Changes in bone marrow fibrosis during momelotinib or ruxolitinib therapy do not correlate with efficacy outcomes in patients with myelofibrosis

Abstract Bone marrow fibrosis (BMF) is a pathological feature of myelofibrosis, with higher grades associated with poor prognosis. Limited data exist on the association between outcomes and BMF changes. We present BMF data from Janus kinase (JAK) inhibitor–naive patients from SIMPLIFY‐1 (NCT01969838), a double‐blind, randomized, phase 3 study of momelotinib vs ruxolitinib. Baseline and week 24 bone marrow biopsies were graded from 0 to 3 as per World Health Organization criteria. Other assessments included Total Symptom Score, spleen volume, transfusion independence status, and hemoglobin levels. Paired samples were available from 144 and 160 patients randomized to momelotinib and ruxolitinib. With momelotinib and ruxolitinib, transfusion independence was achieved by 87% and 44% of patients with BMF improvement of ≥1 grade and 76% and 56% of those with stable/worsening BMF; there was no association between BMF changes and transfusion independence for either arm (momelotinib, p = .350; ruxolitinib, p = .096). Regardless of BMF changes, hemoglobin levels also generally increased on momelotinib but decreased on ruxolitinib. In addition, no associations between BMF changes and spleen (momelotinib, p = .126; ruxolitinib, p = .407)/symptom (momelotinib, p = .617; ruxolitinib, p = .833) outcomes were noted, and no improvement in overall survival was observed with ≥1‐grade BMF improvement (momelotinib, p = .395; ruxolitinib, p = .407). These data suggest that the anemia benefit of momelotinib is not linked to BMF changes, and question the use of BMF assessment as a surrogate marker for clinical benefit with JAK inhibitors.

and hemoglobin levels.Paired samples were available from 144 and 160 patients randomized to momelotinib and ruxolitinib.With momelotinib and ruxolitinib, transfusion independence was achieved by 87% and 44% of patients with BMF improvement of ≥1 grade and 76% and 56% of those with stable/worsening BMF; there was no association between BMF changes and transfusion independence for either arm (momelotinib, p = .350;ruxolitinib, p = .096).Regardless of BMF changes, hemoglobin levels also generally increased on momelotinib but decreased on ruxolitinib.In addition, no associations between BMF changes and spleen (momelotinib, p = .126;ruxolitinib, p = .407)/symptom(momelotinib, p = .617;ruxolitinib, p = .833)outcomes were noted, and no improvement in overall survival was observed with ≥1-grade BMF improvement (momelotinib, p = .395;ruxolitinib, p = .407).These data suggest that the anemia benefit of momelotinib is not linked to BMF changes, and question the use of BMF assessment as a surrogate marker for clinical benefit with JAK inhibitors.

K E Y W O R D S
bone marrow fibrosis, JAK inhibitor, momelotinib, myelofibrosis, ruxolitinib

Leukemic transformation is a potential complication of myelofibrosis
and is almost always fatal, with 20% of primary myelofibrosis cases evolving into acute myeloid leukemia [5,10,11].
BMF is a key histopathological feature and major diagnostic criterion of myelofibrosis and is characterized by increased deposition of reticulin and/or collagen fibers secondary to aberrant immature megakaryocytes, which release excessive amounts of inflammatory and fibrogenic cytokines [3,6,[29][30][31].Bone marrow biopsies are evaluated and scored using updated World Health Organization criteria consisting of 4 escalating grades of severity: grade 0-3 [32].
BMF grading, while not incorporated in conventional prognostic score

Patients
The study design of SIMPLIFY-1, a randomized, head-to-head, double-blind, phase 3 study of momelotinib vs ruxolitinib, has been described elsewhere [27].In brief, 432 JAK inhibitor-naive patients with primary myelofibrosis, post-polycythemia vera myelofibrosis, or post-essential thrombocythemia myelofibrosis were randomized 1:1 to receive momelotinib or ruxolitinib (Figure S1) [27].This JAK inhibitor-naive setting was utilized in the present analysis to minimize confounders from prior treatments.The study was approved by the institutional review boards and independent ethics committees at each study site, and all participants provided written informed consent.Additional signatures were collected for the use of biopsies in other biomarker analyses.

Bone marrow aspirate and biopsy
Bone marrow aspirate and biopsies were performed prior to treatment initiation (baseline), after 24 weeks of randomized treatment (within a window of ± 7 days) with either momelotinib or ruxolitinib, at week 96, and as required to assess response per International Working Group-

Myeloproliferative Neoplasms Research and Treatment and European
LeukemiaNet criteria [40].Patients continued treatment until the completion of this visit.Bone marrow aspirate and biopsy samples were assessed by a local hematopathologist for BMF grading using the revised 2016 World Health Organization classification and diagnostic criteria for myeloproliferative neoplasms; assessment included reticulin (e.g., silver stain) and collagen fibers (e.g., trichrome stain) from grade 0 (normal bone marrow) to grade 3 (diffuse and dense increase in reticulin; coarse bundles of thick fibers consistent with collagen) [32].

Additional analyses
As described previously [27], the rate of ≥35% spleen volume reduction (assessed by magnetic resonance imaging or computed tomography and evaluated by a blinded central reader), the rate of ≥50% reduction in Total Symptom Score (TSS), the rate of transfusion independence, and hemoglobin levels were assessed.For this analysis, baseline and week 24 data were used.

Statistical analysis
χ 2 test was used to assess the association between BMF change and both transfusion independence response and spleen/symptom response at week 24 in each arm.To examine whether BMF change at week 24 was predictive of overall survival (OS), OS was calculated from week 24, analyzed using the Kaplan-Meier analyses, and compared between groups by BMF change at week 24 with log-rank tests and proportional hazard Cox regression models.All analyses were descriptive.

BMF grades at baseline
Of grading trended toward greater spleen volume (Figure 2E) but not higher TSS (Figure 2F).

BMF grade changes from baseline to week 24
In SIMPLIFY-

BMF grade changes and transfusion independence at week 24
Transfusion independence, defined as the absence of transfusions and no hemoglobin levels < 8 g/dL in the 12 weeks before week 24, was achieved by 78% of patients (113/144) in the momelotinib arm vs 53% (85/160) in the ruxolitinib arm in the analysis population with paired BMF grade at baseline and week 24.Of the 31 patients in the momelotinib arm with a ≥1-grade improvement in BMF, 27 (87%) achieved transfusion independence at week 24 (Figure 5A); of the 92 patients with stable BMF, 69 (75%) achieved transfusion independence at week 24; of the 21 patients with worsening BMF, 17 (81%) achieved transfusion independence at week 24.In the ruxolitinib arm, of the 36 patients with a ≥1-grade improvement in BMF, 16 (44%) achieved transfusion independence at week 24 (Figure 5B); of the 94 patients with stable BMF, 48 (51%) achieved transfusion independence at week 24; and of the 30 patients with worsening BMF, 21 (70%) achieved transfusion independence at week 24.Nearly twice as many patients on momelotinib vs ruxolitinib with a ≥1-grade BMF improvement achieved transfusion independence at week 24 (87% [27/31] vs 44% [16/36]).
For those with worsening BMF, the transfusion independence rate was higher in the momelotinib arm compared with the ruxolitinib arm (81% [17/21] vs 70% [21/30]).In patients treated with momelotinib, transfusion independence was achieved regardless of the extent of changes in BMF.

BMF grade changes and hemoglobin at week 24
Hemoglobin levels at baseline and week 24 by changes in BMF grade are shown in Figure 6 and Figure S2.In the momelotinib arm, hemoglobin levels increased regardless of the extent of changes in BMF; changes from baseline to week 24 were 10.7 to 11.7, 11.0 to 11.7, 11.3 to 12.4, and 9.5 to 10.5 (mean level in g/dL) with a ≥1-grade worsening, 1-grade improvement, 2-grade improvement, and 3-grade improvement, respectively.In the ruxolitinib arm, hemoglobin levels generally decreased with both improving and worsening BMF; changes from baseline to week 24 were 11.3 to 10.3, 10.6 to 9.6, and 11.5 to 9.9 (mean level in g/dL) with ≥1-grade worsening, 1-grade improvement, and 2-grade improvement, respectively; there were no patients with a 3-grade improvement in the ruxolitinib arm.
BMF grade improvements were also not associated with OS benefits in patients randomized to momelotinib who achieved transfusion independence at week 24 (HR [improvement vs no improvement], 1.03; p = .950)(Figure 7C).

DISCUSSION
Improvement in BMF has been pursued as an endpoint in clinical trials for myelofibrosis, especially in combination therapies, with a presumed link to disease modification [4,36,38].Although studies have shown that the grade of BMF itself may directly influence OS [34,37,41], The data presented here showed modest improvement in BMF grade after 24 weeks of momelotinib or ruxolitinib treatment, with approximately 20% of patients having a ≥1-grade improvement.In contrast, BMF was stable in > 60% of patients.In both arms, BMF grade change was not associated with symptom or spleen response.
Although such similarities were observed, outcomes related to anemia were different between the two arms.For patients treated with momelotinib, hemoglobin levels increased, and transfusion independence was achieved regardless of BMF changes (improvement or worsening).In contrast, for patients treated with ruxolitinib, those with BMF improvement-even those achieving a 2-grade BMF improvement-did not have improved hemoglobin levels, and worsening BMF grade was generally correlated with decreased hemoglobin levels.Changes in BMF grade also did not impact the achievement of transfusion independence in the ruxolitinib arm.These findings are suggestive of the previously described anemia benefit of momelotinib, a feature of its JAK1/JAK2/ACVR1-mediated mechanism of action, which is not seen with ruxolitinib [42,43].Of note, not achieving transfusion independence was also not associated with changes in BMF grade.
Although SIMPLIFY-1 was completed in 2019, the long-term followup of patients in the study and after rollover to the extended access protocol (NCT03441113) allowed OS assessment.rating additional bone marrow histopathology assessments such as megakaryocyte cell size and clustering, along with improved quantitative approaches for evaluating cytological and topographical bone marrow features, may be informative [46,47].

CONCLUSION
In conclusion, the current analysis showed that both momelotinib and ruxolitinib can improve BMF grade, but the improvement was not associated with OS advantage or better efficacy outcomes; benefits of anemia with momelotinib were observed regardless of BMF changes.
The clinical significance of BMF changes with nontransplant therapies for myelofibrosis should be carefully evaluated to better understand its role in disease modification, the pathogenesis of myelofibrosis, and clinically important outcomes.

6 F I G U R E 3
ruxolitinib arm crossed over to receive open-label momelotinib and patients in the momelotinib arm continued to open-label momelotinib, is shown in Figure 7.The median follow-up for OS was approximately 3 years.OS was analyzed by subgroups of change in BMF F I G U R E 2 Overall SIMPLIFY-1 population by BMF grade at baseline.(A) Hemoglobin.(B) Transfusion status.(C) Platelet counts.(D) IPSS.(E) Spleen volume.(F) TSS.BMF, bone marrow fibrosis; Int1, intermediate 1; Int2, intermediate 2; IPSS, International Prognostic Scoring System; TSS, Total Symptom Score.grades at week 24 within each treatment arm.During the momelotinib open-label extended treatment period, changes in BMF grades at week 24 were not associated with OS benefit in either treatment arm.In the momelotinib-randomized arm, OS rates at 2, 4, and 6 years were approximately 95%, 75%, and 65%, respectively, among patients with worsening BMF; 80%, 60%, and 55%, respectively, among patients with stable BMF; and 85%, 70%, and 70%, respectively, among patients with a ≥1-grade improvement (Figure 7A).The Cox regression model showed no differences in OS in momelotinib-randomized patients between those with a ≥1-grade improvement vs worsening BMF (hazard ratio [HR], 1.22; 95% confidence interval [CI], 0.40-3.74;p = .725).In the ruxolitinib-randomized arm, OS rates at 2, 4, and Change in BMF grade from baseline to week 24 in JAK inhibitor-naive patients in SIMPLIFY-1.(A) Patients treated with momelotinib.(B) Patients treated with ruxolitinib.BMF, bone marrow fibrosis; JAK, Janus kinase.
data, especially in the phase 3 setting, are available on how change or improvement in the grade of BMF potentially correlates with survival or other efficacy outcomes.The current analysis included bone marrow biopsy samples from SIMPLIFY-1, a large, completed phase 3 clinical trial of momelotinib vs ruxolitinib in JAK inhibitor-naive patients with myelofibrosis.The availability of such samples offered a unique opportunity to better understand the impact of two differentiated JAK inhibitors on BMF and the clinical consequences of changes in BMF.The proportions of each baseline BMF grade were similar between treatment arms, likely due to randomization stratification.

F I G U R E 4 F I G U R E 6
Proportion of JAK inhibitor-naive patients in SIMPLIFY-1 who achieved symptom and spleen response by change in BMF grade from baseline to week 24.(A, C) Patients treated with momelotinib.(B, D) Patients treated with ruxolitinib.*Symptom response is defined as achieving a ≥50% reduction in TSS over the 28 days immediately before the end of week 24 compared with baseline.The percentage is calculated using the BMF change category as the denominator (i.e., ≥1-grade improvement, no change, or worsening).† Spleen response is defined as achieving a ≥35% reduction in spleen volume from baseline.The percentage is calculated using the BMF change category as the denominator (i.e., ≥1-grade improvement, no change, or worsening).The p-value was calculated using a χ 2 -test.BMF, bone marrow fibrosis; JAK, Janus kinase; TSS, Total Symptom Score. of first-line JAK inhibitor monotherapy, the results cannot be generalized to myelofibrosis treatments with other mechanisms of action[39].Furthermore, although the patient population was large, the number of patients having BMF improvement of ≥1 grade was low.Finally, although baseline genomic profiling may provide insights into potential biomarkers associated with BMF changes or other clinical endpoints, samples were not available at the time of this analysis.Nonetheless, this represents the most extensive analysis to date of the correlation of BMF changes with other outcome measures, with over 300 paired biopsies and mature clinical data, in JAK inhibitor-naive patients with myelofibrosis.While the data continue to support the key role of ACVR1 inhibition together with JAK1 and JAK2 inhibition in the differentiated anemia benefit derived with momelotinib, they bring into question the use of BMF assessment at week 24 as a surrogate for clinical benefit and disease modification, given the lack of association between changes in BMF grade and OS and other efficacy outcomes.Emerging data from ruxolitinib and ruxolitinib combination therapies have shown approximately 30% of patients having a ≥1-grade improvement in BMF; however, the improvement has not been significantly correlated with survival or other clinical outcomes[36,37], except perhaps in individual cases[44,45].Future analyses of BMF grading incorpo-F I G U R E 5 Proportion of JAK inhibitor-naive patients in SIMPLIFY-1 who achieved transfusion independence response by change in BMF grade from baseline to week 24.(A) Patients treated with momelotinib.(B) Patients treated with ruxolitinib.Transfusion independence response was defined as the absence of red blood cell transfusions and no hemoglobin levels < 8 g/dL in the 12 weeks before week 24.The p-value was calculated using a χ 2 -test.BMF, bone marrow fibrosis; JAK, Janus kinase.Hemoglobin levels at baseline and week 24 in JAK inhibitor-naive patients in SIMPLIFY-1 with worsening BMF grade from baseline to week 24.(A) Patients treated with momelotinib.(B) Patients treated with ruxolitinib.*A total of 3/21 patients were missing week 24 hemoglobin measurement.† A total of 2/30 patients were missing week 24 hemoglobin measurement.BMF, bone marrow fibrosis; JAK, Janus kinase.
Kawashima, Mei Huang, and Bryan Strouse contributed to the study design; Stephen T. Oh, Srdan Verstovsek, Vikas Gupta, Uwe Platzbecker, Timothy Devos, Jean-Jacques Kiladjian, Donal P. McLornan, Andrew Perkins, Maria Laura Fox, Mary Frances McMullin, Adam F I G U R E 7 Overall survival from week 24 by baseline to week 24 BMF grade changes in JAK inhibitor-naive patients.(A) Patients in the momelotinib arm.(B) Patients in the ruxolitinib arm.(C) Patients in the momelotinib arm who achieved transfusion independence.BMF, bone marrow fibrosis; HR, hazard ratio; JAK, Janus kinase; MMB, momelotinib; OS, overall survival; RUX, ruxolitinib.