VI, Paschka P, Roberts ND, et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med. 2016;374(23):2209-21.

10. Kayser S, Dohner K, Krauter J, Kohne CH, Horst HA, Held G, et al. The impact of therapy-related acute myeloid leukemia (AML) on outcome in 2853 adult patients with newly diagnosed AML. Blood. 2011;117(7):2137-45.

11. Smith FO, Alonzo TA, Gerbing RB, Woods WG, Arceci RJ, Children's Cancer G. Long-term results of children with acute myeloid leukemia: a report of three consecutive Phase III trials by the Children's Cancer Group: CCG 251, CCG 213 and CCG 2891. Leukemia. 2005;19(12):2054-62.

12. Connely S, Rau R. The genomics of acute myeloid leukemia in children. Cancer and Metastasis Review. 2020;39(1):189-209.

13. Leak Bryant A, Lee Walton A, Shaw-Kokot J, Mayer D, Reeve B. Patient-reported symptoms and quality of life in adults with acute leukemia: a systematic review. Oncol Nurs Forum. 2015;42(2):E91-E101.

14. Granfeldt Ostgard L, Medeiros B, Sengelov H, Norgaard M, Andersen M, Dufva I. Epidemiology and Clinical Significance of Secondary and Therapy-Related Acute Myeloid Leukemia: A National Population-Based Cohort Study. J Clin Epidemiol. 2015;33(31):3641-9.

15. Daver N, Schlenk RF, Russell NH, Levis MJ. Targeting FLT3 mutations in AML: review of current knowledge and evidence. Leukemia. 2019;33(2):299-312.

16. Arber DA, Orazi A, Hasserjian R, Thiele J, Borowitz MJ, Le Beau MM, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391-405.

17. Dohner H, Estey EH, Amadori S, Appelbaum FR, Buchner T, Burnett AK, et al. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood. 2010;115(3):453-74.

18. Moarii M, Papaemmanuil E. Classification and risk assessment in AML: integrating cytogenetics and molecular profiling. Hematology Am Soc Hematol Educ Program. 2017;2017(1):37-44.

19. Shallis R, Wang R, Davidoff A, Ma X, Zeidan A. Epidemiology of acute myeloid leukemia: Recent progress and enduring challenges. Blood Reviews. 2019;36:70-87.

20. Juliusson G, Antunovic P, Derolf A, Lehmann S, Mollgard L, Stockelberg D, et al. Age and acute myeloid leukemia: real world data on decision to treat and outcomes from the Swedish Acute Leukemia Registry. Blood. 2009;113(18):4179-87.

21. Ostgard LS, Norgaard JM, Severinsen MT, Sengelov H, Friis L, Jensen MK, et al. Data quality in the Danish National Acute Leukemia Registry: a hematological data resource. Clin Epidemiol. 2013;5:335-44.

22. Shah A, Andersson T, Rachet B, Björkholm M, Lambert P. Survival and cure of acute myeloid leukaemia in England, 1971-2006: a population-based study. Br J Haematol. 2013;162((4):509-16.

https://gdt.gradepro.org/app/handbook/handbook.html



23. Meyers J, Yu Y, Kaye JA, Davis KL. Medicare fee-for-service enrollees with primary acute myeloid leukemia: an analysis of treatment patterns, survival, and healthcare resource utilization and costs. Appl Health Econ Health Policy. 2013;11(3):275-86.

24. Renaud L, Nibourel O, Berthon C, Roumier C, Rodriguez C, Frimat C, et al. De Novo and Secondary Acute Myeloid Leukemia, Real World Data on Outcomes from the French Nord-Pas-De-Calais Picardie Acute Myeloid Leukemia Observatory. Blood. 2016;128(22):4013.

25. U.S. Food & Drug Administration. FDA approves new treatment for patients with acute myeloid leukemia 2018 [Available from: https://www.fda.gov/news-events/press-announcements/fda-approves-new-treatment-patients-acute-myeloid-leukemia.

26. European Medicines Agency. EU/3/17/1923 2018 [Available from: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu3171923.

27. Medeiros B. Interpretation of clinical endpoints in trials of acute myeloid leukemia. Leukemia Research. 2018;68:32-9.

28. National Comprehensive Cancer N. The National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines®) - Acute Myeloid Leukemia - Version 3.20202020 14.04.2020. Available from: https://www.nccn.org/professionals/physician_gls/pdf/aml.pdf.

29. Chen X, Xie H, Wood B, Walter R, Pagel J, Becker P, et al. Relation of Clinical Response and Minimal Residual Disease and Their Prognostic Impact on Outcome in Acute Myeloid Leukemia. J Clin Oncol. 2015;10;33(11):1258-64.

30. Walter R, Appelbaum F, Tallman M, Weiss N, Larson R, Estey E. Shortcomings in the Clinical Evaluation of New Drugs: Acute Myeloid Leukemia as Paradigm. Blood. 2010;Oct 7;116(14):2420-9.

31. Short N, Kantarjian H, Garcia-Manero G, Borthakur G, Kadia T, Huang X. A Randomized Phase II Trial of 5-Day Versus 10-Day Schedules of Decitabine for Older Patients with Previously Untreated Acute Myeloid Leukemia. . Blood. 2017;130(Supplement 1):2577.

32. Burnett AK, Milligan D, Prentice AG, Goldstone AH, McMullin MF, Hills RK, et al. A comparison of low-dose cytarabine and hydroxyurea with or without all-trans retinoic acid for acute myeloid leukemia and high-risk myelodysplastic syndrome in patients not considered fit for intensive treatment. Cancer. 2007;109(6):1114-24.

33. Ram R, Amit O, Zuckerman T, Gurion R, Raanani P, Bar-On Y, et al. Venetoclax in patients with acute myeloid leukemia refractory to hypomethylating agents-a multicenter historical prospective study. Ann Hematol. 2019;98(8):1927-32.

34. European Medicines Agency. European Public Assessment Report. Dacogen. International non-proprietary name: decitabine. 2012.

35. European Medicines Agency. European Public Assessment Report. Vidaza. International non-proprietary name: azacitidine. 2015.

36. Wong SS, Wilczynski NL, Haynes RB. Comparison of top-performing search strategies for detecting clinically sound treatment studies and systematic reviews in MEDLINE and EMBASE. Journal of the Medical Library Association : JMLA. 2006;94(4):451-5.

37. Cochrane Collaboration. RoB 2: A revised Cochrane risk-of-bias tool for randomized trials, . 2020.

38. European Network For Health Technology Assessment (EuNetHTA). Comparators & Comparisons: Direct and indirect comparisons 2015 [updated November 2015. Available from: https://www.eunethta.eu/wp-content/uploads/2018/03/Direct_comparators_comparisons.pdf.

39. Cortes JE, Heidel FH, Hellmann A, Fiedler W, Smith BD, Robak T, et al. Randomized comparison of low dose cytarabine with or without glasdegib in patients with newly diagnosed acute myeloid leukemia or high-risk myelodysplastic syndrome. Leukemia. 2019;33(2):379-89.

40. Pfizer INC. Full Clinical Study Report - Protocol B1371003. 2017. 41. Pfizer INC. Final Protocol Amendment 5 - Protocol Number B1371003. 2016. 42. Pfizer INC. Statistical Analysis Plan (SAP). 2016. 43. Dombret H, Seymour JF, Butrym A, Wierzbowska A, Selleslag D, Jang JH, et al. International

phase 3 study of azacitidine vs conventional care regimens in older patients with newly diagnosed AML with >30% blasts. Blood. 2015;126(3):291-9.

44. Kantarjian HM, Thomas XG, Dmoszynska A, Wierzbowska A, Mazur G, Mayer J, et al. Multicenter, randomized, open-label, phase III trial of decitabine versus patient choice, with physician advice, of either supportive care or low-dose cytarabine for the treatment of older patients with newly diagnosed acute myeloid leukemia. J Clin Oncol. 2012;30(21):2670-7.

45. Pfizer INC. Additional requested tables to the BRIGHT AML 1003 study (B1371003), Patient population P2 Unfit, AML only. 2020.

https://www.fda.gov/news-events/press-announcements/fda-approves-new-treatment-patients-acute-myeloid-leukemia

https://www.fda.gov/news-events/press-announcements/fda-approves-new-treatment-patients-acute-myeloid-leukemia

https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu3171923

https://www.nccn.org/professionals/physician_gls/pdf/aml.pdf

https://www.eunethta.eu/wp-content/uploads/2018/03/Direct_comparators_comparisons.pdf



46. Pfizer Inc. Additional analyses to the BRIGHT AML 1003 study (B1371003), Patient population P2 Unfit, AML only. 2020.

47. Pfizer INC. AMCP Dossier: Glasdegib - Version 2.0. 2019. 48. European Network For Health Technology Assessment (EuNetHTA). Endpoints used for REA:

Clinical Endpoints 2015 [updated November 2015. Available from: https://eunethta.eu/wp-content/uploads/2018/02/WP7-SG3-GL-clin_endpoints_amend2015.pdf.

49. Cheson BD, Bennett JM, Kopecky KJ, Buchner T, Willman CL, Estey EH, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21(24):4642-9.

50. Cortes JE, Heidel FH, Hellmann A, Fiedler W, Smith BD, Robak T, et al. Supplement to: Randomized comparison of low dose cytarabine with or without glasdegib in patients with newly diagnosed acute myeloid leukemia or high-risk myelodysplastic syndrome. Leukemia. 2019;33(2):379-89.

51. European Medicines Agency. Dacogen: EPAR - Product Information2019; (10.10.2019). Available from: https://www.ema.europa.eu/en/documents/product-information/dacogen-epar-product-information_en.pdf.

52. Spriggs D, Griffin J, Wisch J, Kufe D. Clinical pharmacology of low-dose cytosine arabinoside. Blood. 1985;65(5):1087-9.

53. Slovak ML, Kopecky KJ, Cassileth PA, Harrington DH, Theil KS, Mohamed A, et al. Karyotypic analysis predicts outcome of preremission and postremission therapy in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Cooperative Oncology Group Study. Blood. 2000;96(13):4075-83.

54. National Comprehensive Cancer N. The National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines®) - Acute Myeloid Leukemia - Version 1.20092009. Available from: http://www.nccn.org/professionals/physician_gls/f_guidelines.asp.

https://eunethta.eu/wp-content/uploads/2018/02/WP7-SG3-GL-clin_endpoints_amend2015.pdf

https://eunethta.eu/wp-content/uploads/2018/02/WP7-SG3-GL-clin_endpoints_amend2015.pdf

https://www.ema.europa.eu/en/documents/product-information/dacogen-epar-product-information_en.pdf

https://www.ema.europa.eu/en/documents/product-information/dacogen-epar-product-information_en.pdf

http://www.nccn.org/professionals/physician_gls/f_guidelines.asp



APPENDIX 1: GUIDELINES FOR DIAGNOSIS AND MANAGEMENT

Table A1. Overview of guidelines used for this assessment

Name of society/organisation issuing guidance

Date of issue

Country/ies to which applicable

Summary of recommendation Level of evidence (A,B,C)/ class of recommendation (I, IIa, IIb, III)

ESMO March 2020

Europe HMAs azacitidine and decitabine are currently the first choice in newly diagnosed unfit AML patients.

IIb

ELN January 2017

Europe Treatment alternatives for patients who are not candidates for standard induction chemotherapy:

Azacitidine;

Decitabine;

LDAC: not recommended in patients with adverse-risk genetics.

Strong recommendation to enrol these patients in clinical trials. Patients who cannot tolerate any antileukemic therapy, or who do not wish any antileukemic therapy:

Best supportive care: including hydroxyurea.

Strong recommendation to enrol these patients in clinical trials

IIb

Source: (2, 3) Abbreviations: ELN=European Leukaemia Net; ESMO=European Society for Medical Oncology; HMAs=hypomethylating agents; LDAC=low-dose cytarabine.



Table A2. Overview of cytogenetic and molecular risk stratification tools used in B1371003, AZA-AML-001 and DACO-016

Genetic risk group

ELN 2010 NCCN 2009 SWOG 2000

Favourable t(8;21)(q22;q22); RUNX1-RUNX1T1 inv(16)(p13.1q22) or t(16;16)(p13.1;q22); CBFB-MYH11 Mutated NPM1 without FLT3-ITD (normal karyotype) Mutated CEBPA (normal

karyotype)

t(15;17) with/without secondary aberrations; t(8;21) lacking del(9q) or complex karyotypes inv(16)/t(16;16)/del(16q),

t(8;21) inv(16) t(16;16)

Normal cytogenetics with isolated NPM1 mutation

Intermediate Intermediate I

Mutated NPM1 and FLT3-ITD (normal karyotype) Wild-type NPM1 and FLT3-ITD (normal karyotype) Wild-type NPM1 without FLT3-ITD (normal karyotype)

Normal +8 +6 -Y del(12p)

Normal +8 only, t(9;11) only

Other abnormalities not listed with better risk and poor risk cytogenetics and molecular mutations

C-KIT in patients with t(8;21) or inv(16)

Intermediate II t(9;11)(p22;q23); MLLT3-MLL

Cytogenetic abnormalities not classified as favorable or adverse

Adverse/Poor/ Unfavourable

-5 or del(5q) -7 abnl(17p) inv(3)(q21q26.2) or t(3;3)(q21;q26.2); RPN1-

EVI1 t(6;9)(p23;q34); DEK-NUP214 t(v;11)(v;q23); MLL rearranged complex karyotype (≥ 3 abnormalities)

-5/ del(5q) -7/ del(7q) abn 3q, 9q, 11q, 20q, 21q, 17p, t(6;9), t(9;22) complex karyotypes (≥ 3 unrelated abn)

-5/ 5q- -7/ 7q- Abnormalities of 11q23, excluding t(9;11) Inv(3) or t(3;3) t(6;9) t(9;22) complex karyotypes (≥ 3 abnormalities)

Normal cytogenetics with isolated FLT3-ITD mutations

Unknown - All other abnormalities -

Source: (17, 53, 54) Abbreviations: p=short arm of the chromosome; q=long arm of the chromosome; t(A;B)=used to denote a translocation between chromosome A and chromosome B. Comment: Classical cytogenetic markers are in bold.



APPENDIX 2: CERTAINTY OF EVIDENCE

Table A3. Risk of Bias 2 Assessment

Domain Signalling question Response Comments

Bias arising from the randomization process

1.1 Was the allocation sequence random? NI No information about random sequence generation was available; allocation by IVRS. 1.2 Was the allocation sequence concealed until participants

were enrolled and assigned to interventions? PY

1.3 Did baseline differences between intervention groups suggest a problem with the randomization process?

N Even though the study initially included both AML and MDS patients, focusing on the AML subgroup only, this did not increase risk of bias per se since random allocation within the AML subgroup was maintained. Existing differences between arms in baseline characteristics within the AML subgroup were not statistically significant (own calculation) and hence did not clearly indicate any systematic issues with the randomisation process.

Risk of bias judgement Low

Bias due to deviations from intended interventions

2.1. Were participants aware of their assigned intervention during the trial?

Y Open-label study.

2.2. Were carers and people delivering the interventions aware of participants' assigned intervention during the trial?

Y

2.3. If Y/PY/NI to 2.1 or 2.2: Were there deviations from the intended intervention that arose because of the experimental context?

NI Due to the open-label design of the study the administration of concomitant therapies could have been influenced by knowledge of the assigned intervention.

2.4 If Y/PY to 2.3: Were these deviations likely to have affected the outcome?

NA

2.5. If Y/PY/NI to 2.4: Were these deviations from intended intervention balanced between groups?

NA

2.6 Was an appropriate analysis used to estimate the effect of assignment to intervention?

Y OS: Full Analysis Set. SAE, Fatal AE, Discontinuation due to AE: even though patients were analysed per protocol for safety outcomes, this is considered appropriate for safety endpoints.

2.7 If N/PN/NI to 2.6: Was there potential for a substantial impact (on the result) of the failure to analyse participants in the group to which they were randomized?

NA

Risk of bias judgement Some concerns



Domain Signalling question Response Comments

Bias due to missing outcome data

3.1 Were data for this outcome available for all, or nearly all, participants randomized?

Y

3.2 If N/PN/NI to 3.1: Is there evidence that result was not biased by missing outcome data?

NA

3.3 If N/PN to 3.2: Could missingness in the outcome depend on its true value?

NA

3.4 If Y/PY/NI to 3.3: Is it likely that missingness in the outcome depended on its true value?

NA


Bias in measurement of the outcome

4.1 Was the method of measuring the outcome inappropriate? N

4.2 Could measurement or ascertainment of the outcome have differed between intervention groups?

N (OS, Fatal AE) PY (SAE, Discontinuations due to AE)

SAE, Discontinuation due to AE: outcome with subjective component, assessed by unblinded investigator.

4.3 Were outcome assessors aware of the intervention received by study participants?

Y (OS, Fatal AE) NA (SAE, Discontinuations due to AE)

Open-label study.

4.4 If Y/PY/NI to 4.3: Could assessment of the outcome have been influenced by knowledge of intervention received?

N (OS, Fatal AE) NA (SAE, Discontinuations due to AE)

4.5 If Y/PY/NI to 4.4: Is it likely that assessment of the outcome was influenced by knowledge of intervention received?

NA

Risk of bias judgement Low (OS, Fatal AE) High (SAE, Discontinuations due to AE)

Bias in selection of the reported result

5.1 Were the data that produced this result analysed in accordance with a pre-specified analysis plan that was finalized before unblinded outcome data were available for analysis?

Y No indication for stopping early for benefit.

5.2 ... multiple eligible outcome measurements (e.g. scales, definitions, time points) within the outcome domain?

N

5.3 ... multiple eligible analyses of the data? N


Overall bias Risk of bias judgement Some concerns (OS, Fatal AE) High (SAE, Discontinuations due to AE)

Abbreviations: AE=adverse event; AML=acute myeloid leukaemia; IVRS=interactive voice response system; MDS=myelodysplastic syndrome; N=no; NA=not applicable; NI=no information; OS=overall survival; PN=probably no; PY=probably yes; SAE=serious AE; Y=yes.



Table A4. GRADE Evidence profile

Certainty assessment № of patients Effect Certainty Importance

№ of studies

Study design

Risk of bias

Inconsistency Indirectness Imprecision Other considerations

Glasdegib + LDAC

LDAC Relative (95% CI)

Absolute (95% CI)*

Overall survival (IVRS) (glasdegib + LDAC: median follow-up 21.7 months vs. LDAC alone 20.1 months)

1 Randomised trial

Seriousa Not serious Not serious Seriousb None 59/78 (75.6%) died during the study Estimated risk of death up to month 6: 39.5%

35/38 (92.1%) died during the study Estimated risk of death up to month 6: 66.5%

HR 0.46

(0.30 to 0.72)

270 fewer deaths per 1.000 up to month 6

(from 385 fewer deaths to 120 fewer deaths)

⨁⨁◯◯ LOW

Critical

Health-related quality of life

- - - - - - - - - - - - Critical

Transfusion independence

- - -- - - - - - - - - - Important

Serious AEs (study period)

1 Randomised trial

Very seriousc

Not serious Not serious Very seriousd

None 59/75 (78.7%)

28/36 (77.8%)

RR 1.01

(0.82 to 1.25)

8 more per 1.000

(from 140 fewer to 194 more)

⨁◯◯◯ VERY LOW

Important

Fatal AEs (study period)

1 Randomised trial

Seriouse Not serious Not serious Seriousf None 22/75 (29.3%)

16/36 (44.4%)

RR 0.66

(0.40 to 1.10)

151 fewer per 1.000

⨁⨁◯◯ LOW

Important



Certainty assessment № of patients Effect Certainty Importance

№ of studies

Study design

Risk of bias

Inconsistency Indirectness Imprecision Other considerations

Glasdegib + LDAC

LDAC Relative (95% CI)

Absolute (95% CI)*


AEs leading to treatment discontinuation (study period)

1 Randomised trial

Very seriousc

Not serious Not serious Seriousf None 23/75 (30.7%)

17/36 (47.2%)

RR 0.65

(0.40 to 1.05)

165 fewer per 1.000


⨁◯◯◯ VERY LOW

Important

Explanations: a Downgraded by 1 level: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) some concerns exist, and even though the B1371003 trial itself was not stopped early for benefit, no phase 3 trial was conducted due to the observed treatment effect in phase 2 of this small, single phase 1b/2 trial. A substantial overestimate of the treatment effect can be considered likely. b Downgraded by 1 level: optimal information size criterion not fulfilled. In phase 2 of the B1371003 trial, 94 OS events have been observed; however, to obtain a statistical power of at least 80 % to detect a true HR of 0.625 (as used within the SAP for sample size planning) at a conventional alpha level of 5 % (2-sided), 160 OS events would have been necessary (7, 8). c Downgraded by 2 levels: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) high risk of bias exists due to the open-label design and endpoint with subjective component. d Downgraded by 2 levels: very few events and CIs around both relative and absolute estimates of effect, that include both appreciable benefit and appreciable harm. e Downgraded by 1 level: according to RoB 2 assessment (see AppendixAPPENDIX 2: Certainty of Evidence) some concerns exist. f Downgraded by 1 level: few events and wide CI including 1.0 but also relative effects of potentially appreciable benefit. * The risk in the intervention group (and its 95 % CI) is based on the observed risk in the comparison group and the estimated relative effect of the intervention (and its 95 % CI). For OS the risk of death up to month 6 within the control group was estimated by the Kaplan-Meier method (i.e., considering censored observations). Abbreviations: AE=adverse event; CI=confidence interval; GRADE=Grading of Recommendations Assessment, Development and Evaluation; HR=hazard ratio; IVRS=interactive voice response system; LDAC=low-dose cytarabine; OS=overall survival; RoB=risk of bias; RR=risk ratio; vs.=versus.



Table A5. Results summary for AEs (any CTCAE grade) occurring in ≥ 20 % of patients separately by MedDRA SOC and PT for BRIGHT AML 1003



LDAC N = 36 n (%)




LDAC N = 36 n (%)



SOC: Gastrointestinal disorders

All PT 58 (77.3) 24 (66.7) 1.16 [0.89; 1.51]; 0.2659

0.11 [-0.07; 0.29]; 0.2476

51 (68.0) 23 (63.9) 1.06 [0.80; 1.42]; 0.6740

0.04 [-0.15; 0.23]; 0.6701

Nausea 27 (36.0) 4 (11.1) 3.24 [1.23; 8.56]; 0.0178

0.25 [0.10; 0.40]; 0.0011

22 (29.3) 4 (11.1) 2.64 [0.98; 7.09]; 0.0542

0.18 [0.04; 0.33]; 0.0141

Diarrhoea 21 (28.0) 9 (25.0) 1.12 [0.57; 2.19]; 0.7411

0.03 [-0.14; 0.20]; 0.7357

13 (17.3) 9 (25.0) 0.69 [0.33; 1.47]; 0.3393

-0.08 [-0.24; 0.09]; 0.3635

Constipation 19 (25.3) 6 (16.7) 1.52 [0.66; 3.48]; 0.3212

0.09 [-0.07; 0.24]; 0.2779

15 (20.0) 5 (13.9) 1.44 [0.57; 3.65]; 0.4426

0.06 [-0.08; 0.21]; 0.4080

Vomiting 18 (24.0) 3 (8.3) 2.88 [0.91; 9.15]; 0.0729

0.16 [0.02; 0.29]; 0.0203

15 (20.0) 3 (8.3) 2.40 [0.74; 7.77]; 0.1439

0.12 [-0.01; 0.24]; 0.0737

SOC: General disorders and administration site conditions

All PT 57 (76.0) 24 (66.7) 1.14 [0.88; 1.48]; 0.3301

0.09 [-0.09; 0.28]; 0.3143

52 (69.3) 22 (61.1) 1.13 [0.84; 1.53]; 0.4110

0.08 [-0.11; 0.27]; 0.3973

Fatigue 21 (28.0) 8 (22.2) 1.26 [0.62; 2.56]; 0.5239

0.06 [-0.11; 0.23]; 0.5044

19 (25.3) 6 (16.7) 1.52 [0.66; 3.48]; 0.3212

0.09 [-0.07; 0.24]; 0.2779

Pyrexia 20 (26.7) 8 (22.2) 1.20 [0.59; 2.46]; 0.6183

0.04 [-0.12; 0.21]; 0.6056

15 (20.0) 8 (22.2) 0.90 [0.42; 1.93]; 0.7860

-0.02 [-0.19; 0.14]; 0.7896

Oedema peripheral 19 (25.3) 7 (19.4) 1.30 [0.60; 2.81]; 0.5007

0.06 [-0.10; 0.22]; 0.4775

17 (22.7) 7 (19.4) 1.17 [0.53; 2.56]; 0.7020

0.03 [-0.13; 0.19]; 0.6936

SOC: Blood and lymphatic system disorders

All PT 53 (70.7) 23 (63.9) 1.11 [0.83; 1.47]; 0.4890

0.07 [-0.12; 0.26]; 0.4791

49 (65.3) 22 (61.1) 1.07 [0.79; 1.46]; 0.6711

0.04 [-0.15; 0.23]; 0.6669

Anaemia 35 (46.7) 15 (41.7) 1.12 [0.71; 1.77]; 0.6262

0.05 [-0.15; 0.25]; 0.6183

33 (44.0) 15 (41.7) 1.06 [0.66; 1.68]; 0.8177

0.02 [-0.17; 0.22]; 0.8158

Febrile neutropenia 26 (34.7) 9 (25.0) 1.39 [0.73; 2.64]; 0.3209

0.10 [-0.08; 0.27]; 0.2866

23 (30.7) 8 (22.2) 1.38 [0.69; 2.78]; 0.3668

0.08 [-0.09; 0.26]; 0.3339

Thrombocytopenia 24 (32.0) 9 (25.0) 1.28 [0.66; 2.46]; 0.4601

0.07 [-0.11; 0.25]; 0.4370

23 (30.7) 9 (25.0) 1.23 [0.63; 2.37]; 0.5442

0.06 [-0.12; 0.23]; 0.5275





LDAC N = 36 n (%)




LDAC N = 36 n (%)



SOC: Infections and infestations

All PT 46 (61.3) 20 (55.6) 1.10 [0.78; 1.56]; 0.5718

0.06 [-0.14; 0.25]; 0.5638

39 (52.0) 19 (52.8) 0.99 [0.68; 1.44]; 0.9386

-0.01 [-0.21; 0.19]; 0.9388

Pneumonia 21 (28.0) 10 (27.8) 1.01 [0.53; 1.91]; 0.9805

0.00 [-0.18; 0.18]; 0.9805

14 (18.7) 9 (25.0) 0.75 [0.36; 1.56]; 0.4373

-0.06 [-0.23; 0.10]; 0.4565

SOC: Investigations

All PT 45 (60.0) 18 (50.0) 1.20 [0.82; 1.75]; 0.3410

0.10 [-0.10; 0.30]; 0.3208

40 (53.3) 17 (47.2) 1.13 [0.75; 1.69]; 0.5560

0.06 [-0.14; 0.26]; 0.5459

Weight decreased 15 (20.0) 1 (2.8) 7.20 [0.99; 52.40]; 0.0513

0.17 [0.07; 0.28]; 0.0013

11 (14.7) 1 (2.8) 5.28 [0.71; 39.33]; 0.1044

0.12 [0.02; 0.22]; 0.0156

SOC: Respiratory, thoracic and mediastinal disorders

All PT 38 (50.7) 22 (61.1) 0.83 [0.59; 1.17]; 0.2844

-0.10 [-0.30; 0.09]; 0.2947

30 (40.0) 19 (52.8) 0.76 [0.50; 1.15]; 0.1906

-0.13 [-0.32; 0.07]; 0.2041

Dyspnoea 16 (21.3) 11 (30.6) 0.70 [0.36; 1.35]; 0.2837

-0.09 [-0.27; 0.08]; 0.3065

13 (17.3) 9 (25.0) 0.69 [0.33; 1.47]; 0.3393

-0.08 [-0.24; 0.09]; 0.3635

Cough 16 (21.3) 6 (16.7) 1.28 [0.55; 2.99]; 0.5692

0.05 [-0.11; 0.20]; 0.5500

12 (16.0) 5 (13.9) 1.15 [0.44; 3.02]; 0.7737

0.02 [-0.12; 0.16]; 0.7678

SOC: Metabolism and nutrition disorders

All PT 43 (57.3) 13 (36.1) 1.59 [0.99; 2.56]; 0.0572

0.21 [0.02; 0.40]; 0.0309

35 (46.7) 11 (30.6) 1.53 [0.88; 2.64]; 0.1303

0.16 [-0.03; 0.35]; 0.0932

Decreased appetite 24 (32.0) 4 (11.1) 2.88 [1.08; 7.68]; 0.0346

0.21 [0.06; 0.36]; 0.0054

15 (20.0) 3 (8.3) 2.40 [0.74; 7.77]; 0.1439

0.12 [-0.01; 0.24]; 0.0737

SOC: Musculoskeletal and connective tissue disorders

All PT 42 (56.0) 11 (30.6) 1.83 [1.08; 3.12]; 0.0256

0.25 [0.07; 0.44]; 0.0079

33 (44.0) 11 (30.6) 1.44 [0.83; 2.51]; 0.1976

0.13 [-0.05; 0.32]; 0.1605

Muscle spasms 16 (21.3) 2 (5.6) 3.84 [0.93; 15.81]; 0.0624

0.16 [0.04; 0.28]; 0.0094

11 (14.7) 2 (5.6) 2.64 [0.62; 11.29]; 0.1905

0.09 [-0.02; 0.20]; 0.1032

SOC: Skin and subcutaneous tissue disorders

All PT 40 (53.3) 13 (36.1) 1.48 [0.91; 2.39]; 0.1138

0.17 [-0.02; 0.37]; 0.0808

33 (44.0) 12 (33.3) 1.32 [0.78; 2.24]; 0.3026

0.11 [-0.08; 0.30]; 0.2727





LDAC N = 36 n (%)




LDAC N = 36 n (%)



SOC: Nervous system disorders

All PT 44 (58.7) 8 (22.2) 2.64 [1.39; 5.01]; 0.0029

0.36 [0.19; 0.54]; <0.0001

38 (50.7) 7 (19.4) 2.61 [1.29; 5.25]; 0.0074

0.31 [0.14; 0.48]; 0.0004

Dizziness 16 (21.3) 3 (8.3) 2.56 [0.80; 8.23]; 0.1145

0.13 [0.00; 0.26]; 0.0490

13 (17.3) 3 (8.3) 2.08 [0.63; 6.84]; 0.2280

0.09 [-0.03; 0.21]; 0.1564

Dysgeusia 18 (24.0) 1 (2.8) 8.64 [1.20; 62.21]; 0.0323

0.21 [0.10; 0.32]; 0.0002

15 (20.0) 1 (2.8) 7.20 [0.99; 52.40]; 0.0513

0.17 [0.07; 0.28]; 0.0013

SOC: Vascular disorders

All PT 23 (30.7) 12 (33.3) 0.92 [0.52; 1.63]; 0.7758

-0.03 [-0.21; 0.16]; 0.7787

16 (21.3) 12 (33.3) 0.64 [0.34; 1.21]; 0.1679

-0.12 [-0.30; 0.06]; 0.1907

SOC: Psychiatric disorders

All PT 23 (30.7) 9 (25.0) 1.23 [0.63; 2.37]; 0.5442

0.06 [-0.12; 0.23]; 0.5275

17 (22.7) 8 (22.2) 1.02 [0.49; 2.14]; 0.9582

0.00 [-0.16; 0.17]; 0.9580

SOC: Renal and urinary disorders

All PT 19 (25.3) 7 (19.4) 1.30 [0.60; 2.81]; 0.5007

0.06 [-0.10; 0.22]; 0.4775

16 (21.3) 5 (13.9) 1.54 [0.61; 3.86]; 0.3617

0.07 [-0.07; 0.22]; 0.3181

SOC: Cardiac disorders

All PT 19 (25.3) 5 (13.9) 1.82 [0.74; 4.49]; 0.1913

0.11 [-0.04; 0.26]; 0.1344

16 (21.3) 4 (11.1) 1.92 [0.69; 5.33]; 0.2105

0.10 [-0.04; 0.24]; 0.1475

SOC: Injury, poisoning and procedural complications

All PT 17 (22.7) 5 (13.9) 1.63 [0.65; 4.07]; 0.2938

0.09 [-0.06; 0.24]; 0.2433

11 (14.7) 5 (13.9) 1.06 [0.40; 2.81]; 0.9132

0.01 [-0.13; 0.15]; 0.9123

Source: (46) Abbreviations: AE=adverse event; CI=confidence interval; LDAC=low-dose cytarabine; PT=preferred term; RD=risk difference; RR=risk ratio; SOC=system organ class.



APPENDIX 3: STUDY DETAILS FOR INDIRECT COMPARISONS

Table A6. Characteristics of the studies included - indirect comparison

Study reference/ID

Study design

Patient population Intervention (number of randomized patients)




AZA-AML-001

Randomised, multicentre, open-label, Phase 3

Newly diagnosed, de novo or secondary AML with >30% BM blasts;

≥65 years;

Not eligible for hematopoietic stem-cell transplantation;

Intermediate- or poor-risk cytogenetics (NCCN 2009 criteria);

ECOG PS≤2.

Azacitidine N = 241

BSC: N = 44;

LDAC: N = 154;

Standard induction therapy: N = 43.

Conventional care regimen: N = 247

BSC: N = 45;

LDAC: N = 158;

Standard induction chemotherapy: N = 44.

Study duration: October 2010 to January 22, 2014 Data cut-off: March 2014. Median Follow-up (both arms combined): 24.4 months Type of analysis: Planned

final analysis

Primary: OS (ITT) Secondary (not controlled for multiplicity):

Estimated 1-year survival rate;

OS in patient subgroups defined by baseline demographic and disease characteristics.

Other:

CR, CRi, PR;

Transfusion independence.

HrQoL: EORTC QLQ-C30 Adverse events (NCI CTCAE Version 4.0)

Indirect comparison: Decitabine vs. Patient choice

DACO-016

Randomized, multicentre open-label, Phase 3

Aged ≥65 years with

newly diagnosed, histologically confirmed de novo or secondary

AML (≥20 % blasts)

and poor- or intermediate-risk cytogenetics (Southwest Oncology Group categorization);

ECOG PS of 0 to 2;

Decitabine N = 242

Treatment choice: N = 243

LDAC: N = 215;

BSC: N = 28.

Study duration: January 2006 to April 2009 Data cut-off (final analysis): October 28, 2008 Median follow-up: not reported Type of analysis: Unplanned follow-up analysis of OS one year

Primary: OS (ITT) Secondary: CR (ITT) Adverse events (NCI CTCAE Version 3.0)



Study reference/ID

Study design

Patient population Intervention (number of randomized patients)




WBC count

≤40,000/mm;

Bilirubin ≤1.5 × ULN;

AST or ALT ≤2.5 ×

ULN;

creatinine clearance

≥40 mL/min;

Life expectancy ≥12

weeks.

after the protocol-specified data cut-off. Data cut-off (updated analysis) 29 October 2010 The updated OS results from the follow-up analysis were used in the Submission Dossier for indirect treatment comparisons

Source: (4, 34, 35, 43, 44) Abbreviations: AML=acute myeloid leukemia; ALT=alanine aminotransferase; AST=aspartate aminotransferase; BM=bone marrow; BSC=Best supportive care; CR=complete remission; CRi=CR with incomplete haematologic recovery; CTCAE=Common Terminology Criteria for Adverse events; ECOG PS=Eastern Cooperative Oncology Group performance status; EORTC QLQ-C30=European Organisation for Research and Treatment of Cancer Core Quality of Life Questionnaire; ITT=intention to treat analysis; LDAC=low-dose cytarabine; N=Number; NCI=National Cancer Institute; OS=overall survival; PR=partial remission; ULN=upper limit of normal; WBC=white blood cell.

Table A7. Characteristics of the interventions and comparators – indirect comparison

Study reference / ID Intervention Comparator(s) Treatment characteristics

AZA-AML-001 Azacitidine 75 mg/m2 per day s.c. for 7 consecutive days per 28-day treatment cycle for at least 6 cycles

BSC: blood product transfusions and antibiotics, with granulocyte colony-stimulating factor for neutropenic infection;

LDAC: 20 mg twice per day for 10 days per 28-day treatment cycle for at least 4 cycles);

IC: cytarabine 100-200 mg/m2 per day by continuous intravenous infusion for 7 days, plus 3 days of either daunorubicin 45-60 mg/m2 per day or idarubicin 9-12 mg/m2 per day) for 1 cycle, followed by up to 2 consolidation cycles for those achieving CR or partial response.

All study participants could receive BSC, including transient use of hydroxyurea (hydroxyurea was not allowed within 72 hours before or after azacitidine administration).

DACO-016

Decitabine 20 mg/m2 per day as a 1-hour intravenous infusion for five consecutive days every 4 weeks

Supportive care;

Cytarabine 20 mg/m2 once per day as a subcutaneous injection for 10 consecutive days every 4 weeks.

Hydroxyurea was allowed until cycle 1 on day 15

Source: (4, 43, 44) Abbreviations: BSC=best supportive care; IC=induction chemotherapy; LDAC=low-dose cytarabine.



Table A8. Information on the course of the studies (including planned duration of follow-up) – indirect comparison

Study reference / ID Outcome category

Planned follow-up Intervention Control

AZA-AML-001 Azacitidine N = 236

LDAC N = 153

Treatment duration [days]

Median [Min; Max] - 191.5 (n.r.) 125.0 (n.r.)

Mean (SD) - n.r. n.r.


Overall survivala Until the date of death from any cause -

Median [Min; Max] 24.4 (n.r.)

Mean (SD) n.r. n.r.

DACO-016 Decitabine N = 242

LDAC N = 215

Treatment duration [days]

Median [Min; Max] - n.r. n.r.

Mean (SD) - n.r. n.r.


Overall survival Until the date of death from any cause or loss to follow-up

-

Median [Min; Max] n.r. n.r.

Mean (SD) n.r. n.r.

Source: (34, 35, 43-45) a In ITT-population (Azacitidine: N = 241 vs. Conventional Care Regimen: N = 247 (158 of whom received LDAC)). Abbreviations: LDAC=low-dose cytarabine; max=maximum; min=minimum; N=number of analysed patients; n.r.=not reported; SD=standard deviation.



Table A9. Baseline characteristics of the study populations (relevant subgroup or whole study population) – indirect comparison

BRIGHT AML 1003 AZA-AML-001 DACO-016


LDAC N = 38

Azacitidine N = 241a

LDAC subgroup N = 158

Decitabine N = 242b

Treatment choice, including BSC

(N = 28) and LDAC (N = 215) N = 243c

Demographic Characteristics

Sex, n (%)

Male 59 (75.6) 23 (60.5) 139 (57.7) 94 (59.5) 137 (56.6) 151 (62.1)

Female 19 (24.4) 15 (39.5) 102 (42.3) 64 (40.5) 105 (43.4) 92 (37.9)

Age (years)

Mean (SD) 76.4 (6.0) 74.8 (4.9) n.r. n.r. 73.14 (5.242) 73.53 (5.668)

Median (range) 77 (64-92) 76 (58-83) 75 (64-91) 75 (65-88) 73 (64-89) 73 (64-91)

Age categories, n (%)

< 75 years 30 (38.5) 15 (39.5) 103 (42.7) 75 (47.5) < 70 years: 71 (29.3) < 70 years: 70 (28.8)

≥ 75 years 48 (61.5) 23 (60.5) 138 (57.3) 83 (52.5) ≥ 70 years: 171 (70.7) ≥ 70 years: 173 (71.2)

Clinical characteristics

Type of AML, n (%)

De novo 38 (48.7) 18 (47.4) 192 (79.7) 135 (85.4) 155 (64.0) 157 (64.6)

Secondary 40 (51.3) 20 (52.6) 49 (20.3) 23 (14.6) 87 (36.0) 84 (34.6)

ECOG PS, n (%)

0-1 36 (46.2) 20 (52.6) 186 (77.2) 123 (77.9) 184 (76.0) 183 (75.3)

2 41 (52.6) 18 (47.4) 55 (22.8) 35 (22.2) 58 (24.0) 60 (24.7)

missing 1 (1.3) 0 0 0 0 0

History of severe cardiac disease, n (%)

no 26 (33.3) 18 (47.4) n.r. n.r. n.r. n.r.

yes 52 (66.7) 20 (52.6) n.r. n.r. n.r. n.r.

Serum creatinine, n (%)

≤1,3 mg/dL 62 (79.5) 32 (84.2) n.r. n.r. n.r. n.r.

> 1,3 mg/dL 15 (19.2) 5 (13.2) n.r. n.r. n.r. n.r.

Missing

1 (1.3) 1 (2.6) n.r. n.r. n.r. n.r.





LDAC N = 38



Decitabine N = 242b


(N = 28) and LDAC (N = 215) N = 243c

ELN cytogenetic risk stratification (according to CRF, per ELN 2010 guidelines)

Favourable 5 (6.4) 3 (7.9) n.r. n.r. n.r. n.r.

Intermediate-I 27 (34.6) 11 (28.9) n.r. n.r. n.r. n.r.

Intermediate-II 21 (26.9) 8 (21.1) n.r. n.r. n.r. n.r.

Adverse 25 (32.1) 16 (42.1) n.r. n.r. n.r. n.r.

Cytogenetic risk, n (%)

according to CRF

(per ELN 2010

guidelinesd)

according to IVRS

according to CRF

(per ELN 2010

guidelinesd)

according to IVRS

per local review

per central review

(per modified

NCCN CPG

2010e)

per local review

per central review

(per modified

NCCN CPG

2010e)

per South West Oncology Group Categorizationf

Good/intermediate 53 (67.9) 49 (62.8) 22 (57.9) 21 (55.3) 159 (66.0)

155 (64.3)

102 (64.6)

104 (65.8)

152 (63.1) 154 (63.6)

Poor 25 (32.1) 29 (37.2) 16 (42.1) 17 (44.7) 82 (34.0) 85 (35.3) 56 (35.4) 54 (34.2) 87 (36.1) 87 (36.0)

Missing 0 0 0 0 0 1 (0.4) 0 0 2 (0.8) 1 (0.4)

Peripheral white blood cell count (103/mm3)

Median (range) 2.47 (0.6-64.0) 4.07 (1.1-45.2) 3.1 (0-33) 2.3 (0-73) 3.10 (0.3-127.0) 3.69 (0.5-80.9)


Median (range) 41 (16-100) 46 (13-95) 70 (2-100) 74 (4-100) 46.6 (3.0 – 100) 45 (0 – 100)

Bone marrow blasts, n (%)

< 20 % 3 (3.8) 1 (2.6) n.r. n.r. n.r. n.r.

≥ 20 - ≤30 % 21 (26.9) 9 (23.7) n.r. n.r. 65 (27.0) 58 (24.1)

> 30 - ≤50 % 20 (25.6) 8 (21.1) n.r. n.r. 67 (27.8) 74 (30.7)

≤ 50% n.r. n.r. 65 (27.0) 27 (17.1) n.r. n.r.

> 50 % 31 (39.7) 17 (44.7) 173 (71.8) 128 (81.0) 105 (43.7) 101 (41.9)

Not reported 3 (3.8) 3 (7.9) 3 (1.2) 3 (1.9) n.r. n.r.





LDAC N = 38



Decitabine N = 242b


(N = 28) and LDAC (N = 215) N = 243c

Hemoglobin (g/dL)

Median (range) 9.1 (6.4-14.0) 9.3 (6.0-14.6) 9.5 (5.0-13.4) 9.3 (5.6-14.4) 9.3 (5.2-15.0) 9.4 (5.0-12.6)

Duration since histopathological diagnosis (months) (days)

Median (range) 0.57 (0.0-3.5) 0.51 (0.1-3.8) 0.3 (0-19.8) 0.3 (-0.2-20.2) 14.0 (3.0-346.0) 15.0 (0.0-398.0)

Source: (34, 35, 39, 40, 43, 44, 46, 47) a Comprised all patients in the azacitidine arm, regardless of their preselected care regimen. Of these 241 patients, 44 were preselected for BSC, 43 were preselected for intensive chemotherapy and 154 were preselected for LDAC. b Comprised all patients in the decitabine arm, regardless of their preselected care regimen. The number of patients preselected for LDAC was not published by Kantarjian 2012. c Comprised all patients in the treatment choice arm, regardless of their preselected care regimen. Of these 243 patients, 28 received BSC and 215 received LDAC. Baseline and disease characteristics for LDAC were available in the publication of Kantarjian 2012, but the complete treatment choice arm is relevant for the indirect comparison. d According to ELN 2010 guidelines: Patients were classified as having poor-risk disease if they had one of the following cytogenetic features: inv(3), t(6;9), 11q23, –5, –5q, –7, abnormal (17p), or complex karyotype (≥3 clonal abnormalities). Patients with none of these features were classified as having good/intermediate- risk disease (good/intermediate cytogenetic risk=favourable, intermediate-I and intermediate-II risk groups according to ELN 2010 (17). e According to NCCN 2009 criteria: intermediate-risk category: +8, t(9;11), normal karyotype. Poor risk category: inv(3), t(3;3), t(6;9), t(9;22),11q23, –5, –5q, –7, -7q or complex karyotype (≥3 clonal abnormalities) (54). f According to South West Oncology Classification; intermediate risk category: 18, 2Y, 16, del(12p), or normal karyotype. Unfavourable risk category: presence of one or more of 25/del(5q), 27/del(7q), inv(3q), abn 11q, 20q, or 21q, del(9q), t(6;9), t(9;22), abn 17p, and complex karyotype defined as 3 or more abnormalities (53). Abbreviations: AML=acute myeloid leukaemia; BSC=best supportive care; CPG=clinical practice guideline; CRF=case report form; ECOG PS=Eastern Cooperative Oncology Group performance status; ELN=European Leukaemia Net; IVRS=interactive voice response system; LDAC=low-dose cytarabine; N=number of patients; NCCN=National Comprehensive Cancer Network; n.r.=not reported; SD=standard deviation.



APPENDIX 4: EVIDENCE GAPS

Table A10. Recommendations for research

Additional evidence generation needs

Research question 1: What is the comparative effectiveness and safety of glasdegib in combination with

LDAC, for the treatment of newly diagnosed de novo or secondary AML in adult patients who are not candidates for standard induction chemotherapy in comparison to other treatment options?

Evidence Limited evidence from a phase 2 subset of one small phase 1b/2 study

Population Newly diagnosed de novo or secondary AML in adult patients who are not candidates for standard induction chemotherapy

Intervention Glasdegib + LDAC

Comparator Decitabine, azacitidine, BSC

Outcome(s) OS, HrQoL, safety

Time stamp 09.07.2020

Study design RCT blinded and powered to show differences, proper indirect evidence or register studies (RWE)

Ongoing studies NCT03416179 (glasdegib + azacitidine vs. azacitidine + placebo), Search conducted on 10th of July 2020

a All comparators were explicitly mentioned by patient organizations. b Including data at 6 and 12 months. c Including data at 90 days. Abbreviations: AE=adverse event; AML=acute myeloid leukaemia; BSC=best supportive care; HrQoL=health-related quality of life; HTA=health technology assessment; LDAC=low-dose cytarabine; OS=overall survival; QT=Q-TWiST; RCT=randomized control trial; SR=study report.

EUnetHTA Joint Action 3 WP4 · 1.1.1 International classification of diseases (Version 10 [ICD-10]...

Documents

Transcript of EUnetHTA Joint Action 3 WP4 · 1.1.1 International classification of diseases (Version 10 [ICD-10]...