Evaluation of Ruxolitinib for Steroid-Refractory Chronic Graft-vs-Host Disease After Allogeneic Hematopoietic Stem Cell Transplantation

Key Points Question Is ruxolitinib an option for patients with steroid-refractory chronic graft-vs-host disease, and what characteristics are associated with treatment response? Findings In this case series of 41 patients with steroid-refractory chronic graft-vs-host disease who were treated with ruxolitinib, heavily pretreated patients could achieve meaningful responses with a favorable safety profile. No lung involvement and haploidentical donors were associated with response to ruxolitinib. Meaning In this study, monotherapy with ruxolitinib was associated with a meaningful response in patients with steroid-refractory chronic graft-vs-host disease, suggesting a possible therapeutic option for a serious disease with no currently accepted standard-of-care treatment.


Introduction
Chronic graft-vs-host disease (cGVHD) is the leading cause of late morbidity and mortality as well as impaired quality of life after allogeneic hematopoietic stem cell transplantation (HSCT). 1 Despite the use of standard prophylaxis, 35% to 70% of recipients develop cGVHD. 2 Established first-line therapy for cGVHD still comprises corticosteroids and calcineurin inhibitors. 3 Approximately half of patients with cGVHD are refractory to corticosteroid therapy. For various second-line therapies or interventions, the response rates range from 30% to 60%, 4,5 and no consensus has been reached regarding the optimal salvage treatment for steroid-refractory (SR)-cGVHD. Therefore, it is essential to identify a promising therapeutic drug for the adequate therapy of SR-cGVHD.
There is preclinical evidence that the pharmacologic inhibition of Janus kinases 1/2 (JAK1/J2) prevents GVHD by blocking interferon-γ receptor and interleukin 6 receptor signaling. JAK1/2 inhibition does not impair donor T-cell alloreactivity. It preserves the beneficial graft vs leukemia effect in vivo, suggesting a potential role for the selective JAK1/2 inhibitor ruxolitinib in the prevention and treatment of GVHD. 6,7 A multicenter retrospective study in centers throughout Europe and the United States first established the feasibility of ruxolitinib for SR-acute GVHD (aGVHD) treatment. 8 Thereafter, the Ruxolitinib for the Treatment of Steroid-Refractory Acute GVHD (REACH-1) study 9 investigated ruxolitinib in treating SR-aGVHD and showed that the overall response rate (ORR) on day 28 was 54.9%. This was the first time such a study was conducted prospectively, and it shed light on later randomized clinical trials. 9 Subsequently, a prospective study showed that combining ruxolitinib with another agent could achieve higher ORRs. 10 At present, there is no standard of care for treating SR-cGVHD because of a lack of available and substantial data from prospective studies. The efficacy of ruxolitinib vs current best available therapy in SR-cGVHD is being evaluated in an ongoing phase 3 trial (REACH-3; NCT03112603).
Based on the limited preclinical and clinical outcomes of ruxolitinib in GVHD published to date, the present study investigated the clinical response to ruxolitinib and its safety profile in patients with SR-cGVHD after allogeneic HSCT. The study aimed to provide informative data for SR-cGVHD management and to measure the benefits and risks for different patient groups.

Methods
This retrospective, single-center case series included 41 consecutive patients who were treated with ruxolitinib for SR-cGVHD after allogeneic HSCT between August 2017 and December 2019 at the First Affiliated Hospital of Zhejiang University School of Medicine. Informed consent was obtained from all recruited patients before ruxolitinib treatment and data collection. For safety and response evaluation, all data were collected from clinical history at the First Affiliated Hospital of Zhejiang University School of Medicine. The study was reviewed and approved by the ethics committee of the First Affiliated Hospital of Zhejiang University School of Medicine. This study followed the reporting guideline for case series. cGVHD according to NIH criteria within 4 weeks after administration, ruxolitinib would be withdrawn.
Response evaluation was conducted according to clinical status in the sixth month after the first ruxolitinib administration. The ORR included complete response (CR) and partial response (PR). CR was defined as the absence of all manifestations of cGVHD; PR was defined as an improvement in cGVHD compared with baseline clinical status and stage according to the NIH consensus and without any progression in any organs or sites. Other situations, including stable disease (SD), defined as no changes, and progressive disease (PD), defined as worsening in at least 1 site or organ, were categorized as treatment failure; discontinuation because of toxic effects from ruxolitinib was not included.
We collected and analyzed the following data: (1) adverse events, such as infections and cytopenia based on NCI-CTCAE version 4.0, were clinically relevant as grade 2 or higher; (2) time to response, defined as the initial use of ruxolitinib to initial response; (3) nonrelapsed mortality (NRM), which was defined as the initial treatment of ruxolitinib until death from any cause except underlying malignant neoplasm relapse or recurrence; (4) overall survival (OS), which was defined as the initial treatment of ruxolitinib until death from any cause; (5) cumulative incidence of cGVHD flare, which was defined as the initial use of ruxolitinib until cGVHD progression; and (6) the cumulative relapse of underlying malignant neoplasm, which was defined as the initial use of ruxolitinib until the first relapse. Patients who were lost to the last follow-up were censored.

Statistical Analysis
Data were analyzed using SPSS statistical software version 22.0.01 (IBM Corp). A 2-tailed P < .05 was considered statistically significant. Univariate comparisons of parameters were performed using the χ 2 test, Fisher exact test, and t test, as appropriate. Variables with P < .20 in univariate analysis were entered into the multivariate model. OS was estimated and plotted using the Kaplan-Meier method.
The log-rank test was applied to compare Kaplan-Meier curves. The proportional-hazards method was used to estimate the cumulative incidence of relapse and NRM. Relapse and NRM were competing risks for each other. R statistical software version 3.4.3 (R Project for Statistical Computing) was used for the competing risk analysis.

Discussion
The treatment strategies for patients with SR-cGVHD have changed. Multiple immunosuppressive agents provide clinicians with new approaches for treating patients with SR-cGVHD with few guidelines and a lack of consensus. Previous studies showed that the ORR of ruxolitinib in refractory cGVHD treatment was approximately 43.5% to 100%, and the CRR was from 3.5% to 13.0%, with a median time to best response ranging from 2 to 4 weeks. 8,[12][13][14] In the present study, the ORR was 70.7% (95% CI, 56.2%-85.3%), and the CRR was 36.6% (95% CI, 21.2%-52.0%). The median time to reach the best response was 2.0 months.
Ruxolitinib showed potential to resolve cGVHD. This study found that compared with patients with moderate cGVHD and less organ involvement, patients with severe and multiple involved organs had similar treatment outcome in terms of not only ORR but also the time to achieve response. Importantly, the treatment response showed no significant difference in patients with different previous lines of second-line agents and the time from cGVHD diagnosis to receiving ruxolitinib. The aforementioned scenario might suggest the use of ruxolitinib in patients with cGVHD regardless of cGVHD severity, the numbers of involved organs, the duration of cGVHD course, or the intensity of previous pharmacological therapies.
A primary aim of this study was to examine the factors associated with treatment response for ruxolitinib. The statistical results showed an association between lung cGVHD and treatment response, which was also reported in a previous study. 13 The multivariate analysis indicated that among patients with cGVHD, lung involvement was associated with a higher risk of treatment failure.
The underlying mechanism of this association has not yet been elucidated. However, ensuing pulmonary fibrosis, symbolized with myofibroblast hyperplasia, is promoted by macrophages, B-cells, and complicated networks of other cells. 15 A 3-phase model was developed for cGVHD in which the third phase is hypothesized to be due to the excessive accumulation of extracellular matrix, causing abnormal fibrosis. 16 For example, interleukin 21 (IL-21) promotes the differentiation of B-cells into plasma cells via the JAK/signal transducer and activator of transcription 3 (STAT3) pathway, leading to antibody secretion and deposition. 17 As an inhibitor of JAK-STAT signaling, ruxolitinib interferes with the activation and differentiation of T-cells and suppresses the activity of macrophages. 18  The cohort showed that male donors, CR before transplantation, skin cGVHD, and moderate cGVHD were associated with prolonged survival. This study found that patients receiving peripheral blood stem cells (PBSC) from male donors had more skin cGVHD events.
The estimated OS was compared between patients with and without a response. That the estimated OS plots were similar might ignore the fact that patients who responded had a higher quality of life, without or with fewer cGVHD events. A recent study demonstrated that the development of cGVHD was not associated with OS. 22 In other words, survival time was comparatively fixed despite the resolution of cGVHD with the treatment of ruxolitinib. In this study, 23 patients had infections, including 11 with lung infection, 5 with CMV DNAemia, 19 with EBV DNAemia, 2 with sepsis, and 1 with HBV reaction. The latest REACH-1 study 9 on aGVHD showed that infection events during ruxolitinib were approximately 80.3%, of which CMV events were the most common (19.7%). Zeiser et al 8 found that the incidence of CMV events was higher in SR-aGVHD than in SR-cGVHD (33.3% vs 14.6%). Collectively, infection in patients with SR-cGVHD was tolerated, but 4 of 9 patients died of infection-related complications, highlighting the importance of antibacterial, antifungal, and antiviral prophylaxis.
Regarding cytopenia, the most common adverse effects related to ruxolitinib in the REACH-1 study were anemia (35.2%), thrombocytopenia (32.4%), and neutropenia (26.8%). The study by Zeiser et al 8 showed that 5 patients with SR-aGVHD were more likely to develop global cytopenias and severe cytopenias compared with patients with cGVHD during the ruxolitinib course. The overall frequency of cytopenias in this study was only 14.6%, and grade 3 to 4 cytopenia occurred in 3 cases (7.3%), which was lower than the values reported in the study by Zeiser et al 8

Limitations
This study has limitations. First, this single-center study lacked sufficient participants and death events to perform a multivariate analysis. For example, we found no statistical significance of OS between patients who did and did not respond. The limited data size dramatically interfered with the P value, which might lead to nonsignificant results, and patients from a single center might contribute to survival bias. It is important to emphasize that because of the study's observational and retrospective nature, the results should be interpreted with caution. Second, some results failed to provide key insights. While we found patients with haploidentical donors might receive more benefit from ruxolitinib compared with matched related donors, the underlying mechanism for this phenomenon remains unclear. Third, although we highlight that ruxolitinib as a single agent could lower expenditure, we could not measure the actual costs of ruxolitinib regarding its dosage and course of treatment.

Conclusions
This case series found that ruxolitinib in patients with SR-cGVHD had an ORR of 70.7% and CRR of 36.6%. Nearly all patients reduced the dose or discontinued the use of concomitant corticosteroids and other immunosuppressive drugs, minimizing their side effects and cost burden. Despite the limited sample size and retrospective nature, the results of this study indicated that patients with no lung involvement and haploidentical relatives as donors were more likely to benefit from ruxolitinib.
Regarding the safety profile, the present study showed that infection events were the most severe adverse effect related to ruxolitinib, highlighting the significance of infection prophylaxis.