Outcomes for ruxolitinib only versus combination with interferon in treating patients with myelofibrosis

TO THE EDITOR:

Myelofibrosis (MF) is the most aggressive of the myeloproliferative neoplasms (MPNs) and is driven by cytokine signal–augmenting mutations in JAK2, CALR, and MPL. MF is commonly associated with burdensome symptoms, splenomegaly, cytopenias, and the propensity for transformation to acute leukemia, all of which compromise patients’ quality of life and increase mortality. MF can arise de novo (primary MF [PMF]) or from preexisting MPNs (secondary MF [SMF]) and both carry a relatively poor prognosis with a median survival of ∼5 and 9 years, respectively.¹ Treatment of MF with JAK inhibitors, like ruxolitinib (RUX), is the standard of care for intermediate and high-risk patients and leads to symptom improvement and spleen volume reduction. However, the Controlled MF Study with Oral JAK Inhibitor Treatment (COMFORT) trials that led to RUX approval did not assess progression-free survival (PFS) or overall survival (OS), and although some post hoc analyses suggested improved OS, this interpretation is controversial.² The lack of significant improvement in PFS and OS outcomes may be because of the inability of RUX to deplete mutated hematopoietic stem cells in MF.³ In contrast, interferon alfa (IFN) has been shown to prolong PFS and/or OS in polycythemia vera^4,5 and is an off-label treatment for MF that may selectively deplete MPN hematopoietic stem cells and improve marrow morphology and fibrosis.^6-9 Because of the lack of appropriately powered randomized, controlled studies, the extent to which IFN affects PFS or OS outcomes in MF is not known.

The combination of the 2 MF treatments, IFN+RUX, is conceptually attractive as a means to enhance clinical efficacy by targeting the malignant clone with IFN while using RUX to improve the cytokine-driven symptom burden of disease and the tolerability of IFN.¹⁰ Based on this rationale, Sørensen et al. tested the combination in 18 patients with MF who were refractory to IFN.¹¹ In this phase 2 trial, 44% of patients achieved either complete (CR) or partial responses (PR), including symptom, spleen size, hematologic, bone marrow, and molecular improvements.¹¹ The combination was also safe and effective in a phase 1/2 trial with high rates of spleen volume reduction and clinical and molecular response in RUX- and IFN-naive patients.¹² However, IFN+RUX has never been directly compared with either agent alone, and thus the efficacy, safety, and PFS/OS of the combination in comparison with RUX monotherapy remains unknown. For this reason and because IFN+RUX has been used to treat MF at the Weill Cornell Medicine (WCM) Silver MPN center since 2012, we conducted a single-center, case-control study to investigate the clinical benefit and tolerability of IFN+RUX in comparison with RUX only. The institutional review board of WCM approved this study, and it was conducted in accordance with the Declaration of Helsinki.

We identified 31 patients with MF who were treated with IFN+RUX by querying our research data repository using previously described methods.⁴ These patients were matched to a control group of 31 patients treated with RUX by propensity-score matching based on the Dynamic International Prognostic Scoring System Plus (DIPSS+) factors at the start of treatment, including age, hemoglobin, platelets, white blood cell count, and peripheral blasts, as well as sex and MF type (PMF vs SMF). We collected data on the treatment dosages and duration, adverse events (AEs), and spleen size by palpation and laboratory values. MF diagnoses were confirmed using the World Health Organization criteria.¹³ The baseline molecular and cytogenetic risks were determined using the Mutation and Karyotype-Enhanced International Prognostic Scoring System for PMF (MIPSS70+).¹⁴ Clinical response and progression were assessed using the revised International Working Group–Myeloproliferative Neoplasms Research and Treatment (IWG-MRT) response criteria.¹⁵ We assessed spleen response (SR50) as ≥50% relative reduction in palpated spleen size, molecular response (MR20) as ≥20% relative reduction in driver mutation allele frequency, and bone marrow response by reticulin score grade reduction as available. OS and PFS were compared using the Kaplan-Meier method and log-rank test to determine statistical significance. Fisher test was used to identify significant differences in AEs and discontinuation rates.

There were no significant differences between the IFN+RUX and RUX cohorts in terms of age, DIPSS+, white blood cell count, hemoglobin, platelets, peripheral blasts, disease duration at the start of treatment, MF type, and driver mutation (Table 1). Among the IFN+RUX patients, 16 received IFN first, followed by addition of RUX, 13 received RUX first, and 2 initiated both treatments simultaneously. IFN+RUX was initiated to improve clinical response (n = 20), for the hypothesized disease-modifying benefit (n = 8), or both (n = 3). The median follow-up duration was similar between the IFN+RUX and RUX groups (5.9 vs 6.6 years, respectively; P = .08). There was no significant difference in the median RUX doses between the IFN+RUX (median, 15 mg/d) and RUX (20 mg/d) groups at any treatment period (supplemental Figure 1). All patients received peginterferon alfa-2a at a median dose of 45 μg/wk throughout (range, 15-135).

The median treatment duration for IFN+RUX and RUX was 1.6 years (range, 0.3-8.2) and 1.7 years (range, 0.2-9.4), respectively. The discontinuation rate was not significantly different between the groups (n = 21, 68% vs n = 18, 58%). In the IFN+RUX group, 14 patients discontinued IFN first, 4 discontinued RUX first, and 3 discontinued both concurrently. Discontinuation in most was attributed to treatment related AEs (8 vs 9 patients in IFN+RUX vs RUX, respectively) or lack of response/disease progression (4 vs 5).

The median PFS was not reached in either cohort, and the 5-year PFS was 81% (IFN+RUX) and 85% (RUX) with no significant difference (Figure 1A). The median OS was 9.7 years (Figure 1B) with no significant difference observed. The proportion of patients who achieved any IWG-MRT response, including clinical improvement, PR, or CR, was comparable between treatment groups at all time points (supplemental Figure 2). Eleven patients (35%) in both groups showed clinical benefit (including clinical improvement, PR, or CR) at any time. Eventually, 14 patients had an allogeneic stem cell transplant (10 IFN+RUX vs 4 RUX; Fisher P = .13).

There were no significant differences in the SR50, MR20, or bone marrow response between the groups. The frequency of SR50 was comparable at 38% IFN+RUX (n = 21) and 30% RUX (n = 20; supplemental Figure 2B). The rates of MR20 were 33% IFN+RUX vs 22% RUX for the 26 patients who underwent molecular testing during treatment (supplemental Figure 2C). There was no significant difference in reticulin grade reduction at any time point, including 6 months before or after treatment, across the 47 patients with at least 1 marrow biopsy during treatment (supplemental Figure 3).

IFN+RUX was equally as well tolerated as RUX with no significant differences in the AEs while on treatment (supplemental Table 1). The most common treatment-emergent AEs included constitutional symptoms (55%), followed by thrombocytopenia (32%) and anemia (31%). Treatment related AEs that led to discontinuation among the IFN+RUX patients included anemia (n = 3), thrombocytopenia (n = 2), mood changes (n = 2), and neuritis (n = 1). In the RUX cohort, these included thrombocytopenia (n = 5) and anemia (n = 4).

Our case-control study of 62 patients with MF provides insight into the effectiveness of IFN+RUX in comparison with the standard of care RUX monotherapy in predominantly patients with intermediate-risk MF. We found comparable clinical response rates as earlier single-arm trials of IFN+RUX,^11,12 but these responses and the survival outcomes were not better than a matched cohort of RUX patients in our experience. There were no significant differences in the AEs or treatment discontinuation rates, supporting the safety profile for IFN+RUX, but the results fell short of providing evidence to justify the financial toxicity of the combination drug costs. However, we acknowledge our nonstandardized treatment approach in predominantly pretreated patients limits any definitive conclusions. The use of IFN+RUX may be justified for select patients; deep MR and SR were observed in patients unresponsive to RUX, and 1 patient experienced >60% reductions in both spleen length and mutation allele frequency, whereas another had >40% reductions in both. These patients were younger and had lower risk at the start of IFN+RUX (52/low risk and 62/intermediate-1 risk, respectively). Such cases highlight the potential of IFN+RUX and support its future investigation, while reaffirming the importance of targeted patient selection for combination therapy use. Although no significant differences in survival or responses were observed, the small cohort size, a third of whom had intermediate-2/high-risk MF, the add-on approach to IFN+RUX, and the high discontinuation rates limit the analysis of potential long-term clinical, molecular, and survival benefits of early and extended IFN+RUX treatment.

The scientific rationale of IFN+RUX, its similar tolerability as RUX observed in this study, and its safety and efficacy in earlier phase 1/2 studies justify further investigation of this combination in a phase 3 randomized clinical trial. For a randomized clinical trial, we suggest the focus should be on treatment-naïve, early-stage PMF or SMF with low to intermediate-1 DIPSS+, and indications for RUX (eg, symptomatic and splenomegaly). This is based on our favorable experience using IFN in early MF.^6,7 In addition to the standard endpoints of spleen response, symptom improvement, and molecular response, we should evaluate PFS/OS in this cohort with a higher likelihood of continuing therapy longer term. The study should be powered for OS as the primary endpoint in comparing these clinically effective, tolerable regimens. Determining the specific patient population that may benefit most and focusing on long-term survival outcomes is essential for future randomized trials of all combination therapies to ensure that the endpoint improvements outweigh the financial burden of such treatments.

Acknowledgments: The authors thank the Johns Family Fund of the Cancer Research & Treatment Fund Inc and MPN Peoria for funding support.

Contribution: G.A.-Z. conceptualized and designed the study; K.E. and A.A. collected data; K.E., A.A., and G.A.-Z. analyzed the data and designed the figure and table; and all authors interpreted data, wrote and reviewed the manuscript, and approved the final submitted version.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Ghaith Abu-Zeinah, Division of Hematology and Medical Oncology, Weill Cornell Medicine, 1300 York Ave, Box 113, New York, NY 10065; email: [email protected].