TRANSFORM-1 phase 3 study: efficacy and safety of navitoclax plus ruxolitinib in patients with untreated myelofibrosis

Key Points

• •In JAK inhibitor–naïve patients with myelofibrosis, NAV+RUX led to a twofold improvement in SVR_35W24 vs PBO+RUX, despite poor prognostic factors.
• •Addition of NAV to RUX did not change symptom severity; TSS at week 24 was not significantly different between the 2 treatment arms.

Visual Abstract

Abstract

The phase 3 TRANSFORM-1 study evaluated ruxolitinib (RUX) in combination with navitoclax (NAV) or placebo (PBO) in Janus kinase inhibitor–naïve adults with intermediate-2 or high-risk myelofibrosis and Eastern Cooperative Oncology Group performance status of ≤2. Patients were randomized 1:1 to NAV (200 mg/d starting dose, or 100 mg escalated to 200 mg/d) or PBO, with RUX dosed per label. The primary end point was ≥35% spleen volume reduction (SVR) at week 24 (SVR_35W24). Secondary end points included change from baseline in total symptom score (TSS) at week 24 and SVR₃₅ at any time. A total of 252 patients (NAV+RUX, n = 125; PBO+RUX, n = 127; median follow-up 20.3 months) were randomized; >80% had intermediate-2 risk, and ∼50% had high-molecular risk disease. SVR_35W24 was achieved by 63.2% with NAV+RUX vs 31.5% with PBO+RUX (P < .0001). Mean change in TSS at week 24 was not significantly different between NAV+RUX and PBO+RUX (−10.2 vs −11.6; P = .2852). SVR₃₅ at any time was achieved in 76.8% with NAV+RUX vs 44.1% with PBO+RUX (nominal P < .0001). A ≥20% variant allele frequency reduction (exploratory end point) occurred in 58.5% (95% confidence interval [CI], 49.0-67.5) with NAV+RUX and 45.5% (95% CI, 36.4-54.8) with PBO+RUX. NAV+RUX showed higher hematologic toxicity vs PBO+RUX (grade 3/4 thrombocytopenia, 54.0% vs 19.2%; grade 3/4 neutropenia, 40.3% vs 8.8%); diarrhea (any grade, 41.9% vs 16.8%) was also more common. Cytopenias were generally manageable and reversible with dose adjustments. This trial was registered at www.clinicaltrials.gov as NCT04472598.

The JAK inhibitors reduce splenomegaly and improve symptom control in patients with myelofibrosis, but usually they do not modify the underlying disease process. Myeloproliferative neoplasms display dependence on BCLxL, leading to the study of navitoclax, a dual BCL2/BCLxL inhibitor, with promising results suggesting disease modification in early-phase trials in combination with ruxolitinib. Pemmaraju et al report a phase 3 randomized trial of navitoclax versus placebo, finding that, while navitoclax doubles the spleen response and generates greater reductions in the variant allele frequency of mutant JAK2, there is no improvement in symptom scores and there is more hematological toxicity. These data highlight the challenges to drug development for this disease.

Introduction

Myelofibrosis (MF) is a rare myeloproliferative neoplasm associated with significant morbidity and mortality.^{1, 2, 3} Signs and symptoms of MF typically include severe fatigue, splenomegaly, weight loss, low-grade fever, bone pain, and night sweats.⁴ The approach to treating MF is generally based upon presence of symptoms and disease risk stratification.^{5, 6, 7} Potential goals of therapy can include symptom control, improvement of anemia, and reduction of spleen size, among others.

Ruxolitinib (RUX), a first-in-class Janus kinase 1/2 inhibitor (JAKi), is considered the standard of care for first-line treatment of most cases of MF, due to its profound impact on MF-associated symptoms and splenomegaly.^{8, 9, 10} Specifically, in a pooled analysis of 2 phase 3 studies that assessed RUX treatment vs placebo or best available therapy in patients with MF, larger spleen size at baseline was associated with shortened survival, whereas reductions in spleen size in responses to RUX treatment were correlated with longer survival.¹¹ In general, clinical trials have shown that, despite significant initial clinical benefit in many cases, patients with MF may lose response to RUX; 50% to 70% of these patients discontinue therapy within 3 to 5 years due to loss of response or intolerance^12,13 and ∼30% become refractory or unresponsive to RUX de novo. Outside the clinical trial setting, duration of RUX treatment may be shorter, although real-world data are limited.

In recent US- and German-based real-world studies of patients with MF treated with RUX, 64% and 48% to 57%, respectively, discontinued RUX, with a median duration of treatment of 11.0 months and 2.1 to 2.9 years in the United States and Germany, respectively.^14,15 In a separate study that reviewed medical records of patients with MF from 6 countries, median duration of RUX treatment was 13.1 months, and 44% of patients experienced inadequate response or disease progression; ∼54% of these patients discontinued RUX.¹⁶ Allogeneic stem cell transplantation is currently the only MF treatment with potential for cure. For those who cannot undergo stem cell transplantation, management is limited to targeting clinical aspects of the disease, such as cytopenias, splenomegaly, and relief of MF-related symptoms. Because symptoms are a major manifestation of the disease and affect quality of life, symptom control (measured by total symptom score [TSS]) is a primary goal in clinical trials, particularly when making comparisons with RUX monotherapy as the standard of care.¹⁷ Although other JAKi, such as fedratinib, pacritinib, and momelotinib, have been approved by the US Food and Drug Administration for specific benefits in patients with MF, their capability for disease modification is limited and most patients do not achieve deep or durable responses.^{18, 19, 20} This gap highlights an unmet need for therapies that address underlying disease biology, extending beyond the management of symptoms and spleen volume. The current landscape calls for new approaches, including combination therapies with RUX as the backbone, to potentially improve long-term outcomes in patients with MF.

Navitoclax (NAV) is an oral inhibitor of antiapoptotic B-cell lymphoma 2 (BCL-2) family member proteins BCL-XL, BCL-2, and BCL-W.²¹ In preclinical studies, the simultaneous inhibition of oncogenic signaling using RUX and prosurvival signaling via NAV within JAK2 V617F–mutated cells resulted in greater antitumor activity when compared with either RUX or NAV alone.^{22, 23, 24} In the phase 2 study, NAV+RUX was evaluated in patients with MF who had previous RUX exposure. This combination demonstrated improvement not only in standard clinical end points, such as spleen volume reduction (SVR) and symptom control but also in biomarkers associated with disease modification, including reduction in variant allele frequency (VAF) and improvements in bone marrow fibrosis (BMF).^{25, 26, 27} Here, we present data from the randomized, double-blind phase 3 TRANSFORM-1 study (ClinicalTrials.gov identifier: NCT04472598) that expanded enrollment and assessed long-term efficacy and safety of NAV+RUX vs placebo plus RUX (PBO+RUX) in JAKi-naïve adults with MF.

Methods

Study and patient population

The phase 3, double-blind, PBO-controlled, multicenter TRANSFORM-1 study (ClinicalTrials.gov identifier: NCT04472598) enrolled adults with intermediate-2 or high-risk JAKi-naïve MF per Dynamic International Prognostic Scoring System Plus risk, with measurable splenomegaly (defined as spleen palpation measurement of ≥5 cm below costal margin, or spleen volume of ≥450 cm³ as assessed centrally by magnetic resonance imaging or computed tomography scan), and Eastern Cooperative Oncology Group performance status of ≤2 (supplemental Figure 1, available on the Blood website). In addition to all patients receiving RUX at labeled dose, patients were randomized 1:1 to receive PBO or NAV (at either 200 mg/d starting dose at a platelet count of >150 × 10⁹/L, or 100 mg/d escalated to 200 mg/d with a platelet count of ≤150 × 10⁹/L).

Study end points

The primary end point for this study was ≥35% SVR at week 24 from baseline (SVR_35W24), as measured centrally by magnetic resonance imaging/computed tomography scan, per International Working Group (IWG) criteria. Secondary end points included change from baseline in TSS at week 24 as measured by Myelofibrosis Symptom Assessment Form version 4.0; ≥35% SVR (SVR₃₅) from baseline at any time during study treatment; duration of SVR₃₅; overall survival (OS); anemia response (per IWG criteria) at any time during study treatment; BMF grade (evaluated centrally), and patient-reported outcomes assessing change in fatigue from baseline (Patient-Reported Outcomes Measurement Information System Fatigue SF 7a) and physical functioning (European Organisation for Research and Treatment of Cancer Quality-of-Life-Questionnaire-Core-30). Exploratory end points included progression-free survival (PFS) per IWG criteria, change in frequency of allelic mutations (ie, JAK2, calreticulin [CALR], and myeloproliferative leukemia oncogene encoding the thrombopoietin receptor [MPL]) from baseline, and ≥50% SVR from baseline at any time. Prespecified exploratory analyses included TSS response rate based on meaningful change threshold of ≥10 points at any time during study treatment; SVR_35W24 across patient subgroups; and the proportion of patients who achieved ≥20% or ≥50% reduction from baseline in VAF of driver genes.

Data analyses

The intent-to-treat population was used for all analyses of baseline characteristics and efficacy and included all patients randomized in the study. The primary end point, SVR_35W24, was analyzed by comparing the SVR_35W24 rate in the intent-to-treat population using the stratified Cochran-Mantel-Haenszel test. All binary secondary end points were analyzed using the stratified Cochran-Mantel-Haenszel test. All continuous secondary end points of patient-reported outcomes were analyzed using a linear mixed effects model. All time-to-event secondary end points were analyzed using a stratified log-rank test. The hazard ratio (HR) for treatment effect was estimated using a stratified Cox proportional hazards model. Survivorship functions were estimated using the Kaplan-Meier method. The safety population consists of all patients who received ≥1 dose of NAV or PBO. Additional methods are detailed in the supplemental Methods.

The independent ethics committee or institutional review board at each study site approved the study protocol, informed consent forms, and recruitment materials before patient enrollment. The studies were conducted in accordance with the International Conference for Harmonisation guidelines, applicable regulations, and the Declaration of Helsinki. All patients provided written informed consent before screening.

Results

Baseline demographics and characteristics

For the TRANSFORM-1 study, 252 patients were enrolled, with a median follow-up of 20.3 (range, 0.0-35.3) months; patients were randomized 1:1 to receive NAV+RUX (n = 125) or PBO+RUX (n = 127; Figure 1). Demographics were generally similar between groups, with >80% of patients with intermediate-2 risk, and almost 50% with high-molecular risk (HMR) mutations (Table 1). Patients in both treatment arms had anemia, with median hemoglobin values of <10 g/dL.

Figure 1.

CONSORT diagram. ∗Due to the sponsor’s decision to discontinue the NAV clinical development program, all active patients have discontinued the study. Patients who were receiving NAV by end of December 2024, deriving clinical benefit per investigators’ assessment, and who intended to continue NAV beyond 30 December 2024, were transitioned to receive extended treatment. Patients who had discontinued the study drug before December 2024 completed the final survival assessment and discontinued the study by 29 January 2025. In the CONSORT diagram, AE is recorded as the primary reason for discontinuation of study treatment. In the AE data set (Table 2), the number of patients who discontinued due to AEs may be different, reflecting the fact that patients may have multiple AEs contributing to discontinuation and, in some cases, an AE may be recorded as a contributing reason for discontinuation even when another reason is selected as the primary reason for CONSORT reporting.

Table 1.

Baseline demographics and clinical characteristics

Characteristic	PBO+RUX (n = 127)	NAV+RUX (n = 125)
Age, median (range), y	69.0 (37-85)	70.0 (42-87)
Male	81 (63.8)	63 (50.4)
White	99 (78.0)	100 (80.0)
Calculated DIPSS+ risk
INT-1	5 (3.9)	8 (6.4)
INT-2	109 (85.8)	104 (83.2)
High	13 (10.2)	13 (10.4)
PLT count∗
≤200 × 109/L	34 (26.8)	31 (24.8)
>200 × 109/L	93 (73.2)	94 (75.2)
Hemoglobin, median (range), g/dL	9.8 (5.6-17.6)	9.9 (4.4-17.5)
Time from MF diagnosis, median (range), mo	6.2 (0.2-198.6)	7.8 (0.2-181.4)
Spleen volume, median (range), cm3	1639.0 (219.4-5663.7)	1441.4 (418.5-8020.4)
Type of MF
Primary MF	72 (56.7)	63 (50.4)
Post-PV MF	25 (19.7)	28 (22.4)
Post-ET MF	30 (23.6)	34 (27.2)
Driver genes†
JAK2 V617F mutation	86 (67.7)	84 (67.2)
CALR frameshift mutation	28 (22.1)	22 (17.6)
MPL W515 mutation	10 (7.9)	14 (11.2)
Negative for JAK2, CALR, and MPL	3 (2.4)	6 (4.8)
HMR	54 (42.5)	59 (47.2)
ASXL1	44 (34.7)	47 (37.6)
SRSF2	16 (12.6)	17 (13.6)
EZH2	11 (8.6)	5 (4.0)
IDH1/2	5 (3.9)	8 (6.4)
U2AF1	6 (4.7)	4 (3.2)
Previous lines of treatment, median (range)‡	1 (1-4)	1 (1-3)
Transfusion dependent§	4 (3.2)	5 (4.0)
TSS score, median (range)	23.7 (6.7-61.6)	20.6 (0.1-60.6)

Data are displayed as n (%) unless stated otherwise.

DIPSS+, Dynamic International Prognostic Scoring System Plus; ET, essential thrombocythemia; INT, intermediate; MPN, myeloproliferative neoplasm; PLT, platelet; PV, polycythemia vera.

^∗Platelet count of ≤150 × 10⁹/L at baseline: PBO+RUX (n = 19 [15.2%]) and NAV+RUX (n = 13 [10.5%]); platelet count of >150 × 10⁹/L at baseline: PBO+RUX (n = 106 [84.8%]) and NAV+RUX (n = 111 [89.5%]).

^†One patient in the NAV+RUX arm had 2 driver-gene mutations.

^‡Treatments given for supportive care or for preexisting MPN (PV or ET) before enrollment; patients did not have previous treatment for MF, with JAKis, BH3-mimetic compounds, or bromodomain and extraterminal motif inhibitors.

^§Transfusion dependency before study enrollment is defined as transfusions of ≥6 units of packed red blood cells, in the 12 weeks before study enrollment, for a hemoglobin level of <85 g/L, in the absence of bleeding or treatment-induced anemia.

Primary and secondary end points

The primary end point was met: 63.2% (n = 79) of patients in the NAV+RUX arm achieved SVR_35W24 compared with 31.5% (n = 40) of patients in the PBO+RUX arm (P < .0001; Figure 2A). The secondary end point was not met: the mean change from baseline in TSS at week 24 was −10.2 (baseline mean, 23.8; week 24 mean, 13.9) in the NAV+RUX arm vs −11.6 (baseline mean, 25.0; week 24 mean, 13.0) in the PBO+RUX arm (P = .2852; Figure 2B). In an additional prespecified exploratory analysis that used an anchor-based definition of a 10-point reduction in TSS as clinically meaningful, the median (proportion of patients who achieved this threshold was 37.6% (95% confidence interval [CI], 29.1-46.7) in the NAV+RUX arm vs 43.3% (95% CI, 34.6-52.4) in the PBO+RUX arm. Considering the significant difference between treatment arms in SVR_35W24, the observed improvement in symptoms was lower than anticipated for the NAV+RUX arm. A post hoc analysis was performed to evaluate potential factors influencing symptom outcomes, including gastrointestinal (GI)–related adverse events (AEs). This analysis revealed patients who received NAV+RUX who did not experience any GI-related AEs had a mean reduction in TSS from baseline at week 24 of −10.9 vs −8.4 for those with ≥1 GI-related AE (supplemental Figure 2). Trends remained similar at subsequent assessments (data not shown).

Figure 2.

SVR_35W24and change from baseline in TSS_W24. (A) Primary end point outcome of the proportion of patients who achieved SVR₃₅ at week 24. (B) Secondary end point outcome of mean change in TSS from baseline at week 24. TSS was calculated based on reporting on the Myelofibrosis Symptom Assessment Form, version 4. BL, baseline.

Additional secondary and exploratory end points

Because the secondary end point of change from baseline in TSS at week 24 was not met, formal statistical testing of other secondary end points was not performed. For these remaining end points, unadjusted P values are reported when available; when P values are not available, results are presented descriptively. SVR₃₅ at any time was achieved by 76.8% (n = 96) of patients in the NAV+RUX arm vs 44.1% (n = 56) of patients in the PBO+RUX arm (nominal P < .0001; Figure 3A). Across nearly all patient subgroups, the combination of NAV+RUX was favored compared with RUX alone in a prespecified exploratory analysis of SVR_35W24 (Figure 3B). Results from secondary and exploratory end point analyses evaluating the temporal dynamics of changes in spleen volume found more patients who received NAV+RUX vs PBO+RUX reached SVR₃₅ and ≥50% SVR from baseline as early as week 12 (Figure 4A); these observed reductions in splenomegaly were also sustained through week 48 in more patients treated with combination therapy compared with RUX alone (Figure 4B). Among patients with HMR, the median duration of SVR₃₅ was not estimable (NE; [95% CI, 27.6 to NE]) in the NAV+RUX arm and 19.4 (95% CI, 13.6 to NE) months in the PBO+RUX arm, indicating longer durability of spleen response for patients who received NAV+RUX (Figure 4C). The Cox proportional hazards model yielded a HR of 0.594 (95% CI, 0.246-1.437), suggesting a trend toward prolonged SVR₃₅ duration in the NAV+RUX arm compared with the PBO+RUX arm; however, the SVR₃₅ duration between arms was not different.

Figure 3.

SVR₃₅at any time and exploratory analysis of SVR_35W24. (A) Secondary end point outcome of the proportion of patients who achieved SVR₃₅ at any time during study treatment. (B) Prespecified exploratory analysis of SVR_35W24 across patient subgroups. DIPSS, Dynamic International Prognostic Scoring System.

Figure 4.

Additional analyses of change from baseline in SVR₃₅ and SVR₅₀, and duration of SVR₃₅. SVR₃₅ (secondary end point) and SVR₅₀ (exploratory end point) by week 12 (A) and week 48 (B). (C) Post hoc analysis of duration of SVR₃₅ among patients with HMR. ∗Unstratified HR. PH, proportional hazards; SVR₅₀, ≥50% SVR.

OS was evaluated to further characterize the overall clinical profile of NAV. Median OS was NE (95% CI, 41.4 to NE) in the NAV+RUX arm and 48.5 (95% CI, 48.5 to NE) months in the PBO+RUX arm (HR, 1.25; 95% CI, 0.75-2.08); the nominal log-rank P value was 0.3867, indicating no difference in OS between the 2 treatment arms. OS was further explored after stratifying by SVR₃₅ responder status in a post hoc analysis to contextualize whether survival outcomes are consistent with the differential spleen response observed between treatment arms. Independent of SVR₃₅ status at week 24, OS did not differ between the 2 treatment arms (supplemental Figure 3A-B).

Anemia response was also evaluated as a secondary end point outcome and was found to be similar between treatment arms. In the NAV+RUX arm (n = 66), 28.8% (95% CI, 18.3-41.3) of patients achieved an anemia response compared with 30.6% (95% CI, 20.2-42.5) in the PBO+RUX arm (n = 72). The response rate difference was −1.4% (95% CI, −16.6 to 13.7), with no difference between arms (nominal P = .8524). A small number of patients were transfusion dependent at baseline; among these patients treated with NAV+RUX (n = 5) and PBO+RUX (n = 4), 1 (20.0%) and 2 (50.0%) patients achieved an anemia response (ie, became transfusion independent), respectively. Hemoglobin response (≥1.5 g/dL increase from baseline without packed red blood cells within 2 weeks or erythropoietin within 4 weeks) was assessed in a post hoc analysis to quantify treatment-related hemoglobin improvement using a consistent definition for the study population, as anemia response per IWG differs by baseline transfusion status. The proportion of patients achieving hemoglobin response at any time was 36.8% in the NAV+RUX arm and 28.4% in the PBO+RUX arm (Figure 5A). NAV+RUX led to a reduction of BMF grade in 57% of patients at any time, compared with 49% of patients in the PBO+RUX group. The median time to first reduction in grade of BMF from baseline was 24.6 (range, 23.1-72.6) weeks for the NAV+RUX group vs 24.7 (range, 23.7-72.3) weeks for the PBO+RUX group.

Figure 5.

Analyses of hemoglobin response, cytopenia onset, and platelet values. (A) Post hoc analysis of the proportion of patients achieving hemoglobin response at any time. (B) Onset of neutropenia by starting dose, (C) onset of thrombocytopenia by starting dose, and (D) platelet values (×10⁹/L) over time. Hemoglobin responses defined as ≥1.5 g/dL increase in hemoglobin from baseline without receiving any red blood cell transfusion within 2 weeks or erythropoietin supplements within 4 weeks. BL, baseline; N. OBS, number observed; SE, standard error.

Fatigue associated with anemia and the ability of patients to perform routine physical activities and carry out daily tasks were also assessed as secondary end point outcomes. Change in fatigue from baseline to week 24 was assessed using the Patient-Reported Outcomes Measurement Information System Fatigue SF 7a. In the NAV+RUX arm (n = 104), the least square (LS) mean change from baseline was −3.75 (95% CI, −5.46 to −2.04), whereas in the PBO+RUX arm (n = 99) it was −3.63 (95% CI, −5.38 to −1.88). The LS mean difference between groups was −0.12 (95% CI, −2.03 to 1.80; nominal P = .9029), indicating similar reductions in fatigue symptoms and impact between patients who received RUX alone or in combination with NAV. Improvements in physical functioning from baseline to week 24 were also similar between treatment arms. The LS mean change from baseline was 7.326 (95% CI, 3.458-11.195) for the NAV+RUX arm (n = 99) and 6.198 (95% CI, 2.238-10.158) for the PBO+RUX arm (n = 95). The between-group LS mean difference was 1.129 (95% CI, −3.278 to −5.536; nominal P = .6152), indicating similar improvement between treatment arms.

As an exploratory end point, median PFS was assessed to evaluate whether NAV+RUX could delay disease progression in patients compared with PBO+RUX. Median PFS was 31.9 (95% CI, 25.5-35.7) months in the NAV+RUX arm and 30.2 (95% CI, 22.9-31.7) months in the PBO+RUX arm (HR, 0.755; 95% CI, 0.496-1.147); the nominal log-rank P value was 0.1822, indicating no difference in PFS between the 2 treatment arms (supplemental Figure 4A). In post hoc analyses, patients with HMR who received NAV+RUX had a lower risk of progression or death compared with patients who received PBO+RUX (HR, 0.450; 95% CI, 0.260-0.779; supplemental Figure 4B). Among patients who achieved SVR₃₅ at week 24, the risk of progression or death was similar between treatment arms (HR, 0.542; 95% CI, 0.284-1.034; supplemental Figure 4C).

To determine the potential disease-modifying activity of NAV+RUX, the proportion of patients achieving 20% and 50% VAF reductions in JAK2, CALR, or MPL were assessed as exploratory end points. At baseline, 70% of patients in the NAV+RUX arm and 63% of patients in the PBO+RUX arm had a high driver-gene burden (JAK2, CALR, or MPL VAF of ≥50%), indicating broadly comparable baseline molecular disease burden between treatment arms. At any time during or after treatment, a ≥20% reduction in VAF was achieved in 58.5% and 45.5% of patients receiving NAV+RUX or PBO+RUX, respectively, whereas a ≥50% reduction in VAF was achieved in 24.6% in the NAV+RUX arm and in 23.1% in the PBO+RUX arm (supplemental Figure 5A-B). Absolute median VAF and median change from baseline at week 12 and week 24 for patients with JAK2, CALR, or MPL driver-gene mutations are shown in supplemental Table 1. Although VAF reduction rates appeared generally similar in both treatment arms, the potential clinical relevance of molecular response was further evaluated in a post hoc analysis of PFS by VAF response category, independent of treatment assignment. Patients who achieved a ≥20% reduction in driver-gene VAF had a median PFS of 31.1 months (95% CI, 25.2-34.2) vs 26.1 months (95% CI, 20.1-31.9) for patients who did not achieve ≥20% VAF reduction (supplemental Figure 5C).

Safety outcomes

Any-grade AEs were reported by all patients in the NAV+RUX arm (100%) and almost all patients in the PBO+RUX arm (97.6%) (Tables 2 and 3), with 90.3% and 79.2% considered grade 3/4 events, respectively. The most common AEs in patients who received NAV+RUX (n = 124) vs patients who received PBO+RUX (n = 125) were thrombocytopenia (any grade, 91.9% vs 54.4%; grade ¾, 54.0% vs 19.2%), anemia (any grade, 67.7% vs 53.6%; grade 3/4, 49.2% vs 44.0%), neutropenia (any grade, 46.0% vs 12.8%; grade 3/4, 40.3% vs 8.8%), and diarrhea (any grade, 41.9% vs 16.8%; grade 3/4, 5.7% vs 0%). The most common serious treatment-emergent AEs (TEAEs) reported in the NAV+RUX arm vs the PBO+RUX arm were infections (16.1% vs 16.8%). As assessed by investigator, serious TEAEs with reasonable possibility of being related to NAV/PBO occurred in 14 (11.3%) and 9 (7.2%) patients in the NAV+RUX and PBO+RUX arms, respectively. The rates of cytopenia were higher in the NAV+RUX arm, and the initial onset of both thrombocytopenia and neutropenia occurred earlier in the NAV-containing arm (Figure 5B-C); platelet counts followed a similar trend over long-term follow-up.

Table 2.

Overview of safety

	PBO+RUX (n = 125)	NAV+RUX (n = 124)
Any AE, n (%)	122 (97.6)	124 (100)
Thrombocytopenia, n (%)	68 (54.4)	114 (91.9)
Anemia, n (%)	67 (53.6)	84 (67.7)
Neutropenia, n (%)	16 (12.8)	57 (46.0)
Diarrhea, n (%)	21 (16.8)	52 (41.9)
Serious TEAEs, n (%)	56 (44.8)	42 (33.8)
Infections	21 (16.8)	20 (16.1)
GI	8 (6.4)	3 (2.4)
CRS	1 (0.8)	0
Anemia	3 (2.4)	3 (2.4)
Neutropenia	0	1 (0.8)
Thrombocytopenia	1 (0.8)	2 (1.6)
TEAEs leading to NAV or PBO discontinuation, n (%)
AML	5 (4.0)	4 (3.2)
Progressive disease of MF	6 (4.8)	3 (2.4)
Other malignancies∗	2 (1.6)	6 (4.8)
Anemia	3 (2.4)	3 (2.4)
Thrombocytopenia	2 (1.6)	6 (4.8)
Neutropenia	0	2 (1.6)
Cardiac disorders	1 (0.8)	2 (1.6)
Infections	1 (0.8)	3 (2.4)
Serious TEAEs leading to NAV or PBO discontinuation, n (%)	11 (8.8)	14 (11.3)
TEAEs leading to RUX discontinuation, n (%)
AML	5 (4.0)	4 (3.2)
Progressive disease of MF	6 (4.8)	3 (2.4)
Other malignancies∗	2 (1.6)	6 (4.8)
Anemia	3 (2.4)	3 (2.4)
Thrombocytopenia	1 (0.8)	2 (1.6)
Neutropenia	0	1 (0.8)
Cardiac disorders	1 (0.8)	2 (1.6)
Infections	1 (0.8)	3 (2.4)
Serious TEAEs leading to RUX discontinuation, n (%)	11 (8.8)	14 (11.3)
TEAEs leading to NAV or PBO interruption, n (%)	54 (43.2)	97 (78.2)
Thrombocytopenia	17 (13.6)	66 (53.2)
TEAEs leading to RUX interruption, n (%)	50 (40.0)	85 (68.6)
Thrombocytopenia	19 (15.2)	61 (49.2)
TEAEs leading to NAV or PBO reduction, n (%)	47 (37.6)	104 (83.9)
Thrombocytopenia	39 (31.2)	88 (71.0)
TEAEs leading to RUX reduction, n (%)	85 (68.0)	112 (90.3)
Thrombocytopenia	66 (52.8)	102 (82.3)
Deaths, n (%)	28 (22.4)	35 (28.2)
Transformation to AML	4	5

Grouped PTs of “thrombocytopenia” or “platelet count decreased” and grouped PTs of “neutropenia” or “neutrophil count decreased” or “febrile neutropenia” for thrombocytopenia and neutropenia, respectively.

AML, acute myeloid leukemia; CRS, cytokine release syndrome; PTs, preferred terms.

^∗Includes diffuse large BCL stage III, lymphoproliferative disorder, non–small cell lung cancer, prostate cancer, squamous cell carcinoma or skin, transformation to AML, diffuse large BCL, and transitional cell carcinoma.

Table 3.

Any-grade and grade 3/4 TEAEs

	PBO+RUX (n = 125) Any grade∗	PBO+RUX (n = 125) Grade 3/4	NAV+RUX (n = 124) Any grade∗	NAV+RUX (n = 124) Grade 3/4
Any TEAE	122 (97.6)	99 (79.2)	124 (100)	112 (90.3)
Thrombocytopenia	68 (54.4)	24 (19.2)	114 (91.9)	67 (54.0)
Anemia	67 (53.6)	55 (44.0)	84 (67.7)	61 (49.2)
Diarrhea	21 (16.8)	0	52 (41.9)	7 (5.7)
Neutropenia	16 (12.8)	11 (8.8)	57 (46.0)	50 (40.3)
COVID-19	29 (23.2)	7 (5.6)	31 (25.0)	2 (1.6)
ALT increased	12 (9.6)	3 (2.4)	26 (21.0)	6 (4.8)
AST increased	16 (12.8)	2 (1.6)	24 (19.4)	3 (2.4)
Headache	12 (9.6)	1 (0.8)	21 (16.9)	0
Fatigue	18 (14.4)	1 (0.8)	21 (16.9)	2 (1.6)
Nausea	9 (7.2)	1 (0.8)	22 (17.7)	1 (0.8)
Dizziness	12 (9.6)	0	17 (13.7)	0
Contusion	9 (7.2)	0	14 (11.3)	0
Dyspnea	20 (16.0)	0	10 (8.1)	0
Pyrexia	14 (11.2)	0	12 (9.7)	0
Constipation	17 (13.6)	1 (0.8)	6 (4.8)	0

Data are reported as n (%). Grouped PTs of “thrombocytopenia” or “platelet count decreased” and grouped PTs of “neutropenia” or “neutrophil count decreased” or “febrile neutropenia” for thrombocytopenia and neutropenia, respectively.

ALT, alanine aminotransferase; AST, aspartate aminotransferase; PTs, preferred terms.

^∗Any grade that occurred in ≥10% of either group.

Although Kaplan-Meier analysis showed similar OS between the NAV+RUX and PBO+RUX arms for patients who did not achieve SVR₃₅ at week 24, a higher number of deaths occurred for patients treated with NAV+RUX vs PBO+RUX (35 vs 28). Of 35 patients who died in the NAV+RUX arm, 10 (8.1%) died ≤30 days after the last dose of NAV. The causes of death 30 days after last dose are unavailable, because investigators were not required to report the reasons.

The relative median dose intensity of RUX was 49.4% (range, 11.4%-106.1%) for NAV+RUX and 73.6% (range, 24.5%-125.1%) for PBO+RUX, which roughly corresponds to an average RUX dose of ∼10 mg twice daily and ∼15 mg twice daily, respectively. Acute myeloid leukemia and progressive disease of MF were the most common TEAEs for NAV/PBO or RUX discontinuation in either arm; additional TEAEs and serious TEAEs leading to NAV/PBO or RUX discontinuation are detailed in Table 2, as well as TEAEs leading to NAV/PBO and RUX interruption and reduction, respectively. The proportion of patients with ≥2 NAV/PBO dose reductions for any reason was 45.2% (n = 56) in the NAV+RUX arm and 11.2% (n = 14) in the PBO+RUX arm, with dose reductions and interruptions for NAV/PBO most commonly due to thrombocytopenia; none were due to bleeding (Table 2).

Discussion

In the phase 3, double-blind, PBO-controlled, multicenter TRANSFORM-1 study, the primary end point was met, with a significantly higher proportion of patients in the NAV+RUX arm achieving SVR_35W24 when compared with the PBO+RUX arm (63.2% vs 31.5%). Nearly all patient subgroups analyzed showed a trend in improved SVR₃₅ at any time for NAV+RUX vs RUX alone. In patients with HMR mutations, SVR₃₅ responses were durable but did not differ significantly between treatment arms, providing further evidence that there is an unmet need for efficacious treatments in patients with MF with HMR mutations. It is worth noting that spleen size reduction reportedly correlated with survival in a pooled analysis of patients treated with RUX.¹¹ Thus, the significantly higher proportion of patients in the NAV+RUX arm who achieved SVR_35W24 when compared with the PBO+RUX arm may be clinically meaningful.

Despite significant SVR with NAV+RUX treatment (compared with PBO+RUX), symptom improvement was comparable between the 2 treatment arms. Therefore, the secondary end point of TSS was not met, and no additional formal statistical testing was undertaken. Nonetheless, several considerations deserve attention. Although symptom improvement was comparable between the 2 arms, the dosage of RUX was lower in the NAV+RUX arm vs the PBO+RUX arm, because cytopenia required dose reduction of both NAV and RUX. It is also important to note that this study had a relatively low baseline TSS at enrollment, which may result in patients having a narrower range to detect treatment-related improvements. In addition, RUX, even as a monotherapy, is known to induce a high degree of symptom improvement.¹¹

Furthermore, in post hoc analyses, patients who received NAV+RUX and did not experience any GI-related AEs had a mean reduction in TSS from baseline at week 24 of −10.9, compared with −8.4 in those who experienced ≥1 GI-related AEs. Although this difference may suggest a modest effect of GI toxicity on symptom severity, patients in the PBO+RUX arm who experienced GI-related AEs had a mean TSS reduction of −14.0, achieving the meaningful change threshold (patients in the PBO+RUX arm without GI-related AEs had a mean TSS reduction of −10.8). These findings suggest GI toxicity did not uniformly affect clinically meaningful symptom improvement, given that some groups achieved the threshold despite GI events. Moreover, most GI AEs observed in this study were low grade and manageable; accordingly, it is anticipated that sustained spleen responses and resolution of GI events over the longer term may further improve symptom outcomes, even in patients in whom changes were not clearly detectable in this study.

Importantly, this trial highlights broader challenges within MF clinical research, notably the limitations of current JAKi therapies for disease modification and the inadequacy of SVR and TSS as end points for evaluating combination strategies.²⁸ In MF, SVR and symptom scores have traditionally served as surrogate measures of treatment benefit.¹⁷ However, their ability to fully capture the complexity of disease biology, especially in studies testing combination regimens or novel mechanism of action, is increasingly recognized as insufficient. For example, JAKi alone often produces substantial symptomatic relief, leading to a ceiling effect in which additional improvement may not be detectable, and symptom scores can be confounded by overlapping drug-related AEs. Furthermore, these end points do not necessarily reflect changes in disease course to underlying pathology. Therefore, although the TRANSFORM-1 trial did not meet the TSS end point (adding NAV to RUX also did not change symptom severity), its results provide valuable insight into the limitations of current evaluation frameworks and call for a critical reexamination of outcome measures in future MF studies.

Indeed, although several of the assessed efficacy measures favored NAV+RUX over PBO+RUX, they did not extend PFS or translate into improved OS in the overall population within the current follow-up period. Nonetheless, a notable reduction in risk of progression or death among patients with HMR receiving NAV+RUX was observed; and a ≥20% reduction in VAF was achieved in 58.5% of patients in the NAV+RUX arm and 45.5% in the PBO+RUX arm. Although the relationship between these markers (and others, including BMF) and survival outcomes is currently unknown,²⁸ these data suggest that some subgroups may derive particular benefit and that meaningful molecular responses could translate into improved survival given sufficient follow-up (recognizing that the 24-week time point used in this trial is relatively short for evaluating PFS and long-term clinical benefit in MF and that longer follow-up will be essential to fully assess the impact of study drug). Therefore, it remains possible that the benefits of VAF reduction could translate into a survival advantage with longer follow-up, and that molecular response could serve as a valuable end point in future studies.

The open-label, phase 2, multicenter, REFINE trial (ClinicalTrials.gov identifier: NCT03222609) evaluated the efficacy and safety of NAV alone or in combination with RUX in patients with primary or secondary MF. In cohort-1A, patients with suboptimal response to RUX who received NAV add-on therapy achieved SVR_35W24 in 27%, BMF improvement in 33% of evaluable patients, and anemia response in 64%.²⁶ In cohort 3 (which enrolled fewer patients than the randomized, double-blind, phase 3 TRANSFORM-1 trial) JAKi-naïve patients treated with NAV+RUX achieved SVR_35W24 regardless of poor prognostic features (eg, aged ≥75 years, high Dynamic International Prognostic Scoring System Plus score, and HMR mutations).²⁹ Similarly, the phase 3 MANIFEST-2 study demonstrated that pelabresib plus RUX resulted in significantly higher SVR_35W24 rates than PBO+RUX, whereas TSS improvement remained comparable between arms, again illustrating the challenges of interpreting symptoms scores as sole measures of benefit.³⁰ Overall, these trials highlight that the addition of novel agents to RUX can enhance certain outcomes, such as spleen response or disease modification markers, without adversely affecting symptoms. Combination therapies may allow patients to remain on treatment longer by retaining efficacy, even as current end points such as TSS may not fully reflect their clinical value. Again, highlighting that the incorporation of molecular and disease-modifying markers alongside traditional end points may advance MF drug development.

The safety profile of NAV+RUX was generally consistent with expectations for combination therapy in MF, although hematologic AEs, including thrombocytopenia, neutropenia, and anemia, were observed more frequently and with earlier onset compared with PBO+RUX. Of note, thrombocytopenia was the most common AE, affecting nearly all patients on NAV+RUX and presenting as grade 3/4 in more than half. Despite NAV+RUX leading to earlier onset of thrombocytopenia, platelet counts improved over time and eventually recovered to levels similar to the median counts observed in the PBO+RUX arm, reflecting reversibility and effective management through dose adjustments. Dose interruptions and reductions were more common with NAV+RUX, largely in response to thrombocytopenia, although importantly, none were due to bleeding.

GI AEs, such as diarrhea, were reported more frequently in the NAV+RUX arm; however, most episodes were low grade. The rate of serious GI AEs was also lower in the NAV+RUX arm (2.4%) than the PBO+RUX (6.4%) arm, although no formal statistical testing was performed. Regarding serious AEs, neutropenia was reported in 1 patient in the NAV+RUX arm and thrombocytopenia was reported in 1 patient in each arm.

Conclusions

In JAKi-naïve patients with MF, NAV+RUX led to a twofold improvement in SVR_35W24 vs PBO+RUX. The addition of NAV to RUX did not change symptom severity, with similar reductions in TSS observed between treatment arms. NAV+RUX showed higher hematologic toxicity vs PBO+RUX, but cytopenias were generally manageable and reversible with dose adjustments.

Conflict-of-interest disclosure: N.P. reports board of directors/management role with Dan’s House of Hope; consulting/scientific advisory board/speaking role with AbbVie, Aplastic Anemia & MDS International Foundation, Aptitude Health, Astellas Pharma US, Blueprint Medicines, Bristol Myers Squibb (BMS), CancerNet, CareDx, Celgene, Cimeio Therapeutics AG, ClearView Healthcare Partners, CTI BioPharma, Curio Science, Dava Oncology, EUSA Pharma, Harborside Press, IMedEX, Immunogen, Intellisphere, Magdalen Medical Publishing, Medscape, Menarini Group, Neopharm, Novartis Pharmaceuticals, OncLive, Pacylex, Patient Power, PeerView Institute for Medical Education, Pharma Essentia, Physician Education Resource; a leadership role with American Society of Hematology Committee on Communications and ASCO Cancer.Net editorial board; licenses in Karger Publishers; and uncompensated role with HemOnc Times/Oncology Times. A.J.M. has received honoraria for consulting and speaker fees from Novartis, Celgene/BMS, AbbVie, CTI BioPharma, MD Education, Sierra Oncology, Medialis, MorphoSys, Ionis, Medscape, Karyopharm, Sensyn, Incyte, Galecto, Pfizer, Relay Therapeutics, GSK, Alethiomics, Opna, and Gilead; has received research funding from Celgene/BMS, Novartis, Roche, Alethiomics, Opna, Agios, Synox, and Galecto; and is cofounder and equity holder in Alethiomics Ltd, a spin-out company from the University of Oxford. T.C.P.S. reports consulting/advisory roles for Novartis, BMS, GSK, AbbVie, and Rain Oncology; and research funding from Imago Biosciences, and CellCentric. F. Palandri reports acting as consultant for, and receiving honoraria from, Novartis, BMS Celgene, GSK, Amgen, MorphoSys, AbbVie, Karthos, AOP Pharma, Sierra Oncology, and CTI BioPharma. S.K. has received research grant/funding from Geron, Janssen, AOP Pharma, and Novartis; has received consulting fees from AbbVie, Pfizer, Incyte, Ariad, Novartis, AOP Pharma, BMS, Celgene, Geron, Janssen, CTI BioPharma, Roche, Bayer, Protagonist, and PharmaEssentia; has received payment or honoraria from AbbVie, Pfizer, Incyte, Ariad, Novartis, AOP Pharma, BMS, Celgene, Geron, Janssen, CTI BioPharma, Roche, Bayer, Protagonist, PharmaEssentia, AstraZeneca, iOMEDICO, and GSK; has received travel/accommodation support from Alexion, Novartis, BMS, Incyte, AOP Pharma, CTI BioPharma, Pfizer, Celgene, Janssen, Geron, Roche, AbbVie, GSK, Sierra Oncology, Protagonist, Karthos, iOMEDICO, and AstraZeneca; had a patent issued for a bromodomain and extraterminal motif protein inhibitor at Rheinisch-Westfälische Technische Hochschule Aachen University; and has participated on advisory boards for Pfizer, Incyte, Ariad, Novartis, AOP Pharma, BMS, Celgene, Geron, Janssen, CTI BioPharma, Roche, Bayer, GSK, Sierra Oncology, AbbVie, Protagonist, and PharmaEssentia. R.D. reports an advisory role for Novartis; and research funding (clinical) from Novartis, AbbVie, BMS, Pfizer, Lilly, AstraZeneca, Allogene, and BeiGene. J.-J.K. reports advisory roles for AbbVie, AOP Health, Novartis, BMS, GSK, and Incyte. C.B. reports consultant roles for AbbVie, Alexion, Genentech, Gilead, GSK, Karyopharm, Novartis, OncoVerity, PharmaEssentia, and Taiho. S.K.T. reports research support from Karyopharm Therapeutics; and consultancy for, and honoraria for advisory boards from, Karyopharm Therapeutics, Novartis, GSK, AbbVie, Genentech, and Sobi. S.-P.Y. reports acting as an investigator for AbbVie sponsored trials. M.T.G.C. reports consulting or advisory roles for Novartis, Pfizer, GSK, and AbbVie. A.M. reports honoraria for advisory boards and speaker fees from Novartis and AbbVie; and research funding from Novartis and AbbVie. S.-S.Y. reports grants from Roche-Genentech, Yuhan Pharma, and Kyowa Kirin; consulting or advisory roles for Amgen, Novartis, Janssen, Takeda, PharmaEssentia, and Handok Pharma; and data and safety monitoring board role with PharosiBio, Chugai Pharma, and CKD Pharma. K.K. reports honoraria from Novartis, Takeda, and PharmaEssentia. T.D. reports consultancy for, and honoraria for advisory boards from, GSK, Incyte, Novartis, and BMS. A.P. reports consultancy for advisory boards from Novartis; and speaker bureau support from Novartis, outside the submitted work. A.R. reports honoraria from Roche, Pfizer, AbbVie, Swixx, Sobi, Novartis, and Pfizer; consulting or advisory roles for Roche, Swixx, Sobi, Johnson & Johnson, and AstraZeneca; and travel, accommodations, and expenses (reimbursed by any for-profit health care company) from Roche. O.G.S. reports honoraria for advisory boards from AbbVie, Janssen, Astellas Pharma, AstraZeneca, BMS, and Pfizer. A.B. reports advising Novartis, Swedish Orphan Biovitrum, AbbVie (local in Serbia), Takeda (local in Serbia); and speakers bureau role with AstraZeneca and Novartis. R.M., J.G.H., B.M., Q.J., A.S.C., E.J.P., and J.P. are AbbVie employees and may own stock. F. Passamonti reports honoraria for lectures and advisory boards from Novartis, BMS/Celgene, Sierra Oncology, AbbVie, Janssen, Roche, AOP Orphan, Karyopharm, Kyowa Kirin, MEI, and Sumitomo. The remaining authors declare no competing financial interest.