Efficacy and safety of ruxolitinib in Asian patients with myelofibrosis

Abstract

Myelofibrosis is characterized by progressive cytopenias, bone marrow fibrosis, splenomegaly and severe constitutional symptoms. In the phase 3 Controlled Myelofibrosis Study with Oral JAK Inhibitor Treatment (COMFORT) studies, ruxolitinib, a potent Janus kinase 1 (JAK1)/JAK2 inhibitor, provided substantial improvements in splenomegaly, symptoms, quality-of-life measures and overall survival compared with placebo or best available therapy. No assessments of the efficacy and safety of ruxolitinib have been conducted in Asian patients. Here, we describe results from an open-label, single-arm, phase 2 trial evaluating ruxolitinib in Asian patients with myelofibrosis (n = 120). The primary endpoint was met, with 31.7% of patients achieving a ≥ 35% reduction from baseline spleen volume at week 24. As measured by the 7-day Myelofibrosis Symptom Assessment Form v2.0, 49% of patients achieved a ≥ 50% reduction from baseline in total symptom score. Adverse events were consistent with those seen in the COMFORT studies. Ruxolitinib was well tolerated in Asian patients with myelofibrosis and provided substantial reductions in splenomegaly and improvements in symptoms.

Keywords::

• Myelofibrosis
• JAK1/JAK2 inhibitor
• phase 2
• Asian patient population

Introduction

Myelofibrosis (MF) is the most debilitating subtype of the classic myeloproliferative neoplasms, which also include polycythemia vera (PV) and essential thrombocythemia (ET) [1,2]. It can develop de novo (primary MF, PMF) or present in patients following the development of PV (post-PV MF, PPV-MF) or ET (post-ET MF, PET-MF), and is characterized by progressive bone marrow fibrosis, worsening cytopenias and extramedullary hematopoiesis, often leading to hepatosplenomegaly [1]. Patients with MF often experience severe disease-related complications including thrombosis, anemia, microvascular issues and significant constitutional symptoms such as fatigue, night sweats, weight loss, bone pain and fevers [3]. Symptoms such as early satiety, dyspnea, and pain and discomfort due to enlarged spleen size are associated with splenomegaly, which presents in 89% of patients with PMF [3,4]. As a result of these complications, patients with MF often have a substantially reduced quality of life (QOL) and life span (median survival, 5 years) [3,4]. Based on the presence of factors such as age ≥ 65 years, hemoglobin level < 10 g/dL and constitutional symptoms, patients are categorized as having low- (no risk factors), intermediate-1- (one risk factor), intermediate-2- (two risk factors) or high- (≥ 3 risk factors) risk disease [4].

The Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signaling network is central to the regulation of many cellular responses that are mediated by cytokines and growth factors (i.e. hematopoiesis and inflammation). Dysregulation of the JAK/STAT pathway is a characteristic feature contributing to the pathogenesis of MF [5]. Interestingly, somatic mutations in JAK2 are present in approximately 55–65% of patients with PMF and PET-MF, and in 96% of patients with PPV-MF.

A phase study and two phase 3 studies conducted in the United States and Europe led to the approval of ruxolitinib, a potent and selective JAK1/JAK2 inhibitor. In the phase 3 studies, Controlled Myelofibrosis Study with Oral JAK Inhibitor Treatment (COMFORT)-I and COMFORT-II, ruxolitinib showed superiority over placebo and best available therapy (BAT), respectively, providing significant reductions in splenomegaly and symptom burden and improving QOL measures and overall survival [6–9]. In COMFORT-II, the primary and secondary endpoints were achieved, with 28% and 32% of ruxolitinib-treated patients achieving a ≥ 35% reduction from baseline spleen volume at 48 and 24 weeks, respectively, compared with no patients in the BAT arm at either time point. Spleen responses were durable, with 73% of patients maintaining their response at a median of 48 weeks from their initial response (median duration had not been reached at a median follow-up of 151 weeks). Long-term ruxolitinib treatment (median, 151 weeks) was associated with a significant survival benefit over BAT; patients in the ruxolitinib group had a 52% reduction in the risk of death compared with those in the BAT treatment group (hazard ratio [HR], 0.48; 95% confidence interval [CI], 0.25–0.85; p = 0.009) [8].

No evaluations of the efficacy and tolerability of ruxolitinib have been conducted in patients with MF in Asian countries. Furthermore, only a small number of Asian patients or volunteers have been enrolled in any study of ruxolitinib. Here we describe results from an open-label, single-arm, phase 2 trial evaluating ruxolitinib in Asian patients with splenomegaly (≥ 5 cm below costal margin) and intermediate-2- or high-risk PMF, PPV-MF or PET-MF.

Materials and methods

Eligibility criteria

Patients 18 years of age or older with World Health Organization-defined PMF or International Working Group for Myelofibrosis Research and Treatment (IWG-MRT)-defined PPV-MF or PET-MF and palpable splenomegaly measuring ≥ 5 cm below the costal margin were eligible for this study, irrespective of JAK2 mutation status [4,10]. Eligible patients had at least two IWG-MRT risk factors (age > 65 years, constitutional symptom burden, hemoglobin < 10 g/dL, white blood cells > 25 × 10⁹/L and circulating blasts ≥ 1%) indicating an intermediate-2- or high-risk disease state [4]. In addition, patients were required to have an Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2 (scale: 0–5, with 0 indicating a patient who is completely active and 5 indicating death [a higher score represents increasing disability]) and appropriate bone marrow reserves [11]. A circulating blast count < 10%, ineligibility for stem cell transplant and no prior JAK inhibitor treatment were also requirements for participation.

Study design

This multicenter study was conducted in China, Japan, Korea and Taiwan. Patients with baseline platelet counts of 100–200 × 10⁹/L received 15 mg ruxolitinib twice daily (bid), and those with baseline platelet counts > 200 × 10⁹/L received 20 mg bid ruxolitinib. For optimal dosing, a standardized dose regimen was used to determine dose modifications. Dose increases were permitted at week 4 for patients who demonstrated reduced efficacy as demonstrated by a < 40% reduction from baseline spleen length. These patients were required to have a platelet count of at least 150 × 10⁹/L and an absolute neutrophil count of at least 1 × 10⁹/L at every post-baseline assessment. The study protocol included mandatory dose decreases or interruptions for safety. Dosing was held if platelet counts decreased to < 50 × 10⁹/L or if the absolute neutrophil count decreased to ≤ 0.5 × 10⁹/L during treatment. Dosing was also held if patients had treatment-related grade 3 or 4 non-hematologic toxicities while receiving ruxolitinib. Dose interruption due to a hematologic adverse event (AE) or abnormal laboratory value was monitored at least once a week for 4 weeks and for 4-week intervals thereafter until the event resolved or stabilized. Delay in dosing of ≥ 8 weeks required discontinuation from the study, except in cases of interruptions due to splenectomy, where the dose could be restarted if patients recovered from the surgery within 12 weeks.

Endpoints and assessments

The primary endpoint of this study was the proportion of patients who achieved a ≥ 35% reduction from baseline spleen volume at week 24 as measured by magnetic resonance imaging (MRI; or computed tomography for patients who could not undergo MRI); all scans were read by a central reader. The secondary endpoints included the percentage of patients who achieved a ≥ 50% reduction in total symptom score at week 24 as measured by the 7-day modified MF Symptom Assessment Form (MFSAF) version 2.0, the absolute and percent changes in QOL measures as assessed by the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30) at week 24, the best response rate for patients who achieved a ≥ 35% reduction in spleen volume and the spleen response duration, which was defined as the time interval between the first ≥ 35% reduction in spleen volume from baseline and any of the following events: death, splenic irradiation, splenectomy, leukemic transformation or an increase in spleen volume of ≥ 25% from baseline [12,13].

The MFSAF v2.0 diary, provided as a paper-based questionnaire, assesses several symptoms including early satiety, inactivity, night sweats, bone and muscle pain, abdominal discomfort and pain, and itching. Symptom scores (0–10) indicate severity, with 0 representing “absence” and 10 representing “worst imaginable.” The MFSAF-Total Symptom Score (TSS) includes the sum of all symptoms except inactivity; assessments of TSS included patients with a non-zero and non-missing TSS at baseline and a non-missing score at week 24 (n = 101). The baseline TSS is the average of the daily scores throughout the 7-day baseline interval, and the week-24 TSS is the 28-day average of the daily scores prior to the week-24 visit. The EORTC QLQ-C30 consists of 30 questions that measure physical, role, cognitive, emotional and social functional scales; fatigue, pain and nausea/vomiting symptom scales; and a global health status/QOL scale. A higher score is associated with a better level of functionality and QOL.

Two sensitivity analyses were performed to determine the effect of a stricter definition of end date on the duration of spleen response. The redefined end dates were the first scan that no longer showed a ≥ 35% reduction and that showed a ≥ 25% increase over nadir (the first sensitivity analysis) and the first scan that was > 25% over nadir (second sensitivity analysis). Change from baseline in spleen volume, spleen size (assessed by palpation), ECOG performance status and transfusion dependency, as well as time to first response in patients achieving a ≥ 35% reduction from baseline spleen volume at any time during the study, were also explored.

Safety

The occurrence of all AEs, including serious AEs (SAEs), was collected. The severity of each AE and its relationship to ruxolitinib treatment was also reported. Hematology, coagulation and urinalysis results, blood chemistry and vital signs were regularly monitored, and physical examinations, electrocardiography (ECG) monitoring and regular assessments of body weight were also performed. A follow-up visit was completed 30 days after the last ruxolitinib dose was received.

Statistical analysis

This study had a group sequential design and included one interim analysis to show efficacy at the interim time point (cut-off, 29 October 2012). The interim analysis was performed when the first 50 patients had either completed week 24 or discontinued treatment. An alpha-spending function approximating the O’Brien–Fleming stopping boundary was used to determine the boundary of the single-sample binomial test; this boundary was not crossed at interim analysis. A one-sided single-sample binomial test with a type I error of 0.025 was used for the primary analysis.

The final analysis (cut-off, 7 June 2013) includes all patients. The boundaries at the final analysis were recalculated according to the actual number of enrolled patients (n = 120) and the alpha spending function used at the interim analysis to generate a cumulative type I error of < 2.5%. The best response rate (the proportion of patients who achieved a ≥ 35% reduction from baseline spleen volume at any time during the study) was estimated with a 95% CI; the duration of best response was evaluated by a Kaplan–Meier estimate. The percentage of patients who achieved a ≥ 50% decrease in MFSAF v2.0 TSS at week 24 was estimated with a 95% CI; time to first ≥ 50% reduction from baseline TSS as well as time to first ≥ 35% reduction in spleen volume were assessed by a Kaplan–Meier estimate.

Role of the funding source

This study was sponsored by Novartis Pharmaceuticals and conducted in compliance with the guidelines from the US Food and Drug Administration's Office of Good Clinical Practice and the principles of the Declaration of Helsinki. The study protocol was reviewed by each participating center's institutional review board and written informed consent was obtained from each patient. The first draft of the manuscript was provided by the first author with the assistance of a medical writer funded by the sponsor; the first author made the final decision for publication. All authors and sponsor representatives reviewed the manuscript and attest to the integrity of the data and adherence to the planned protocol and statistical analyses.

Results

Patient characteristics

Complete patient demographics and disposition are shown in Tables I and II. The study population included 120 patients who were predominantly Chinese (52.5%) and Japanese (25%; Table I); the median age was 61 years (range, 25–80 years). The majority of patients (66.7%) had PMF (PPV-MF, 17.5%; PET-MF, 15.8%), with a median time since initial MPN diagnosis of 5.22 years. Sixty-nine percent and 31% of patients were classified as having intermediate-2- and high-risk disease, respectively. Most patients (56.7%) had an ECOG score of 0, 40% had an ECOG score of 1. A baseline hemoglobin level of < 10 g/dL (median, 9.80 g/dL) was reported in 56.7% of patients and 15.8% of patients were considered transfusion-dependent. Hydroxyurea (55.8%), interferons (30%), and thalidomide and lenalidomide (24.2%) were the most commonly reported previous therapies.

Table I. Patient demographics and baseline characteristics.

Characteristic	All patients (n = 120)
Age, median (range), years	61.0 (25.0–80.0)
Female, n (%)	62 (51.7)
Ethnicity, n (%)
Chinese	63 (52.5)
Japanese	30 (25.0)
Korean	17 (14.2)
Taiwanese	10 (8.3)
MF subtype, n (%)
PMF	80 (66.7)
PPV-MF	21 (17.5)
PET-MF	19 (15.8)
Time since diagnosis*, median (range), years	5.22 (0.1–30.8)
Spleen volume, median (range), cm^3	2159.0 (658–7849)
Palpable spleen size below costal margin, median (range), cm	15.0 (5.0–45.0)
ECOG performance status, n (%)
0	68 (56.7)
1	48 (40.0)
2	4 (3.3)
Prior hydroxyurea therapy, n (%)	67 (55.8)
Prior splenic radiotherapy, n (%)	3 (2.5)
Hemoglobin, median (range), g/dL	9.80 (4.6–18.0)
Category, n (%)
< 10 g/dL,	68 (56.7)
> 10 g/dL,	52 (43.3)
Platelet count, median (range), × 10^9/L	231.5 (96.0–1257.0)
Category, n (%)
100–200 × 10^9/L	46 (38.3)
> 200 × 10^9/L	74 (61.7)
Circulating blasts, median (range), %	1.0 (0.0–8.0)
Category, n (%)
< 1%	41 (34.2)
≥ 1%	79 (65.8)
Transfusion, any, n (%)	19 (15.8)
Red blood cells, n (%)	17 (14.2)
Platelets, n (%)	0
Other blood component, n (%)	2 (1.7)
JAK2 V617F mutation status, n (%)
Negative	16 (13.3)
Positive	41 (34.2)
Not assessed^†	63 (52.5)

MF, myelofibrosis; PMF, primary myelofibrosis; PPV-MF, post-polycythemia vera myelofibrosis; PET-MF, post-essential thrombocythemia myelofibrosis; ECOG, Eastern Cooperative Oncology Group.

*The diagnosis date recorded for patients with PPV-MF and PET-MF was the date of diagnosis of PV or ET.

^†Mutation status was not assessed in patients enrolled in China.

Table II. Patient disposition.

	All patients (n = 120)
Ongoing, n (%)	93 (77.5)
Discontinued, n (%)	27 (22.5)
Duration of exposure, median (range), months	8.44 (0.5–21.7)
Primary reason for discontinuation, n (%)
Adverse event(s)	11 (9.2)
Disease progression	9 (7.5)
Death	3 (2.5)
Patient withdrew consent	3 (2.5)
Protocol deviation	1 (0.8)

The median daily doses for the 15 mg (n = 46) and 20 mg (n = 74) bid starting-dose groups were 20.64 mg/day and 36.11 mg/day, respectively. The median duration of follow-up was 8.44 months (range, 0.5–21.7 months, Table II). For most patients the daily dose decreased over time, stabilizing after week 8 at approximately 10 mg and 15 mg for the 15 mg bid and 20 mg bid starting-dose groups, respectively (Figure 1). In Japanese patients, there was a slight increase in the frequency of AE-induced dose reductions or interruptions, which resulted in a lower dose intensity in this cohort that stabilized after week 8.

Figure 1. Mean dose intensity over time by week and starting dose of ruxolitinib. Boxes represent median and interquartile range, whiskers (lines) represent 5th and 95th percentiles.

The majority of discontinuations were due to AEs (9.2%, n = 11) and disease progression (7.5%, n = 9; Table II). For patients who discontinued prior to week 24 (n = 19), AEs (n = 9), disease progression (n = 5), death (n = 3), deviation from protocol (n = 1) and withdrawal of consent (n = 1) were the causes. Overall, 101 patients were evaluable for spleen response at week 24, and 96 were evaluable for TSS assessment. At data cut-off (7 June 2013), 22.5% of patients (n = 27) discontinued treatment and 77.5% (n = 93) were still receiving ruxolitinib.

Efficacy

Spleen response

The primary endpoint was achieved with 31.7% of patients (n = 38) achieving a ≥ 35% reduction from baseline spleen volume at week 24 (95% CI, 23.3–40.0%; one-sided p = 0.0007, the target number of responders at the final analysis was ≥ 33 patients) [Figure 2A]]. The median percent reduction from baseline spleen volume at week 24 was 26.9% [Figure 2(B)]. At any time during the study, 38.3% of patients (n = 46) had a ≥ 35% reduction in spleen volume from baseline (best response rate; 95% CI, 29.6–47.0%), with a median percent reduction from baseline spleen volume of 30.8% [Figures 2(A) and 2(B)]. The median time to the first ≥ 35% reduction in spleen volume was observed at 12.43 weeks (95% CI, 12.14–12.86 weeks); the first MRI assessment did not occur until the week-12 visit and regular assessments occurred at 12-week intervals thereafter.

Figure 2. (A) Percentage of patients who achieved a ≥ 35% reduction from baseline in spleen volume after treatment with ruxolitinib. (B) Percentage change from baseline in spleen volume at week 24 for individual patients who received ruxolitinib. Only patients with assessment at both baseline and week 24 are included; 19 patients (15.8%) discontinued from the study prior to spleen volume assessment at week 24. (C) Probability of maintenance of spleen response in patients treated with ruxolitinib.

Although almost all patients experienced a decrease from baseline spleen volume at some point during the study, a small proportion of patients (five of 111 patients who had an evaluable imaging assessment post-baseline; 4.5%) experienced an increase from baseline as their greatest percent change. The median duration of response could not be determined, because only one patient experienced a defined loss of response (death). The probability of a response lasting > 24 weeks was 0.96 (95% CI, 0.76–0.99) [Figure 2(c)]. The Kaplan–Meir estimates of duration of response 24 weeks after the first response were 0.86 (95% CI, 0.66–0.94) and 0.82 (95% CI, 0.63–0.92) for the first and second sensitivity analyses, respectively. The median duration of response was not estimated for either analysis due to too few events (first analysis, seven events; second analysis, nine events).

Ruxolitinib treatment was also associated with reductions (mean, − 40.7%) in palpable spleen length. Mean reductions in spleen length were observed in the patient population at each time point, with the mean percent change from baseline spleen length ranging from − 37.0% to − 40.8% between week 4 and week 48.

Symptom burden and functioning/QOL measures

Approximately half of evaluated patients (49%, 47 of 101) experienced a ≥ 50% improvement from baseline TSS at 24 weeks (95% CI, 39.0–59.0%) [Figure 3(A)]; improvement was seen across individual MFSAF-assessed symptoms [Figure 3(B)]. The median percent change from baseline TSS was − 47.2% at week 24. In particular, abdominal discomfort (3.5), early satiety (2.6) and night sweats (2.8), symptoms with the highest mean scores at baseline, improved substantially, with − 30.6%, − 35.5% and − 49.9% reductions from baseline, respectively [Figure 3(B)]. Ruxolitinib treatment resulted in rapid symptom reduction, with a median time to first ≥ 50% reduction in TSS occurring at 4.57 weeks (95% CI, 4.14–8.14 weeks).

Figure 3. (A) Percentage of patients treated with ruxolitinib who achieved a ≥ 50% reduction in total symptom score. The percentage of patients achieving ≥ 50% reduction from baseline is calculated based on the total number of patients having non-zero total symptom score at baseline and non-missing score at that time point as the denominator. (B) Mean percentage change from baseline at week 24 in individual symptom scores (Myelofibrosis Symptom Assessment Form v2.0) in patients treated with ruxolitinib.

The baseline score for global health status/QOL (59.4 out of 100) was the lowest among the EORTC QLQ-C30 functional/QOL scale scores, indicating a reduced QOL and level of functioning prior to study entry. Patients reported high baseline individual symptom scores for fatigue (mean 33.1), insomnia (21.4), dyspnea (19.2), appetite loss (18.1) and pain (17.8). The mean percentage changes in baseline score at week 24 were − 8.2%, − 12.1%, + 4.7%, − 41.4%, and − 16.3% for fatigue, insomnia, dyspnea, appetite loss, and pain, respectively (reduction of symptoms is demonstrated by a negative change in score). For global health/QOL, the mean change was + 5.2, which, although not clinically meaningful (clinical relevance was defined as a 10-point difference in score), indicated a modest improvement in functioning.

Exploratory analyses

The mean (standard deviation) monthly rate of packed red blood cell transfusion was 0.23 (0.684) units/month at baseline and 0.5 (1.242) units/month at any time after ruxolitinib treatment. The mean (standard deviation) monthly rate of platelet transfusion increased from 0 units/month at baseline to 0.16 (1.477) units/month on treatment. Although a small subset of patients (n = 16) had a slightly worse ECOG score (increase from 0 or 1 to ≥ 2) over the course of the study, the majority of patients reported ECOG performance statuses similar to baseline.

JAK2 V617F allele burden and bone marrow fibrosis grade were not assessed for all patients. A modest decrease (median, − 5.26%) in allele burden was observed in evaluable patients (n = 31) at week 24; two patients experienced a reduction from baseline allele burden of > 20%. Of the 37 patients who had marrow fibrosis assessments at week 48, four patients experienced an improvement from baseline measurements, and four demonstrated worsening in fibrosis density. In general, patients experienced an increase in body weight over time, with a mean change of 2.87 kg at week 24.

Safety

Almost all patients (96.7%) experienced one or more AEs (any grade) over the course of the study. The incidence of grade 3 or 4 AEs was 65.8%. The most commonly reported AEs included anemia (60%), decreased platelet count (30.8%), thrombocytopenia (26.7%) and diarrhea (25.8%) and were predominantly grade 1 or 2, with the exception of anemia (Table III). The most frequent (≥ 5%) grade 3 or 4 AEs included anemia (44.2%), decreased platelet count (8.3%), thrombocytopenia (6.7%) and herpes zoster (5%). Anemia, decreased platelets and thrombocytopenia were the most commonly reported (≥ 20%) AEs suspected to be related to ruxolitinib treatment, and occurred in 58%, 28% and 27% of patients, respectively. AEs of decreased platelet count and thrombocytopenia were collected as separate categories per the treating investigator and coded as such; patients may have experienced an AE in each of the two categories. Overall, 24.2% of patients (n = 29) experienced at least one SAE, which included pyrexia (six patients), pneumonia (four patients), herpes zoster (four patients), cardiac failure (three patients), lung infection (three patients), anemia (two patients) and renal failure (two patients).

Table III. Frequent adverse events (≥ 10% in all grades) regardless of relationship to study drug.

Adverse events	All patients (n = 120)
	All grades, n (%)	Grade 3/4, n (%)
Any adverse event	116 (96.7)	79 (65.8)
Anemia	72 (60.0)	53 (44.2)
Platelet count decrease	37 (30.8)	10 (8.3)
Thrombocytopenia	32 (26.7)	8 (6.7)
Diarrhea	31 (25.8)	5 (4.2)
Upper respiratory tract infection	21 (17.5)	3 (2.5)
Alanine aminotransferase increase	18 (15.0)	2 (1.7)
Pyrexia	18 (15.0)	1 (0.8)
Aspartate aminotransferase increase	16 (13.3)	2 (1.7)
Cough	14 (11.7)	0
Herpes zoster	14 (11.7)	6 (5.0)
Nasopharyngitis	13 (10.8)	0
Constipation	12 (10.0)	0
Gamma-glutamyltransferase increase	12 (10.0)	2 (1.7)
Headache	12 (10.0)	0

AEs necessitating dose reduction or interruption (65%) were reported frequently, while those requiring discontinuation were reported infrequently (10%, n = 12). Decreased platelet count (27.5%; grade , 5.8%) thrombocytopenia (23.3%; grade , 6.7%) and anemia (11.7%; grade , 10.8%) were the most common causes for dose reductions and interruptions. Japanese patients had a higher incidence of anemia-induced dose reductions or interruptions compared with non-Japanese patients in this study (30% vs. 6%).

Patients commonly reported new or worsening hematologic abnormalities (Table IV), with the most common grade abnormalities being decreased hemoglobin (55.7%), lymphocytes (19.5%), and platelet count (15.3%). Anemia, decreased platelet count, thrombocytopenia and diarrhea were the most frequently reported AEs across treatment periods (Table V).

Table IV. New or worsening hematologic parameters.

	Worsening from baseline to
	Any grade, n/total (%)*	Grade 3, n/total (%)*	Grade 4, n/total (%)*
Hemoglobin	82/97 (84.5)	54/97 (55.7)	—
Platelets	79/118 (66.9)	16/118 (13.6)	2/118 (1.7)
Leukocytes	19/118 (16.1)	0/118 (0)	1/118 (0.8)
Lymphocytes	50/118 (42.4)	22/116 (19)	1/118 (0.8)
Neutrophils	28/118 (23.7)	7/118 (5.9)	2/118 (1.7)

*Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 was used for grading. Total = number of patients evaluable after baseline, who had less than the respective grade at baseline. n = number of patients who had less than the respective grade at baseline, and progressed to that grade after baseline. Missing baseline values were excluded. Patients are counted only for the worst grade observed after baseline. Laboratory assessments performed more than 30 days after discontinuation of study treatment are not summarized. The percentage of patients with new or worsening abnormalities is calculated based on the number of patients with evaluable post-baseline results who had less than the respective grade at baseline as the denominator.

Table V. Frequent adverse events by treatment period.

	Adverse event, %
	Week 0–4 (n = 120)*	Week > 4–8 (n = 119)*	Week > 8–12 (n = 118)*	Week > 12–24 (n = 114)*	Week > 24–36 (n = 102)*	Week > 36–48 (n = 69)*	Week > 48–60 (n = 39)*	Week > 60–72 (n = 32)*	Week > 72 (n = 18)*
Any adverse event	59	60	60	75	52	62	72	69	50
Anemia	11	24	27	26	18	15	28	22	11
Platelet count decrease	10	13	12	11	10	12	5	6	11
Thrombocytopenia	10	8	7	12	3	12	15	9	0
Diarrhea	10	1	5	10	3	3	3	5	6

*n⁵ risk set at the start date of each time interval. If a patient is still on study (including AE monitoring period) at the start date of a time interval, that patient is counted in the denominator.

A total of seven patients, four with ≥ 3% peripheral blasts at baseline, experienced leukemic transformation (blasts > 20%), including one patient who experienced leukemic transformation after the time of data cut-off (mean follow-up, 303 days). Over the course of the study, including the first 30 days after the last dosage, six patients (5%) died; one patient died 48 days after the last ruxolitinib dose. Of these deaths (n = 7), two were due to hepatorenal syndrome and cardiopulmonary failure (one each) and five were due to infection (e.g. sepsis, pneumonia/lung infection); one death due to sepsis and one due to hepatorenal syndrome were suspected by the investigator to be related to ruxolitinib treatment.

Discussion

In this multinational, open-label, phase 2 study, ruxolitinib provided a substantial improvement in splenomegaly and symptoms in Asian patients with PMF, PPV-MF or PET-MF. Patients in this study had baseline characteristics generally representative of the overall population of patients with MF, with only minor differences compared with patients in the COMFORT studies. At baseline, patients in this study had a smaller median spleen volume (2159 cm³ vs. 2598 cm³ and 2408 cm³ in COMFORT-I and -II, respectively) and a lower median hemoglobin level (9.80 g/dL vs. 10.5 g/dL and 10.6 g/dL), and a larger proportion of patients had PMF (67% vs. 45% and 53%) and intermediate-2-risk disease (69% vs. 41% and 40%). Although no direct comparisons can be made, a similar number of patients in this and the COMFORT-II study (31.7% vs. 31.9%) achieved at least a 35% reduction in spleen volume from baseline at the week 24, and the majority of patients (95%; 106 of 111 evaluable patients) experienced a reduction in spleen size at any time during the study (median: 31% vs. 28% in COMFORT-II). Additionally, reductions in spleen volume were comparable across Asian ethnicities. Ruxolitinib treatment resulted in rapid spleen responses, with a median response time similar to that observed in COMFORT-II (12.4 weeks vs. 12 weeks). Median time to response coincided with the first MRI assessment, further supporting that response typically occurred early in the treatment period. It is possible that the actual time to response was shorter, because the first MRI assessment did not occur until week 12.

In addition to improving splenomegaly, ruxolitinib provided significant reductions in symptom burden as measured by the MFSAF v2.0. Forty-nine percent of evaluable patients achieved a ≥ 50% improvement in baseline TSS at week 24, which is comparable to results observed in COMFORT-I (45.9%). Rapid improvements in TSS were seen across all symptoms (night sweats, itching, abdominal discomfort, early satiety, abdominal pain, inactivity and muscle/bone pain), with a median time similar to COMFORT-I (4.57 weeks vs. 4.4 weeks). Compared with patients in the COMFORT studies, observed baseline individual subscale scores of the EORTC QLQ-C30 were higher for global health status/QOL and role and physical functioning and also for MF-related symptoms, including fatigue, pain, dyspnea and appetite loss, indicating a better level of functioning but an increased symptom burden. Overall, the change in EORTC QLQ-C30 was not clinically meaningful (mean change, + 5.2; clinical relevance defined as a 10-point difference in score), but a trend for improvement was observed.

Based on the predicted frequency of leukemic transformation in patients with MF (0.06 patients per patient-year) [14–16], approximately six transformations would be predicted for this study population (n = 120). In this study, seven patients (mean follow-up, 303 days), all of whom had long disease histories since initial MPN diagnosis (3–25 years) and baseline blast counts ≥ 1% (four patients had ≥ 3% blast counts at baseline), experienced leukemic transformations; these factors are associated with increased incidence (20–30%) of transformation and poor leukemia-free survival [17–19].

Given the role of JAK2 in hematopoiesis, myelosuppression is an identified consequence of JAK inhibition. As such, anemia (60%), decreased platelet count (31%) and thrombocytopenia (27%) were the most commonly reported hematologic AEs, and were the AEs suspected most frequently (≥ 20%) to be related to ruxolitinib treatment (58%, 28% and 27%, respectively). Though no direct comparisons can be made between this phase 2 study and the phase 3 COMFORT studies, it can be noted that in this study the proportion of patients who developed anemia was relatively larger (60% vs. 31% and 41% in COMFORT-I and COMFORT-II) and the frequency of anemia-induced dose reductions or interruptions was higher (12% vs. 6% and 5%), both of which could be attributed to lower median baseline hemoglobin levels in this study population. The percentage of new or worsening hemoglobin abnormality from baseline, however, was similar to the COMFORT trials (85% vs. 82%). Despite the increased occurrence in this population, anemia was manageable with blood transfusion and dose adjustments, and was the cause of discontinuation in only one patient. Overall, AEs leading to discontinuation occurred infrequently (10%; 12 patients), and approximately 24% of patients experienced ≥ 1 SAE, all of which resolved (excluding patients who died). Subgroup analyses of the safety of ruxolitinib showed similar profiles across Asian ethnicities.

In conclusion, ruxolitinib represents an effective therapeutic option for Asian patients with MF and provides substantial improvements in splenomegaly and constitutional symptoms. As evidenced by the low discontinuation rate and the occurrence of predictable and manageable AEs in our study, ruxolitinib is well tolerated in these patients, and has a safety profile similar to what has been observed in the COMFORT studies.

Acknowledgements

The authors wish to thank Dr. Mineo Kurokawa for his contributions as a study investigator (University of Tokyo Hospital, Tokyo, Japan). The authors would also like to thank the patients and their families for their participation, and medical staff from participating institutions for their contributions to this study. Editorial support was provided by Kia C. Walcott, PhD, and was funded by Novartis.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.