Abstract
Chronic graft-versus-host disease (cGVHD) remains a major barrier to a successful hematopoietic stem cell transplantation (HSCT); in cases refractory to first-line therapy with steroids, there is no standard of care for second-line therapy. As such, ruxolitinibis a promising drug in this scenario. We retrospectively analyzed its efficacy and safety as treatment of steroid-refractory cGVHD in 35 patients from two transplant centers, with the longest followup described to date. Evaluated patients had a median of 3 organs affected (1-7), with most (64%) having moderate cGVHD. The median number of previous therapy lines was two (1-6). Overall response rate was 89% (complete response, 26%) after a median of 4 weeks of therapy. Median follow-up was 43 months (range 11-59). At follow-up, of the 27 patients still alive, 18 patients (67%) are free of any immunosuppression, and 6 (22%) are using ruxolitinib as the only immunosuppressive drug. Failure-free survival was 77.1%, 68.6%, 54% and 51,4% at 6, 12, 24 and 36 months, respectively. Median overall survival (OS) was not reached. Toxicities were mostly hematological and resolved after dose reduction in most cases. In conclusion, our data, which represent the cohort of patients with cGVHD treated with ruxolitinib with the longest follow-up to date, support the use of this drug as a safe and effective option for refractory cGVHD.
Introduction
Graft-versus-host disease (GVHD) represents a major complication after allogeneic hematopoietic stem cell transplantation (HSCT) [1]. Chronic GVHD (cGVHD) is a significant cause of morbidity, increased risk of infection, and reduced quality of life (QoL) after transplantation, affecting approximately 40% of patients. First-line systemic therapy consists of systemic corticosteroids, and response rates are variable. Secondand further-line treatment options for steroid-refractory cGVHD are needed in 40-60% of cases, and most of the therapies available are less than ideal, with low response rates and significant toxicities [2].Ruxolitinib, a JAK1/2 inhibitor, appears to be a promising therapy for GVHD, as reported in a few reports with a small number of patients [3-6]. The mechanism of action seems to be related to suppression of the proinflammatory signaling that mediates tissue damage and by the promotion of tolerogenic regulatory T (T-reg) cells [7]. In preclinical models, the drug showed to reduce the incidence and severity of GVHD in vivo while preserving graft-versus-leukemia (GVL) effects [8].
In the context of acute GVHD (aGVHD), recently published phase 2 and phase 3 trials confirmed good responses, with overall response rates at day 28 between 50-60% [9, 10]. In the context of cGVHD, Zeiser et al recently presented preliminary results of a prospective trial comparing ruxolitinib versus best available therapy showing promising results – the best overall response rate was 76%, and overall response rates at week 24 was 50%, significantly better than the control group [11]. Retrospective studies describing the feasibility and efficacy of ruxolitinib use in refractory cGVHD, despite demonstrating overall response rates ranging from 40% to 80%, mostly have a limited median follow-up time, making a reliable assessment of duration of response and long-term outcomes difficult [5,6, 11-15].We have previously reported the efficacy of the JAK1/2 inhibitor ruxolitinib in corticosteroidrefractory cGVHD in 20 patients (BMT, 2018) [16]. Here, we report on an extended follow-up of 35 patients from two transplant centers. The aim of this analysis is to provide long-term efficacy and safety data.
Materials and Methods
Patients
The study was approved by the local ethics committees. All patients treated with ruxolitinib for steroid-refractory or steroid-dependent cGvHD in two transplant centers were analyzed. Diagnosis of cGVHD was performed according to the National Institutes of Health (NIH) consensus criteria [17]. Steroid-refractory and steroid-dependent cGVHD were defined as any disease that failed to respond to previous immunosuppressive therapy with steroids, at a dose of ≥ 0.5 mg/kg/day for at least 4 weeks, or were unable to taper it with or without additional immunosuppressive drugs.
Response Criteria and Assessments
Treatment responses were evaluated at one, three, six and twelve months after starting ruxolitinib, and every year thereafter. Responses were categorized according xenobiotic resistance to the NIH 2014 Criteria [18] as a complete response (CR), partial response (PR) or lack of response (unchanged, mixed response, or progression). CR to ruxolitinib is defined as a resolution of all disease manifestations in every organ or site involved. PR is defined as improvement in at least 1 organ or site, without progression. Lack of response is defined as disease progression in any organ or site, Epoxomicin or outcomes that did not meet criteria for CR or PR, or in which additional agents were used to control GVHD after ruxolitinib treatment. Overall response rate is defined as the proportion of patients who achieved either a complete or partial response at anytime point.Data were collected on common adverse events related to ruxolitinib, including cytopenias and infections. These were described according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE).
Statistical analysis
Probability of failure-free survival (FFS) and overall survival (OS) were calculated with the Kaplan-Meier method. OS data was assessed from the first dose of ruxolitinib until the last follow-up, or death. FFS was defined by the absence of other line of therapy for GVHD, nonrelapse mortality and disease relapse during ruxolitinib treatment. The software SPSS, version 21.0 (SPSS Inc., Chicago, IL, USA) R (version 3.5.0) and RStudio, (version 1.2.5033), were used for statistical analysis.
Results
Patients Characteristics
Our cohort was comprised of 35 patients receiving off-label ruxolitinib as salvage therapy for steroid-refractory or steroid-dependent cGVHD between March 2016 and May 2020. Median age was 54 years (range, 23 to 73 years). Patients underwent HSCT between January 2010 and September 2019, most commonly for an underlying diagnosis of acute myeloid leukemia (n=14), acute
lymphoblastic leukemia (n=7) and myelodysplastic syndrome (n=7). Detailed characteristics of the patients including gender, age, diagnosis, donor type, stem cell source, conditioning regimen and GvHD prophylaxis are illustrated in Table 1. The median time to ruxolitinib initiation after cGVHD onset was 17 months (range, 1 to 71 months) . Median followup after starting ruxolitinib was 43 months (range, 11 to 59 months).
Baseline Characteristics of cGVHD
According to the NIH 2014 severity classification, patients treated with ruxolitinib had moderate (23/35; 66%) or severe (12/35; 34%) cGVHD before ruxolitinib was started. Most patients (n = 22/63%) had steroid-refractory cGvHD, compared to 13 (37%) patients with steroid-dependent cGvHD. The median number of organ sites affected was 3 (range, 1-7). Skin was the most commonly involved organ, with 80% of patients (n=28) affected, followed by mouth (n=23/66%) and eyes (n=19/54%). The gastrointestinal (GI) tract, liver, lungs, genital tract and joints were affected in 23%, 20%, 20%, 17% and 9% of the patients, respectively. The median number of previous immunosuppressive lines of treatment before ruxolitinib was 2 (range, 1-6). At the time of ruxolitinib introduction, all 35 patients were sob steroid in varying doses (0.1 to 1mg/kg/day); in 8/35 patients, a second therapy was in use (calcineurin inhibitor in 3 patients, mycophenolate mofetil in 2, and methotrexate, etanercept and ibrutinib in each of the 3 remaining patients). Detailed data are illustrated in Table 1.Median dose of ruxolitinib was 15 mg, ranging from 5 mg up to 30 mg twice daily. Initial dose was 10 mg twice daily with progressive increase in cases of refractoriness or partial response and doses adjustments in cases of toxicity.
Efficacy
1. Response
Overall response rate was 89% (31/35), with 63% (22/35) PR and 26% (9/35) CR. Only four patients showed no response to ruxolitinib treatment. Median time to response was 4 (1–25) weeks after initiation of ruxolitinib. Among the 31 patients with a response to ruxolitinib, steroids were discontinued in 81% (n=25); in 19% (n=6), a reduction > 50% on the dose of the steroids was possible. The additional drugs present at ruxolitinib initiation were discontinued in all responding patients, in a median of 8 weeks (3-32). In 3 patients, additional therapies (steroid and methotrexate) were initiated before a clear response to ruxolitinib was observed – these patients presented a later response to therapy (12, 20 and 24 months), and the additional drug was suspended after the objective response to ruxolitinib.
Regarding 4 patients who did not respond to ruxolitinibat anytime, all had severe and steroidrefractory cGVHD, with a median number of 4 organs affected (three patients had severe lung involvement), and a median of five lines of therapy before ruxolitinib was started. In addition, median time since cGVHD start and ruxolitinib treatment was 21 months in these patients. Among those who responded to ruxolitinib, the median number of affected organs was 3, the median number of previous therapies was 2, and median time to ruxolitinib introduction was 7 months. The median duration of ruxolitinib treatment was 12 months (range: 2 to 51 months).Response rates across subgroups are illustrated in Figure 1.
Response rates across affected organs are illustrated in Figure 2. We found no significant differences in response rates in cGVHD involving skin, mouth, eye. We observed lower response rates in patients with lung involvement. Despite the small number of patients, cGVHD affecting the genital tract and joints seems to have higher rates of CR to ruxolitinib.At follow-up, among the 27 patients still alive, 6 patients (22%) are on treatment with ruxolitinib and remain in PR (5 cases) or in CR (1 case), while 21 patients (78%) stopped the drug, after achieving a sustained response (15 cases), changing to ibrutinib (2 cases), due to toxicity (2 cases), or due to non-availability of the drug (2 cases). Regarding the 15 patients in which ruxolitinib was suspended due to a sustained response, 8 achieved a CR and 7 achieved the criteria for PR. In these cases of PR, involvement related to GVHD was considered residual,and there would be no additional benefit to maintain the drug.At the time of this analysis, 18 patients (67%) are free of any immunosuppression, and 6 (22%) are in use of ruxolitinib as the only immunosuppressive drug.
Regarding 19 patients in whom ruxolitinib was discontinued for other reasons than absence of response /progression of cGVHD, GVHD relapse was observed in 7 patients, within a median of 5 months (range, 1-9) after drug suspension. Additional therapies in these patients were: ibrutinib (2 patients no responses), methotrexate (3 patients – partial response in 2 and no response in 1 patient) and rituximab (2 patients – both achieved a partial response). In 5 patients the drug was reintroduced, and a new objective response was achieved in all cases.FFS rate was 77.1%, 68.6%, 56% and 51,4% at 6, 12, 24 and 36 months, respectively, with a median of 27 months (Figure 3). Reasons for failure included initiation of a new systemic therapy (N: 8), relapse of the underlying disease (N: 2) and death (N: 7).
2. Toxicities and Infectious Complications
Hematological toxicities were commonly seen in our cohort: 11 patients (31%) had cytopenia using ruxolitinib. Neutropenia > grade 3 was observed in 5 patients, anemia > grade 3 was observed in 5 patients and thrombocytopenia > grade 3 was observed in 2 patients. In 7 cases, dose reduction led to resolution of cytopenia, while in 2 patients no dose modifications were made to avoid loss of the clinical benefit of the drug, with close monitoring of possible cytopenia-related symptoms.Overall, 2 patients permanently discontinued the drug due to refractory or recurrent cytopenia despite dose reductions.
During the follow-up time on ruxolitinib, bacterial infections occurred in 20% of patients, and at least one viral infection in 40%. There were 8 upper airway infections (half due to Influenza A), 6 cytomegalovirus reactivations (1 refractory to ganciclovir), 2 cases of herpes simplex with oral lesions, 1 case of herpes zoster and 1 case of polyomavirus hemorrhagic cystitis. Patients with herpes infections were on acyclovir prophylaxis. Four invasive fungal infections were diagnosed: 2 probable aspergilloses and 2 Pneumocystis jirovecii (PJP) pneumonia (both of the patients were on prophylaxis). No cases of mycobacteria disease or colonization were observed. There were 4 deaths attributable to infection – 1 due to PJP pneumonia and 3 due to bacterial infections.
3. Disease Relapse and Overall Survival
Relapse of the underlying disease occurred in 2 patients (6%) – one with a diagnosis of refractory acute myeloid leukemia with complex monosomal karyotype and the second with a Ph positive acute lymphoblastic leukemia. Both patients were on ruxolitinib at the time of relapse (use for about 2 years and 9 months, respectively), having achieved a CR regarding cGVHD. In one case, ruxolitinib was withdrawn at the time of relapse with no change in cGVHD signs and symptoms. In the second case, ruxolitinib was initially maintained, since GVHD was well controlled, but was also withdrawn after achieving a new CR and molecular remission with blinatumomab and the addition of ponatinib.OS at 43 months was 77%; median OS was not reached (Figure 4). Three patients died from severe cGVHD, 4 patients from infectious complications and 1 from ischemic heart disease.
Discussion
Ruxolitinib, through JAK1/2 inhibition, appears to inhibit multiple levels of the immune response involved in cGVHD. A clinical beneficial effect of ruxolitinib in refractory cGVHD has been described since 2015 with the retrospective series published by Zeiser et al [3]. Since then, retrospective reports have also shown good results, and recently, preliminary results from the REACH 3 study have brought more robust evidence for the use of ruxolitinib in this scenario. [11,12,13] The present series presents data on the longest follow-up reported to date, with a median of 43 months. Our results are slightly better than those previously reported – we found an overall response rate of 89% with most responses being partial. We observed good responses particularly in cases involving the skin, mouth, eyes, joints and genital tract. Organs such as the liver and GI tract appeared to have variable responses; in cases of lung involvement, the response to ruxolitinib was much less impressive. It seems that ruxolitinib should be introduced early in the course of the disease, as responses might be better when the drug is used as second-line therapy after steroid failure, or to allow for steroid tapering.
The prior retrospective studies that described the use of ruxolitinib in cGVHD demonstrated lower response rates, both overall and complete response. Zeiser et al [3], published the first experience with ruxolitinib in 41 patients with steroid-refractory cGVHD, showing an overall response rate of 85.4% (CR: 14.6%), with an excellent tolerability profile. Another cohort of 46 patients was reported by Modi et al., demonstrating lower overall response rates of 48% and 43% at 6 months and 12 months, respectively; the 1-year probability of treatment failure-free survival was 54.2% [12]. Recently, a multicenter Spanish survey reported results of 56 patients with refractory cGVHD on ruxolitinib after a median of 3 lines of therapy, with an overall response rate of 57.1% after a median of 4 weeks of treatment. Median follow-up of this cohort was around six months, and OS at 1 year was 81% [13]. Regarding rates of steroid dose reduction, the few studies that bring this information demonstrate rates around 57 to 89% [3, 6, 13] – our results seem to be similar, with all responding patients being able to reduce the dose or discontinue steroid. Our rate of steroid discontinuation of 81% is promising, and probably reflect the higher rate of CR reported in comparison to the other studies.
Primary results of the REACH3 Trial, a prospective phase 3 open-label and randomized study comparing ruxolitinib to best available therapy in steroid-refractory/steroiddependent cGVHD were presented an overall response rate at week 24 of 50% vs 26% was demonstrated, with longer failure-free survival and greater symptom improvement with ruxolitinib. The best overall response with ruxolitinib was 76% [11].Comparing with our results, we highlight a couple differences in patients and disease characteristics of the two cohorts. At first, in REACH 3, 59% of the patients had severe cGVHD (our cohort: 35%), 73% steroid-refractory GVHD and 27% steroid-dependent GVHD (our cohort: 63% steroid-refractory and 37% steroid-dependent GVHD). Also, we used higher doses (median dose 15 mg twice daily, maximum dose of 30 mg) than REACH 3, with good tolerance in most cases. These differences might have contributed to the better results we had.
Although this and other studies [5,6,12-14] show response rates between 40% and 80%, follow-up of patients is generally short, with a maximum of 12 months. Thus, duration of response, need for association of immunosuppressants and long-term outcomes are variables not yet well known with use of ruxolitinib for cGVHD. To our knowledge, the longest follow-up reported was the updated long-term results of the series by Zeiser at al. [4]. After a median follow-up of 24 months, the 1-year OS was 92.7%, with an estimated median OS not yet reached. GVHD relapse or progression after PR/CR was observed in 36% of the patients, and 24% of patients have an ongoing response and are free from immunosuppression. Response to retreatment with ruxolitinib or any immunosuppressive therapy was observed in 86% of cases. In our cohort, with a median follow-up of 43 months, 18 patients (67%) are free of any immunosuppression, and 6 (22%) are in use of ruxolitinib as the only immunosuppressive drug.
It is not clear for how long ruxolitinib should be used in patients who do respond, especially after achieving CR. Our initial idea was to use ruxolitinib for long periods of time, based on experience with former strategies, on the good objective response observed with the drug,and a favorable toxicity profile. However, in addition to possible adverse events and risk of
infections, there is a fear of long-term immunosuppression that may be associated to loss of the GVL effect and increased risk of leukemia relapse. Recently, in patients with CR, we tried to discontinue the drug about 1 year after its introduction. The median duration of ruxolitinib treatment in our cohort was 12 months. We observed a very low relapse rate (6%),
corroborating preclinical data for preservation of the GVL effect with the drug. Regarding loss of response after discontinuing ruxolitinib, ˜40% of the patients had a relapse of GVHD; in most cases, ruxolitinib was reintroduced and in all, a second response was achieved. It is also not well defined if there is a need for a drug taper and at what speed it should be performed. Following sociology of mandatory medical insurance the same approach in classical therapies for cGVHD, we tend to perform a slow taper over 3 to 6 months.Although most of the patients responding in our series achieved a PR (63%), in cGVHD this is an important endpoint, since QoL seems to be significantly improved, and many adverse events are eliminated by the withdrawal of steroids and other
immunosuppressive drugs.Limitations of our trial include the limited number of patients, the retrospective nature of the analysis, and fact that only two centers were included. Also, QoL was not formally evaluated.
In summary, our cohort of patients with cGVHD receiving ruxolitinib as salvage therapy in a real-world practice setting with the longest follow-up reported to date demonstrated an overall response rate of 89%, with an OS of 75% in a follow-up of 43 months. These favorable results with an oral drug with minimal adverse events make ruxolitiniba promising therapeutic option. Costs remain a major concern, especially in developing countries and in public health systems. Randomized trials comparing ruxolitiniband other second-line therapies are currently underway and the final results of REACH3 study are expected to confirm evidence regarding the true efficacy and impact of ruxolitinib in steroid-refractory cGVHD.
