Ten years after ruxolitinib approval for myelofibrosis: a review of clinical efficacy

Abstract

Myelofibrosis (MF) is a chronic myeloproliferative neoplasm characterized by splenomegaly, abnormal cytokine expression, cytopenias, and progressive bone marrow fibrosis. The disease often manifests with burdensome symptoms and is associated with reduced survival. Ruxolitinib, an oral Janus kinase (JAK) 1 and JAK2 inhibitor, was the first agent approved for MF. As a first-in-class targeted treatment, ruxolitinib approval transformed the MF treatment approach and remains standard of care. In addition, targeted inhibition of JAK1/JAK2 signaling, a key molecular pathway underlying MF pathogenesis, and the large volume of literature evaluating ruxolitinib, have led to a better understanding of the disease and improved management in general. Here we review ruxolitinib efficacy in patients with MF in the 10 years following approval, including demonstration of clinical benefit in the phase 3 COMFORT-I/II trials, real-world evidence, translational studies, and expanded access data. Lastly, future directions for MF treatment are discussed, including ruxolitinib-based combination therapies.

Keywords:

• Myelofibrosis
• ruxolitinib
• myeloproliferative neoplasm
• Janus kinase

Introduction

Myelofibrosis (MF) is a chronic myeloproliferative neoplasm (MPN) [1]. Common clinical manifestations include cytopenias, progressive bone marrow fibrosis, abnormal cytokine expression, and splenomegaly [2,3]. These manifestations combined lead to burdensome symptoms [4] and reduced patient survival [5,6]. The central component of MF pathogenesis involves constitutive activation of the Janus kinase (JAK) and signal transducer and activator of transcription (STAT) pathway, a key regulator of inflammatory cytokine signaling and hematopoiesis [7,8]. Somatic driver mutations in JAK2, CALR, or MPL activate JAK-STAT pathway signaling and perpetuate cytokine signaling in the absence of ligand binding [8,9]. Although allogeneic stem cell transplantation is a potentially curative option, many patients with MF are not eligible for this procedure due to their performance status and existing comorbidities [3,10].

The efficacy and safety of ruxolitinib, an oral JAK1 and JAK2 inhibitor, in MF were demonstrated in the randomized, placebo-controlled phase 3 COMFORT-I trial [11], and the COMFORT-II trial, in which patients were assigned to receive ruxolitinib or best available therapy that included, for example, hydroxyurea (HU) or glucocorticoids [12]. Based on data from these 2 pivotal trials, ruxolitinib was approved by the US Food and Drug Administration (FDA) in 2011 for treatment of adults with intermediate (int) or high-risk MF, including primary MF, post–polycythemia vera MF, and post–essential thrombocythemia MF. Subsequently, in 2014, ruxolitinib was approved for use in patients with polycythemia vera who had inadequate response to or were intolerant of HU, with approvals for use in steroid-refractory acute graft-versus-host disease (GVHD) and chronic GVHD after failure of systemic therapy following in 2019 [13,14].

Here we review the efficacy of treatment with ruxolitinib for patients with MF 10 years after FDA approval and provide a brief overview of future directions in MF, including a focus on the outlook for ruxolitinib-based combination therapy.

Efficacy of ruxolitinib

Spleen volume and MF symptoms

Ruxolitinib treatment efficacy in patients with MF was first demonstrated in the randomized, double-blind, placebo-controlled phase 3 COMFORT-I study [11]. The primary endpoint was the proportion of patients with ≥35% reduction in spleen volume from baseline at Week 24. Of 309 patients enrolled, 155 received ruxolitinib and 154 received placebo. At the Week 24 time point, 41.9% of patients in the ruxolitinib group achieved the primary endpoint compared with 0.7% in the placebo group (odds ratio [OR], 134.4; 95% CI, 18.0, 1004.9; p < 0.001). Additionally, 45.9% of patients in the ruxolitinib group had ≥50% reduction from baseline in total symptom score at Week 24 compared with 5.3% in the placebo group (OR, 15.3; 95% CI, 6.9, 33.7; p < 0.001). Most patients who received ruxolitinib in COMFORT-I had improvements in MF-related symptoms, whereas symptoms worsened in the majority of patients who received placebo.

In the 48-week, randomized COMFORT-II trial (ruxolitinib, n = 146; best available therapy, n = 73), 28% of patients with MF who received ruxolitinib achieved the primary endpoint of ≥35% reduction in spleen volume at Week 48 compared with 0% of patients who received best available therapy (p < 0.001) [12]. Corresponding percentages of those achieving spleen response at Week 24 were 32% and 0%, respectively (p < 0.001). MF-associated symptoms, including loss of appetite, dyspnea, fatigue, insomnia, and pain, were reduced at Week 48 in patients who received ruxolitinib and worsened in patients who received best available therapy. At primary data endpoints in both COMFORT-I and -II, discontinuation rates in the ruxolitinib groups (14 and 38%, respectively) were lower than those in the placebo (24%) and best available therapy (58%) groups. Although ruxolitinib discontinuation rates vary across subsequent studies conducted in a wide range of settings, real-world evidence is generally consistent with the COMFORT data [15].

Separate 5-year analyses of COMFORT-I and COMFORT-II showed that ruxolitinib was associated with long-lasting reductions in spleen size, with the median duration of ≥35% spleen volume reduction from baseline being ∼3.2 years with maintenance of long-term ruxolitinib therapy in both trials [16,17]. Additionally, reductions in spleen volume and length at Week 24 with ruxolitinib treatment were correlated with longer overall survival (OS) in a pooled analysis of the COMFORT trials [18]. However, in a post hoc analysis from the prospective phase 3b expanded-access JUMP study in patients with Dynamic International Prognostic Scoring System intermediate-2 or high risk, spleen response status at Week 24 was not correlated with OS [19]. Nonetheless, correlations between spleen volume reduction and OS in patients with MF who were treated with ruxolitinib have been observed in a real-world treatment setting [20], reinforcing the importance of the spleen volume endpoint.

Survival with ruxolitinib treatment

Results from the COMFORT studies also demonstrated prolonged survival with ruxolitinib treatment. In the 24-week COMFORT-I study, ruxolitinib treatment was associated with a significant improvement in survival compared with placebo after 4 additional months of follow-up (hazard ratio [HR], 0.50; 95% CI, 0.25, 0.98; p = 0.04) [11] and at a 5-year analysis (HR, 0.69; 95% CI, 0.50, 0.96; p = 0.025) despite permitted crossover from placebo to ruxolitinib treatment [16]. In the 48-week COMFORT-II study, OS was not significantly different between ruxolitinib and best available therapy after 2 additional months of follow-up (HR, 1.01; 95% CI, 0.32, 3.24); however, crossover of patients from best available therapy to ruxolitinib limited interpretation of survival data [12]. An important limitation of this analysis is that neither COMFORT study was powered to evaluate OS at the primary endpoint. In an attempt to address this limitation, a pooled analysis corrected for crossover from the control arms of COMFORT-I and COMFORT-II, and ruxolitinib was associated with prolonged survival (HR, 0.29; 95% CI, 0.13, 0.63; p = 0.01). The Kaplan-Meier–estimated OS at Week 144 was 78% for patients treated with ruxolitinib and 31% in the rank-preserving structural failure time (RPSFT)-adjusted control arm, corresponding to a 47% reduction in absolute risk with ruxolitinib compared with control when adjusting for treatment crossover [18]. In a separate pooled analysis of the COMFORT studies, there was a 30% reduction in risk of death among patients randomized to ruxolitinib compared with control treatment (HR, 0.70; 95% CI, 0.54, 0.91; p = 0.0065), even before correcting for crossover [21]. The survival benefit of ruxolitinib was further reinforced after correcting for crossover with the RPSFT model or censoring at crossover (Figure 1) [21].

Figure 1. Five-Year Overall Survival Data in the Pooled COMFORT-I and COMFORT-II studies. (A) Corrected for crossover with the RPSFT model. (B) Censored at crossover. HR: hazard ratio; NE: not evaluable; OS: overall survival; RPSFT: rank-preserving structural failure time. Reprinted with permission from Verstovsek et al. [21].

Consistent with clinical trial evidence, a real-world analysis of US Medicare patients with MF showed that patients treated with ruxolitinib had a significantly lower risk of mortality than those who were not exposed to ruxolitinib (HR, 0.61; 95% CI, 0.45, 0.83; p = 0.002) [22]. Similarly, a single-center study covering ruxolitinib treatment experience over a 20-year period (2000–2020) reported that patients with MF treated with ruxolitinib had significantly longer median OS than patients treated with other agents (ruxolitinib, 84 months; investigational agents, 45 months; other, 17 months; p < 0.001) [23]. In addition, results of a prospective real-world study using data from the ERNEST registry of patients with MF showed that median OS was significantly longer among patients treated with ruxolitinib compared with those treated with HU (6.7 vs 5.1 years; p < 0.001) [24].

Importantly, multiple pooled analyses of COMFORT data have demonstrated an OS benefit with ruxolitinib compared with control treatment regardless of anemia status at baseline, development of anemia during study treatment, or transfusion status at Week 24 [21,25,26]. Indeed, development of anemia during study treatment was not prognostic for OS [25,26]. In agreement with these findings in clinical trial settings, a real-world analysis demonstrated that hemoglobin decreases 6 months after ruxolitinib initiation were also shown to not affect OS [20].

The data described above are not without limitations. For example, both COMFORT trials enrolled relatively high-risk MF patients; thus, the OS effect of ruxolitinib in lower-risk MF remains unclear. In addition, a meta-analysis of OS data from COMFORT-I and -II found that ruxolitinib significantly improved OS at 51 weeks of follow-up, when compared with placebo, but did not detect a statistically significant difference compared with best available therapy [27]. Ultimately, a randomized controlled trial specifically designed and powered to evaluate OS improvement in MF would be required for definitive evidence. However, the preponderance of data from clinical trials and real-world analyses indicates that patients with MF treated with ruxolitinib have demonstrable improvements in OS compared with those treated with other therapies.

Disease modification and biomarkers of ruxolitinib treatment

There is no consensus on which criteria constitute disease modification during treatment of MF [28–30]. Disease burden in MF was traditionally assumed to correlate with histopathologic or molecular parameters such as bone marrow fibrosis and JAK2V617F allele burden [31]. However, this assumption has been challenged as recent evidence demonstrated that changes in fibrosis are not correlated with efficacy outcomes [32] and a correlation between clinical benefit and MF treatment-related changes in allele burden has not been established in the literature. Initially, this was taken as evidence of an absence of disease modification by ruxolitinib, but the definition of disease modification has evolved [30]. More recently, cytokine reduction has become considered a more representative marker of disease modification, as there is an established association with clinical benefit [31,33].

Results from the COMFORT studies demonstrated that ruxolitinib has at least partial disease-modifying activity in MF. Patients receiving ruxolitinib in COMFORT-I and -II had reductions in levels of plasma C-reactive protein and proinflammatory cytokines interleukin (IL)-6 and tumor necrosis factor (TNF)-α and increased levels of leptin and erythropoietin [11,12]. A reduction in proinflammatory cytokine levels was also reported in the phase 1/2 study of ruxolitinib in patients with MF; notably, the cytokine reduction was correlated with improvements in MF-related symptoms [33]. However, although ruxolitinib leads to a reduction in inflammatory cytokine levels, it does not achieve full normalization, suggesting that other aberrant signaling pathways such as nuclear factor kappa B (NF-κB) play a key role in MF-associated inflammation [34].

Ruxolitinib treatment also reduces JAK2V617F allele burden in some patients. In COMFORT-I, JAK2V617F allele burden mean reduction was 10.9% at Week 24 and 21.5% at Week 48 in the ruxolitinib group compared with 3.5% and 6.3%, respectively, in the placebo group, although few patients were evaluable at Week 48 (13 in the ruxolitinib group and 9 in the placebo group). In COMFORT-II, among evaluable patients, approximately one-third of JAK2V617F-positive patients had a > 20% reduction from baseline in absolute allele burden at Weeks 168 and 192, while 25.5 and 16.7% had any increase in allele burden at the same time points, respectively [17]. Among the few patients treated with ruxolitinib who had an increase in allele burden, highest absolute change from baseline during treatment was <10%. A reduction in JAK2V617F allele burden was also observed in the phase 1/2 study of ruxolitinib in patients with MF [33].

Finally, treatment with ruxolitinib has been shown to reverse or delay bone marrow fibrosis in some patients. Patients randomized to ruxolitinib had improved (15.8%), stable (32.2%), or worsening (18.5%) fibrosis at their last assessment in the 5-year follow-up of COMFORT-II (the remainder had missing data); for patients randomized to best available therapy, the corresponding proportions were 2.7, 17.8, and 5.5%, respectively [17]. In a long-term (up to 66 months) analysis from the phase 1/2 study of ruxolitinib in MF, patients treated with ruxolitinib had greater odds of bone marrow fibrosis improvement or stabilization compared with best available therapy [35]. These observations were extended in an analysis of patients treated with ruxolitinib in COMFORT-I, which showed that most patients had improvements or stabilization from baseline in bone marrow fibrosis [36].

One limitation of these results is the post hoc or pooled nature of the analyses.

Ruxolitinib versus hydroxyurea

National Comprehensive Cancer Network Guidelines (NCCN Guidelines®) recommend cytoreductive therapy with HU for some patients with lower-risk symptomatic MF but do not recommend HU for patients with higher-risk MF [10]. Although JAK inhibitors now form the cornerstone of treatment for patients with MF, HU is still commonly used in routine clinical practice as cytoreductive therapy. Two retrospective studies reported that HU and interferon were used as frequently as ruxolitinib to treat MF during the post-ruxolitinib approval time period, including HU and interferon use in the first-line setting and in patients with int-2 risk disease [37,38]. Clinical trial and real-world experience support ruxolitinib as a superior first-line treatment option for patients with higher-risk MF compared with HU. In COMFORT-II, ruxolitinib provided greater clinical benefit compared with best available therapy, among which HU was the most commonly used agent [12]. As mentioned above, results of a prospective real-world study using data from the ERNEST registry of patients with MF showed that ruxolitinib was associated with prolonged median survival versus HU (6.7 vs 5.1 years; p < 0.001). There was no difference in median OS regardless of whether ruxolitinib was used as first-line treatment (6.4 years) or after HU (7.8 years; p = 0.99) [24].

Timing of ruxolitinib treatment initiation

Evidence suggests that earlier treatment with ruxolitinib may provide additional clinical benefit versus delaying treatment. First, delaying ruxolitinib treatment can directly delay a clinical benefit. In the COMFORT-I study, the majority of int-2 and high-risk patients who initially received placebo had an increase in spleen volume from baseline at Week 48, whereas the majority of patients who received ruxolitinib had a decrease in spleen volume [11]. Patients treated with ruxolitinib also had a mean improvement in total symptom score at Week 24, compared with a mean worsening for patients treated with placebo. Similarly, in COMFORT-II, nearly all patients treated with ruxolitinib had decreases in spleen volume from baseline to Week 48, compared with approximately half of patients treated with best available therapy [12]. However, patients who crossed over from placebo or best available therapy to ruxolitinib because of worsening symptoms and spleen growth had a reduction from baseline in spleen volume that was maintained over the study [16,17]. Furthermore, OS among patients initially randomized to ruxolitinib in the COMFORT trials was greater than the crossover group in a 5-year pooled analysis, underlining the benefits of earlier intervention with ruxolitinib [21]. These results were bolstered by findings from a pooled post hoc analysis of COMFORT data, showing improved clinical outcomes for patients who initiated ruxolitinib earlier (≤12 vs >12 months from diagnosis), including fewer cytopenia events, better spleen response, and longer OS [39]. Finally, an Italian study of 408 patients with MF treated with ruxolitinib (NCT01493414 or treated off-study) showed that a > 2-year delay in ruxolitinib initiation was associated with a lower probability of spleen response [40]. The data supporting early treatment with ruxolitinib are limited by being from post hoc and retrospective analyses.

Patients classified with lower-risk (i.e. int-1) disease can also derive clinical benefit from treatment with ruxolitinib. The ROBUST trial (NCT01558739), an open-label, phase 2 study in the United Kingdom, enrolled patients with int-1, int-2, or high-risk disease (Table 1 [15,41-46]). At Week 48, a ≥ 50% reduction in spleen length was achieved by 50% of patients with int-1, 15% with int-2, and 48% with high-risk disease. A ≥ 50% reduction in MF Symptom Assessment Form Total Symptom Score (MF-SAF TSS) at Week 48 was achieved by 21% of patients overall (int-1, 21%; int-2, 23%; high-risk, 19%). Similarly, in the phase 3b JUMP study (NCT01493414), a ≥ 50% reduction from baseline in spleen length was achieved by 64% of patients with int-1 at Week 24 and by 61% of patients at Week 48 [43]. Responses on the Functional Assessment of Cancer Therapy–Lymphoma total score (FACT-Lym TS) were achieved by 30% to 40% of patients at each time point; responses on the Functional Assessment of Chronic Illness Therapy (FACIT)–Fatigue scale were achieved by 34% to 47% at each time point. These clinical benefits exemplify the advantage of starting treatment with ruxolitinib early, before onset of worse anemia levels typically seen in int-2 or high-risk patients at baseline, which are prognostic for worse OS [26].

Table 1. Key Ruxolitinib-based clinical studies following the COMFORT trials.

Trial name	Study participants, N	Phase (NCT#)	Objective	Main inclusion criteria	Main efficacy results	Safety overview
REALISE [42]	Patients with anemia N = 51	2 (NCT02966353)	To assess a novel dosing strategy for RUX in patients with anemia	Age ≥18 y Palpable spleen Hb <10 g/dL ECOG PS 0–2 Peripheral blood blasts <10%	Week 24 PSL50, 56% Week 24 MF-SAF TSS, median change from BL, −55%	Hematologic grade ≥3 AEs: Thrombocytopenia, 20% Anemia, 31% Hb levels remained stable through study RBC transfusion requirement: RBC-TD at BL, decreased RBC-TI at BL, stable
JUMP [15,43]	Expanded access, including patients with PLT <100 × 10^9/L N = 2233	3b (NCT01493414)	To assess safety and efficacy of RUX, including in patients with low PLT counts	Age ≥18 y PLT, ≥50 × 10^9/L IPSS risk criteria: high, int-2, int-1 Int-1-risk: PS ≥5 cm from costal margin	Overall population
					Week 24 PSL50, 57% Week 96 OS, 87%	Hematologic grade ≥3 AEs: Thrombocytopenia, 19% Anemia, 35%
					PLT <100 × 10^9/L
					Week 24 PSL50, 38% Week 24 FACT-Lym TS response rate, 40% Week 24 FACIT-Fatigue scale response rate, 45% Week 96 OS, 75%	Thrombocytopenia, 54% Anemia, 36%
					PLT ≥100 × 10^9/L
					Week 24 PSL50, 57% Week 24 FACT-Lym TS response rate, 54% Week 24 FACIT-Fatigue scale response rate, 49% Week 96 OS, 90%	Thrombocytopenia, 17% Anemia, 35%
ROBUST [41]	Int-1 or higher-risk MF N = 48 (Int-1-risk, n = 14; int-2-risk, n = 13, High-risk, n = 21)	2 (NCT01558739)	To evaluate safety and efficacy of RUX, including in patients with int-1 risk disease	Age ≥18 y PLT, ≥100 × 10^9/L IPSS risk criteria: high, int-2, int-1 Int-1-risk: PS ≥5 cm from costal margin Peripheral blood blasts <10% ECOG PS 0–2	Week 24 treatment success^b rate: Overall, 54% Int-1-risk, 57% Int-2-risk, 31% High-risk, 67% Week 48, MF-SAF TSS response rate: Overall, 21% Int-1-risk, 21% Int-2-risk, 23% High-risk, 19%	Hematologic grade ≥3 AEs: Thrombocytopenia, 13% Anemia, 21%
EXPAND [44]	PLT ≤100 × 10^9/L S1, N = 44 S2, N = 25	Ib (NCT01317875)	To establish RUX MSSD in patients with low PLT	Age ≥18 y PLT: S1, ≤100 × 10^9/L and ≥75 × 10^9/L S2, ≤75 × 10^9/L and ≥50 × 10^9/L IPSS risk criteria: high, int-2, int-1 PS ≥5 cm from costal margin ECOG PS 0–2	MSSD established as RUX 10 mg bid Week 48, spleen response rate, S1, 33% S2, 30% Week 24 TSS mean change from BL, S1, −7.7 S2, −3.9	Hematologic grade ≥3 AEs: S1 Thrombocytopenia, 35% Anemia, 20% S2 Thrombocytopenia, 78% Anemia, 17%
[45]	Rechallenge N = 13 (2nd rechallenge: n = 4)	Retrospective case series	To assess if RUX rechallenge results in durable responses	RUX re-treatment at least once 2012–2017	Primary treatment	-
					Constitutional symptom^a improvement, 92% Spleen size reduction, 85% (mean reduction, 10.9 cm)
					1st rechallenge
					Constitutional symptom improvement, 92% Spleen size reduction, 69% (mean reduction, 7.1 cm)
					2nd rechallenge
					Constitutional symptom improvement, 100% Spleen size reduction, 100%
[46]	Rechallenge N = 26 (OPEN, n = 12; HIRE, n = 14)	Retrospective case series	To describe RUX treatment patterns and clinical outcomes with RUX rechallenge after interruption	MF diagnosis 2012–2016 (OPEN) or 2013–2017 (HIRE)	Symptom improvement, 46% (OPEN), 43% (HIRE) Spleen size reduction, 40% (OPEN), 33% (HIRE)	-

AE: adverse event; bid: twice daily; BL: baseline; ECOG PS: Eastern Cooperative Oncology Group performance status; FACIT-Fatigue: Functional Assessment of Chronic Illness Therapy–Fatigue scale; FACT-Lym TS: Functional Assessment of Cancer Therapy–Lymphoma total score; Hb: hemoglobin; HIRE: HealthCore Integrated Research Environment claims and provider data repository; int: intermediate; IPSS: International Prognostic Scoring System; MF: myelofibrosis; MF-SAF TSS: MF-Symptom Assessment Form Total Symptom Score; MSSD: maximum safe starting dose; OPEN: Cardinal Health Oncology Provider Extended Network; OS: overall survival; PLT: platelet count; PS: palpable splenomegaly; PSL50; ≥50% reduction in palpable spleen length; RBC: red blood cell; RUX: ruxolitinib; S1: stratum 1; S2: stratum 2; TD: transfusion dependent; TI: transfusion-independent; TSS: Total Symptom Score.

^aUnexplained fever, night sweats, fatigue, and weight loss; ^b Defined as the proportion of patients achieving ≥50% reduction in spleen length and/or a ≥ 50% decrease in the MF-SAF TSS.

Peritransplant ruxolitinib

Current guidelines recommend continuing JAK inhibitor therapy including ruxolitinib near to the start of conditioning therapy prior to HCT for continued improvement of splenomegaly and MF-related symptoms [47]. The safety and efficacy of ruxolitinib use peritransplant was evaluated in a multicenter prospective study [48]. Most (>90%) patients who received a short course of ruxolitinib were able to proceed with their scheduled transplant, all had engraftment, and relapse rate was low. Only 3 patients had return of their MF symptoms after cessation of ruxolitinib. 25% of patients developed serious infections, and as discussed below for ruxolitinib, patients treated with ruxolitinib and those undergoing HCT should be monitored for infection.

To address the issue of appropriate timing of HCT in patients that are responding well to ruxolitinib treatment, a pilot, open-label study evaluated the safety of peritransplantation ruxolitinib treatment for MF. Patients were enrolled at 2 dose levels of ruxolitinib, 5 mg twice daily (bid) (n = 6) and 10 mg bid (n = 12), which was administered with a fludarabine and melphalan conditioning regimen and prophylactic tacrolimus and sirolimus for graft-versus-host disease. All 18 patients completed ruxolitinib treatment. Overall, most adverse events (AEs) were grade 1/2, and none of the serious adverse events (AEs) (19 events in 8 patients) were considered related to the study intervention. The addition of ruxolitinib to peritransplant regimens was considered safe and well tolerated, albeit with the caveat that this was based on data from an early phase trial.

Ruxolitinib dose optimization

The recommended starting dose of ruxolitinib is based on platelet count [13]. Nevertheless, it is important to ensure adequate ruxolitinib dosing to maximize clinical benefit, with dose titration for patients with confounding factors, including hepatic/renal impairment, anemia, low platelet counts, or coadministration with strong cytochrome P450 3A4 (CYP3A4) inhibitors (Table 2). The recommended starting dose is 20 mg bid for patients with platelet count >200 × 10⁹/L; 15 mg bid for patients with 100 to 200 × 10⁹/L; and 5 mg bid for patients with 50 to <100 × 10⁹/L [13] (in the European Union, 10 mg bid is recommended for platelet counts 75 to <100 × 10⁹/L). The open-label phase 1b EXPAND study (NCT01317875) later established that 10 mg bid was the maximum safe starting dose in patients with intermediate- or high-risk MF with platelet count 50 to 99 × 10⁹/L [44].

Table 2. Ruxolitinib dosing guidance based on baseline platelet counts, renal or hepatic insufficiency, coadministration with strong CYP3A4 inhibitors, or anemia.

Parameter	Recommended ruxolitinib starting dose
Prescribing information recommendations
PLT count^a: >200 × 10^9/L 100 to 200 × 10^9/L 50 to <100 × 10^9/L	20 mg bid 15 mg bid 5 mg bid
Moderate/severe renal or mild/moderate/severe hepatic impairment with PLT: >150 × 10^9/L 100 to 150 × 10^9/L 50 to <100 × 10^9/L	No dose adjustment 10 mg bid 5 mg qd
End-stage renal disease with PLT: >200 × 10^9/L 100 to 200 × 10^9/L	20 mg once after dialysis session 15 mg once after dialysis session
Coadministration with strong CYP3A4 inhibitors and PLT: ≥100 × 10^9/L 50 to <100 × 10^9/L	• 10 mg bid 5 mg qd
REALISE Study [42]
Anemia	10 mg bid for 12 wk regardless of baseline PLT count; patients with ANC >0.5 × 10^9/L and Hb ≥6.5 g/dL were eligible for dose increases after 12 wk 15 mg bid from wk 12 if PLT 100 to ≤200 × 10^9/L 20 mg bid from wk 16 if PLT ≥200 × 10^9/L 25 mg bid from wk 20 if PLT ≥200 × 10^9/L and patients did not achieve 50% reduction in palpable spleen length

ANC: absolute neutrophil count; bid: twice daily; CYP3A4: cytochrome P450 3A4; Hb: hemoglobin; PLT: platelet; qd: once daily.

^aDoses may be titrated in individual patients based on safety and efficacy outcomes.

Support for ruxolitinib dosing and titration in patients with lower platelet counts (50 to 100 × 10⁹/L) comes from 2 studies. In an open-label phase 2 study (NCT01348490), patients with MF and platelet counts 50 to 100 × 10⁹/L initiated ruxolitinib at 5 mg bid, with 48% (25/52) titrated up to ≥10 mg bid by Week 24 [49,50]. A ≥ 35% reduction in spleen volume was achieved by 17% of patients overall, including 73% who were titrated to 10 mg bid. In addition, 35% of patients overall achieved a ≥ 50% reduction in MF-SAF TSS by Week 24, including 44% who were titrated to 10 mg bid. In a second study, the phase 3 expanded-access JUMP study [43], initial ruxolitinib dosing was based on platelet counts (50 to <100 × 10⁹/L, 5 mg bid; 100 to 200 × 10⁹/L, 15 mg bid; >200 × 10⁹/L, 20 mg bid). Decreases in spleen length from baseline were sustained with long-term ruxolitinib treatment: at Week 96, ≥25% and ≥50% reductions from baseline in spleen length were achieved by 81% and 67% of all patients, respectively, including 65% and 29% of patients in the lowest platelet count cohort [43].

Unlike with platelet levels, the ruxolitinib prescribing information does not provide guidance on dosing based on anemia or hemoglobin levels [13]. To address this unmet need, the phase 2 REALISE study (NCT02966353) evaluated an alternative ruxolitinib dosing strategy in 51 anemic patients with MF, in which anemia was defined as hemoglobin levels <10 g/dL [42] (Table 2). Patients received ruxolitinib 10 mg bid for 12 weeks, after which those with higher platelet counts were eligible for dose increases up to 25 mg bid [42]. Overall, 56% of patients met the primary endpoint (≥50% reduction in spleen length by Week 24), including 6/9 (66.7%) with baseline transfusion dependence and 7/15 (47%) who received a dose increase at 12 weeks because of insufficient response [42]. Median (range) percentage change from baseline in best MF-SAF 2.0 score was −44 (−89 to 200) [42]. Anemia led to dose interruption or adjustment in 11.8% of patients, and 66.7% of patients received transfusions, although transfusion requirements decreased during the study for those who were transfusion-dependent at baseline. Median hemoglobin levels were stable throughout the study, and only 1 patient discontinued due to anemia.

The observational, longitudinal, real-world RUXOREL-MF study demonstrated the importance of dose optimization in maximizing ruxolitinib clinical benefit [20]. A model of OS 6 months after initiation of treatment with ruxolitinib demonstrated that shorter survival was associated with ruxolitinib doses ≤20 mg bid at baseline, Month 3, and Month 6; palpable spleen length reduction ≤30% from baseline at Months 3 and 6; requirement for red blood cell transfusions at Month 3 and/or 6; or requirement for red blood cell transfusions at all time points. Findings from the model also suggested that a suboptimal starting dose of ruxolitinib could prevent patients from reaching the recommended 20-mg bid dose, highlighting the importance of using the 20-mg bid dose from the beginning if possible [20]. In agreement with these findings, an analysis from COMFORT-I showed that clinically meaningful improvements in spleen volume and MF symptoms were most apparent at ruxolitinib doses ≥10 mg bid [51,52].

Ruxolitinib discontinuation

Ruxolitinib discontinuation may be appropriate in patients with MF because of treatment intolerance, a failure to respond or maintain a stable response (these patients may be candidates for combination therapy [see later section on this topic]), or disease progression to blast phase [53,54]. Conversely, it may be inappropriate to discontinue ruxolitinib due to other medical concerns such as SARS-CoV-2 infection or COVID-19 treatment [55,56]. Notably, an increased risk of death was reported among patients with MF who discontinued ruxolitinib due to COVID-19 infection, indicating that continuing treatment with ruxolitinib, if feasible, appears advisable in this situation [56]. Coadministration of the COVID-19 treatment nirmatrelvir/ritonavir and ruxolitinib has not been studied; however, ruxolitinib is metabolized by the enzyme CYP3A, which nirmatrelvir is a strong inhibitor of [57]. Based on evidence that coadministration of ruxolitinib with CYP3A inhibitors increases ruxolitinib exposure and prolonged the half-life, the ruxolitinib dose should be halved and administered twice daily for coadministration with nirmatrelvir/ritonavir [57]. Additionally, patients with thrombocytopenia or anemia can generally be effectively managed with dose interruptions/modifications or transfusions, as outlined in the section above, and thus should not necessarily discontinue treatment. Patients with baseline predictors of ruxolitinib discontinuation (int-2/high-risk disease, platelet count <100 × 10⁹/L, transfusion dependency, and unfavorable karyotype) [58] may warrant closer attention for potential dose interruptions or modifications and need for supportive treatment.

Ruxolitinib discontinuation and MF symptom rebound

In rare cases, discontinuation of ruxolitinib has been associated with severe, life-threatening AEs attributed to MF symptom rebound and cytokine storm. In a multivariable analysis, platelet count <100 × 10⁹/L and splenomegaly at ruxolitinib discontinuation were identified as independent risk factors. Probability of ruxolitinib re-initiation was significantly higher among patients who did versus did not develop MF symptom rebound after ruxolitinib discontinuation. Per the prescribing information, ruxolitinib should be discontinued gradually, with tapering. MF symptom rebound after ruxolitinib discontinuation may become rarer in real-world practice with the availability of 2 other approved JAK inhibitors, fedratinib [59,60] and pacritinib [61,62], which provide an opportunity to switch between JAK inhibitors upon treatment failure or intolerance [63,64]. Both fedratinib and pacritinib inhibit JAK2 with limited effects on JAK1, which plays a key role in inflammatory cytokine signaling and may mitigate the effects of JAK inhibitor discontinuation [62,65–69].

Ruxolitinib re-treatment

Re-treatment with ruxolitinib has been associated with clinical benefit [43,45,46]. A post hoc analysis of JUMP data from 207 patients who had ≥1 dose interruption of ≥7 days showed that 68.2% of patients achieved a ≥ 50% reduction in palpable spleen length at any time [43]. Of these patients, 58% achieved this spleen length response before treatment interruption, and 42% achieved this after treatment interruption. Compared with spleen length at the time of treatment restarting, 24% of patients achieved a further 50% reduction, 67% remained stable, and 9% had a ≥ 50% increase in spleen length. In a case series of 13 patients who underwent ruxolitinib rechallenge after an inadequate response or a loss of response, 69% had a mean reduction in spleen size of 7.1 cm, and 92% had constitutional symptom improvement (unexplained fever, night sweats, fatigue, and weight loss). In addition, a retrospective chart review from 2 US national healthcare databases evaluated patients who restarted ruxolitinib treatment after discontinuation (primarily because of AEs) [46]. Reasons for restarting ruxolitinib included lack of alternative treatment, resolution of AEs, worsening symptoms, and increased spleen size. After ruxolitinib was restarted, patients experienced clinical benefit as evidenced by improvement in symptoms (44% of evaluable patients) and reduction in spleen size (37%).

Side effects

Hematologic AEs, primarily anemia and thrombocytopenia, are the most common AEs observed with ruxolitinib, occurrence of which is often early in treatment, with hemoglobin and platelet levels reaching a nadir within 8–12 weeks. After the nadir, hemoglobin levels rebound toward baseline and platelet levels reach a new steady state [11,12,16]. As discussed in the sections above, both anemia and thrombocytopenia can generally be managed with dose modifications or interruptions, are usually not cause for discontinuation, and are not prognostic for clinical benefit with ruxolitinib [11,12].

Rates of major adverse cardiovascular events are low with oral ruxolitinib in MF, and although infection and secondary malignancies have been reported, no clear association has been established [15,16]. Overall, a meta-analysis of clinical trial data from 6 randomized controlled studies did not find an increased risk of infection with ruxolitinib treatment in patients with MPNs at either the early treatment stage (OR [95% CI]: 1.23 [0.91, 1.67]) or treatment extension (OR [95% CI]: 0.53 [0.36, 0.79]) [70]. However, herpes zoster infections have been reported with ruxolitinib, and the meta-analysis did detect an increased risk during both early treatment (OR [95% CI]: 7.39 [1.33, 41.07]) and treatment extension (OR [95% CI]: 5.23 [1.46, 18.79]). Non-live, subunit herpes zoster vaccine should be considered for patients receiving ruxolitinib. It is recommended that patients are assessed for infection, treated appropriately and promptly, and that ruxolitinib therapy be avoided in patients with active, serious infections until infections have resolved [13]. Regarding lymphomas, a retrospective chart review of European patients with MPNs who had received JAK2 inhibitors (ruxolitinib, gandotinib, fedratinib, or momelotinib) raised the possibility of an association between JAK inhibitor treatment of MF and an elevated frequency of aggressive B-cell lymphomas [71]. However, this was refuted in a larger American retrospective review of 2583 patients with MPNs, which did not find a significant difference in the incidence of lymphoma when comparing patients who received versus did not receive JAK inhibitor therapy. Moreover, there is a markedly higher risk for post-MPN lymphoid malignancies among patients with MF overall compared with the general population [72].

Although cases of nonmelanoma skin cancer (NMSC) have been reported with ruxolitinib, whether ruxolitinib increases risk for NMSC remains a topic of debate and active research. The highest-quality evidence comes from COMFORT-I and II. The 5-year analysis of COMFORT-I showed that the exposure-adjusted rate of NMSC was similar between ruxolitinib and placebo, and in COMFORT-II, the rates were 6.1 and 3.0/100 patient-years for ruxolitinib and best available therapy, respectively [16,17]. Taken together with retrospective analyses [73–75], these data support consideration of skin cancer monitoring in patients taking JAK inhibitors.

Overall, nonhematologic AEs with ruxolitinib in clinical trials (COMFORT, JUMP, ROBUST, REALISE, and EXPAND) were predominantly grade 1/2 [11,15,41,42,44], with real-world experience generally aligning with this safety profile [76,77].

Other JAK inhibitors

Additional JAK inhibitors have been approved or are in late-stage clinical development for treatment of MF (Table 3 [77-89]). Fedratinib was approved in 2019 for treatment of adults with int-2 or high-risk primary or secondary MF [59]. In 2022, pacritinib was approved for patients with intermediate- or high-risk primary or secondary MF with platelet counts <50 × 10⁹/L [10]. In addition to inhibiting JAK kinases, pacritinib has a potential additive anti-inflammatory effect through inhibition of IL-1 receptor-associated kinase 1 (IRAK-1)– and IL-1–related signaling [90]. Momelotinib, which is in development for treatment of MF, is an inhibitor of JAK1 and JAK2 as well as cytokine-driven Activin A Receptor Type 1 (ACVR1; i.e. activin receptor-like kinase-2 [ALK2]) signaling to suppress hepcidin expression, leading to increased iron availability for erythropoiesis that has the potential to improve anemia [80,91]. Data from a phase 2 study in transfusion-dependent patients with MF treated with momelotinib reported that 41% of patients had a transfusion-independent response lasting ≥12 weeks [92]. Circulating hepcidin levels decreased acutely after each dose, and there was a numerical decrease over the 24-week dosing period. At baseline and throughout the duration of the study, hepcidin levels were lower in patients who achieved red blood cell transfusion independence.

Table 3. Mechanism of action and pharmacokinetics of JAK inhibitors for MF.

Agent	Target	Pharmacokinetics	Pharmacodynamics	Week 24 efficacy
Ruxolitinib	JAK1/JAK2 [78]	Half-life: 3 h [79] Cmax: 649–4780 nM^a AUC: 2610–16,600 nM·h^a	Maximum pSTAT3 inhibition, 72–95%^a	COMFORT-I [11] SVR35, 42% TSS50, 46% COMFORT-II [12] SVR35, 32%
Momelotinib	JAK1/JAK2 ALK2 [80]	Half-life: 3.7 h^b [81] Cmax: 768 nM^b AUC: 4049 ng·h/mL^b	—	SIMPLIFY-1 [82] SVR35 MMB, 27% RUX, 29% TSS50 MMB, 28% RUX, 42% SIMPLIFY-2 [83] SVR35 MMB, 7% TSS50 MMB, 26% MOMENTUM [39] SVR35 MMB, 23% TSS50 MMB, 25%
Fedratinib	JAK2, FLT3 [66]	Half-life: 14.3–62.1 h^c [68] Cmax: 5.0–1760 nM^c AUC: 70.0–29,300 ng·h/mL^c	Mean pSTAT3 inhibition, −24% to +58%^d	JAKARTA [84] SVR35, 36% (400 mg) and 40% (500 mg) TSS50, 36% (400 mg) and 34% (500 mg)
Pacritinib	JAK2, IRAK1, CSF1R, FLT3, ALK2 [69,85,86]	Half-life: 1–4 days [87] Cmax: 6.9–7.9 nM^e AUC: 133–177 μg·h/mL^e	Maximum pSTAT3 inhibition, 32%–53%^f	PERSIST-1 [61] SVR35, 19% TSS50, 19% PERSIST-2 [88] SVR35, 18% TSS50, 25% PAC203 [89] SVR35, 9% (200 mg bid), 2% (100 mg bid), and 0% (100 mg qd) TSS50, 7% (200 mg bid), 7% (100 mg bid), and 8% (100 mg qd)

ALK2: activin receptor-like kinase-2; AUC: area under the curve; bid: twice per day; C_max: maximum plasma concentration; FLT3: Fms-like tyrosine kinase 3; JAK: Janus kinase; MMB: momelotinib; qd: once per day; RUX: ruxolitinib; SVR35: ≥35% spleen volume reduction from baseline; TSS50: ≥50% reduction in Total Symptom Score from baseline.

^aSteady-state parameters in healthy participants from a dose range 15 mg every 12 h to 100 mg every 24 h.

^bSingle 200 mg dose in healthy participants.

^cAscending single dose range 10–680 mg in healthy participants.

^dAscending single dose range 10–680 mg in healthy participants; PD assessments at 3 h post dose, mean pSTAT3 inhibition calculated as 100 – mean percentage of pSTAT3 remaining.

^eDay 15 assessments for dose range 100–300 mg in patients with lymphoma.

^fDay 1 assessments for dose range 100–600 mg in patients with relapsed lymphoma.

Although clinical benefit has been observed with fedratinib, pacritinib, and momelotinib, longer follow-up times will be required to assess effects on OS. Until such data are available, no conclusions can be drawn regarding OS, and the OS benefit observed with ruxolitinib should not be extended to these other JAK inhibitors with varying mechanisms of action.

A network meta-analysis of efficacy endpoints for ruxolitinib, fedratinib, pacritinib, and momelotinib indicated that the choice of JAK inhibitor may depend on use as first- or second-line therapy and the risk of developing anemia or thrombocytopenia [93]. In the first-line setting, patients treated with ruxolitinib, momelotinib, and fedratinib had similar spleen volume reductions, with pacritinib treatment being less efficacious in the first-line setting than ruxolitinib. No significant differences were observed between ruxolitinib and fedratinib regarding symptom improvements. Symptom improvement was not evaluated for the other 2 agents, and the meta-analysis did not include OS as an endpoint. The occurrence of anemia and/or thrombocytopenia was significantly different according to the JAK inhibitor used. Anemia occurred less frequently in patients treated with momelotinib than in patients treated with ruxolitinib, fedratinib, or pacritinib. In contrast, there were fewer occurrences of thrombocytopenia with fedratinib compared with ruxolitinib, momelotinib, or pacritinib.

Profiling levels of 148 biomarkers using a panel of 12 human cell culture systems revealed that ruxolitinib, fedratinib, momelotinib, and pacritinib each had a distinct biomarker profile [94]. Ruxolitinib was the most broadly active, demonstrating antiproliferative effects on B cells and T cells, down-regulation of inflammation-related signaling, and immunomodulatory activities. The distinct phenotypic outcomes associated with each JAK inhibitor reflect differential effects on JAK2, other JAKs, and kinases of other classes and may inform further exploration of clinical outcomes.

Overall historical perspective

As discussed in the sections above, ruxolitinib has demonstrated efficacy in the treatment of MF including improvements in spleen volume, symptom response, and OS. Being the first approved agent and first widely used targeted agent for MF, the introduction of ruxolitinib has been transformative. Since its approval, ruxolitinib has become the cornerstone of MF treatment and the standard of care [95].

Determining the true impact of ruxolitinib on real-world outcomes among patients with MF is challenging, given general changes in MF diagnostic criteria and practice during this time frame, but several studies have made the attempt. In a phase 1/2 study, long-term outcomes of 107 patients with MF who were treated with ruxolitinib were analyzed, and survival relative to 310 matched historical control patients with MF was assessed [96]. Compared with historical controls, OS was significantly better among patients treated with ruxolitinib (HR = 0.58, p = 0.005). Underlying this difference was a highly significantly longer OS among high-risk patients treated with ruxolitinib compared with controls (HR = 0.50, p = 0.006). A caveat of this analysis is that it compared patients managed in a clinical trial setting with historical controls managed in real-world clinical practice. A separate study aimed to avoid this caveat in a real-world retrospective analysis of US patients with intermediate- or high-risk MF divided into 3 groups based on ruxolitinib approval status at diagnosis and ruxolitinib exposure: preapproval (index year 2010–2011; all ruxolitinib-unexposed, n = 278); post-approval (index year 2012–2017; n = 1399; ruxolitinib-unexposed, n = 1127; ruxolitinib-exposed, n = 272) [22]. Compared with the preapproval cohort, patients exposed to ruxolitinib had significantly prolonged survival (HR, 0.36; p < 0.001). Not only was OS longer among ruxolitinib-exposed patients compared with the historical cohort but also compared with the contemporary, ruxolitinib-unexposed cohort that also benefited from contemporary improvements in patient management (HR, 0.61; p = 0.002). Similar findings were reported by a separate single-center, retrospective analysis [23]. Historical comparisons on spleen size and symptom improvement are not available, but based on results from COMFORT-II, it is clear ruxolitinib provides superior benefit than the best available therapies available at the time [12].

In addition to establishing higher expectations for clinical outcomes, the development of ruxolitinib also elevated the importance of assessing and managing MF-related symptoms and provided the tools to do so. The MF-SAF was created to properly evaluate ruxolitinib effects and is now widely available as a validated disease-specific tool [97,98]. Assessing patient-reported outcomes (PROs) with the MF-SAF and other PRO tools is expected not only in clinical trials but also in real-life patient management settings [31,99,100].

In addition to establishing PROs as clinical trial endpoints, results from the COMFORT trials challenged previous assumptions on MF pathogenesis [31]. Heralding in the JAK inhibition era of MF therapy, the success of ruxolitinib prompted increased interest in better understanding MF pathophysiology to improve patient management [95,96]. As a result, additional research has led to the development of novel therapies, for use either alone or in combination with JAK inhibitors.

Future directions: ruxolitinib-based combination therapy

There remains an unmet need for patients with MF following ruxolitinib discontinuation. A retrospective review of 107 patients in the INCB18424 phase 1/2 study showed that 36% of patients had acquired ≥1 additional gene mutation at the time of ruxolitinib discontinuation compared with the time of ruxolitinib initiation [101]. Patients with clonal evolution (i.e. those that had acquired mutations) had significantly shorter survival after discontinuation compared with those without clonal evolution. Several additional retrospective studies have also reported that patients with MF have increased morbidity and poorer overall outcomes after discontinuing ruxolitinib treatment [54,58]. Although salvage therapies were shown to provide some benefit in these analyses, overall the results underlined the need for additional treatment strategies in patients who discontinue ruxolitinib [54,58].

Despite the development of myriad new therapies and increased focus on combination treatment rather than monotherapy, inhibition of the JAK-STAT pathway remains a cornerstone of the MF treatment paradigm [28]. Exploratory and phase 2 clinical trials have evaluated ruxolitinib in combination with various agents, including pelabresib, parsaclisib, azacitidine, navitoclax, selinexor, danazol, luspatercept, and thalidomide (Table 4 [102-112]) [113]. Many of these therapeutic approaches are also being investigated in ongoing phase 3 trials, including combinations of ruxolitinib with pelabresib (MANIFEST-2 [NCT04603495]), parsaclisib (LIMBER-313 [NCT04551066], LIMBER-304 [NCT04551053]), navitoclax (TRANSFORM-1 [NCT04472598], TRANSFORM-2 [NCT04468984]), and luspatercept (INDEPENDENCE [NCT04717414]). These ongoing phase 3 trials have the potential to deliver illuminating data that could expand available treatment options for patients with MF. The importance of optimizing treatment strategies that combine ruxolitinib with other agents, determining how best to use these new therapeutic modalities, and developing new endpoints to assess their efficacy will likely increase as the MF treatment landscape continues to evolve [28,114].

Table 4. Key clinical trials of ruxolitinib-based combination therapies.

Agent (class) Study participants, N	Phase (NCT#)	Main inclusion criteria	Main efficacy results	Safety overview
RUX + pelabresib [103] (BET inhibitor) JAKi-experienced: N = 86 JAKi-naive: N = 84	2 (NCT02158858; MANIFEST)	Age ≥18 y Peripheral blood blasts <10% ECOG PS ≤2 DIPSS ≥ int-2 ANC ≥1 × 10^9/L JAKi-experienced: PLT, ≥75 × 10^9/L JAKi-naive: • PLT, ≥100 × 10^9/L	JAKi-experienced
			Week 24 SVR35, 20%^a Week 24 TSS50, 37% Conversion from TD to TI, 16%	Hematologic grade ≥3 AEs: Thrombocytopenia, 33% Anemia, 19%
			JAKi-naive
			Week 24 SVR35, 68% Week 24 TSS50, 56%	Hematologic grade ≥3 AEs: Thrombocytopenia, 12% Anemia, 35%
RUX + parsaclisib [104] (PI3Kδ inhibitor) N = 74	2 (NCT02718300)	Age ≥18 y Peripheral blood blasts ≤10% ECOG PS ≤2 Suboptimal response to RUX ANC ≥0.5 × 10^9/L PLT, ≥50 × 10^9/L	Week 12 SVR35, 0% (qd/qw) and 6% (qd) Week 12 spleen volume reduction ≥10%, 33% (qd/qw) and 59% (qd)	Dose interruptions due to AEs: Parsaclisib, 18/33 and 10/20 patients in qd/qw and qd RUX, 4/33 and 4/20
RUX + azacitidine [105] (hypomethylating agent) N = 46	2 (NCT01787487)	Age ≥18 y DIPSS int-1 or -2 ECOG PS ≤2 ANC ≥1 × 10^9/L PLT, ≥50 × 10^9/L	ORR, 33/46 (72%) (IWG-MRT criteria) 7/33 responders (21%) achieved response after AZA addition Patients with BL splenomegaly >5 cm BCM Week 24 > 50% reduction in spleen length, 62% (21/34) Improvement in BM at Month 24 (n = 14): Reticulin fibrosis, 8 (57%) Collagen, 7 (50%) Osteosclerosis, 4 (29%)	Hematologic grade ≥3 AEs: Anemia (46%) Thrombocytopenia (30%) Neutropenia (14%) 8% discontinued because of drug-related toxicity (neutropenia, n = 2; anemia, n = 1; thrombocytopenia, n = 1)
RUX + navitoclax [106] (BCL inhibitor) N = 34 (combination treatment cohort)	2 (NCT03222609; REFINE)	Age ≥18 y Peripheral blood blasts ≤10% ECOG PS ≤2 DIPSS: int or high-risk MF Suboptimal response to RUX PLT, ≥100 × 10^9/L	Week 24 SVR35, 26.5% Week 24 TSS50, 30% BMF improved by 1–2 grades in 33% OS not reached at median follow-up of 21.6 mo	Hematologic grade ≥3 AEs: Thrombocytopenia, 56% Anemia, 32% Discontinuations due to AEs: Navitoclax, 5 patients (15%) RUX, 2 (6%) patients
RUX + danazol [107] (androgen) N = 14	2 (NCT01732445)	Age ≥18 y ECOG PS ≤2 Anemia (Hb <10 g/dL) ANC ≥1 × 10^9/L PLT, ≥50 × 10^9/L	IWG-MRT clinical improvement, 21.4% ORR, 7.1% 4/5 (80%) JAKi-naive patients had stable or increasing Hb Of 9 patients who had received JAKi, 5 (56%) had stable or increasing Hb and 8 (89%) had stable or increasing PLT	Hematologic grade ≥3 AEs: 71% (n = 10) Nonhematologic grade ≥3 AEs: 14% (n = 2)
RUX + luspatercept [108,109] (TGFβ superfamily receptor ligand trap) N = 79	2 (NCT03194542)	Age ≥18 y ECOG PS ≤2 Anemia ANC ≥1 × 10^9/L PLT, ≥50 × 10^9/L	Mean Hb increase ≥1.5 g/dL from BL: NTD, no RUX = 3/20 (15%); NTD + RUX = 8/14 (57%) RBC-TI ≥12 weeks during study: TD no RUX = 4/21(19%); TD + RUX = 8/22 (36%) ≥50% reduction in RBC transfusion burden: TD, no RUX = 8/21 (38%), TD + RUX = 10/22 (46%)	TRAEs in ≥5% of patients: Hypertension, 13% Bone pain, 9% Diarrhea, 5% 10% discontinued because of drug-related toxicity
RUX + sotatercept [109] (TGFβ superfamily receptor ligand trap) Combination cohort: n = 9 (sotatercept monotherapy: n = 24)	2 (NCT01712308)	Age ≥18 y Hb <10 g/dL Sporadic RBC transfusions, or TD	ORR = TI + Hb increase ≥1.5 g/dL from BL for ≥12 consecutive weeks without RBC transfusion 6/17 (35%) in sotatercept monotherapy cohort	TRAEs: Grade 2 bilateral lower limb pain, n = 2 (1 patient in each cohort) Hypertension, n = 1
		RUX combination cohort
		≥6 months RUX with stable dose for ≥2 months	Response in 1/8 (12.5%)
RUX + thalidomide [110] (immunomodulatory agent) N = 23 (n = 15 evaluated)	2 (NCT03069326)	Age ≥18 y ECOG PS ≤2 ANC ≥1 × 10^9/L PLT, ≥50 × 10^9/L Suboptimal response, or refractory to RUX single-agent therapy RUX treatment for ≥3 months, and stable dose for ≥4 weeks before enrollment	Cycle 3: Significant increase from BL in PLT Significant increase in PLT in patients with thrombocytopenia at BL Trend toward increase in Hb over time Spleen volume reduction in 6/11 (55%) ORR, 9/15 (60%) ClinImp, 7/15 (46.6%) PLT response, 6/8 (75%) with thrombocytopenia at BL	Nonhematologic grade ≥3 AEs: Limb edema, diverticulitis, hypertension, syncope (n = 1 each) Thromboembolic event and grade 3 neutropenia, n = 1
RUX + prednisone, thalidomide, and danazol [111] N = 72 (n = 53 in combination therapy group)	Pilot study (ChiCTR1900025219)	Age ≥18 y SV ≥450 cm^3 Peripheral blood blasts <10% ECOG PS ≤2 DIPSS: int-1, int-2, or high-risk ANC ≥1 × 10^9/L PLT, ≥50 × 10^9/L	Combo therapy group: Spleen reductions ≥50%, 44.9% Symptom response, 41.7%	20.8% had new or worsening anemia (grade ≥3, 13%) 18.9% had new or worsening thrombocytopenia (grade ≥3, 1.8%)
RUX + selinexor [112] (Nuclear export inhibitor) N = 10 (n = 3 in selinexor 40 mg qw + RUX group, n = 7 in selinexor 60 mg qw + RUX group)	1/2 (NCT04562389)	Age ≥18 y Peripheral blasts ≤5% and bone marrow blasts ≤10% ECOG PS ≤2 DIPSS: int-1, int-2, or high-risk ANC ≥1.5 × 10^9/L PLT, ≥100 × 10^9/L	Week 12 SVR35, 4/5 evaluable patients	No dose-limiting TEAEs No grade ≥3 neutropenia or thrombocytopenia

AE: adverse event; ANC: absolute neutrophil count; AZA: azacitidine; BCL: B-cell lymphoma; BCM: below costal margin; BET: bromodomain and extra-terminal; BL: baseline; BM: bone marrow; BMF: bone marrow fibrosis; ClinImp: clinical improvement; DIPSS: Dynamic International Prognostic Scoring System; ECOG PS: Eastern Cooperative Oncology Group performance status; Hb: hemoglobin; int: intermediate; IWG-MRT: International Working Group for Myelofibrosis Research and Treatment; JAKi: Janus kinase inhibitor; MF: myelofibrosis; NTD: non-transfusion dependence; ORR: overall response rate; OS: overall survival; PI3K: phosphoinositide 3 kinase; PLT: platelet; qd: once daily; qw: once weekly; RBC: red blood cell; RUX: ruxolitinib; SV: spleen volume; SVR35: ≥35% SV reduction from baseline; TD: transfusion dependent; TEAE: treatment-emergent adverse event; TGF: transforming growth factor; TI: transfusion independence; TRAE: treatment-related adverse event; TSS50: ≥50% reduction in Total Symptom Score from baseline.

^aFor SVR35 rates at Week 24, discontinued patients were included as nonresponders, and ongoing patients that had not yet reached Week 24 were excluded.

Conclusion

In the more than 10 years since FDA approval of ruxolitinib for treatment of MF, a host of clinical trials, pooled analyses, and real-world data have demonstrated its efficacy, including effects on spleen response, MF symptom resolution, and OS benefits compared with other therapies. Despite the proven benefits, ruxolitinib treatment is sometimes delayed or avoided in favor of HU or other therapeutic approaches. This is in contrast to evidence suggesting that earlier administration of ruxolitinib may be beneficial to patients, as well as other data demonstrating the poor outcomes that occur in patients who discontinue treatment with ruxolitinib. Over the next 10 years and beyond, the importance of optimizing therapeutic strategies, including how best to use ruxolitinib as monotherapy or in combination with other agents for further disease modification, will continue to be a topic of utmost importance.

Disclosure statement

Naveen Pemmaraju: Board of directors/management for Dan’s House of Hope. Consultant for AbbVie, Astellas Pharma US, Inc., Cimeio Therapeutics AG, CTI BioPharma, Immunogen, Intellisphere, LLC., Patient Power, PharmaEssentia, Protagonist Therapeutics, and Total CME. Scientific/advisory committee member for Blueprint Medicines, Cancer.Net, CareDx, CTI BioPharma, Incyte, Menarini Group, and Pacylex. Speaker/preceptorship with AbbVie, Aplastic Anemia & MDS International Foundation, Curio Science LLC, Dava Oncology, Intellisphere, LLC., Magdalen Medical Publishing, Medscape, Novartis Pharmaceuticals Corp, Physicians Education Resource, Protagonist Therapeutics, and Targeted Oncology. Research support from US Department of Defense (DOD).

Prithviraj Bose: Research support from Incyte, BMS, CTI, Morphosys, Kartos, Blueprint, Cogent, Astellas, Pfizer, NS Pharma, Promedior, Disk Medicine, Ionis, and Telios. Honoraria from Incyte, BMS, CTI, Sierra (now GSK), Blueprint, Cogent, Novartis, Karyopharm, AbbVie, Pharma Essentia, Morphosys, and Kartos.

Raajit Rampal: Consulting fees from Incyte, Celgene/BMS, Blueprint, AbbVie, Promedior, CTI, Stemline, Galecto, Pharmaessentia, Protagonist, Constellation/Morphosys, Novartis, Sierra Oncology/GSK, and Servier. Research funding from Constellation Pharmaceuticals, Incyte, Zentalis, and Stemline Therapeutics.

Aaron Gerds: Consulting or advisory role with Incyte, AstraZeneca/MedImmune, CTI, Apexx Oncology, and Celgene; received research funding from Pfizer, CTI, Incyte Corporation, Roche/Genentech, Gilead Sciences, Imago Biosciences, Sierra Oncology, and Celgene; equity ownership in Samus Therapeutics.

Angela Fleischman: Speakers bureau for CTI BioPharma and Bristol Myers Squibb, speakers bureau and consultation for PharmaEssentia, and advisory board for Sierra Oncology.

Srdan Verstovsek: Research support from Incyte, Roche, NS Pharma, Celgene, Gilead, Promedior, CTI BioPharma Corp., Genentech, Blueprint Medicines Corp., Novartis, Sierra Oncology, PharmaEssentia, AstraZeneca, Italfarmaco, and Protagonist Therapeutics. Consulting for Constellation, Sierra Oncology, Incyte, Novartis, Celgene, and BMS.

Additional information

Funding

The cost of development of this article and publication fees were funded by Incyte Corporation (Wilmington, DE, USA). Writing assistance was provided by Samantha Locke, PhD, an employee of ICON (Blue Bell, PA, USA), and was funded by Incyte.