Dynamic and Time-to-Event Analyses Demonstrate Marked Reduction in Transfusion Requirements for Janus Kinase Abstract Inhibitor-Naïve Myelofibrosis Patients Treated with Momelotinib Compared Head to Head with Ruxolitinib #1663
Ruben A Mesa, MD, FACP1, John Catalano, MBBS, FRACP, FRCPA2, Francisco Cervantes, MD3, Timothy Devos, MD, PhD4, Miklós Egyed, MD, PhD5, Jason Gotlib, MD6, Jean-Jacques Kiladjian, MD, PhD7, Donal P. McLornan, MB, BCh (Hons), MRCP, PhD, FRCPath8, Kazuya Shimoda, MD, PhD9, Elisabeth Coart, PhD10, Koenraad D’Hollander, MD, MSC10, Rafe Donahue, PhD11 and Mark M. Kowalski, MD, PhD11 1Mays Cancer Center, UT Health San Antonio Cancer Center, San Antonio, TX; 2Frankston Hospital, Frankston, VIC, Australia; 3Hematology Department, Hospital Clinic, IDIBAPS, University of Barcelona, Barcelona, Spain; 4Department of Microbiology and Immunology, University Hospitals Leuven (UZ Leuven), Leuven, Belgium; 5Department of Hematology, Somogy County Mór Kaposi General Hospital, Kaposvár, Hungary; 6Division of Hematology, Stanford Cancer Institute, Stanford University School of Medicine, Stanford, CA; 7Centre d’Investigations Cliniques (INSERM CIC 1427), AP-HP, Hopital Saint-Louis, Université de Paris, Paris, France; 8Guy’s and St Thomas’ NHS Foundation Trust, London, United Kingdom; 9Department of Gastroenterology and Hematology, Faculty of Medicine, University of Miyazaki, Miyazaki, Japan; 10IDDI, Louvain-la-Neuve, Belgium; 11Sierra Oncology, Vancouver, BC, Canada
Background Momelotinib’s Mechanism of Action
Anemia is a Critical Prognostic Hallmark of Myelofibrosis Figure 1: Anemia Predicts Poor Survival in Myelofibrosis Momelotinib Uniquely Inhibits JAK1, JAK2 and ACVR1 Figure 2: Momelotinib Uniquely Inhibits All Three Myelofibrosis Figure 3: Momelot inib Inhibits ACVR1 Leading to Positive Standard Analyses of the SIMPLIFY-1 Data Demonstrate Disease Drivers Anemia Benefits Momelotinib’s Anemia Benefits vs Ruxolitinib • Approximately 60% of patients with myelofibrosis (MF) are 10 • Momelotinib (MMB) is a potent inhibitor of JAK1, JAK2 and ACVR1 (Figures 2 and 3) providing benefits on all three hallmarks anemic and 45% are transfusion dependent (TD) within 1 Interleukins Interferons Ligand BMP2, BMP6 • SIMPLIFY 1 (S1) is a Phase 3, head-to-head comparison of MMB No anemia of MF: constitutional symptoms, anemia and splenomegaly, Plasma Iron Deficiency Plasma Iron Normalization year of diagnosis (Tefferi 2014), with most progressing to 8 Median survival 7.9 years vs ruxolitinib (RUX) in JAKi treatment naïve myelofibrosis patients transfusion dependency over time. without the myelosuppressive properties of approved JAK Cytokine Receptors EPOR / MPL ACVR1 with a 24-week double-blind treatment period. Mild anemia inhibitors (JAKi). Median survival 4.9 years BMP2, BMP6 BMP2, BMP6 6 • Relevant baseline characteristics were well balanced, including • Transfusion dependency and moderate/severe anemia Moderate anemia JAK2 JAK2 JAK2 Median survival 3.4 years • The complex and inter-related drivers of anemia in MF include JAK1 ACVR1 ACVR1 median hemoglobin levels (10.5 g/dL and 10.3 g/dL) and (Figure 1) are acknowledged to be critical negative prognostic Severe anemia bone marrow fibrosis fundamental to the disease, splenic Survival percentages of TD and TI patients (25% and 24%; and 68% and factors in MF (Elena 2011, Nicolosi 2018). 4 Median survival 2.1 years sequestration of red blood cells (RBCs), and anemia of 70%) for the MMB and RUX groups, respectively. STAATT P SMAD1,5SM 1,5 Fe2+ inflammation, including the inflammatory cytokine-mediated STATT P P Fe2+ SMAD1,5 • Accordingly, the burden of receiving transfusions as supportive P SMAD1,5 • A variety of standard landmark analyses of anemia benefit were 2 inhibition of residual erythropoietic bone marrow capacity. P care, as well as the complexities and complications of Increased Decreased conducted in S1, demonstrating consistently positive benefits in • Anemia of inflammation is a complex disorder, driven through hepcidin hepcidin transfusions, have a significant impact on both quality of life P<0.0001 favor of MMB: 0 both the direct and indirect effects of cytokines and increased and overall survival in MF. 0 5 10 15 20 25 30 35 synthesis of the iron regulatory hormone hepcidin controlled via MMB RUX p-value References: Tefferi et al, Mayo Clin Proc. 2014; Elena et al, Haematologica 2011; Nicolosi M et al; Leukemia 2018 Years ACVR1-directed SMAD signaling. Aberrant activation of hepcidin transcription Decreased hepcidin transcription Decreased inflammation Reduced extramedullary Decreased hepcidin Nicolosi M et al, Leukemia 2018 and aberrant cytokine hematopoiesis improves transcription restores iron via hyperactivated ACVR1 signaling results in restores iron homeostasis and increases hemoglobin % TI at Week 24 67 49 < 0.001 • Increased hepcidin leads to iron sequestration in monocytes and signaling reduces splenomegaly homeostasis and increases profound functional iron deficiency anemia leading to an array of anemia benefits constitutional symptoms hemoglobin leading to an macrophages, resulting in perturbed iron homeostasis and a array of anemia benefits % TD at Week 24 30 40 0.019 characteristic iron-restricted anemia (Ganz 2013, Langdon 2014). % TD à TI (rolling 12-week) 49 29 0.0455 References: Ganz et al, Physiol Rev. 2013; Langdon at al, Am J Hematol. 2014
Methods and Results Conclusions
Novel Dynamic Analyses of Transfusion Burden • Uniquely amongst the JAKi class, momelotinib • Landmark or other “static” analyses alone do not Figures 4B and 4C: Immediate and Sustained Reduction completely describe the patient burden of transfusions or in Transfusion Burden on Momelotinib Zero-Inflated Negative Binomial Methods Mean Cumulative Function Methods Time-to-Loss of Transfusion Independence potently inhibits JAK1, JAK2 and ACVR1, resulting fully elucidate the differences in therapeutic options for Methods in a significant mechanistically-driven range clinicians. • To further analyze transfusion burden in S1, a A proportional hazards recurrent events model of myelofibrosis benefits, including an array of Kaplan-Meier Survival Function Estimates for Time to Third RBC Unit zero-inflated negative binomial (ZINB) model, analyzing recurrent data was employed to assess the • The duration of TI response in S1 was determined • Retrospective analyses of S1 data were performed using With Number of Patients at Risk and 95% Confidence Limits encompassing baseline covariates, was fit to the relative cumulative transfusion burden between groups by a KM analysis of time to loss of TI (Figure 7). positive anemia outcomes. a variety of novel dynamic anemia benefit endpoints to 1.00 transfusion data. in a dynamic model across the duration of treatment: explore the relative burden of transfusions in patients • ‘Loss of TI’ was defined by the requirement for RBC • Current approaches to analyzing anemia benefit treated with momelotinib vs ruxolitinib. • This ZINB model was employed to compare the • Transfusions of RBC units were considered as transfusion or hemoglobin < 8.5 g/dL at any time. in myelofibrosis have relied on assessments 81% proportions of patients with zero transfusion burden recurrent events, examined with and without of transfusion dependence and transfusion 0.75 (i.e. transfusion free) and the mean transfusion rates patients’ baseline characteristics as covariates. between treatment groups (Figure 5). independence by landmark analysis. • The model focused on the hazard, or risk, of 54% 0.50 • Zero-inflated models are useful when a large undergoing an RBC unit transfused. • The novel dynamic and time-to-event methods Time-to-Transfusion Method proportion of individuals has zero events, as is the Figure 7: Transfusion Independence is Durable described here support momelotinib as a preferred case for patients in S1, while other individuals have • The outcome from this recurrent events model is a Kaplan-Meier (KM) estimates of time-to-first, time-to-third, hazard ratio (HR) comparing one treatment to the on Momelotinib treatment option for patients with myelofibrosis, and time-to-fifth RBC unit(s) transfused were employed 0.25 counts ranging to several dozen over the same time period. other and a “mean cumulative function” (MCF) that demonstrating significant anemia benefits to determine and compare relative ‘transfusion events’ Treatment describes the average cumulative number of events between groups: Proportion with fewer than 3 units transfused Momelotinib Censored Obs. 1.00 including substantively reduced transfusion burden Ruxolitinib Logrank p < .0001 • There are two components to the distribution of (RBC unit transfusions) for patients in each group +++ ++++++++++++ +++ 0.00 ++ ++ ++ ++++ compared directly with ruxolitinib, specifically: counts: the “zero” component and the “non-zero” (Figure 6). +++ Momelotinib 215 184 173 161 149 1451 87 0.75 +++++++++ • The first RBC unit transfused provides an alternate ++ ++++++++++++ ++++++++++++++++ Ruxolitinib 217 199 162 131 112 1061 77 component. Such distributions require the zero and ++++++++++++ +++++++++ relative assessment of the proportion of patients who are ++++ ++ +++ +++ • Immediate and sustained benefit manifests 0 4 8 12 16 20 24 non-zero components to be modeled separately. transfusion free over the course of treatment (Figure 4A). 0.50 Time from randomization to third unit (weeks) in an overall reduced transfusion burden on • When greater than zero, the number of RBC units transfused momelotinib. Survival probability Survival 0.25 can be considered to represent ‘transfusion events’; Figure 6: An Average Patient Received Twice as Many RBC assuming two units of RBCs per typical transfusion, the • Patients receiving momelotinib had a significantly Transfusions on Ruxolitinib at Any Timepoint third and fifth RBC units transfused represent the second Figure 5: ZINB Model Demonstrates Increased Odds 0.00 reduced chance of receiving one transfusion, Kaplan-Meier Survival Function Estimates for Time to Fifth RBC Unit 0 40 80 120 160 and third ‘transfusion event’ respectively (Figure 4B and 4C). With Number of Patients at Risk and 95% Confidence Limits of Zero Transfusions on Momelotinib Time in weeks two ‘transfusion events’ (≥ 3 RBC units) or three
1.00 ‘transfusion events’ (≥ 5 RBC units). 0.9 Average Cumulative Number of Units Transfused MMB 172 (100) 116 (67) 69 (40) 9 (5) 3 (2) Figure 4A: More Patients Require No Transfusions on Momelotinib Strata 0.8 6 Number at risk: n (%) 83% Momelotinib • Odds of receiving zero transfusions, in a 0.7 Ruxolitinib Kaplan-Meier Survival Function Estimates for Time to First RBC Unit 0.75 Odds of 5 covariate ZINB model, were more than With Number of Patients at Risk and 95% Confidence Limits 0.6 zero transfustion This analysis demonstrated that the median time to loss of TI 62% 9.3 times higher nine times higher on momelotinib. 1.00 0.5 on momelotinib 4 was not reached for MMB-treated patients, with a follow up Higher cumulative transfusion period exceeding 3 years. 0.50 0.4 units with ruxolitinib • At any given time, the mean cumulative number
Probability 3 0.3 of RBC units received for a typical patient
0.75 Number of Units 73% 0.2 2 receiving momelotinib is approximately half 0.25 0.1 of that for patients receiving ruxolitinib (HR = Treatment 1 0.0 0.50 Proportion with fewer than 5 units transfused Momelotinib Censored Obs. Ruxolitinib Momelotinib 0.522). Thus, for patients on momelotinib half as 46% Ruxolitinib Logrank p < .0001 0 MOMENTUM 0.00 0 4 8 12 16 20 24 many RBC units are transfused at any time as Momelotinib 215 194 181 168 154 1461 87 Ruxolitinib 217 199 179 154 122 1101 77 Time on Study (weeks) compared to ruxolitinib. 0.25 0 4 8 12 16 20 24 The outcomes of the covariate* ZINB model demonstrate Phase 3 Clinical Trial Proportion with no units transfused Treatment Time from randomization to fifth unit (weeks) that a typical patient in S1 had an 82% chance of receiving • A sustained and durable period of TI was Momelotinib Censored Obs. no transfusions when receiving MMB vs only a 33% chance Ruxolitinib Logrank p < .0001 The hazard ratio for an RBC unit transfused for patients maintained over long-term treatment on 0.00 when receiving RUX. • Momelotinib’s suite of anemia benefits will be further Momelotinib 215 174 163 155 147 1441 87 receiving momelotinib was approximately one-half that momelotinib. evaluated in MOMENTUM, a Phase 3 clinical trial intended Ruxolitinib 217 174 128 115 106 1031 77 Similar results hold when examining a burden of fewer than three The odds of zero RBC units transfused were 9.3 times for ruxolitinib patients (HR 0.522; p < 0.0001) for models 0 4 8 12 16 20 24 both with and without patients’ baseline characteristics to support potential registration of momelotinib for • Importantly, these meaningful anemia benefits and fewer than five RBC units transfused to Week 24. In both higher on MMB than on RUX (p < 0.0001). the treatment of previously JAKi treated patients with Time from randomization to first unit (weeks) as covariates. cases a strong treatment effect was observed, demonstrating myelofibrosis. reported from the SIMPLIFY 1 Phase 3 trial that the proportion of patients with fewer transfusions was were achieved while also maintaining clinically The KM time-to-first RBC unit transfused analysis indicated an greater in the MMB arm compared with RUX: * Covariates were disease diagnosis (PMF, post-PVMF, post-ET MF), bone marrow fibrosis grade, • In addition to assessments of constitutional symptoms, and number of RBC units transfused in the eight weeks prior to randomization (0, 1-3, >4). comparable benefits on splenomegaly and immediate and sustained MMB treatment effect compared to landmark anemia rates (i.e. transfusion independence) and RUX (log-rank p < 0.0001). • Odds of receiving fewer than three transfusions was constitutional symptoms as directly compared 3.7 times higher on MMB (81%) compared to RUX splenomegaly, MOMENTUM will provide an opportunity In this model, patients randomized to MMB were more likely to (54%, p < 0.0001). to further evaluate associations between anemia benefit, to ruxolitinib. receive no transfusions (73%) compared to patients randomized transfusion burden and patient reported measures of • Odds of receiving fewer than five transfusions was 3.0 times clinical benefit. to RUX (46%). The odds of receiving no transfusions during higher on MMB (83%) compared to RUX (62%, p < 0.0001). treatment was 3.2 times higher on MMB than on RUX.
Information Acknowledgments For more information, email [email protected] The SIMPLIFY-1 and -2 studies were sponsored by Gilead Sciences. As the current or visit www.sierraoncology.com compound sponsor, Sierra Oncology would like to thank all participating patients and their families as well as participating study sites.