CADTH Reimbursement Review

Ruxolitinib (Jakavi)

Sponsor: Novartis Pharmaceutical Canada Inc.

Therapeutic area: Acute graft versus host disease

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Submitted for Review

aGvHD = acute graft-versus-host disease; NOC = Notice of Compliance; SR = steroid refractory.

^aThe reimbursement request differs from the Health Canada–approved indication, in that it states “steroid refractory or dependent” rather than “inadequate response to corticosteroids or other systemic therapies.”

Source: Product monograph.¹

Introduction

Hematopoietic stem cell transplants (HSCTs) provide stem cells to patients whose bone marrow has been destroyed by disease, chemotherapy, or radiation.² The 2 main types of stem cell transplants (SCTs) are autologous and allogeneic (alloSCTs). Although alloSCT has curative potential, there is a risk that the donor’s stem cells could die or be destroyed by the patients’ body before settling in the patient’s bone marrow, or the donor’s immune cells may attack healthy cells in the patient’s body; the latter is called graft-versus-host disease (GvHD).² GvHD is a multi-system disorders in which donor-derived immune cells initiate an adverse immune reaction to tissues, cells, and organs in the transplant recipient, leading to tissue damage, organ failure, or death.³ GvHD has been found to be the leading cause of morbidity and nonrelapse mortality (NRM) in patients after alloSCT, and affects up to 70% of patients who receive SCTs.^3,4 From 2008 to 2019, an estimated 13,033 transplants were performed in Canada, of which 5,672 were alloSCTs.⁵

Acute GvHD (aGvHD) typically occurs within 100 days after alloSCT, and typically affects the skin, liver, and intestines.⁶ aGvHD occurs in 30% to 50% of patients who undergo alloSCT, with 14% to 36% developing severe aGvHD (i.e., grade III to IV).⁷ Prognosis among patients with aGvHD is less favourable, with an estimated 3-year survival rate of 54%.⁸ Common symptoms of aGvHD are skin rash, burning and redness of the skin on the palms of the hands or soles of the feet, blisters and peeling skin, diarrhea, persistent nausea and vomiting, cramping or abdominal pain, enlarged liver, liver tenderness, abnormal liver enzymes or liver failure, and increased levels of serum bilirubin.⁹

The choice of initial treatment for patients with aGvHD depends on multiple factors, including the severity of symptoms, the type of prophylactic regimen used, and the importance of a graft-versus-tumour (GVT) effect.⁷ Grade I aGvHD (i.e., skin involvement of 50% or less of the body surface area without liver or gastrointestinal [GI] tract involvement) is often managed with topical treatments (e.g., tropical steroids) and adjustment of prophylactic treatments (e.g., calcineurin inhibitors [CNIs], such as cyclosporine in combination with methotrexate). Patients with grade II or higher aGvHD receive high-dose systemic glucocorticoids (e.g., methylprednisolone or prednisone) in addition to the care received by patients with grade I aGvHD. In patients who respond, steroid treatment is continued for several weeks until gradually being tapered to prevent a flare of aGvHD. Approximately 25% to 40% of patients achieve complete responses with glucocorticoids⁷; however, treatment is associated with significant side effects, which affect patients’ quality of life and increase susceptibility to infection.¹⁰ Approximately 25% to 50% of patients become refractory to steroids and are considered to have steroid-refractory (SR)-aGvHD. There is currently no standardized second-line therapy for patients with aGvHD who have an inadequate response to steroids in Canada.

In the absence of sufficient evidence to guide second-line treatment selection, factors that influence the choice include the experience of the treating physician, the types of aGvHD prophylaxis used, and the risk of potential toxicities and worsening pre-existing comorbidities.⁷ According to the clinical experts consulted by CADTH, available second-line options in Canada include extracorporeal photopheresis (ECP), mycophenolate mofetil (MMF), etanercept, infliximab, mammalian target of rapamycin (mTOR) inhibitors (i.e., everolimus or sirolimus), antithymocyte globulin (ATG), and interleukin-2 receptors. It was noted by the clinical experts that ATG is also often used as prophylaxis rather than aGvHD treatment. The clinical experts expressed challenges with currently available therapies in this heavily pre-treated target population, including no responses or only PRs (response rate was estimated by the clinical experts to be approximately 50%, but the durable response rate, in which patients still show a response at about 2 months, to be only 10% to 30%). According to the clinical experts, responses in this patient population are important because they enable the tapering of steroids, which mitigates long-term side effects (e.g., osteoporosis, hypertension, hyperglycemia, diabetes, and bone and joint health) and risk of infection. It was emphasized by the clinical experts that opportunistic infection and organ damage from aGvHD are major causes of NRM in SR-aGvHD.

There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profile that improve health related quality of life (HRQoL), reduce disease symptoms of aGvHD, enhance patient’s performance status, and improve overall survival. They noted that a convenient oral route of administration would help improve adherence and reduce hospital-based resource use.

Ruxolitinib is a Janus-associated kinase (JAK) inhibitor which has received market authorization by Health Canada for the treatment of SR or dependent aGvHD in adult and pediatric patients 12 years and older. Ruxolitinib underwent review by Health Canada through the expedited pathway (Orbis). The sponsor’s requested reimbursement criteria for ruxolitinib for aGvHD differ from the Health Canada indication in that the sponsor’s criteria state, “inadequate response to corticosteroids or other systemic therapies” rather than “steroid refractory or dependent” aGvHD.

This current CADTH review focuses on aGvHD. Ruxolitinib is concurrently being reviewed by CADTH in the chronic GvHD (cGvHD) setting for the treatment of cGvHD in patients 12 years and older who have an inadequate response to corticosteroids or other systemic therapies. Ruxolitinib received a positive conditional CADTH recommendation in March 2016 for the treatment of patients with polycythemia vera (PV) who are resistant to or intolerant of hydroxyurea, according to the modified European LeukemiaNet Criteria used in the RESPONSE trial, and who have a good performance status. Ruxolitinib is available as 5 mg, 10 mg, 15 mg, and 20 mg tablets. The recommended starting dose for ruxolitinib for aGvHD is 5 mg administered orally twice daily. An increase in dose to 10 mg twice daily is recommended after at least 3 days of treatment if the absolute neutrophil count (ANC) and platelet count are not decreased by 50% or more relative to the first day of dosing with ruxolitinib. The product monograph states that tapering of ruxolitinib in the setting of aGvHD may be considered in patients with a response and after they have discontinued corticosteroids. A tapering of ruxolitinib — reducing the dose to 50% every 2 months — is recommended; in the event that signs or symptoms of GvHD reoccur during or after the taper, re-treatment with ruxolitinib should be considered.¹

The objective of this CADTH report is to perform a systematic review of the beneficial and harmful effects of ruxolitinib (10 mg twice daily oral tablets) for the treatment of aGvHD in patients 12 years and older who have an inadequate response to corticosteroids or other systemic therapies.

According to the information in the product monograph and confidential information in the Health Canada Reviewer Report, Health Canada considered reviewing REACH 1 as the pivotal study and using the safety data from REACH 2 as supportive evidence for the aGvHD indication.^1,11 This decision is the result of uncertainties around the REACH 2 trial that were identified by the FDA upon the review of raw data from REACH 2 as part of a sponsor-proposed label update of the FDA-approved ruxolitinib indication for aGvHD, which was based on REACH 1 trial data. The sponsor subsequently withdrew the supplemental New Drug Application for the REACH 2 update; in its submission, the sponsor noted that the proposed label update could not be completed based on REACH 2 trial data. Specifically, the 3 issues identified by the FDA concerned the following: missing and incomplete information in the case report forms related to organ staging of aGvHD, which precluded the FDA from conducting their own analyses and confirming the investigators’ efficacy assessments; the frequency of protocol deviations related to investigator-determined organ staging; and the information regarding prior therapies (i.e., prophylactic versus therapeutic aGvHD treatments). Health Canada concluded that as a result of the “integrity issues in the REACH 2 study” identified by the FDA, the efficacy results of REACH 2 could not be “verified.”¹¹ Therefore, Health Canada considered REACH 1 to be the pivotal study, and safety and pharmacokinetic data from REACH 2 to be supportive evidence in their assessment of the aGvHD indication. The sponsor clarified with CADTH that the initial decision to submit the REACH 2 trial as the pivotal trial for the proposed CADTH reimbursement request remains unchanged.³ This CADTH review report does not include a review or critical appraisal of the issues raised by the FDA, as CADTH has no access to the FDA’s assessment, and reviewing raw data is not in the mandate of CADTH.

Health Canada based its efficacy assessment of the REACH 1 trial on the FDA-evaluable population (49 patients who failed steroids alone), which is a subset of the REACH 1 trial’s full efficacy-evaluable patients (N = 71). According to the Health Canada Clarifax, the following were excluded from the FDA-evaluable population: 9 patients who failed steroids plus additional therapies before randomization, and 13 patients who did not meet the SR criteria because of suboptimal dosing or duration of corticosteroid treatment before randomization.¹² The HC Clarifax noted that the available data are considered insufficient to support the proposed ruxolitinib indication for patients with aGvHD who failed 1 systemic treatment in addition to corticosteroids (± CNIs).¹² Of note, this CADTH clinical review does not report on the subset of 49 patients in the FDA-evaluable population; it presents results for all 71 efficacy-evaluable patients from the REACH 1 trial. Results for the 49 patients in the FDA-evaluable population are reported in the product monograph.¹

The Health Canada Reviewer Report stated that the assessment of the safety of ruxolitinib for aGvHD was informed by the full safety-evaluable population (N = 71) of the REACH 1 trial and by safety results observed in the REACH 2 trial.¹ Safety results reported in this CADTH clinical review are based on the 71 patients in the REACH 1 trial and on safety results from the REACH 2 trial. The Health Canada Reviewer Report noted that patients with SR or steroid-dependent aGvHD who meet the eligibility criteria in REACH 1 are considered acceptable.¹¹ The REACH 1 eligibility criteria for SR and steroid-dependent aGvHD align with the current National Comprehensive Cancer Network guidelines for hematopoietic cell transplant, version 2.2020,¹³ the Alberta Bone Marrow and Blood and Cell Transplant Program Standard Practice Manual (2021),¹⁴ and the European Society for Blood and Marrow Transplantation (EBMT)-National Institutes of Health (NIH)-Center for International Blood and Marrow Transplant Research (CIBMTR) Task Force position statement,¹⁵ with minor variations. Steroid dependence has been defined as the inability to taper prednisone under 2 mg/kg per day after initially successful treatment for at least 7 days or as the recurrence of aGvHD activity during steroid taper.¹¹

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups that responded to CADTH’s call for patient input and by clinical experts consulted by CADTH for the purpose of this review.

Patient Input

Eight patient groups — Lymphoma Canada, the Lymphoma & Leukemia Society of Canada, Chronic Lymphocytic Leukemia Canada, Myeloma Canada, the Aplastic Anemia & Myelodysplasia Association of Canada, Canadian Myeloproliferative Neoplasms Research Foundation, the Canadian Chronic Myelogenous Leukemia (CML) Network, and Cell Therapy Transplant Canada (CTTC) — created 1 joint patient input document for this review. The input was based on an online survey, and responses from 68 patients were reported in the patient input. Sixty patients reported having received an SCT, 6 patients reported not having received an SCT, and 2 patients did not answer the question. Of the 60 patients who received an SCT, 49 reported having received an alloSCT. Fifty-three patients had experienced GvHD after their SCT. Data on the type of GvHD were available for 45 of the 53 patients with GvHD: 13% experienced aGvHD, 24% experienced cGvHD, and 62% experienced both acute and chronic GvHD. Twenty patients reported receiving ruxolitinib treatment.

Overall, survey respondents indicated that they had long-lasting GvHD symptoms (3 to 5 years for 26% of respondents, and more than 5 years for 28% of respondents). To manage their GvHD, the respondents reported requiring numerous medical consultations, hospital stays, and nights away from home. Respondents reported a range of GvHD symptoms that had a significant impact on daily activities and were detrimental to quality of life. Respondents highlighted problems with interruption of life goals and accomplishments (career, school), difficulty sleeping, and impacts on their mental health (stress, anxiety, worry, and concentration problems), and finances. Other common symptoms experienced by respondents included burning and redness of the skin on the palms of the hands or soles of the feet, rashes that could spread over the entire body, blisters and peeling skin, skin problems (such as dryness, rash, itching, peeling, darkening, hard texture, and feeling tight), enlarged liver, liver tenderness, abnormal liver enzymes or liver failure, jaundice, dry eyes that may have a burning or gritty feeling, dry mouth with or without mouth ulcers, diarrhea, loss of appetite, stomach cramps, vomiting, weight loss, pain in muscles and joints, mobility issues, infections, and difficulty breathing.

According to the patient input received, respondents expected new drugs or treatments to address the following key outcomes: OS, GvHD symptoms, quality of life, and severity of side effects. Additionally, the administration of treatment in the outpatient setting (rather than requiring an overnight hospital stay), access to treatment locally (rather than requiring extensive travel), coverage of treatment by insurance or drug plans, and treatments recommended by health care professionals were perceived to be very important by respondents. Respondents who had direct experience with ruxolitinib indicated that, overall, ruxolitinib was an effective treatment, improved their quality of life, and had tolerable side effects, and they would take again if recommended by their physician and would recommend it to other patients.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical experts consulted by CADTH noted that there are currently no Health Canada–authorized standard-care regimens specifically for patients with SR-aGvHD in Canada. According to the clinical experts consulted by CADTH, available second-line options in Canada include ECP, MMF, etanercept, infliximab, mTOR inhibitor (e.g., sirolimus or sirolimus), and ATG. It was noted by the clinical experts that ATG was often used as prophylaxis rather than a treatment for aGvHD. There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve HRQoL, reduce symptoms of aGvHD, enhance performance status, and improve OS. They added that a convenient oral route of administration would help improve adherence and reduce hospital-based or ambulatory centre resource use. They stated that ruxolitinib is used as add-on to immunosuppressive regimens of corticosteroids with or without CNIs in patients 12 years and older with grades II to IV SR-aGvHD as per the REACH 2 trial. They agreed that ruxolitinib, as a therapy for SR-aGvHD, would likely shift the current treatment paradigm. The clinical experts consulted by CADTH agreed that patients who meet the inclusion and exclusion criteria from the REACH 2 trial should be eligible for ruxolitinib therapy. The clinical experts identified patients with grade IV aGvHD, who have the highest risk of death from aGvHD, as being most in need of ruxolitinib therapy. Patient subgroups that would potentially benefit least from ruxolitinib may include patients with refractory vomiting or ileus who are not able to take an oral drug such as ruxolitinib, and patients with thrombocytopenia, especially those with clinical bleeding, who may be a challenge to treat with ruxolitinib and may receive an alternative second-line drug instead. Patients with active uncontrolled infections or non-aGvHD cytopenia are a challenge to treat with ruxolitinib or with other available second-line therapy options; ruxolitinib should be used with caution and may require dose adjustment in these patients. The clinical experts consulted by CADTH felt that it would be reasonable to generalize the REACH 2 trial results to patients who received 2 or more systemic treatments for aGvHD and to leave it to the discretion of the treating physician to apply some flexibility in terms of using ruxolitinib in patients with overlap syndrome and patients with grade I aGvHD.

In the opinion of the clinical experts consulted by CADTH, an accurate assessment of response in patients with aGvHD is based on the NIH consensus criteria, which was used in the REACH 2 trial. Response to treatment is usually assessed daily for inpatients and weekly for outpatients. The clinical experts indicated that the most clinically meaningful responses to treatment include improvements in OS (survival beyond 1 year after alloSCT), overall response (complete or partial response), improvements in HRQoL and performance status, and the ability to taper corticosteroids.

Clinician Group Input

Two clinician groups provided input: CTTC, which contained input from 8 clinicians; and Ontario Health Cancer Care Ontario Complex Malignant Hematology, which contained input from 2 clinicians. The views of the clinician groups were, overall, consistent with the clinical experts consulted by CADTH, indicating that, based on the evidence from the REACH 2 trial, ruxolitinib would likely become the dominant first-line therapy for SR-aGvHD. The outcomes assessed in the REACH 2 trial were judged to be applicable to Canadian clinical practice and reflective of clinically meaningful responses. It was noted by both input groups that ruxolitinib is not as immunosuppressive as other available therapies. Clinicians from Ontario Health Cancer Care Ontario noted the following drawbacks of currently available therapies: IV administration, which requires patients to be at the hospital; side effects and broad immune suppression; and the high price and delivery costs of treatment. Input from CTTC noted that a Health Canada–approved and provincially funded therapy for SR-aGvHD would be an important step forward in the target setting, because existing therapies offer low response rates and high rates of toxicity. According to input from CTTC, experience with ruxolitinib (accessible through compassionate access programs) and real-world effectiveness appear similar to what was observed in the REACH 2 trial, which had low rates of toxicity.

Drug Program Input

The Formulary Working Group identified the following jurisdictional implementation issues: relevant comparators, consideration for initiation of therapy, consideration for discontinuation of therapy, considerations for prescribing of therapy, and system issues and economic considerations. The clinical experts consulted by CADTH weighed evidence from the REACH 2 trial and considered the REACH 1 trial and other clinical considerations to provide responses to the Provincial Advisory Group’s drug program implementation questions. Refer to Table 4 for more details.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

The REACH 2 trial is a completed, international, multi-centre, open-label, randomized, phase III trial that compared ruxolitinib (10 mg administered orally twice daily) with investigator’s choice of best available therapy (BAT) — i.e., ATG, ECP, mesenchymal stem cells (MSCs), methotrexate, MMF, mTOR inhibitors (everolimus or sirolimus), etanercept, or infliximab — in patients 12 years and older with grade II to IV SR-aGvHD. Patients continued to receive their systemic immunosuppressive regimen of corticosteroids with or without CNIs. Staging of aGvHD was based on the NIH criteria of Harris et al. (2016).¹⁶ A total of 309 patients were randomized in a 1:1 ratio to receive ruxolitinib or BAT. The primary outcome was overall response rate (ORR) at day 28, and the key secondary outcome was the rate of durable ORR at day 56. Additional secondary outcomes were OS, failure-free survival (FFS), ORR at day 14, duration of response (DOR), best overall response (BOR), HRQoL assessed with the Functional Assessment of Cancer Therapy-Bone Marrow Transplant (FACT-BMT) and the 5-level EQ-5D (EQ-5D-5L) instruments, event-free survival (EFS), NRM, incidence of malignancy relapse or progression, cumulative steroid dose up to day 56, incidence of cGvHD, and safety.

The REACH 1 trial is a completed, open-label, single-arm, multi-centre phase II trial that evaluated the efficacy and safety of ruxolitinib in combination with corticosteroids in patients with grade II to IV SR-aGvHD. The severity grading of aGvHD was based on the NIH criteria by Harris et al. (2016).¹⁶ A total of 71 patients were enrolled to received ruxolitinib 5 mg orally twice daily; then, if hematologic parameters were stable and no treatment-related toxicity was observed after the first 3 days of treatment, the dose could be increased to 10 mg orally twice daily. The primary outcome was ORR at day 28 and the key secondary outcome was DOR at month 6. Additional secondary outcomes were OS, FFS, ORR at day 14, DOR at month 3, NRM, incidence of malignancy relapse or progression, relapse rate, relapse-related mortality rate, and safety.

The REACH 2 trial enrolled male and female patients 12 years and older who had undergone alloSCT, had evidence of myeloid and platelet engraftment (ANC > 1,000/mm³ and platelet count > 20,000/mm³), and were diagnosed with grade II to IV aGvHD, defined by the NIH consensus criteria, that was determined to be corticosteroid-refractory, according to the protocol-defined criteria. Overall, the REACH 1 trial had similar inclusion criteria. However, there were slight variations in the definition of corticosteroid refractoriness (criterion C, as outlined in Table 4) and engraftment. Both studies excluded patients who had received more than 1 systemic treatment for SR-aGvHD, a clinical presentation resembling de novo overlap syndrome (defined by Jagasia et al. [2015])¹⁷ or active uncontrolled infection. REACH 2 explicitly excluded patients with multi-focal leuko-encephalopathy, whereas REACH 1 did not. In the REACH 2 trial, the mean ages for the ruxolitinib and BAT groups were, respectively, 48.1 (standard deviation [SD] = 16.30) and 50.9 (SD = 14.9) years. The majority of patients (ruxolitinib versus BAT groups) were 18 to 65 years of age (83.1% versus 81.3%), but a few were 12 years to younger than 18 years (3.2% versus 2.6%). Most patients were male (59.7% versus 58.7%). The aGvHD grade at baseline was mostly grade III (44.2% versus 43.9%), followed by grade II (32.5% versus 34.8%) and grade IV (19.5% versus 20.6%). The most common criteria for SR-aGvHD were failure to achieve a response after 7 days (46.8% versus 40.6%), failure on steroid taper (30.5% versus 31.6%), and progression after at least 3 days (22.7% versus 27.7%).¹⁸ The ruxolitinib group had a higher proportion of patients with aGvHD organ involvement at baseline of the skin (60.4% versus 47.7%) and liver (24.0% versus 16.1%), and a lower proportion of patients with aGvHD organ involvement at baseline of the upper GI tract (18.2% versus 23.9%) and lower GI tract (62.3% versus 74.2%). Demographic characteristics and disease and alloSCT history at baseline in the REACH 1 trial were similar, overall, to those in the REACH 2 trial. As in the REACH 2 trial, the majority of patients in the REACH 1 trial were 18 to 65 years (81.7%), and the distribution of aGvHD grades was similar in the 2 trials, with the majority of patients having grade III aGvHD (46.5%), followed by grade II (31.0%) and grade IV (22.5%). As in REACH 2 trial, the most common criteria for SR-aGvHD were no aGvHD improvement after 7 days of primary treatment (40.8%), failing corticosteroid taper (36.2%), and progression after 3 days or primary treatment (23.9%). Most patients in the 2 trials received grafts from identical human leukocyte antigen (HLA)-matched donors (60.2% in REACH 2 and 63.4% in REACH 1).¹⁹

This CADTH review is based on REACH 2 data from the primary analysis (July 25, 2019), the updated secondary analysis (January 6, 2020), and the final analysis (April 23, 2021), which was conducted once all patients had completed the study.²⁰ The REACH 1 data in this CADTH review are based on the final data cut-off date of June 5, 2019, which is also the study completion date.²¹

Efficacy Results

The key efficacy results from the REACH 2 and REACH 1 trials are summarized in Table 2.

As of the primary analysis, the median duration of follow-up for OS in the REACH 2 trial was 5.04 months in the ruxolitinib group and 3.58 months in the BAT group. Median OS was 11.14 months, or 339 (95% confidence interval [CI], 186 to not evaluable [NE]) days, in the ruxolitinib group, compared with 6.47 months, or 197 (95% CI, 114 to 458) days, in the BAT group, with a stratified hazard ratio (HR) of 0.83 (95% CI, 0.60 to 1.15).²¹ The OS results at the secondary analysis were, overall, consistent with those in the primary analysis of the REACH 2 trial. In the REACH 1 trial, median OS was 232.0 (95% CI, 93.0 to 675.0) days at the final analysis.¹⁹

As of the primary analysis in the REACH 2 trial, the number of patients who experienced a FFS event (i.e., hematologic disease relapse or progression, NRM, or the addition of new systemic aGvHD treatment) was 84 (54.5%) and 119 (76.8%) in the ruxolitinib and BAT groups, respectively. Median FFS was 4.99 months and 1.02 months in the ruxolitinib and BAT groups, respectively, with a HR of 0.46 (95% CI, 0.35 to 0.60). The FFS results of the secondary analyses were, overall, consistent with those of the primary analysis in the REACH 2 trial.¹⁸ In the REACH 1 trial, 60 (84.5%) patients experienced an event (i.e., underlying malignancy relapse or progression [n = 3], death [n = 22], addition of new systemic aGvHD treatment [n = 28], or signs or symptoms of cGvHD [n = 7]). The median FFS was 85.0 (95% CI, 42.0 to 158.0) days.¹⁹

In the REACH 2 trial, ORR at day 28 was only analyzed at the primary analysis and was not reassessed at the secondary or final analyses. As of the primary analysis, the REACH 2 trial met its primary objective. The proportion of patients who achieved an overall response at day 28 was 62.3% (n = 96) (95% CI, 54.2% to 70.0%) in the ruxolitinib group and 39.4% (n = 61) (95% CI, 31.6% to 47.5%) in the BAT group, with a stratified odds ratio of 2.64 (95% CI, 1.65 to 4.22). The proportion of patients who achieved a complete response (CR) or particle response (PR) was 34.4% (n = 53) and 27.9% (n = 43), respectively, in the ruxolitinib group, and 19.4% (n = 30) and 20.0% (n = 31), respectively, in the BAT group.¹⁸ The REACH 1 trial met the predetermined threshold for a positive study outcome (lower limit of the 95% CI for ORR at day 28 ≥ 40%). The proportion of patients who achieved an overall response at day 28 was 56.3% (n = 40) (95% CI, 44.0% to 68.1%). The proportion of patients who achieved a CR, a very good partial response (VGPR), or a PR was 19 (26.8%), 6 (8.5%), or 15 (21.1%), respectively.¹⁹

In the REACH 2 trial, durable ORR at day 56 was only analyzed at the primary analysis and was not reassessed at the secondary, or final analyses. As of the primary analysis, the proportion of patients who achieved a durable ORR at day 56 was 39.6% (n = 61) in the ruxolitinib group and 21.9% (n = 34) in the BAT group, with a stratified odds ratio of 2.38 (95% CI, 1.43 to 3.94) in favour of ruxolitinib.¹⁸ Durable ORR at day 56 was not assessed in the REACH 1 trial.¹⁹

As of the primary analysis in the REACH 2 trial, among the patients who achieved a CR or PR at or before day 28, median DOR was 168 days (range = 22 to 423) in the ruxolitinib group and 101 days (range = 10 to 289) in the BAT group. Results for DOR at the secondary and final analyses were consistent with DOR results at the primary analysis.¹⁸ In the REACH 1 trial, the median DOR for responding patients at any time point was 345.0 (95% CI, 154.0 to not NE) days, with a median follow-up time of 128.5 days (range = 3 to 805 days).¹⁹ The 6-month event-free probability for DOR in responding patients (i.e., PR, VGPR, or CR) at any time point was 62.1% (95% CI, 45.8% to 74.8%).¹⁹

In the REACH 2 trial, BOR was only analyzed at the primary analysis and was not reassessed at the secondary or final analyses. At the primary analysis, the proportion of patients who had achieved BOR by day 28 in the ruxolitinib group was 81.8% (95% CI, 74.8% to 87.6%) and 60.6% (95% CI, 52.5% to 68.4%) in the BAT group, with an odds ratio of 3.07 (95% CI, 1.80 to 5.25).¹⁸ In the REACH 1 trial, the proportion of patients who achieved BOR at any time point was 76.1% (95% CI, 64.5% to 85.4%).¹⁹

In the REACH 2 trial, as of the primary analysis, a higher proportion of patients had tapered off corticosteroids in the ruxolitinib group (21.4%; 95% CI, 15.2% to 28.8%) than in the BAT group (14.8%; 95% CI, 9.6% to 21.4%).¹⁸ The proportions of patients with a relative dose intensity (RDI) of 50% or less (ruxolitinib and BAT group) were 29.2% and 24.5%, and with a RDI of more than 50% were 68.8% and 74.8%.¹⁸ The results for cumulative steroid dosing until day 56 at the secondary and final analyses were, overall, consistent with those at the primary analysis in the REACH 2 trial.¹⁸ In the REACH 1 trial, the proportion of patients who were still receiving ruxolitinib and who had tapered off (discontinued) corticosteroids was 6.9% at day 56, 34.8% at day 100, and 61.1% at day 180. The proportion of patients with a decrease of 50% or more in corticosteroid dose relative to day 1 (or day 2) increased from 23.2% on day 14 to 55.8% on day 28 and 100.0% on day 100.¹⁹

Harms Results

The key harms outcomes reported in the REACH 2 (secondary analysis) and REACH 1 (final analysis) trials are summarized in Table 2.

In the REACH 2 trial, as of the secondary analysis, the percentage of patients reporting at least 1 treatment-emergent adverse event (TEAE) was 99.3% in the ruxolitinib group and 98.7% in the BAT group. The most commonly reported TEAEs (ruxolitinib versus BAT groups) were anemia (40.1% versus 32%), thrombocytopenia (36.8% versus 20.7%), cytomegalovirus infection (30.9% versus 26.7%), neutropenia (24.3% versus 14.7%), and edema peripheral (24.3% versus 21.3%). In the REACH 1 trial, as of the final analysis, all patients in the REACH 1 trial experienced at least 1 TEAE (100.0%). The most commonly reported TEAEs in REACH 1 were similar to those reported in REACH 2, and included anemia (64.8%), thrombocytopenia (62.0%), hypokalemia (49.3%), neutropenia (47.9%), and peripheral edema (46.5%).¹⁹

In the REACH 2 trial, the percentage of patients who experienced at least 1 TEAE of grade 3 or higher in the ruxolitinib and BAT groups was 91.4% and 87.3%, respectively. The most commonly reported TEAEs of grade 3 or higher (ruxolitinib versus BAT groups) were anemia (35.5% versus 24.0%), thrombocytopenia (33.6% versus 16.7%), neutropenia (21.7% versus 12.0%), platelet count decrease (17.8% versus 15.3%), and white blood cell count decrease (13.2% versus 8.7%).¹⁸ In the REACH 1 trial, TEAEs of grade 3 or higher occurred in 97.2% of patients. The most commonly reported TEAEs of grade 3 or higher in REACH 1 were similar to those reported in REACH 2, and included thrombocytopenia (53.5%), anemia (50.7%), neutropenia (42.3%), and hyperglycemia (19.7%).¹⁹

In the REACH 2 trial, the percentage of patients experiencing at least 1 serious TEAE was 66.4% in the ruxolitinib group and 53.3% in the BAT group. The most common serious TEAEs were (ruxolitinib versus BAT) sepsis (7.9% versus 7.3%), pyrexia (6.6% versus 4.0%), septic shock (6.6% versus 5.3%), and diarrhea (5.3% versus 2.0%).¹⁸ In the REACH 1 trial, the percentage of patients experiencing serious TEAEs was 83.1%. The most commonly reported serious TEAEs in REACH 1 were similar to those reported in REACH 2, and included sepsis (12.7%), pyrexia (11.3%), respiratory failure (11.3%), and lung infection (7.0%).¹⁹

In the REACH 2 trial, the percentage of patients who discontinued treatment due to TEAEs in the ruxolitinib group was 27.0% and in the BAT groups was 9.3%. The most commonly cited TEAEs contributing to treatment discontinuation were neutropenia (n = 4; 2.6%), sepsis (n = 4; 2.6%), anemia (n = 3; 2.0%), and thrombocytopenia (n = 3; 2.0%) in the ruxolitinib group, and sepsis (n = 1; 0.7%), anemia (n = 1; 0.7%), thrombocytopenia (n = 1; 0.7%), and platelet count decrease (n = 1; 0.7%) in the BAT group.¹⁸ In the REACH 1 trial, TEAEs led to discontinuation of ruxolitinib treatment in 32.4% of patients. The most commonly reported TEAEs leading to discontinuation of ruxolitinib were sepsis (5.6%), acute kidney injury (2.8%), and respiratory failure (2.8%).¹⁹

In the REACH 2 trial, on-treatment deaths occurred in 28.3% and 24.0% of patients in the ruxolitinib and BAT groups, respectively. The most common cause of death was the study indication of aGvHD (including aGvHD and related complications) in 21 (13.8%) and 21 (14.0%) patients in the ruxolitinib and BAT groups, respectively.¹⁸ In the REACH 1 trial, there were 35.2% (n = 25) of patients who had died during treatment with ruxolitinib or within 30 days of their last dose. The most common cause of death was “other” (25.4%, n = 18), which included underlying GvHD, multi-organ failure, pulseless electrical activity arrest, and respiratory failure, many of which were counted as fatal TEAEs.¹⁹

In the REACH 2 trial, serious infections were reported in 38.2% and 30.0% of patients in the ruxolitinib and BAT groups, respectively, and serious infections of grade 3 or higher in 38.2% and 28.7% of patients, respectively. The percentage of patients experiencing at least 1 infection TEAE of any grade was 80.9% and 69.3% in the ruxolitinib and BAT groups, respectively.¹⁸ In the REACH 1 trial, there were 58 patients (81.7%) with at least 1 TEAE infections and infestation, 36 of whom experienced serious TEAE infections and infestations.¹⁹

In the REACH 2 trial, 1 patient in each of the ruxolitinib and BAT groups reported experiencing bradycardia of any grade. No patients reported bradycardia of grade 3 or higher.¹⁸ In the REACH 1 trial, 2 patients reported experiencing bradycardia of any grade, and 1 patient-reported bradycardia of grade 3 or higher.¹⁹

In the REACH 2 trial, cytopenia TEAEs of any grade (ruxolitinib versus BAT group) included anemia (40.8% versus 34.0%), thrombocytopenia (56.6% versus 36.7%), leukopenia (46.7% versus 32.0%), and other cytopenias (8.6% versus 6.0%). Cytopenia TEAEs grade 3 or higher were of special interest (ruxolitinib versus BAT group), and included anemia (36.2% versus 25.3%), thrombocytopenia (50.7% versus 32.0%), leukopenia (42.8% versus 27.3%), and other cytopenias (5.9% versus 4.7%).¹⁸ In the REACH 1 trial, cytopenia TEAEs of any grade included anemia (64.8%), neutropenia (47.9%), and thrombocytopenia (62.0%). Cytopenia TEAEs of grade 3 or higher included anemia (50.7%), neutropenia (42.2%), and thrombocytopenia (53.5%).¹⁹

In the REACH 2 trial, lipid abnormality events of any grade were reported in 9.9% and 7.3% of patients in the ruxolitinib and BAT groups. Lipid abnormality events of grade 3 or higher were reported in the ruxolitinib and BAT groups by 3.9% and 2.7% of patients, respectively.¹⁸ Lipid abnormalities were not reported in the REACH 1 trial.

In the REACH 2 trial, the safety profile for the 9 adolescents was, overall, similar to that of the study safety set.¹⁸ REACH 1 did not include any adolescents.¹⁹

Table 2: Summary of Key Results From Pivotal and Protocol Selected Studies

BOR = best overall response; CI = confidence interval; CR = complete response; DOR = duration of response; FFS = failure-free survival; HR = hazard ratio; NA = not applicable; NE = not evaluable; NR = not reported; ORR = overall response rate; OS = overall survival; PR = partial response; RDI = relative dose intensity; SAE = serious adverse event; TEAE = treatment-emergent adverse event; VGPR = very good partial response; vs. = versus; WDAE = withdrawal due to adverse event.

^aUpon request to the sponsor, the median OS follow-up time was not provided.

^bParticipants with no observed death or loss to follow-up were censored at the last date they were known to be alive.

^cHR and 95% CI are obtained from the stratified Cox proportional hazards model using the Wald test.

^dP value nominal.

^eThe 95% CI for the response rate was calculated using the Clopper-Pearson exact method.

^fOdds ratio and 95% CI are calculated using the stratified Cochran-Mantel-Haenszel test.

^gThe event was defined as the progression of aGvHD or the addition of systemic therapies for aGvHD after day 28. The competing risks included death without prior observation of aGvHD progression and onset of cGvHD.

^hMedian and quartiles are provided using Kaplan–Meier method.

ⁱRDI includes days of zero dose in the calculation.

Sources: Clinical Study Report (REACH 2),¹⁸ Clinical Study Report (REACH 1).¹⁹

Critical Appraisal

There are insufficient data to describe how the requested reimbursement criteria match the patient population in the REACH 2 trial. The sponsor was asked for clarification on the number of patients in the REACH 2 trial who had an inadequate response to corticosteroids, an inadequate response to other systemic therapies, or an inadequate response to corticosteroids and other systemic therapies, but reported that such data are not available.²² Because an inadequate response to corticosteroids was an eligibility criterion of the REACH 2 trial, it follows that all patients in the trial had an inadequate response to corticosteroids; it also follows that data for patients who only had an inadequate response to other systemic therapies and not to steroids were not available from the REACH 2 trial. However, the number of patients who had an inadequate response to other systemic therapies in addition to an inadequate response to corticosteroids remains unclear. It is not known if patients who are refractory to 1 therapy, as opposed to multiple therapies, would respond differently to ruxolitinib. The clinical experts consulted by CADTH agreed that the difference between patients who either have an inadequate response to corticosteroids alone or to multiple therapies would be unlikely to influence the treatment effect of ruxolitinib.

The REACH 2 trial had an open-label design, so the investigator and the study participants were aware of their treatment status, which increases the risk of detection and performance bias. This had the potential to bias results and outcomes in favour of ruxolitinib if the assessor (investigator or patient) believed the study drug was likely to provide a benefit. Subjective outcomes (i.e., adverse outcomes and patient-reported outcomes) may be at particular risk of bias with an open-label design. Furthermore, the underlying complexity of aGvHD has been acknowledged as a key challenge for the design and analysis of clinical trials in the current target setting, and may contribute to subjective inter-physician variability in response assessments. To mitigate the impact of this bias, the investigators used standardized criteria (i.e., aGvHD disease evaluation and response-assessment criteria were done in accordance with the standard NIH criteria of Harris et al. [2016])¹⁶ to evaluate responses. However, no independent review committee was used to evaluated responses. Overall, the magnitude and direction of this bias remain unclear. Although imbalances were noted for a few baseline characteristics (e.g., prior therapy of steroids plus CNIs plus an aGvHD prophylaxis; organ involvement of the skin, liver, and upper and lower GI tracts; time from diagnosis of underlying disease to transplant and time from diagnosis of underlying disease to screening), they were unlikely to influence clinical outcomes, according to the clinical experts consulted by CADTH. Patients in the BAT group who experienced disease progression, mixed response, or no response were allowed to add or initiate a new systemic therapy up to day 28 without proceeding to discontinuation; however, this was considered a failure of initial BAT. The clinical experts consulted by CADTH noted that changing or initiating new systemic aGvHD therapies is reflective of clinical practice. It was felt by the clinical experts that changes to the BAT treatment up the day 28 were unlikely to affect OS results, given the similar efficacy and similar responses achieved with various BAT therapies. Addition to or change of systemic therapy was treated as treatment failure and, therefore, did not affect ORR at day 28 or the FFS outcomes. Crossover of patients in the BAT group to the ruxolitinib group after day 28 may have biased the OS and EFS outcomes. Patients in the BAT group could cross over to the ruxolitinib group if they failed to meet the primary end point (CR or PR at day 28), lost the response thereafter, and met the criteria for progression, mixed response, or no response, which necessitated new additional systemic immunosuppressive treatment. Overall, 49 patients in the BAT group crossed over to the ruxolitinib group. Crossover of patients in the BAT group may have prolonged survival beyond what would have occurred had the patients only received their randomized study treatment. During the randomized treatment phase (i.e., the period from day 1 to week 24 or end of treatment [EOT]), the median duration of treatment was close to twice as long with ruxolitinib as with BAT, at 82.5 days (range = 8 to 396) and 45.5 days (range = 2 to 218), respectively. A safety comparison between the study groups over that period may have been biased against ruxolitinib. Additionally, the investigator’s choice of BAT may have influenced the safety profile in the BAT group, as the toxicity profiles of BAT treatments differ. The interpretation of results for the EQ-5D-5L and the FACT-BMT scales (i.e., the ability to assess trends over time and to make comparisons across treatment groups) is limited by the significant decline in patients available to provide assessments over time. It was noted that few patients in the trial were younger than 18 years. The clinical experts supported the generalization of the study results to patients younger than 18 years, as the management of these patients is similar to the management of adults in clinical practice, the safety profile of ruxolitinib in these patients was similar to the overall safety set, and there is no biologic rational to assume that outcomes with ruxolitinib would be different between adult and adolescents with SR-aGvHD.

REACH 1

The sponsor was asked for clarification on the number of patients in the REACH 1 trial who had an inadequate response to corticosteroids, an inadequate response to other systemic therapies, or an inadequate response to corticosteroids and other systemic therapies, and reported that 42 patients were refractory to steroids alone and 29 patients were refractory to steroids and 1 additional systemic therapy (i.e., 1 systemic treatment in addition to corticosteroids (± CNIs) for aGvHD was allowed in the REACH 1 trial).²² The sponsor was asked about the specific types of additional systemic therapies received by the 29 patients in REACH 1 who were refractory to 1 additional systemic therapy, but no additional data were provided beyond the information shown in Table 16.²² It is not known if patients who are refractory to 1 therapy, as opposed to multiple therapies, would respond differently to ruxolitinib. The clinical experts consulted by CADTH agreed that the difference between patients who have an inadequate response to corticosteroids alone or to multiple therapies would be unlikely to have an impact on the treatment effect of ruxolitinib.

Phase II (randomized or nonrandomized) trials document safety outcomes and investigate whether the estimate of effect for a new drug is large enough to use it in confirmatory phase III trials. Phase II trials may not accurately predict harm and/or the effectiveness of treatments. There are numerous examples of phase III trials with results that did not support the phase II trial results.²³ Interpretation of time‐to‐event end points, such as OS, is limited in single‐arm studies. The nonrandomized design makes it a challenge to interpret OS events attributable to ruxolitinib, because all patients received the same treatment. The noncomparative design of the REACH 1 trial precludes the ability to compare the relative therapeutic benefit or safety of ruxolitinib with currently available therapies in Canadian clinical practice. All patients in the REACH 1 trial received at least 1 concomitant medication. For instance, CNIs and glucocorticoids were received by 88.7% and 45.1% of patients, respectively. Given the uncontrolled design of the REACH 1 trial, the effect of concomitant treatments on the overall study outcome cannot be determined. Outcomes such as observed responses, durability of responses, and survival may have been influenced by the concomitant use of steroids or other therapies. The REACH 1 trial had an open-label design in which the investigator and the study participants were aware of their treatment status, which increased the risk of detection and performance bias. This had the potential to bias results in favour of ruxolitinib if the assessor (investigator or patient) believed the study drug was likely to provide a benefit. Furthermore, the underlying complexity of aGvHD and its nonspecific presentation have been acknowledged as a key challenge in the design and analysis of clinical trials in the current target setting, and may contribute to subjective inter-physician variability in response assessments. To mitigate the impact of this bias, the investigators used standardized criteria (i.e., aGvHD disease evaluation and response-assessment criteria were done in accordance with the standard NIH criteria of Harris et al. [2016]¹⁶) to evaluate responses. No formal statical significance and hypotheses testing was performed, so no P values were reported. Point estimates with 95% CIs were reported to estimate the magnitude of the treatment effect. The REACH 1 trial did not collect data on patient-reported outcomes. The input provided by the patient advocacy and registered clinician groups, as well as by the clinical experts consulted by CADTH, agreed that improvements in HRQoL and aGvHD symptom severity are important treatment goals for the target population. aGvHD has been found to be the leading cause of morbidity in patients who have undergone alloSCT and who have a multitude of symptoms and various degrees of severity.³

Indirect Comparisons

No indirect treatment comparisons were included in the sponsor’s submission to CADTH or identified in the literature search.

Other Relevant Evidence

The section on other relevant evidence included:

• 1 additional relevant study (Moiseev et al. [2020]²⁴) included in the sponsor’s submission to CADTH that reported results for ruxolitinib in adults and children with SR-aGvHD

• a brief summary of methods and results of post hoc analyses of the REACH 2 trial that were applied in the submitted pharmacoeconomic model

• a list of ongoing trials, presented in Table 43.

Moiseev et al. (2020) Study

Description of the Study

Moiseev et al. (2020)²⁴ was a prospective, single-centre, open-label study conducted in Russia that included 75 patients with acute (n = 32) or chronic (n = 43) SR-GvHD. In the study sample of adults and children, about half the participants were children (53% had acute and 39% had chronic GvHD). The median ages in the acute and chronic GvHD groups were 17 years (range = 1 to 67) and 21 years (range = 2 to 62), respectively. Adult and children weighing more than 40 kg received ruxolitinib at a starting dose of 10 mg twice a day, children weighing less than 40 kg received 0.15 mg/kg twice a day. Previous treatments were continued if the attending physician considered it necessary. Ruxolitinib was stopped if there were signs of GvHD progression. The primary end point was ORR. ORR for acute and chronic GvHD was assessed in accordance with the joint statement criteria of Martin et al. (2009)²⁵ and the NIH criteria of Lee et al. (2015),²⁶ respectively. The secondary end points included OS, toxicity, relapse, and infection complications.

Efficacy Results

The ORR was 75% (95% CI, 57% to 89%) in the aGvHD group and 81% (95% CI, 67% to 92%) in the cGvHD group. OS was 59% (95% CI, 49% to 74%) in the aGvHD group and 85% (95% CI, 70% to 93%) in the cGvHD group. In patients with aGvHD and cGvHD, there were no significant differences between adults and children in any of the outcomes, including ORR (aGvHD P = 0.31; cGvHD P = 0.35) and survival (aGvHD P = 0.44; cGvHD P = 0.12).

Harms Results

The most common adverse event (AE) was hematologic toxicity, with 79% and 44% of grade III to IV neutropenia occurring in the acute and chronic GvHD groups, respectively. There were no significant differences in toxicity between adults and children.

Critical Appraisal

Given the single-arm observational design, interpretation of the study results is limited. Because of the lack of a comparator group and blinding, it is difficult to determine the effectiveness of the treatment on the study outcomes. Given the relatively small sample size of patients with aGvHD (n = 32), the generalizability of these results may be limited. Moreover, as this trial was conducted in Russia, there may be limitations to the generalizability of these findings to the Canadian context.

Relevance for CADTH Review

In the REACH 3 trial, patients 12 to 18 years represented a small proportion of the study sample (3.6%). In the study by Moiseev et al. (2020),²⁴ approximately 50% of the study sample was younger than 18 years. Hence, this additional study supplements the evidence for ruxolitinib in patients younger than 18 years.

Post Hoc Analyses of the REACH 2 Trial

Several post hoc analyses of the REACH 2³ trial were conducted, and the results were applied to the submitted pharmacoeconomic model. High-level summaries of the methods and results of the post hoc analyses were provided by the sponsor and were summarized by CADTH, and key critical appraisal points were added by the CADTH review team. The post hoc analyses included OS by response, DOR by response at day 28, duration of treatment by response at day 28, duration of treatment by individual initial BAT, duration of treatment from randomization, and resource use by study group for initial hospitalization and response at day 28 for readmissions. The CADTH review team was unable to conduct a rigorous evaluation of the conduct and reporting of the post hoc analyses, as only a high-level summary of methods was provided by the sponsor. Overall, the CADTH methods team concluded that results from post hoc analyses are considered exploratory and hypotheses-generating only. Because of the lack of formal inferential statistical testing, the ability to interpret results of such analyses is significantly limited.

Conclusions

One phase III, open-label, multi-centre randomized controlled trial (RCT) (REACH 2) and 1 single-arm phase II trial (REACH 1) were included in this CADTH review. The REACH 2 trial demonstrated statistically significant improvements in ORR at day 28 and in the rate of durable ORR at day 56 in patients treated with ruxolitinib, compared those treated with BAT. The improvements in the response outcomes of the magnitude observed in the REACH 2 trial were considered clinically meaningful by the clinical experts consulted by CADTH. Other secondary outcomes — DOR, BOR, FFS, and steroid use — were supportive of the observed ORR day 28 benefit with ruxolitinib. The open-label design of the trial and reliance on the assessment of trial outcomes by local investigators may have introduced bias that is difficult to quantify. The results of HRQoL, EQ-5D-5L, and FACT-BMT measures remain uncertain because of several important limitations. The actual degree of OS benefit with ruxolitinib is uncertain, given the risk of potential bias that arises from the crossover of patients in the BAT group to the ruxolitinib group and the limited follow-up time. The clinical experts consulted by CADTH noted that no new safety concerns were observed with ruxolitinib. Although the REACH 1 trial achieved the predetermined threshold for a positive outcome (lower limit of the 95% CI for ORR at day 28 ≥ 40%) in patients who received ruxolitinib, there was uncertainty regarding the magnitude of clinical benefit directly attributable to ruxolitinib because of limitations associated with the study design, including the single-arm, open-label design with no formal statical significance testing and the relatively small sample size (N = 71).

Introduction

Disease Background

HSCTs provide stem cells to patients whose bone marrow has been destroyed by disease, chemotherapy, or radiation.² The 2 main types of SCT are autologous and allogeneic transplants. alloSCTs use stem cells from a matched related or unrelated donor, whereas autologous SCTs use the patient’s own stem cells.² Between 2006 and 2014, nearly 1 million SCTs were performed worldwide, approximately 40% of which were alloSCTs.³ alloSCT can be used for the treatment of malignant and nonmalignant hematologic diseases.³ According to a Canadian population-based cohort study,²⁷ 547 alloSCTs were performed in Ontario between 2012 and 2015. Between 2008 and 2019, an estimated 13,033 transplants were performed in Canada, 5,672 of which were alloSCTs⁵; in the year from 2018 to 2019, 2,843 transplants were recorded in Canada.⁵ Most transplants were conducted in patients with plasma cell disorders (30.28%), non-Hodgkin’s lymphoma (20.07%), and acute myeloid leukemia (15.93%). Although alloSCT has curative potential, there are risks that the donor’s stem cells will die or be destroyed by the patient’s body before settling in the patient’s bone marrow or that the donor’s immune cells will attack healthy cells in the patient’s body; the latter is called GvHD.^2,3

GvHD is a multi-system disorder in which the donor-derived immune cells initiate an adverse immune reaction in the transplant recipient’s tissues, cells, and organs, leading to tissue damage, organ failure, or death. GvHD has been found to be the leading cause of morbidity and NRM in patients after alloSCT,⁴ affecting up to 70% of patients who receive HSCTs.³ GvHD is estimated to be responsible for 21% to 31% and 31% to 40% of deaths in patients who received a transplant from a HLA-matched sibling and from an unrelated donor, respectively.³

GvHD has a multitude of syndromes that are defined by clinical manifestations, according to NIH consensus criteria, rather than the time of onset (i.e., before or after day 100 of transplant, as was used previously).^28,29 GvHD is typically classified as acute or chronic, depending on the set of distinct clinical manifestations. Overlap syndrome may also occur, in which diagnostic or distinctive features of aGvHD and cGvHD appear together.⁶ cGvHD typically occurs 100 days or more after alloSCT and can last a few months or a lifetime, affecting almost any part of the body.⁶ aGvHD typically occurs within 100 days after alloSCT and often affects the skin, liver, and intestines.⁶

Acute GvHD occurs in 30% to 50% of patients who undergo alloSCT, with 14% to 36% developing severe aGvHD (i.e., grade III).⁷ Prognosis among patients with aGvHD is poor, with an estimated 3-year survival rate of 54%⁸; only 25% to 30% of patients with grade III aGvHD and 1% to 2% of patients with grade IV aGvHD experience long-term (> 2 years) survival.⁷ Grade I aGvHD has not been found to significantly affect long-term survival.³ It has been suggested that patients with SR-aGvHD are at an elevated mortality risk, with an estimated 2-year survival rate of 17%. Patients with SR-GvHD have a high mortality from infections.⁷ In a retrospective study of 127 patients with SR-aGvHD, 4-year infection-related mortality and OS were 46% and 15%, respectively; the 1-year incidence of bacterial, viral, and fungal infections was 74%, 65%, and 14%, respectively.⁷

Among the several factors thought to affect the incidence and severity of aGvHD, 1 of the most important is alloSCT from HLA-nonidentical or unrelated donors.³⁰ Additional risk factors include the advanced age of patient and/or donor, a female donor for male recipient, peripheral blood as the stem cell source, the type of GvHD prophylaxis, and recipient seropositivity for cytomegalovirus.⁷

Clinical manifestations of aGvHD typically affect the skin, GI tract, and liver. Common symptoms of aGvHD are skin rash, burning and redness of the skin on the palms of the hands or soles of the feet, blisters and peeling skin, diarrhea, persistent nausea and vomiting, cramping or abdominal pain, enlarged liver, liver tenderness, abnormal liver enzymes or liver failure, and increased levels of serum bilirubin.⁹ Commonly, 1 of the first clinical symptoms of aGvHD is a maculopapular skin rash with gradually worsening manifestations.³¹ The degree of skin involvement is graded according to the degree and severity of lesions as stage 1 to 4. GI symptoms typically involve the upper and lower tracts. Diagnosis is typically confirmed upon review of tissues with upper endoscopy, rectal biopsy, or colonoscopy. The degree of GI involvement is graded according to the severity of diarrhea as stage 1 to 4. Liver involvement is commonly observed in patients who also exhibit symptoms of skin and GI aGvHD. Liver biopsy is used to confirm GvHD of the liver. Early signs of hepatic involvement include abnormal liver function tests, such as a serum liver rise in conjugated bilirubin and alkaline phosphatase. The degree of liver involvement is graded according to the serum total bilirubin levels as stage 1 to 4. The stages of skin, GI, and liver involvement are combined to determine the overall severity grade of aGvHD.³¹

Over the past 40 years, several grading systems for aGvHD have been developed; the most commonly used include the initial Glucksberg grade (grades I to IV)³² and the International Bone Marrow Transplant Registry (IBMTR) grading system (grades A to D).^33,31 These grading systems have developed over time; for example, the Glucksberg system has been updated by an NIH working group to include persistent nausea with histologic evidence of GvHD in stage 1 upper GI aGvHD (Przepiorka et al. [1995]³⁴). A more recent widely adopted grading system is the NIH consensus grading,¹⁶ which allows measurement of the frequency of stools or the stool volume when staging lower GI involvement (stages 0 to 4). The NIH criteria have been refined and tested for clarity and ease of use by an international GvHD research consortium (the Mount Sinai Acute GvHD Consortium) to standardize the compilation of clinical data from multiple organ systems for clinical research (Harris et al. [2016]¹⁶) (see Table 46 in Appendix 3 for the updated NIH criteria).³ Standardized approaches to grading of aGvHD that have been developed through international expert consensus, such as the NIH criteria, increase the uniformity of aGvHD symptom capture and may help to reduce variability in the diagnosis and grading of aGvHD between transplant centres.^7,31

Diagnosis of aGvHD may be straightforward in patients who present with classical symptoms, such as rash, diarrhea with abdominal cramps, and rising serum bilirubin concentration in the first few months after alloSCT.³¹ However, in many cases, aGvHD presents with less obvious signs and symptoms, and competing causes of abnormalities must be excluded. For example, skin rash can be caused by a variety of drugs that these patients are often treated with, diarrhea may be caused by infection, and hyperbilirubinemia is a common side effect of multiple drugs.³¹ Alternative diagnoses may be excluded on biopsy of the involved tissue; however, biopsies may lack sensitivity and specificity,⁷ and liver biopsy may pose a significant risk of major bleeding, given that most patients with aGvHD are thrombocytopenic.³¹ Currently, there are no established biomarkers for the diagnosis or prognosis of aGvHD.³¹ The diagnosis of aGvHD comes down to a careful integration of all available clinical information.³

About 5% of all patients with aGvHD are 12 to 18 years.³ The management and treatment of aGvHD is similar in adolescents and adults.³ aGvHD is a major cause of morbidity and mortality after alloSCT in patients of all ages.³⁵ Adults and children with grade III to IV aGvHD have a 2-year survival rate of just 27% to 35%.³ The clinical experts consulted by CADTH agreed that there is a significant unmet need in adolescents with aGvHD.

Standards of Therapy

Treatment options generally aim for the immunosuppression of donor T cells, which can cause aGvHD. However, the same cells are likely responsible for elimination of residual malignant cells’ GVT effect). The management of patients with aGvHD must therefore achieve a balance between the benefits of reducing GvHD and the potential harms from a reduced GVT effect.⁷ The choice of initial treatment for patients with aGvHD depends on multiple factors, including the severity of symptoms, the type of prophylactic regimen used, and the importance of a GVT effect.⁷ Grade I aGvHD (i.e., skin involvement of 50% or less of the body surface area without liver or GI tract involvement) is managed with topical treatments (e.g., tropical steroids) and adjustment of prophylactic treatments (e.g., MMF or CNIs such as cyclosporine). Patients with grade II or higher aGvHD (i.e., skin involvement of greater than 50% of the body surface area with liver or GI tract involvement) receive high-dose systemic glucocorticoids (e.g., methylprednisolone or prednisone) in addition to the care that patients with grade I aGvHD receive. In patients who respond, steroid treatment is continued over several weeks until it is gradually tapered to prevent a flare of aGvHD. Approximately 25% to 40% of patients achieve CRs with glucocorticoids⁷; however, treatment is associated with significant side effects that can affect a patient’s quality of life and increase susceptibility to infection.¹⁰ Approximately 25% to 50% of patients become refractory to steroids and are considered to have SR-aGvHD. In general, patients whose disease progresses 3 to 5 days after the initial start of systemic steroids or who show no response after 5 to 7 days are considered to have SR-aGvHD; the exact definition of SR-aGvHD may vary by centre.⁷

There is currently no standardized second-line therapy for patients with aGvHD who have an inadequate response to steroids in Canada. Evidence from patients with SR-aGvHD has mostly been obtained in retrospective, single-arm, phase II studies. Comparisons of data across studies is a challenge because of the small number of enrolled patients, the heterogenous patient populations, and the lack of standardized end points.⁷ The weighted average 6-month survival estimate across 25 studies that evaluated second-line therapies for aGvHD was reported to be 0.49 by the American Society of Blood and Morrow Transplantation; however, given the significant heterogeneity across study populations and designs, interpretation of this estimate is significantly limited.³⁶ In the absence of sufficient evidence to guide the selection of second-line treatment, factors that influence the treatment choice include the experience of the treating physician, the type of aGvHD prophylaxis used, and the risk of potential toxicities and worsening pre-existing comorbidities.⁷ According to the clinical experts consulted by CADTH, available second-line options in Canada include ECP, MMF, etanercept, infliximab, mTOR inhibitors (i.e., everolimus or sirolimus), ATG, and interleukin-2 receptors. The clinical experts noted that available therapies are currently used off-label. In the absence of proven treatment options, there is inter-province variability in standard practices and access to therapies. The clinical experts explained the challenges with currently available therapies in this heavily pre-treated target population, including no responses or only PRs (response rate was estimated by the clinical experts to be approximately 50%, but the durable response rate, in which patients still show a response at about 2 months, to be only 10% to 30%). According to the clinical experts, responses in this patient population are important to enable the tapering of steroids to mitigate long-term side effects (e.g., osteoporosis, hypertension, hyperglycemia, diabetes, and bone or joint health) and the risk of infection. It was emphasized by the clinical experts that opportunistic infection and organ damage related to aGvHD are major causes of NRM in patients with SR-aGvHD.

There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve HRQoL, reduce the symptoms of aGvHD, enhance performance status, and improve OS. They added that a convenient oral route of administration would help improve adherence and reduce hospital-based resource use.

Drug

Ruxolitinib is a JAK inhibitor that mediates the signalling of a number of cytokines and growth factors important for hematopoiesis and immune function.¹ Ruxolitinib binds to and inhibits protein tyrosine kinases JAK 1 and JAK2, which may lead to a reduction in inflammation and an inhibition of cellular proliferation.³⁷ Ruxolitinib has received Health Canada market authorization for the treatment of SR or dependent aGvHD in patients 12 years and older. The sponsor’s requested reimbursement criteria for ruxolitinib for aGvHD differ from the Health Canada indication, in that the sponsor’s criteria specify “inadequate response to corticosteroids or other systemic therapies,” rather than “steroid refractory or dependent” aGvHD. Concurrent with this CADTH review for aGvHD, ruxolitinib is being reviewed by CADTH in the cGvHD setting for the treatment of cGvHD in patients 12 years and older who have an inadequate response to corticosteroids or other systemic therapies. Ruxolitinib has 2 Health Canada–approved indications: the treatment of splenomegaly and/or its associated symptoms in adults with primary myelofibrosis (also known as chronic idiopathic myelofibrosis), post-PV myelofibrosis, or post-essential thrombocythemia myelofibrosis; and the control of hematocrit in adults with PV resistant to or intolerant of a cytoreductive drug. Ruxolitinib received a positive conditional CADTH recommendation in March 2016 for the treatment of patients with PV who are resistant to or intolerant of hydroxyurea, according to the modified European LeukemiaNet Criteria used in the RESPONSE trial, and who have a good performance status. Ibrutinib is the only Health Canada–approved therapy for GvHD³; it is indicated in the chronic setting for the treatment of patients with steroid-dependent or SR-cGvHD.³⁸ However, ibrutinib has not been reviewed by CADTH and is currently not publicly reimbursed in Canada for the current target indication; it is available through private drug insurance.

After being granted priority review with orphan product designation, the FDA-approved ruxolitinib in September 2021 for cGvHD after the failure of 1 or 2 lines of systemic therapy in patients 12 years and older, based on evidence from the phase III REACH 3 trial.³⁹ In May 2019, after granting priority review, the FDA-approved ruxolitinib for SR-aGvHD in patients 12 years and older based on evidence from the phase II REACH 1 trial.⁴⁰ In addition to the GvHD setting, the FDA has approved ruxolitinib for intermediate or high-risk myelofibrosis, including primary myelofibrosis, post-PV myelofibrosis, and post-essential thrombocythemia myelofibrosis in adults, and for PV in adults who have had an inadequate response to or are intolerant of hydroxyurea. After being granted priority review with orphan product designation, the FDA-approved belumosudil in July 2021 for patients 12 years and older with cGVHD after failure of at least 2 lines of systemic therapy based on evidence from the phase II KD025 to 213 trial.⁴¹ The European Medicines Agency has approved ruxolitinib for the treatment of disease-related splenomegaly or symptoms in adults with primary myelofibrosis (also known as chronic idiopathic myelofibrosis), post-PV myelofibrosis, or post-essential thrombocythemia myelofibrosis, and for the treatment of adults with PV who are resistant to or intolerant of hydroxyurea.

Ruxolitinib is available as 5 mg, 10 mg, 15 mg, and 20 mg tablets. The recommended starting dose for ruxolitinib for aGvHD is 5 mg administered orally twice daily. A dose increase to 10 mg twice daily is recommended after at least 3 days of treatment if ANCs and platelet counts are not decreased by at least 50% relative to the first day of dosing with ruxolitinib. The product monograph also states that tapering of ruxolitinib may be considered in patients with a response after they have discontinued corticosteroids. Tapering of ruxolitinib, by reducing the dose to 50% every 2 months, is recommended; in the event that signs or symptoms of GvHD recur during or after the taper, re-treatment with ruxolitinib should be considered.¹ Table 3 summarizes key characteristics of ruxolitinib.

According to information in the product monograph and confidential information in the Health Canada Reviewer Report, Health Canada considered reviewing REACH 1 as the pivotal study, and using safety data from the REACH 2 trial as supportive evidence for the aGvHD indication.^1,11 This decision is the result of uncertainties that were identified by the FDA when it reviewed raw data from the REACH 2 trial as part of a sponsor-proposed label update for the FDA-approved indication of aGvHD, which had already been approved based on REACH 1 data. The sponsor’s submission noted that the proposed label update based on REACH 2 trial data could not be completed, and the sponsor subsequently withdrew the supplemental New Drug Application for REACH 2. Specifically, the 3 issues identified by the FDA concerned the following: missing and incomplete information in the case report forms related to organ staging of aGvHD, which precluded the FDA from conducting their own analyses and confirming the investigators’ efficacy assessments; the frequency of protocol deviations related to investigator-determined organ staging; and information regarding prior therapies (i.e., prophylactic versus therapeutic aGvHD treatment). Health Canada concluded that as a result of “integrity issues in the REACH 2 study” identified by the FDA, the efficacy results of REACH 2 could not be “verified.”¹¹ Therefore, Health Canada upheld its initial decision to consider REACH 1 as the pivotal study, and to use safety and pharmacokinetic data from the REACH 2 trial as supportive evidence in their assessment of the aGvHD indication. The sponsor clarified with CADTH that the initial decision to submit the REACH 2 trial as the pivotal trial for the proposed CADTH reimbursement request remains unchanged.³ This CADTH report does not include a review or critical appraisal of the issues raised by the FDA, as CADTH has no access to the FDA’s assessment and reviewing raw data are not in the mandate of CADTH.

Health Canada based their efficacy assessment of the REACH 1 trial on the FDA-evaluable population (49 patients who failed steroids alone), which is a subset of the REACH 1 trial’s full efficacy-evaluable patients (N = 71). According to the Health Canada Clarifax, the following subsets of patients were excluded from the FDA-evaluable population: 9 patients who failed steroids plus additional therapies before randomization, and 13 patients who did not meet the SR criteria because of suboptimal dosing or duration of corticosteroid treatment before randomization.¹² The Health Canada Clarifax noted that the available data are considered insufficient to support the proposed ruxolitinib indication for patients with aGvHD who failed 1 other systemic treatment in addition to corticosteroids (± CNIs).¹² This CADTH clinical review report does not report on the subset of 49 patients in the FDA-evaluable population; it presents results for all 71 efficacy-evaluable patients from the REACH 1 trial. Results for the 49 patients in the FDA-evaluable population are reported in the product monograph.¹

The Health Canada Reviewer Report stated that assessment of the safety of ruxolitinib for aGvHD was informed by the full safety-evaluable population (N = 71) of the REACH 1 trial and by safety results observed in the REACH 2 trial.¹ Safety results reported in this CADTH clinical review are based on 71 patients in the REACH 1 trial and safety results from the REACH 2 trial. The Health Canada Reviewer Report noted that patients with SR or steroid-dependent aGvHD as per the eligibility criteria in REACH 1 are considered acceptable.¹¹ The REACH 1 eligibility criteria for SR and steroid-dependent aGvHD align with the current National Comprehensive Cancer Network guidelines for hematopoietic cell transplantation, version 2.2020,¹³ the Alberta Bone Marrow and Blood and Cell Transplant Standard Practice Manual (2021),¹⁴ and the EBMT-NIH-CIBMTR Task Force position statement,¹⁵ with minor variations. Steroid dependence has been defined as the inability to taper prednisone under 2 mg/kg per day after initially successful treatment for at least 7 days or as the recurrence of aGvHD activity during steroid taper.¹¹

Table 3: Key Characteristics of Ruxolitinib

aGvHD = acute graft-versus-host disease; ANC = absolute neutrophil count; JAK = Janus-associated kinase; SAE = serious adverse event; SR = steroid refractory.

^aHealth Canada–approved indication.

Sources: Product monograph,¹ National Library of Medicine.³⁷

Stakeholder Perspectives

Patient Group Input

Eight patient groups — Lymphoma Canada, the Lymphoma & Leukemia Society of Canada, Chronic Lymphocytic Leukemia Canada, Myeloma Canada, the Aplastic Anemia & Myelodysplasia Association of Canada, the Canadian Myeloproliferative Neoplasms Research Foundation and the CML Network, MPN Canadian Research Foundation, and CTTC — created 1 joint patient input for this review. The input was based on an online survey with 68 respondents. Sixty patients reported having received an SCT, 6 patients reported not having received an SCT, and 2 patients did not answer the question. Of the 60 patients who received an SCT, 49 reported having received an alloSCT. Fifty-three patients had experienced GvHD after their SCT. Data on type of GvHD were available for 45 of the 53 patients with GvHD: 13% experienced aGvHD, 24% experienced cGvHD, and 62% experienced both acute and chronic GvHD. Twenty patients reported receiving ruxolitinib.

Respondents indicated that they had long-lasting GvHD symptoms (3 to 5 years for 26% of respondents and more than 5 years for 28% of respondents). To manage their GvHD, the respondents reported requiring numerous medical consultations, hospital stays, and nights away from home. Respondents reported a range of GvHD symptoms that had a significant impact on daily activities and were detrimental to quality of life. Respondents highlighted the interruption of life goals and accomplishments (career, school), difficulty sleeping, and impacts on their mental health (stress, anxiety, worry, and concentration problems), and finances. Other commonly experienced symptoms reported by respondents included burning and redness of the skin on the palms of the hands or soles of the feet, rashes that could spread over the entire body, blisters and peeling skin, skin problems (such as dryness, rash, itching, peeling, darkening, hard texture, and feeling tight), enlarged liver, liver tenderness, abnormal liver enzymes or liver failure, jaundice, dry eyes that may have a burning or gritty feeling, dry mouth with or without mouth ulcers, diarrhea, loss of appetite, stomach cramps, vomiting, weight loss, pain in muscles and joints, mobility issues, infections, and difficulty breathing.

According to the patient input received, respondents expected new drugs or treatments to improve the following key outcomes: OS, GvHD symptoms, quality of life, and severity of side effects. Additionally, the ability to receive treatment in the outpatient setting (rather than requiring an overnight hospital stay), having access to treatment locally (rather than requiring extensive travel), coverage of treatment by insurance or drug plans, and treatments recommended by health care professionals were perceived to be very important by respondents. Respondents who had direct experience with ruxolitinib indicated that, overall, ruxolitinib was an effective treatment, improved their quality of life, had tolerable side effects, and they would take again if recommended by their physician and would recommend it to other patients.

Clinician Input

Input From the Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol; assisting in the critical appraisal of clinical evidence; interpreting the clinical relevance of the results; and providing guidance on the potential place in therapy). The following input was provided by 3 clinical specialists with expertise in the diagnosis and management of aGvHD.

Unmet Needs

The clinical experts consulted by CADTH noted that there are currently no standard-care regimens specific to patients with SR-aGvHD in Canada. The clinical experts noted that, in the absence of proven treatment options, there is inter-province variability in standard practices and access to therapies. Available therapies are currently used off-label and, because of the lack of large-scale, positive, randomized, prospective studies, no consensus could be reached on the optimal second-line therapy for SR-aGvHD.³ There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve HRQoL, reduce the symptoms of aGvHD, enhance performance status, and improve OS. The added that a convenient oral route of administration would help improve adherence and reduce hospital-based resource use. The clinical experts expressed challenges with currently available therapies in this heavily pre-treated target population, including no responses or only PRs (response rate was estimated by the clinical experts to be approximately 50%, but the durable response rate, in which patients still show a response at about 2 months, to be only 10% to 30%) and intolerance to currently available treatments. According to the clinical experts, responses in this patient population are important to enable the tapering of steroids to mitigate long-term side effects (e.g., osteoporosis, hypertension, hyperglycemia, diabetes, and bone or joint health) and the risk of opportunistic infection. It was emphasized by the clinical experts that opportunistic infection and organ damage from aGvHD are major causes of NRM in SR-aGvHD.

Place in Therapy

Ruxolitinib can be used as an add-on to an immunosuppressive regimen of corticosteroids with or without CNIs in patients 12 years and older with grade II to IV SR-aGvHD, according to NIH criteria used in the REACH 2 trial. The clinical experts agreed that ruxolitinib would likely shift the current treatment paradigm. They noted that ruxolitinib works as a JAK inhibitor and blocks the JAK-STAT pathway, and that its mechanism of action is novel in the context of other immunosuppressives used in the management of aGvHD, potentially offering synergy with other therapies. The clinical experts highlighted the differential effects of ruxolitinib on different T cell subsets, which help prevent post-transplant relapse. The trial excluded patients who received more than 1 systemic treatment for SR-aGvHD, patients with overlap syndrome, and patients with grade I SR-aGvHD. The clinical experts consulted by CADTH felt that it would be reasonable to generalize the REACH 2 trial results to patients who received 2 or more systemic treatments for aGvHD. They noted that ruxolitinib has a novel mechanism of action in the context of other second-line immunosuppressives, with the potential to offer synergy with other therapies. As well, given the acceptable safety profile of ruxolitinib, they felt that it would be reasonable to leave it to the discretion of the treating physician to apply some flexibility in terms of using ruxolitinib in patients with overlap syndrome and patients with grade I aGvHD.

Patient Population

Overall, the clinical experts consulted by CADTH agreed that patients who meet the inclusion and exclusion criteria of the REACH 2 trial should be eligible for ruxolitinib. Although they agreed that there is currently insufficient evidence to identify which patient subgroups would most likely show a response to ruxolitinib, the clinical experts identified patients with grade IV aGvHD, who have the highest risk of death from aGvHD, as the subgroup most in need of ruxolitinib therapy. Subgroups that would potentially benefit least from ruxolitinib may include patients with refractory vomiting or ileus who are not able to take an oral drug such as ruxolitinib, and patients with thrombocytopenia, especially those with clinical bleeding, who may be a challenge to treat with ruxolitinib and may receive an alternative second-line drug instead. Patients with active uncontrolled infections or non-aGvHD cytopenia are a challenge to treat with ruxolitinib or other available second-line therapy options; ruxolitinib should be used with caution and may require dose adjustment in these patients.

The clinical experts agreed that patients considered to have SR-aGvHD would be possible candidates for ruxolitinib. They noted that aGvHD is a complex clinical diagnosis made on a daily basis by experienced bone marrow transplant physicians. In general, the experts agreed that the diagnosis of aGvHD is made on clinical grounds (patients with grade II to IV aGvHD are generally not asymptomatic), although tissue diagnosis (biopsy) confirmation may be pursued. They noted that SCT programs are highly specialized and well supported by histopathologists. The clinical experts agreed that a potential for misdiagnosis exists, given the underlying complexity of aGvHD and the nonspecific presentation of aGvHD, so a differential diagnosis must be considered. However, misdiagnosis is minimized in Canada because patients who have undergone alloSCT are followed in specialized clinics.

Assessing Response to Treatment

In the opinion of the clinical experts consulted by CADTH, an accurate assessment of response in aGvHD is based on the NIH consensus criteria used in the REACH 2 trial. Given the underlying complexity of aGvHD evaluation, the experts noted that treatment response can be formally assessed, but it is a challenge to undertake outside the structure of a clinical trial. Clinical assessments in Canadian clinical practice to evaluate response to treatment examine the following areas: skin rash; nausea or vomiting; stool consistency, diarrhea volumes, or abdominal pain; liver function tests; and performance status. In addition, the ability to reduce the corticosteroid dose without exacerbating symptoms and signs of aGvHD is indictor of response to treatment.

The experts noted that response to treatment is usually assessed daily for inpatients and weekly for outpatients.

The clinical experts indicated that the most clinically meaningful responses to treatment include improvements in OS (survival more than 1 year after alloSCT) overall response (CR or PR), improvements in HRQoL and performance status, and the ability to taper corticosteroids.

Discontinuing Treatment

In the opinion of the clinical experts consulted by CADTH, ruxolitinib should be discontinued if a patient experiences aGvHD disease progression, lacks response after 4 to 6 weeks on treatment and either an alternative treatment is being introduced or further treatment is deemed futile and a palliative care approach is pursued, or experiences intolerable toxicity (e.g., severe and refractory thrombocytopenia, cytopenia, or severe neurologic toxicity). The clinical experts agreed that, based on their clinical experience, ruxolitinib generally works quickly (within a few weeks).

Prescribing Conditions

In the opinion of the clinical experts consulted by CADTH, ruxolitinib is an oral drug that can be self-administered in a patient’s home. Patients are assessed and managed in the SCT follow-up clinic. All assessments and prescriptions should be managed by providers familiar with GvHD. Generally, patients with aGvHD are medically unwell to the point of requiring hospitalization in a bone marrow transplant unit in the hospital. Occasionally, patients may be managed as an outpatient, such as with higher doses of steroids and a second-line drug (like ruxolitinib). Patients who respond to treatment are generally able to transition to outpatient care.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups.

The information in this section is a summary of 2 inputs provided by the registered clinician groups that responded to CADTH’s call for clinician input for the purpose of this review. The full clinician inputs received can be found in Appendix 6.

Two clinician group inputs were provided: 1 from CTTC (which contained input from 8 clinicians) and 1 from Ontario Health Cancer Care Ontario Complex Malignant Hematology (which contained input from 2 clinicians). The views of the clinician groups were, overall, consistent with the clinical experts consulted by CADTH, indicating that, based on the evidence from the REACH 2 trial, it is likely that ruxolitinib will become the dominant first-line therapy for SR-aGvHD. The outcomes assessed in the REACH 2 trial were judged to be applicable to Canadian clinical practice and reflective of clinically meaningful responses. It was noted by both input groups that ruxolitinib is not as immunosuppressive as other available therapies. The clinicians from Ontario Health Cancer Care Ontario noted the drawbacks of currently available therapies, such as the IV administration (which requires patients to be at the hospital), side effects and broad immune suppression, and the high price and delivery costs of treatments. The CTTC noted that a Health Canada–approved and provincially funded therapy for SR-aGvHD would be an important step forward in the present target setting, because existing therapies offer low response rates and high rates of toxicity. According to input from the CTTC, experience with ruxolitinib (accessible through compassionate access programs) and real-world effectiveness appear similar to what was observed in the REACH 2 trial, with low rates of toxicity.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Table 4: Summary of Drug Plan Input and Clinical Expert Response

AE = adverse event; aGvHD = acute graft-versus-host disease; ATG = antithymocyte globulin; BMT = bone marrow transplant; CIBMTR = Center for International Blood and Marrow Transplant Research; CNI = calcineurin inhibitor; EBMT = European Society for Blood and Marrow Transplantation; ECP = extracorporeal photopheresis; GvHD = graft-versus-host disease; IL 2r = interleukin-2 receptor; MMF = mycophenolate mofetil; NIH = National Institutes of Health; PR = partial response; SCT = stem cell transplant; SR = steroid refractory.

Clinical Evidence

The clinical evidence included in the review of ruxolitinib is presented in 3 sections. The first section, the systematic review, includes clinical pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as studies that were selected according to an a priori protocol. For the second section, no indirect evidence that met the selection criteria specified in the review was identified. The third section includes additional relevant studies that were considered to address important gaps in the evidence in the systematic review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of ruxolitinib (10 mg twice daily oral tablets) for the treatment of aGvHD in patients 12 years and older who have an inadequate response to corticosteroids or other systemic therapies.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important by patients, clinicians, and drug plans.

Of note, the systematic review protocol presented here was established before the granting of a Notice of Compliance from Health Canada.

Table 5: Inclusion Criteria for the Systematic Review

AE = adverse event; aGvHD = acute graft-versus-host disease; ATG = antithymocyte globulin; BOR = best overall response; cGvHD = chronic graft-vs.-host disease; ECP = extracorporeal photopheresis; EFS = event-free survival; FFS = failure-free survival; HRQoL = health-related quality of life; IL-2r = interleukin-2 receptor; MMF = mycophenolate mofetil; mTOR = mammalian target of rapamycin; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival; PR = partial response; RCT = randomized controlled trial; SAE = serious adverse event; SR = steroidrefractory; vs. = versus; WDAE = withdrawal due to adverse event.

^aThese outcomes were identified as being of particular importance to patients in the input received by CADTH from a patient group.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy, according to the PRESS Peer Review of Electronic Search Strategies checklist.⁴²

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946—) via Ovid and Embase (1974—) via Ovid. All Ovid searches were run simultaneously as a multi-file search. Duplicates were removed using Ovid deduplication for multi-file searches, followed by manual deduplication in Endnote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were ruxolitinib and GvHD. Clinical trials registries were searched: the NIH’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. See Appendix 1 for the detailed search strategies.

The initial search was completed on September 2, 2021. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee on April 27, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature resource.⁴³ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by a review of bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

A focused literature search for network meta-analyses was run in MEDLINE All (1946–) for GvHD. No limits were applied.

Findings From the Literature

A total of 245 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of Included Studies

AE = adverse event; aGvHD = acute graft-versus-host disease; alloSCT = allogeneic stem cell transplant; ANC = absolute neutrophil count; ATG = antithymocyte globulin, BAT = best available therapy; cGVHD = chronic graft-vs.-host disease; CMV = cytomegalovirus; CNI = calcineurin inhibitor; CR = complete response; DLI = donor lymphocyte infusion; DOR = duration of response; ECG = electrocardiogram; ECOG PS = Eastern Cooperative Oncology Group performance status; ECP = extracorporeal photopheresis; EFS = event-free survival; EOT = end of treatment; EQ-5D-5L = 5-level EQ-5D; FACT-BMT = Functional Assessment of Cancer Therapy-Bone Marrow Transplant; GI = gastrointestinal; GvHD = graft-versus-host disease; ICU = intensive care unit; JAK = Janus-associated kinase; MAGIC = Mount Sinai Acute GvHD Consortium; MMF = mycophenolate mofetil; MSC = mesenchymal stromal cell; mTOR = mechanistic (formerly mammalian) target of rapamycin; MTX = methotrexate; NA = not applicable; NIH = National Institutes of Health; NRM = nonrelapse mortality; ORR = overall response rate; PR = partial response; RCT = randomized controlled trial; SR = steroid refractory; vs. = versus.

^aRecipients of nonmyeloablative, myeloablative, and reduced-intensity conditioning are eligible.

^bRecipients of nonmyeloablative and myeloablative transplants are eligible.

^cUse of growth factor supplementation and transfusion support was allowed during the trial. Evident myeloid and platelet engraftment was to be confirmed in the 48 hours before study treatment start.

^dUse of growth factor supplementation and transfusion support was allowed during the trial.

^eActive uncontrolled infection included significant bacterial, fungal, viral, or parasitic infection requiring treatment. Infections are considered controlled if appropriate therapy has been instituted and, at the time of screening, no signs of progression are present. Progression of infection is defined as hemodynamic instability attributable to sepsis, new symptoms, worsening physical signs, or radiographic findings attributable to infection. Persisting fever without other signs or symptoms will not be interpreted as progressing infection.

^fAn active uncontrolled infection is defined as hemodynamic instability attributable to sepsis or new symptoms, worsening physical signs, or radiographic findings attributable to infection. Persistent fever without signs or symptoms will not be interpreted as an active uncontrolled infection.

^gInfections are considered controlled if appropriate therapy has been instituted and, at the time of screening, no signs of progression are present. Progression of infection is defined as hemodynamic instability attributable to sepsis, new symptoms, worsening physical signs, or radiographic findings attributable to infection. Persist fever without other signs or symptoms will not be interpreted as progressing infection.

^hConcomitant use of CNIs and steroids is allowed. Medications used for aGvHD prophylaxis (i.e., started before the diagnosis of aGvHD) that failed to prevent aGvHD in a patient before randomization could not be selected as BAT in the same patient.¹⁸

ⁱPatients who met crossover criteria and received ruxolitinib could continue corticosteroids and CNIs for aGvHD treatment, per standard of care, with discontinuation required of any other systemic immunosuppressive treatment before crossover, unless it was used for aGvHD prophylaxis (i.e., started before the diagnosis of aGvHD).

^jTwo additional reports were included: the clinical study reports for the REACH 2¹⁸ and REACH 1¹⁹ trials from the submission to CADTH.

Sources: Zeiser et al. (2020),¹⁰ Clinical Study Report (Reach 2),¹⁸ Clinical Study Report (REACH 1),¹⁹ Jagasia et al. (2020),⁴⁴ ClinicalTrials.gov (REACH 1),⁴⁵ ClinicalTrials.gov (REACH 2),²⁰ Protocol (REACH 2),³ Protocol (REACH 1),³ Statistical analysis plan (REACH 1),³ additional information request.²¹

Description of Studies

REACH 2

The REACH 2 trial is a completed, international, multi-centre, open-label, randomized, phase III trial comapring ruxolitinib (10 mg administered orally twice daily) with investigator choice of BAT — i.e., ATG, ECP, MSCs, methotrexate, MMF, mTOR inhibitors everolimus or sirolimus), etanercept, or infliximab — in patients 12 years and older with grade II to IV SR-aGvHD. Patients continued to receive their systemic immunosuppressive regimen of corticosteroids with or without CNIs. Staging of aGvHD was based on the NIH criteria of Harris et al. (2016)¹⁶ (see Table 46 in Appendix 3 for NIH consensus criteria for aGvHD grading). This CADTH review is based on data from the primary analysis (July 25, 2019), the updated secondary analysis (January 6, 2020), and the final analysis (April 23, 2021). The final analysis was conducted once all patients had completed the study. The REACH 2 trial was sponsored by Novartis.³

The primary end point of the trial was the ORR at day 28. Patients in this international trial were randomized at 105 sites in 22 countries, which are listed in Table 6. The majority of sites were in Europe. REACH 2 enrolled 14 patients who were randomized at 8 Canadian sites. A total of 309 patients were randomized (using an interactive voice response system) in a 1:1 ratio to receive ruxolitinib or BAT. Enrolment occurred between April 12, 2017 and May 30, 2019.¹⁰ Randomization was stratified by aGvHD grade (grade II versus grade III versus grade IV).¹⁸

The study design is depicted in Figure 2. The study consisted of 4 main periods: the screening period (lasting for 28 days), the treatment period (day 1 to week 24 or EOT), the long-term follow-up period (from EOT to month 24), and the safety follow-up.¹⁸ The end of the study was to occur when all patients had completed month 24 (i.e., the end of the long-term follow-up observation period), unless the patient withdrew consent. During the treatment period, the EOT visit was to occur at week 24, or earlier if the patient met any study treatment discontinuation criteria.¹⁸

The primary end point, ORR, was assessed at day 28. Between day 28 and week 24, patients in the BAT group could cross over to the ruxolitinib group if they failed to meet the primary end point or lost their response after having achieved ORR and met the criteria for progression, mixed response, or no response, requiring new additional systemic immunosuppressive treatment for aGvHD, and did not show any signs or symptoms of cGvHD. During the randomized treatment period, aGvHD disease and safety assessments were planned every week from day 1 to day 56, and every 28 days thereafter until week 24.³

Figure 2: Study Design of REACH 2 Trial

aGvHD = acute graft-versus-host disease; BAT = best available therapy; EOT = end of treatment; ICF = informed consent form; SR = steroid refractory.

Source: Protocol (REACH 2).³

REACH 1

The REACH 1 trial is a completed, open-label, single-arm, multi-centre phase II trial that evaluated the efficacy and safety of ruxolitinib in combination with corticosteroids in patients with grade II to IV SR-aGvHD. The severity grading of aGvHD was based on the NIH criteria of Harris et al. (2016)¹⁶ (see Table 46 in Appendix 3). This CADTH review is based on the final data cut-off date of June 5, 2019, and the study was completed on June 5, 2019.²¹ The REACH trial was sponsored by Incyte-Corporation.¹⁹

The primary end point was ORR at day 28. Patients in this multi-centre trial were enrolled at 26 centres in 17 American states. A total of 71 patients were enrolled to received ruxolitinib (5 mg orally twice daily; if hematologic parameters were stable and no treatment-related toxicity was observed after the first 3 days of treatment, the dose could be increased to 10 mg orally twice daily). Enrolment occurred from December 2016 to December 2017.¹⁹

The study consisted of 3 phases: the screening phase (lasting up to 28 days), the treatment phase, and the follow-up phase. During the follow-up, patients were followed for safety (final follow-up was 30 to 35 days after EOT) and for survival until death, withdrawal of consent, or the end of the study, whichever occurred first. The end of the study was to occur when 75% of patients achieved 2-year NRM, the patient died, or the patient was lost to follow-up.¹⁹

Assessments for aGvHD staging were performed weekly for the first 8 weeks after enrolment, every 28 days thereafter, on days 100, 180, and 365, and at the EOT visit.¹⁹ Safety was evaluated from screening though to 30 to 35 days after EOT by monitoring the incidence and severity of AEs and serious adverse events (SAEs).³

Populations

Inclusion and Exclusion Criteria

The key inclusion and exclusion criteria used in the REACH 2 and REACH 1 trials are described in Table 6.

Briefly, the REACH 2 trial enrolled male or female patients 12 years and older who had undergone alloSCT, had evidence of myeloid and platelet engraftment (ANC > 1,000/mm³ and platelet count > 20,000/mm³), and were diagnosed with grade II to IV aGvHD, as defined by the NIH consensus criteria, and which was determined to be corticosteroid-refractory per protocol-defined criteria.

The REACH 1 trial had, overall, similar inclusion criteria. However, corticosteroid refractoriness (criterion C in Table 4) included a slight variation, in that failure to taper corticosteroids in REACH 1 included patients who began corticosteroids at a lower dose (at least 1 mg/kg methylprednisolone) but developed new GvHD, as opposed to patients who failed to taper the methylprednisolone dose to less than 0.5 mg/kg per day (or equivalent prednisone dose < 0.6 mg/kg per day) for a minimum 7 days in REACH 2. The clinical experts consulted by CADTH noted that criteria A and B are most commonly used in clinical practice and were consistent in REACH 2 and REACH 1. According to the clinical experts, the differences in criterion C would be unlikely influence overall study results. There was a slight variation in the definition of engraftment; patients in REACH 1 did not require adequate platelet engraftment, but patients in REACH 2 did. Patients in both trials required evidence of myeloid engraftment.

Both studies excluded patients who had received more than 1 systemic treatment for SR-aGvHD, presented with a clinical presentation resembling de novo overlap syndrome (defined by Jagasia et al. [2015]¹⁷) or active uncontrolled infection. REACH 2 explicitly excluded patients with multi-focal leuko-encephalopathy, whereas REACH 1 did not.

Baseline Characteristics

The demographic characteristics and disease and alloSCT history at baseline of patients in the REACH 2 trial are summarized in Table 7, Table 8, Table 9, and Table 10.

REACH 2

The mean ages for the ruxolitinib and BAT groups were, respectively, 48.1 (SD = 16.30) and 50.9 (SD = 14.9) years. The majority of patients (ruxolitinib versus BAT) were 18 to 65 years of age (83.1% versus 81.3%), but there were a few patients 12 to 18 years or younger (3.2% versus 2.6%). Most patients were male (59.7% versus 58.7%) and identified as White (72.1% versus 65.8%). The most common primary diagnosis was malignant leukemia or myelodysplastic syndrome (83.8% versus 78.1%), and most patients had received a peripheral blood alloSCT (87.0% versus 76.1%). The aGvHD grade at baseline was mostly grade III (44.2% versus 43.9%), followed by grade II (32.5% versus 34.8%) and grade IV (19.5% versus 20.6%). The most common criterion for SR-aGvHD was failure to achieve a response after 7 days (46.8% versus 40.6%), followed by failure on steroid taper (30.5% versus 31.6%), and progression after at least 3 days (22.7% versus 27.7%).¹⁸

In terms of prior aGvHD therapy, most patients had received steroids plus CNIs (49.4% versus 49.0%), followed by steroids plus CNIs plus another systemic aGvHD treatment (37.0% versus 31.6%). The ruxolitinib group had a higher proportion of patients who received steroids plus CNIs plus an aGvHD prophylaxis as prior therapy (26.6% versus 19.4%) and patients with aGvHD organ involvement at baseline in the skin (60.4% versus 47.7%) and liver (24.0% versus 16.1%), and a lower proportion of patients with aGvHD organ involvement at baseline in upper GI (18.2% versus 23.9%) and lower GI (62.3% versus 74.2%). Patients in the ruxolitinib group had a longer mean time from diagnosis of underlying disease to screening (2.16 years versus 1.72 years) and from diagnosis of underlying disease to transplant (713.07 days versus 553.29 days).¹⁸

Table 7: Summary of Demographic Baseline Characteristics in REACH 2, Full Analysis Set

BAT = best available therapy; SD = standard deviation.

Source: Clinical Study Report (REACH 2).¹⁸

Table 8: Summary of Disease History by Treatment in REACH 2, Full Analysis Set

BAT = best available therapy; CIBMTR = Center for International Blood and Marrow Transplant Research; MDS = myelodysplastic syndrome; PV = polycythemia vera; SD = standard deviation.

Source: Clinical Study Report (REACH 2).¹⁸

Table 9: Summary of Transplant-Related History by Treatment in REACH 2, Full Analysis Set

BAT = best available therapy; CMV = cytomegalovirus; HCT = hematopoietic cell transplantation; SD = standard deviation.

Source: Clinical Study Report (REACH 2).¹⁸

Table 10: Summary of aGvHD Disease History by Treatment in REACH 2, Full Analysis Set

aGvHD = acute graft -vs.-host disease; BAT = best available therapy; CNI = calcineurin inhibitors; GI = gastrointestinal; SD = standard deviation; SR = steroid refractory.

Source: Clinical Study Report (REACH 2).¹⁸

REACH 1

Patients’ demographic characteristics and disease and alloSCTs history at baseline in the REACH 1 trial were, overall, similar to those in the REACH 2 trial and are summarized in Table 11, Table 12, Table 13, and Table 14. As in the REACH 2 trial, the majority of patients in the REACH 1 trial were 18 to 65 years of age (81.7%), identified as White (93.0%), had received a peripheral blood alloSCT (80.3%), and the most common primary diagnosis was malignant leukemia or myelodysplastic syndrome (71.9%). The distribution of aGvHD grades was similar in the 2 trials, with grade III aGvHD in the majority of patients (46.5%), followed by grade II (31.0%) and grade IV (22.5%) in REACH 1. As in REACH 2 trial, the most common criterion for SR-aGvHD in REACH 1 was no aGvHD improvement after 7 days of primary treatment (40.8%), followed by failing corticosteroid taper (36.2%) and progression after 3 days or primary treatment (23.9%). The mean time from diagnosis of underlying disease to screening was similar in the 2 trials; it was 2.15 years in the REACH 1 trial, and 2.16 years versus 1.72 years in the ruxolitinib and BAT groups in the REACH 2 trial, respectively. The mean time from transplant to randomization was 82.93 days in the REACH 2 trial and 110.8 days in the REACH 1 trial. Most patients in the 2 trials received grafts from identical HLA-matched donors (60.2% in the REACH 2 trial and 63.4% in the REACH 1 trial).¹⁹ Differences in patient’s baseline characteristics between the REACH 2 and REACH 1 trials were noted for the following categories. The mean age of patients was slightly higher in the REACH 1 trial, at 52.9 years (SD = 14.18), than in the overall REACH 2 trial population, at 49.5 years (SD = 15.69). No patient in the REACH 1 trial was younger than 18 years, but in the REACH 2 trial overall, 2.9% of patients were 12 years to younger than 18 years. Most patients in the REACH 2 trial were male (59.2%), compared with 49.3% in the REACH 1 trial.¹⁹

In terms of prior aGvHD therapy, as in the REACH 2 trial, all patients had received corticosteroids as prior anti-GvHD treatment. However, the percentage of patients who had additionally received prior CNIs as aGvHD treatment was lower in the REACH 1 trial (23.9%) than in the REACH 2 trial (50.0% in the ruxolitinib group and 49.0% in the BAT group). In the REACH 1 trial, all patients had received treatment for GvHD prophylaxis before enrolment, compared with 72.1% of patients in the REACH 2 trial. The REACH 1 trial had a higher proportion of patients who had received prior prophylactic treatment with CNIs (97.2%) than the REACH 2 trial (65.6% in the ruxolitinib group and 60.0% in the BAT group).¹⁹

Table 11: Summary of Demographic Baseline Characteristics in REACH 1, Efficacy-Evaluable Population

BMI = body mass index; BSA = body surface area; ECOG PS = Eastern Cooperative Oncology Group performance status; SD = standard deviation.

Source: Clinical Study Report (REACH 1).¹⁹

Table 12: Summary of Baseline aGvHD Disease Characteristics in REACH 1, Efficacy-Evaluable Population

aGvHD = acute graft-versus-host disease; BSA = body surface area; GI = gastrointestinal; MAGIC = Mount Sinai Acute GvHD Consortium; SD = standard deviation; SR = steroid refractory.

Source: Clinical Study Report (REACH 1).¹⁹

Table 13: Summary of Cancer History in REACH 1, Efficacy-Evaluable Population

ALL = acute lymphoblastic leukemia; AML = acute myeloid leukemia; CLL = chronic lymphocytic leukemia; CR = complete response; MDS = myelodysplastic syndrome; PR = partial response; SD = standard deviation.

^a“Other” includes multiple myeloma, mixed B-cell myeloid acute leukemia, unclassifiable acute leukemia, diffuse large B-cell lymphoma (with subsequent T-cell prolymphocytic leukemia), chronic myelogenous leukemia, chronic myelomonocytic leukemia, plasma cell leukemia, essential thrombocythemia (with secondary myelofibrosis), and myeloproliferative neoplasm.

Source: Clinical Study Report (REACH 1).¹⁹

Table 14: Summary of alloSCT History in REACH 1, Efficacy-Evaluable Population

alloSCT = allogeneic stem cell transplant; CMV = cytomegalovirus; GvHD = graft-versus-host disease; HLA = human leukocyte antigen; PBSC = peripheral blood stem cell; SD = standard deviation.

^aTime since stem cell transplant = first date of ruxolitinib – transplant date + 1.

^bTime since diagnosis of GVHD = first date of ruxolitinib – GVHD diagnosis date + 1.

Source: Clinical Study Report (REACH 1).¹⁹

Prior aGvHD Treatment

REACH 2

Almost all patients in the full analysis set received treatment for aGvHD before enrolment in the trial, and the proportions and types of therapies were balanced across groups, as summarized in Table 15.

The most common prior systemic treatments (ruxolitinib versus BAT) were glucocorticoids (mainly methylprednisolone, prednisolone, prednisone, and methylprednisolone sodium succinate) (96.8% versus 96.1%), followed by corticosteroids (weak; mainly methylprednisolone, prednisolone, and methylprednisolone sodium succinate) (87.7% versus 89.0%;), corticosteroids (plain; mainly methylprednisolone, prednisolone, and prednisone) (87.0% versus 80.6%), and corticosteroid combinations for the treatment of acne (mainly methylprednisolone and methylprednisolone sodium succinate) (68.2% versus 76.1%). The most commonly used CNIs (50.6% versus 49.0%) included ciclosporin (37.0% versus 27.7%) and tacrolimus (14.9% versus 21.3%).¹⁸

The sponsor was asked for clarification on the number of patients in the REACH 2 trial who had an inadequate response to corticosteroids, an inadequate response to other systemic therapies, or an inadequate response to corticosteroids and other systemic therapies, and noted that such data are not available.²² Because an inadequate response to corticosteroids was an eligibility criterion in the REACH 2 trial, it follows that all patients had an inadequate response to corticosteroids. However, the proportion of patients who had an inadequate response to other systemic therapies in addition to having an inadequate response to corticosteroids remains unclear (for more details, refer to the REACH 2 internal validity section in Critical Appraisal).

REACH 1

As in the REACH 2 trial, all patients in the REACH 1 trial received prior systemic therapy for aGvHD (Table 16), and all patients received prior systemic therapy with glucocorticoids (81.7% of patients received methylprednisolone and 57.7% received prednisone). The number of patients who received CNIs (predominantly tacrolimus [19.7%], with or without methotrexate) as prior systemic therapy was lower in the REACH 1 trial (23.9%) than in the REACH 2 trial.¹⁹

The sponsor was asked for clarification on the number of patients in the REACH 1 trial who had an inadequate response to corticosteroids, an inadequate response to other systemic therapies, or an inadequate response to corticosteroids and other systemic therapies, and reported that 42 patients were refractory to steroids alone and 29 patients were refractory to steroids and 1 additional systemic therapy (receipt of 1 systemic treatment in addition to corticosteroids [± CNIs] for aGvHD was allowed in the REACH 1 trial).²² The sponsor was asked about the specific types of additional systemic therapies received by the 29 patients in the REACH 1 trial who were refractory to 1 additional systemic therapy, but no information beyond that in Table 16 was provided.²²

Table 15: Prior aGvHD Treatment in REACH 2, Full Analysis Set

aGvHD = acute graft-versus-host disease; ATC = Anatomical Therapeutic Chemical; ATG = antithymocyte immunoglobulin; BAT = best available therapy; CNI = calcineurin inhibitor; ECP = extracorporeal photopheresis; MSC = mesenchymal stromal cell.

Note: A medication or therapy can appear in more than 1 ATC class.

Source: Clinical Study Report (REACH 2).¹⁸

Table 16: Prior aGvHD Treatment in REACH 1, Efficacy-Evaluable Population

aGvHD = acute graft-versus-host disease; ATG = antithymocyte immunoglobulin; CNI = calcineurin inhibitor; SD = standard deviation.

^aDate of last exposure to corticosteroids as prior anti-GVHD therapy – date of first exposure to corticosteroids as prior anti-GVHD therapy + 1; includes methylprednisolone, methylprednisolone sodium succinate, and prednisone.

Source: Clinical Study Report (REACH 1).¹⁹

Prior aGvHD Prophylaxis Treatment

REACH 2

Approximately 2-thirds of patients (72.1%) in the full analysis set of the REACH 2 trial received aGvHD prophylaxis before enrolment into the trial, and the proportions and types of therapies were balanced across groups, as summarized in Table 17. The most common prior prophylactic treatments (ruxolitinib versus BAT) were CNIs (mainly ciclosporin) (65.6% versus 60.0%), followed by other ophthalmologicals (mainly ciclosporin) (56.5% versus 42.6%), glucocorticoids (mainly methylprednisolone) (19.5% versus 20.0%), and corticosteroids (plain; mainly methylprednisolone) (18.8% versus 13.5%).¹⁸

REACH 1

In the REACH 1 trial, all patients received GvHD prophylaxis before enrolment, compared with 72.1% of patients in the REACH 2 trial (see Table 18). The REACH 1 trial had a higher proportion of patients who received prior prophylactic treatment with CNIs (97.2%) than the REACH 2 trial (ruxolitinib versus BAT = 65.6% versus 60.0%). After CNIs, the most common prior prophylactic therapies in the REACH 1 trial were selective immunosuppressants (mainly MMF) (71.8%) and folic acid analogues (mainly methotrexate) (22.5%).¹⁹

Table 17: Prior aGvHD Prophylaxis Treatment in REACH 2, Full Analysis Set

aGvHD = acute graft-versus-host disease; ATC = Anatomical Therapeutic Chemical; ATG = antithymocyte immunoglobulin; BAT = best available therapy; CNI = calcineurin inhibitor; MMF = mycophenolate mofetil.

Source: Clinical Study Report (REACH 2).¹⁸

Table 18: Prior aGvHD Prophylaxis Treatment in REACH 1, Efficacy-Evaluable Population

aGvHD = acute graft-versus-host disease; ATG = antithymocyte immunoglobulin; CNI = calcineurin inhibitor; GvHD = graft-versus-host disease.

Source: Clinical Study Report (REACH 1).¹⁹

Interventions

REACH 2

Patients randomized to the REACH 2 trial were allocated to either the ruxolitinib group or the BAT group. Treatments and doses are described in Table 19. For treatment management based on patient response at day 28, refer to Table 20.

During the randomized treatment period in the REACH 2 trial, aGvHD disease and safety assessments were planned every week from day 1 up to day 56, and every 28 days thereafter until week 24. In case the taper of ruxolitinib was not completed by week 24 because of aGvHD flare or safety concerns, patients continued to be assessed (every 8 weeks from week 24 to week 48, and every 12 weeks thereafter) until taper of ruxolitinib was complete. All patients who discontinued treatment (regardless of when it occurred) entered the long-term survival period and were followed for long-term data on survival, any relapse or progression of the underlying hematologic disease, NRM, any occurrence of graft failure, EFS, any occurrence of cGvHD, and occurrence of any second primary malignancies. During the long-term survival follow-up period, follow-up occurred at 6, 9, 12, 18, and 24 months from randomization (day 1), as applicable, once the treatment period was completed. During the crossover period, patients followed the same treatment and taper schedule as patients originally randomized to ruxolitinib treatment; assessment occurred every week up to day 56 after ruxolitinib initiation, and every 28 days thereafter until week 24. A 30-day safety follow-up assessment was completed after the final dose of ruxolitinib or BAT for all patients.³

REACH 1

All patients enrolled in the REACH 1 trial received ruxolitinib. Treatments and doses are described in Table 19.

Assessments for aGvHD staging were performed weekly for the first 8 weeks after enrolment, every 28 days thereafter, on days 100, 180, and 365, and at the EOT visit. Unscheduled visits were allowed at the investigator's discretion.³ Safety was evaluated from screening to 30 to 35 days after EOT by monitoring the incidence and severity of AEs and SAEs.³ The treatment period started on the day the patient received the first dose of the study drug and ended when the patient permanently discontinued the study treatment, per the investigator’s decision. Patients who experienced disease progression or started a new GvHD therapy, entered the survival follow-up period and were assessed at least every 8 weeks (± 7 days) for new GvHD therapy, survival, withdrawal of consent, or the end of the study, whichever occurred first.³ Patients who terminated study treatment for reasons other than progression of GvHD entered the disease status follow-up period and were assessed every 28 days to monitor disease status until GvHD progression, relapse of malignancy, death, or the end of the study.¹⁹

Table 19: Treatment Regimens in the REACH 2 and REACH 1 Trials

AE = adverse events; aGvHD = acute graft-versus-host disease; ATG = antithymocyte globulin; BAT = best available therapy; cGvHD = chronic graft-vs.-host disease; CNI = calcineurin inhibitor, CR = complete response, DLI = donor lymphocyte infusion; ECP = extracorporeal photopheresis; EOT = end of treatment; IEC = independent ethics committee; IRB = institutional review board; MSC = mesenchymal stromal cell; mTOR = target of rapamycin; PR = partial response.

^aStable hematologic parameters are defined as the absence of a ≥ 50% decrease in platelet counts and/or ANC relative to day 1.

^bPatients on ruxolitinib were allowed to continue ruxolitinib outside the study if they met the study discontinuation criteria, responded to ruxolitinib at day 28 (or crossover day 28), and were assessed, per investigator, to be deriving clinical benefit from ruxolitinib; these patients do not enter the long-term follow-up period.

^cA patient may be withdrawn from study treatment as follows:

• If, during the course of the study, a subject is found not to have met eligibility criteria, then the medical monitor, in collaboration with the investigator, will determine whether the subject should be withdrawn from the study.

• If a patient is noncompliant with study procedures or study drug administration in the investigator's opinion, the sponsor should be consulted for instruction on handling the patient.

Source: Protocol (REACH 2).³

Table 20: Treatment Management Based on Patient Response at Day 28 in REACH 2

aGvHD = acute graft-versus-host disease; BAT = best available therapy; CNI = calcineurin inhibitor, CR = complete response, EOT = end of treatment; PR = partial response.

Source: Protocol (REACH 2).³

Tapering of Therapies

REACH 2

Tapering of corticosteroids, CNIs, and ruxolitinib was done in 2 steps: first systemic corticosteroids were tapered upon documented CR or PR, then CNIs and ruxolitinib were tapered.

Taper of corticosteroids could be initiated at day 7 in patients who achieved a CR or PR. Taper consisted of a 10% dose reduction every 5 days up to approximately day 56 to allow a 7 to 8 week taper. It was anticipated that the taper of immunosuppressive medications would be complete by week 24. If the taper occurred before day 56, the dose of corticosteroids could be re-escalated at the investigator’s discretion and was not considered a treatment failure as long as the criteria for treatment failure were not met (criteria on the treatment of aGvHD flare are listed in the following). Taper of corticosteroids was to be repeated once patients achieved PR or CR, in which case the taper of corticosteroids could be prolonged beyond day 56, delaying the taper of CNIs and of ruxolitinib.³

Tapering guidelines per the study protocol for CNIs indicate that taper could be initiated once patients with documented CR or PR were off systemic corticosteroids. Taper consisted of a 25% dose reduction per month, or could follow a taper schedule per institutional practice.³

Taper guidelines for ruxolitinib indicated that, starting at day 56, taper could be initiated as required once patients with documented CR or PR were off systemic corticosteroids. Taper consisted of a 50% dose reduction every 2 months (56 days) from 10 mg to 5 mg administered orally twice daily. In patients with systemic aGvHD response (i.e., no worsening of aGvHD signs or symptoms), further taper included a second 50% dose reduction, to 5 mg orally once daily, for an additional 2 months (56 days), before discontinuation of ruxolitinib. The taper of ruxolitinib was to be complete by week 24 (or crossover week 24), except if prolonged tapering was indicated because of an aGvHD flare or other safety concerns, in which case the taper would be initiated no later than week 24 (or crossover week 24) and completed by the end of the patient’s study (no later than week 96; up to approximately 2 years from randomization). If GvHD flare occurred during ruxolitinib taper after day 56, patients could have had the ruxolitinib dose increased back to initial dose, and ruxolitinib taper repeated as needed until the patient’s end of study (up to approximately 2 years form randomization). Should a tapering strategy not be in the best interest of the patient, or should the taper be completed before crossover week 24, the patient must still follow the assigned visit evaluation schedule, including all safety and efficacy assessments, until crossover week 24.³

REACH 1

Tapering of corticosteroids and ruxolitinib was done in 2 steps: first, systemic corticosteroids were tapered per instructional guidelines at a rate appropriate for the resolution of GvHD manifestation; and second, taper of ruxolitinib could be attempted.³

Corticosteroid taper was done per instructional guidelines at a rate that was appropriate for the resolution of GvHD manifestation. Although there were no protocol guidelines for corticosteroid taper, recommendations were provided (see Table 21).³

Per-protocol guidelines for ruxolitinib indicated that after day 180, a taper of 1 dose level (e.g., 10 mg twice daily to 5 mg twice daily to 5 mg once daily) could be initiated once patients who achieved either a CR or VGPR were off systemic corticosteroids for at least 8 weeks. Ruxolitinib taper could be initiated during ongoing, concomitant use of ani-infective medication, GvHD prophylaxis medications (including CNIs), transfusion support, and topical steroid therapy. A subsequent dose level reduction of ruxolitinib was allowed after an additional 56 days during which patients did not experience any grade 2 or higher hematologic toxicity related to ruxolitinib or symptoms of an active infection.³

Investigators who wished to initiate taper or ruxolitinib earlier than proposed in the guidelines could do so conditional on consultation with and approval from the sponsor’s medical monitor.³

Table 21: Recommended Corticosteroid Administration in REACH 1, Days 1 to Day 28

Source: Protocol (REACH 1).³

Treatment of aGvHD Flare

REACH 2

aGvHD flare was defined as any increase in symptoms of aGvHD sustained for more than 24 hours after an initial response (CR or PR) that required re-escalation of immunosuppressive therapy for aGvHD (e.g., corticosteroids, CNIs, BAT, and/or ruxolitinib dosing). An aGvHD flare was not considered a treatment failure as long as no change or addition of another systemic treatment was required. If addition or initiation of a new systemic therapy was required because it was not possible to taper corticosteroids to a dose below methylprednisolone 0.5 mg/kg per day (or equivalent < 0.6 mg/kg per day of prednisone) for a minimum 7 days or to re-escalate corticosteroids to methylprednisolone more than 2 mg/kg per day (or equivalent > 2.5 mg/kg per day of prednisone), patients were considered to have experienced treatment failure. See the Tapering of Therapies section for instruction on the treatment of aGvHD flare.³ The sponsor was asked about the number of aGvHD flares experienced by patients in the trial, but explained that those data were not available.²¹

REACH 1

Upon aGvHD flare, the dose of corticosteroids could be re-escalated at the investigator’s discretion. An aGvHD flare was not considered a treatment failure as long as the dose did not exceed the initial starting dose, the flare was not unresponsive to the re-escalation, and there were no multiple flares.³

Upon aGvHD flare, the dose of ruxolitinib could be re-escalated by incremental dose levels (e.g., 5 mg once daily to 5 mg twice daily to 10 mg twice daily), provided hematologic thresholds were met. An aGvHD flare was not considered a treatment failure as long as no additional systemic therapy was required (including the restarting of corticosteroids). Upon treatment failure, ruxolitinib would be discontinued.³

If aGvHD signs or symptoms recurred at a later time, after patients had completely tapered off ruxolitinib, ruxolitinib could be restarted at the discretion of the investigator and the patient would resume the assessment schedule using the original day 1 as a reference point.³

Dose Modifications

REACH 2

Dose interruptions or reductions were permitted for patients who did not tolerate the protocol-specified dosing schedule. For ruxolitinib, patients could not receive less than 5 mg once daily and not more than 10 mg twice daily. Once a dose or schedule modification occurred, titration back up to the original dose or schedule was permitted. If a ruxolitinib or BAT dose interruption exceeded 14 days, the study treatment had to be discontinued or the sponsor’s medical monitor contacted.³

REACH 1

Dose reductions or modifications of ruxolitinib were allowed based on AEs, clinical evaluation, and laboratory assessments. Patients could not receive less than 5 mg once daily and not more than 10 mg twice daily. If the ruxolitinib dose interruption exceeded 14 days, the study treatment had to be discontinued or the sponsor’s medical monitor contacted.³

Concomitant Medications

REACH 2

Routine administration of transfusion support and systemic immunosuppressive regimens of corticosteroids, CNIs (cyclosporine or tacrolimus), and topical corticosteroid therapy were allowed and were administered per institutional guidelines. Routine administration of antibiotics, anti-infectives, and immunizations as prophylactic therapy was allowed per institutional guidelines. Other systemic medications for aGvHD could be continued only if used as aGvHD prophylaxis (i.e., started before the diagnosis of aGvHD).³

REACH 1

Routine administration of systemic immunosuppressive regimens of corticosteroids, CNIs (cyclosporine or tacrolimus), and topical corticosteroid therapy was allowed per institutional guidelines. Routine administration of antibiotics, anti-infectives, and immunizations as prophylactic therapy was allowed per institutional guidelines. Other systemic medications for aGvHD could be continued only if used as aGvHD prophylaxis (i.e., started before the diagnosis of aGvHD).³

Outcomes

A list of end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 22. These end points are summarized in the following text. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Table 22: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

AE = adverse event; BOR = best overall response; cGvHD = chronic graft-vs.-host disease; DOR = duration of response; EFS = event-free survival; EQ-5D-5L = 5-level EQ-5D; FACT-BMT = Functional Assessment of Cancer Therapy-Bone Marrow Transplant; FFS = failure-free survival; HRQoL = health-related quality of life; ICU = intensive care unit; NR = not reported; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival.

Sources: Statistical Analysis Plan (REACH 2),³ Statistical Analysis Plan (REACH 1).³

Overall Survival

REACH 2

OS was a secondary outcome in the REACH 2 trial and was defined as the time from date of randomization to time of death from any cause.¹⁸

REACH 1

OS was a secondary outcome in the REACH 1 trial and was defined as the time from date of enrolment (first date of ruxolitinib treatment) to death from any cause.³

Failure-Free Survival

REACH 2

FFS was a secondary outcome in the REACH 2 trial and was defined as the time from date of randomization to date of hematologic disease relapse or progression, NRM, or addition of new systemic aGvHD treatment. For this analysis, the local investigator’s review of hematologic disease relapse or progression, according to protocol-defined criteria, was used.³

REACH 1

FFS was a secondary outcome in the REACH 1 trial and was defined as the interval from date of the first dose of ruxolitinib to date of underlying malignancy relapse or progression, death, addition of new systemic aGvHD treatment, or signs or symptoms of cGvHD.¹⁹

ORR at Day 28

REACH 2

ORR at day 28 was the primary outcome of the REACH 2 trial. ORR at day 28 was defined as the proportion of patients with CR or PR, according to the standard NIH criteria of Harris et al. (2016),¹⁶ at day 28.³ Response was assessed relative to the disease evaluation of aGvHD at baseline. aGvHD disease grading by the investigator was used for randomization and all analyses.³

CR was defined as a score of 0 for aGvHD grading (see Table 46 in Appendix 3 for aGvHD staging criteria of Harris et al. [2016]¹⁶) in all evaluable organs, which indicates complete resolution of all signs and symptoms of aGvHD in all evaluable organs without administration of additional systemic therapies for any earlier progression, mixed response, or nonresponse of aGvHD.³

PR was defined as a improvement of 1 stage, per staging criteria of Harris et al. (2016),¹⁶ in 1 or more organs involved with aGvHD signs or symptoms without progression in other organs or sites and without administration of additional systemic therapies for an earlier progression, mixed response, or nonresponse of aGvHD.³

Mixed response was defined as improvement of at least 1 stage, per aGvHD staging criteria of Harris et al. (2016),¹⁶ in the severity of aGvHD in at least 1 organ, accompanied by progression in another organ or development of signs or symptoms of aGvHD in a new organ.³

Progression was defined as worsening in 1 or more organs by 1 or more stages, per aGvHD staging criteria of Harris et al. (2016),¹⁶ without improvement in any involved organ.³

A patient was not considered a responder at day 28 in the event of any of the following:¹⁸

• missing aGvHD assessment at baseline or day 28

• no CR or PR at day 28

• additional systemic therapy for aGvHD before day 28.

REACH 1

ORR at day 28 was the primary end point in the REACH 1 trial. ORR at day 28 was defined as the proportion of patients with CR, VGPR, or PR, per Center for International Blood and Marrow Transplant Research (CIBMTR) modifications to the International Bone Marrow Transplant Registry response index (CIBMTR [2009],⁴⁶ Martin et al. [2009],²⁵ Harris et al. [2016]¹⁶) at the day 28 response assessment (± 2 days) and before the start of new anti-aGvHD therapy, if applicable. Response was assessed relative to the disease evaluation of aGvHD at study day 1. aGvHD disease grading by the investigator was used for randomization and all analyses.^3,19

CR is defined as a score of 0 for the aGvHD grading in all evaluable organs without administration of additional systemic therapies for any earlier progression, mixed response, or nonresponse of aGvHD.^3,19

VGPR is defined as follows:

• for skin, no rash or residual erythematous rash involving less than 25% of the body surface, without bullae (residual faint erythema and hyperpigmentation do not count)

• for liver, total serum bilirubin concentration below 2 mg/dL or below 25% of baseline at enrolment

• for gut —

  • tolerance of food or enteral feeding

  • predominantly formed stools

  • no overt GI bleeding or abdominal cramping

  • no more than occasional nausea or vomiting.

PR is defined as improvement of 1 stage in 1 or more organs involved with aGvHD symptoms without progression in other organs or sites without administration of additional systemic therapies for any earlier progression, mixed response, or nonresponse of aGvHD.¹⁹

Mixed response is defined as improvement in 1 or more organs with deterioration in another organ that manifests symptoms of aGvHD or development of symptoms of aGvHD in a new organ.¹⁹

Progression of disease is defined as deterioration in at least 1 organ without any improvement in others.¹⁹

A patient was not considered a responder at day 28 in the event of any of the following:¹⁹

• missing aGvHD assessment at baseline or day 28

• no CR, VGPR, or PR at day 28

• additional systemic therapy for aGvHD before day 28 (including the need to re-escalate steroid dose above the day 1 dose).

Overall, the definition for ORR at day 28 was similar in the REACH 2 and REACH 1 trials. However, the REACH 1 trial assessed VGPR based on Martin et el. (2009),²⁵ which defined a response state that approximated CR, with some qualifications. Martin et al. (2009)²⁵ suggests that a benefit of the VGPR outcome is the inclusion of minor intermittent clinical abnormalities that may not be due to aGvHD, as opposed to CR that is strictly defined as a resolution of all signs and symptoms of aGvHD. The Martin et al. (2009)²⁵ notes that VGPR should not be a substitute for CR in clinical trials and that it’s use will have to be validated through experience.²⁵

ORR at Day 14

REACH 2

ORR at day 14 was a secondary outcome in the REACH 2 trial. ORR at day 14 was defined as the proportion of patients with CR or PR at day 14. Definitions of CR, PR, mixed response, and progression are the same as for ORR at day 28. ORR at day 14 and at day 56 was derived as it was for ORR at day 28. The local investigator’s review of aGvHD assessment data was used for this analysis.¹⁸

REACH 1

ORR rate at days 14, 56, and 100 was a secondary outcome in the REACH 1 trial. ORR at days 14, 56, and 100 was defined as the proportion of patients achieving a CR, VGPR, or PR at days 14, 56, and 100, respectively. Definitions of CR, VGPR, and PR are the same as for ORR at day 28. The investigator’s review of aGvHD assessment data was used for this analysis.¹⁹

Rate of Durable ORR at Day 56

REACH 2

Rate of durable ORR at day 56 was the key secondary outcome of the REACH 2 trial. Rate of durable ORR at day 56 was defined as the proportion of all patients in each group who achieved a CR or PR at day 28 and maintained a CR or PR at day 56. Response was assessed relative to the last disease evaluation of aGvHD before or at the start of crossover treatment (ruxolitinib). The local investigator’s review of aGvHD assessment data was used for this analysis.¹⁸

A patient was not considered a durable responder at day 56 in the event of the following:¹⁸

• not a responder at day 28

• missing aGvHD assessment at day 56

• no CR or PR at day 56

• additional systemic therapy for aGvHD before day 56.

REACH 1

Rate of durable ORR at day 56 was not reported in the REACH 1 trial.¹⁹

Duration of Response

REACH 2

DOR, a secondary outcome in the REACH 2 trial, was assessed in responders (CR or PR at day 28) only, and was defined as time from first response until aGvHD progression or until the addition of systemic therapies for aGvHD on or after day 28.¹⁸

The first documented response of CR or PR (i.e., the start date of response) could have occurred before or on day 28. If the start date of response was before day 28, there should have been no progression or addition of systemic therapies for aGvHD between the response start date and day 28. The end date of the DOR interval was defined as the date of aGvHD progression or the date of addition of systemic therapies for aGvHD after day 28.³

REACH 1

DOR at month 6 was a key secondary outcome in the REACH 1 trial and was performed once all patients had completed the day 180 visit. This end point was assessed in patients who had at least 1 response measurement (i.e., PR, VGPR, or CR on or before the start of new anti-GvHD therapy, if appliable). The DOR interval was defined as the difference between the end of response (progression or death) and the start of response (PR or better). Definitions of CR, VGPR, PR, and progressive disease are the same as for ORR at day 28. The investigator’s review of aGvHD assessment data was used for this analysis.¹⁹

DOR at month 3 was a secondary outcome in the REACH 1 trial and was performed once all patients had completed the day 84 visit. DOR at month 3 was derived in the same way as DOR at month 6.¹⁹

Additional analyses were performed as reported in the Clinical Study Report for the DOR using the day 28 response (i.e., patients who had CR, VGPR, or PR at the day 28 assessment or other response assessments within 2 days prior or after day 28, before the start of new anti-GvHD therapy, if applicable) and time to first response; however, these supportive analyses were not pre-specified a priori in the statistical analysis plan of the REACH 1 trial.¹⁹

Best Overall Response

REACH 2

BOR was designated a secondary outcome in the REACH 2 trial and defined as the proportion of patients who achieved overall response (CR or PR) at any time point up to and including day 28 and who had no additional systemic therapy for aGvHD before the time point.¹⁸

REACH 1

The BOR rate was not pre-specified a priori in the statistical analysis plan of the REACH 1 trial, but was assessed as an additional supportive analysis of the primary end point. It was defined as the proportion of patients who achieved a CR, VGPR, or PR at any time point before starting a new anti-aGvHD therapy.¹⁹

Health-Related Quality of Life

REACH 2

The HRQoL outcomes measured in the REACH 2 trial included the FACT-BMT and EQ-5D-5L instruments as secondary outcomes. Neither an analysis plan or objective nor a minimally important difference (MID) for the FACT-BMT and EQ-5D-5L instruments were specified a priori in the statistical analysis plan; it was noted, however, that the scores for each scale were calculated according to the respective user’s guides^47,48 of the instruments. Scores for each scale (mean, SD, median, minimum, and maximum) and changes from baseline to each visit were measured and summarized descriptively.³ The FACT-BMT and EQ-5D-5L questionnaires were administered at baseline and every week in the first 2 months, and every 4 weeks thereafter until EOT. During the crossover period (after completion of all assessments at cycle 7 day 1), disease assessments were planned at the same frequency as during the randomized treatment period. The FACT-BMT was not administered to patients younger than 18 years.³ A detailed discussion and critical appraisal of the HRQoL measures is provided in Appendix 4.

The FACT-BMT instrument is a self-administered questionnaire that combines assessments of 2 tools: the Functional Assessment of Cancer Therapy-General (FACT-G) questionnaire, which assesses the effects of cancer therapy on physical, social/family, emotional, and functional well-being on 23 items; and the bone marrow transplant subscale, which assesses specific bone marrow transplant-related concerns on 23 items.^3,49 The higher the score, the better the quality of life.⁵⁰

No study was found that assessed psychometric properties related to the validity or reliability of the 3-level EQ-5D (EQ-5D-3L) in patients with aGvHD. A MID for the FACT-BMT was not identified in the literature for patients with aGvHD.

The EQ-5D is a generic, utility-based measure of HRQoL. The EQ-5D is a self-administered 2-part questionnaire, consisting of the EQ-5D descriptive system and the EQ-5D visual analogue scale (VAS).⁴⁷ For the descriptive system of the EQ-5D, respondents are asked to indicate their health status that day (i.e., a 1-day recall) on 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). The rating on each dimension is combined to create a descriptive health profile. The EQ-5D-5L was created by the EuroQol Group in 2009 to enhance the instrument’s sensitivity and to reduce the ceiling effects of the EQ-5D-3L.⁴⁷ The EQ-5D VAS is a distinct component of the EQ-5D questionnaire. The VAS score is determined by asking respondents to rate their health that day on a vertical line, with anchors (end points) labelled “Worst imaginable health state” at 0 and “Best imaginable health state” at 100. Although the EQ-5D index score reflects societal preferences for the health state, the VAS captures the individual’s own value or judgment of his or her current health state. EQ-5D VAS scores are not used to create utility scores but provide complementary information to the EQ-5D-5L index score.⁴⁷

No study was found that assessed psychometric properties related to the validity or reliability of the EQ-5D-3L in patients with aGvHD. A MID for the EQ-5D-5L was not identified in the literature for patients with aGvHD.

REACH 1

HRQoL outcomes were not included in the REACH 1 trial.

Event-Free Survival

REACH 2

EFS was a secondary outcome in the REACH 2 trial and was defined as the date of randomization to the date of hematologic disease relapse or progression, graft failure, or death from any cause. The local investigator’s review of hematologic disease relapse or progression, according to protocol-defined criteria, was used for this analysis.³

REACH 1

EFS was not assessed in the REACH 1 trial.

Nonrelapse Mortality

REACH 2

NRM was a secondary outcome in the REACH 2 trial and was defined as the time from date of randomization to date of death not preceded by hematologic disease relapse or progression.¹⁸

REACH 1

NRM was a secondary outcome in the REACH 1 trial and was defined as the proportion of patients whose death was the result of causes other than malignancy relapse at months 6, 9, 12, and 24.¹⁹

Incidence of Malignancy Relapse or Progression

REACH 2

The incidence of malignancy relapse or progression was a secondary outcome in the REACH 2 trial and was defined as the time from date of randomization to hematologic malignancy relapse or progression. Malignancy relapse or progression was assessed according to local institutional practices.³

REACH 1

The incidence of malignancy relapse or progression was not assessed in the REACH 1 trial. However, relapse rate and relapse-related mortality rate were 2 secondary outcomes in the REACH 1 trial and were defined as the proportion of patients who experienced underlying malignancy relapse and the proportion of patients who experienced malignancy relapse with a fatal outcome, respectively.³

Cumulative Steroid Dosing Until Day 56

REACH 2

Cumulative steroid dosing until day 56 was a secondary outcome in the REACH 2 trial and assessed the cumulative steroid dose up to day 56 or discontinuation of randomized treatment for each patient.¹⁸

REACH 1

Average and cumulative steroid dosing on days 28, 56, 100, and 180 was an exploratory outcome in the REACH 1 trial and assessed the cumulative steroid dose up to days 28, 56, 100, and 180 or corticosteroid discontinuation.

In addition, the use of immunosuppressive medications during ruxolitinib treatment was an exploratory outcome in the REACH 1 trial and assessed the proportion of patients taking immunosuppressive medications at certain time points while on ruxolitinib.³

Incidence of cGvHD

REACH 2

The incidence of cGvHD was a secondary outcome in the REACH 2 trial and was defined as the time from randomization to onset of cGvHD.¹⁸

REACH 1

The incidence of cGvHD was an exploratory outcome in the REACH 1 trial and was defined as the proportion of patients who experienced cGvHD.³

Resource Use

REACH 2

The assessment of resource use was a secondary outcome in the REACH 2 trial and was captured for use in a post-study health economics analysis. Resource-use data collected included time to discharge (only for patients starting treatment while hospitalized), measured as the start of treatment to discharge), the number of readmissions to a hospital inpatient unit for any reason by inpatient setting (e.g., intensive care unit, general ward) measured as the number of accesses to hospital for any reason that required at least 1 overnight stay and the duration of readmissions by inpatient setting measured as the number of overnight stays for each access to hospital that required an admission.^3,18

REACH 1

Resource use was not assessed in the REACH 1 trial.

Safety

REACH 2

AEs that occurred or worsened after patients’ informed consent were to be recorded in the AE case report forms. Abnormal laboratory values or test results observed in patients only constituted AEs if they were associated with clinical signs or symptoms, were considered clinically meaningful, required therapy (e.g., hematologic abnormality requiring transfusion or hematological stem cell support), or required changes to the study drug. Components of study end points (i.e., worsening of study indication [aGvHD], including occurrence of aGvHD flare, occurrence of cGvHD, or relapse or recurrence of underlying disease [including fatal outcomes]) were not reported as SAEs and were reported on electronic case report forms other than AE electronic case report forms.³

AEs were assessed and graded according to the Common Terminology Criteria for Adverse Events (CTCAE), version 4.03. If a toxicity was not included in the CTCAE v4.03 criteria, it was graded on a scale of 1 to 5, as follows: 1 = mild, 2 = moderate, 3 = severe, 4 = life-threatening, and 5 = death related to the AE.³

All AEs were documented by the investigator from the date a patient signed informed consent to at least 30 days after the final dose of the study treatment.

The REACH 2 trial included the following parameters for the analysis of AEs: AEs by system organ class and preferred term, severity (based on CTCAE grade), type of AE, relationship to study treatment, seriousness (SAEs and non-SAEs), AEs of special interest, death, AEs leading to treatment discontinuation, AEs leading to dose interruption or adjustment, and AEs leading to fatal outcome.³

The presentation of AEs was performed for the following 4 mutually exclusive categories: pre-treatment period, on-randomized treatment period, on-crossover treatment period, and post-treatment period.³

REACH 1

Safety was designated a secondary outcome in the REACH 1 trial. A TEAE was any AE that was described for the first time or involved the worsening of a pre-existing event after the first dose of the study drug until 30 days after the final dose of the study drug.³

AEs were assessed and graded according to the CTCAE v4.03. If a toxicity was not included in the CTCAE v4.03 criteria, it was graded on a scale of 1 to 4 as follows: 1 = mild, 2 = moderate, 3 = severe, 4 = life-threatening.³

The REACH 2 trial included the following outcomes for the analysis of AEs: AEs by MedDRA preferred term and system organ class, incidence of AEs, relationship to study treatment, seriousness (SAEs and non-SAEs), AEs of special interest, deaths, AEs leading to treatment discontinuation, AEs leading to dose interruption or adjustment, and AEs leading to fatal outcome.³ Safety was assessed with vital signs, Eastern Cooperative Oncology Group (ECOG) performance status, physical examination, 12-lead electrocardiogram, and clinical laboratory assessments.¹⁹

Statistical Analysis

Details related to statistical analyses of efficacy end points are summarized in Table 23 for the REACH 2 trial and Table 24 for the REACH 1 trial.

Sample-Size Determination

REACH 2

The trial sample size of 308 patients was determined based on the primary end point of ORR at day 28.³ Assuming an odds ratio of 1.63 or higher for the primary outcome, the study would have 90% power to demonstrate a statistically significant difference at a 1-sided alpha of 0.025. Several assumptions drove the sample-size determination. First, ORR at day 28 for the BAT group was expected to be 58%, as suggested in the study by Martin et al. (2012),³⁶ based on aggregated results of 29 studies that evaluated available secondary therapies for aGvHD. Second, for the BAT group, it was expected that stratum-specific rates were 69%, 59%, and 50% for grades II, III and IV, respectively, based on the assumed aGvHD grade II:III:IV ratio of 0.2:0.4:0.4. No rationale was provided in the sponsor’s submission for the assumption of the aGvHD grade ratio. Third, it was expected that treatment with ruxolitinib would result in an 18% improvement in ORR, corresponding to an increase in ORR to 75%. No rational was provided in the sponsor’s submission for the assumed ORR benefit with ruxolitinib. If the response rates in grades II, III, and IV in the BAT group were, respectively, 69%, 59%, and 50% (57% overall), corresponding response rates of 78%, 70%, and 62% (68% overall) or more in the ruxolitinib group would achieve statistical significance.³

For the analysis of the key secondary outcome — durable ORR at day 56 — a total of 308 patients was considered sufficient. Assuming an odds ratio of 1.59 or more in the ruxolitinib group, the study would have 90% power to demonstrate a statistically significant difference at a 1-sided alpha of 0.025. Several assumptions drove the sample-size determination. First, durable ORR at day 58 for the BAT group was expected to be approximately 35%, based on a study by van Groningen et al. (2016)⁵¹ that evaluated responses with second-line therapy in 21 patients with severe aGvHD. Second, for the BAT group, it was expected that stratum-specific rates would be 45%, 36%, and 30% for grades II, III, and IV, respectively, based on the assumed aGvHD grade II:III:IV ratio of 0.2:0.4:0.4. No rationale was provided in the sponsor’s submission for the assumption of the aGvHD grade ratio. Third, it was expected that treatment with ruxolitinib would result in a 20% improvement in ORR, corresponding to an increase in durable ORR to 55%. No rationale was provided in the sponsor’s submission for the assumed ORR benefit with ruxolitinib. If the response rates in grades II, III, and IV in the BAT group were, respectively, 45%, 36%, and 30% (35% overall), corresponding response rates of at least 57%, 47%, and 41% (47% overall) in the ruxolitinib group would achieve statistical significance.³

REACH 1

The trial sample size of 70 patients was determined based on the primary end point of ORR at day 28. Assuming an ORR of 60%, the study would have a greater than 90% probability of having a 95% CI with a lower limit of 40% or more. If 37 or more of 70 patients responded (i.e., if the lower limit of the 95% CI for ORR at day 28 exceeded 40%), it was predetermined that the trial results would be considered positive. The minimum clinically meaningful proportion of patients with an objective response was considered to be a 20% improvement over historic data. No rationale was provided in the sponsor’s submission for the assumed minimum clinically meaningful improvement, and no benchmark for historic ORR was provided.^3,19

Interim Analysis

REACH 2

No formal interim analysis was planned a priori in the statistical analysis plan.³ A hierarchical testing procedure was applied for the primary (ORR at day 28) and key secondary end points (ORR at day 56) in the primary analysis (July 25, 2019 data cut-off date), which was planned for when all patients had completed the day 56 visit or discontinued earlier.¹⁸ Formal statistical significance testing with full alpha was only planned a priori for the primary analysis. An updated secondary analysis for secondary end points (January 6, 2020, data cut-off date) was planned for when all patients had completed approximately 6 months of treatment or discontinued earlier. The end of the study occurred when all patients had completed the study (up to 24 months from randomization), unless the patient withdrew consent; at the time of the final analysis (April 23, 2021),⁵² secondary end points were analyzed.

REACH 1

An interim analysis for futility was planned after 35 patients completed the day 28 visit. Enrolment in the REACH 1 trial could have been terminated if the lower boundary of futility was crossed. To calculate the lower boundary of futility, a group sequential-design method for 1 sample binary outcome data and the spending function of Hwang-Shih-DeCani (−4) were used. The study was to be terminated if no more than 15 patients responded at the time of the interim analyses (for the null hypothesis, P = 0.4; for the alternative hypothesis, P = 0.6, and a 1-sided binomial test with alpha of 0.025). This was to provide a 70.03% probability for the response rate of 40% at the interim analysis.³ At the interim analysis, 21 of 35 patients (60.0%) achieved an overall response at day 28, and the study proceeded as planned.¹⁹ No formal data monitoring committee was to review the results of the futility analysis.³

If 37 or more patients responded at the final analysis, it was predetermined that the trial results would be positive, based on 89.88% power for the response rate of 60% at the final analysis, with a type I error of 0.0189.³

The final analysis, when the predetermined threshold (i.e., lower limit of the 95% CI for ORR at day 28 was 40% or greater) was to be assessed, was planned for when 75% of patients had achieved 2-year NRM, died, or were lost to follow-up, whichever occurred first. The final analysis was conducted at the data cut-off date of June 5, 2019. The results of the final analysis have been provided to CADTH by the sponsor and are the focus of this CADTH review.

An earlier data analysis occurred at the July 2, 2018, data cut-off date, which was not pre-specified a priori in the statical analysis plan of the REACH 1 trial. According to the sponsor, the July 2, 2018 data cut-off date was requested by the FDA to provide data for 6-month DOR and 6-month NRM once all patients had completed the 6-month (day 180) study visit, started a new treatment for aGVHD, or discontinued the study treatment.²¹ The results of this earlier data cut-off date were published by Jagasia et al. (2020).⁴⁴ This CADTH review will focus on the results of the final analysis.

Primary Outcome

REACH 2

The primary outcome in the REACH 2 trial was ORR at day 28. A brief overview of the statistical methods used for the primary outcome is provided in Table 23.

To compare ORRs between the 2 study groups, a Cochrane-Mantel-Haenszel chi-square test, stratified by the randomization factor (i.e., aGvHD grade II versus III versus IV) was used at a 1-sided 2.5% level of significance. The primary analysis was based on the full analysis set, per the intention-to-treat principle. ORR was also summarized using descriptive statistics (N, %) and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³). P value, odds ratio, and 95% Wald confidence limits were calculated using the stratified Cochran-Mantel-Haenszel test. Patients with missing assessments were considered nonresponders.³

Supportive analyses for ORR at day 28 were planned as follows:

• ORR at day 28 assessed with the same analysis conventions as the primary efficacy analysis, using all patients in the per-protocol set

• a detailed description of organ-specific response for all organs at day 28, using shift tables of aGvHD stage by organ and treatment group to compare baseline and day 28 value

• logistic regression model to estimate treatment effect adjusted for key baseline and prognostic factors (covariates may include age, sex, race, aGvHD grade, source of graft, criteria for SR-aGvHD, and prior aGvHD therapy in addition to treatment).

A sensitivity analysis was performed using the full analysis set to assess the impact of stratification by comparing the study groups with the Fisher’s exact test. ORR was summarized using descriptive statistics (N, %) and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³).³

In a supportive analysis, ORR at crossover day 28 was summarized descriptively and defined as the proportion of crossover patients with CR or PR at crossover day 28, according to the standard NIH criteria of Harris et al. (2016).¹⁶ Using the local investigator’s review of aGvHD assessment, summary statistics (N, %) with 2-sided exact binomial 95% CIs were presented.³

REACH 1

The primary outcome in the REACH 1 trial was ORR at day 28. A brief overview of statistical methods used for the primary outcome is provided in Table 24.

No formal statistical tests were performed. The 95% CI for ORR was estimated using the exact method for binomial distribution. The primary analysis was performed when the final patient completed the day 28 visit or withdrew from the study. Patients with insufficient response data at day 28 (e.g., death, discontinuation, missing visit) were considered nonresponders.³

An additional analysis was performed, as reported in the Clinical Study Report, for the BOR; however, this supportive analysis was not pre-specified in the statistical analysis plan of the REACH 1 trial. The 95% CI for BOR was estimated using the exact method for binomial distribution.³

Key Secondary Outcome

REACH 2

Durable ORR at day 56 was the key secondary outcome in the REACH 2 trial. A hierarchical testing procedure was applied in which the durable ORR at day 56 was only tested if the ORR at day 28 was statistically significant. The key secondary outcome was tested at the same time as the primary outcome (i.e., at the primary analyses, when all patients had completed their day 56 visit or discontinued the study earlier). To compare durable ORRs at day 56 in the 2 study groups, a Cochrane-Mantel-Haenszel chi-square test, stratified by the randomization factor (i.e., aGvHD grade II versus III versus IV) was used at a 1-sided 2.5% level of significance. The analysis was based on the full analysis set, per the intention-to-treat principle. Durable ORR at day 56 was also summarized using descriptive statistics (N, %) and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³). P value, odds ratio, and 95% Wald confidence limits were calculated using the stratified Cochran-Mantel-Haenszel test. Patients with missing assessments were considered nonresponders. Missing data were handled the same way they were in the primary outcome.

As an additional analysis, durable ORR at crossover day 56 was analyzed and defined as the proportion of all crossover patients who achieved a CR or PR at crossover day 28 and maintained a CR or PR at crossover day 56. Response was assessed relative to the last assessment of aGvHD before or at the start date of crossover treatment (ruxolitinib). Durable ORR at crossover day 56 was summarized using descriptive statistics (N, %) and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³), based on the crossover set.³

REACH 1

DOR at month 6 was the key secondary outcome in the REACH 2 trial. Six-month DOR was assessed when all patients had completed the day 180 visit. A Kaplan–Meier (KM) plot of DOR was presented with its 95% CI; 95% CI was estimated using the method of Brookmeyer and Crowley (1982).⁵⁴ Reasons for censoring included the patient still responding at the data cut-off date and discontinuation at the time of last valid response assessment.³

Additional analyses were performed, as reported in the Clinical Study Report, for the DOR using the day 28 response and time to first response; however, these supportive analyses were not pre-specified a priori in the statistical analysis plan of the REACH 1 trial.³

Secondary Outcomes

REACH 2

All analyses for other secondary end points were noncomparative in nature and were not included in formal hypothesis testing. All analyses were based on the full analysis set.

The ORR at day 14 was a secondary outcome in the REACH 2 trial. ORR at day 14 was also summarized using descriptive statistics (N, %) by treatment group and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³). Odds ratio and 95% Wald confidence limits were calculated using the stratified Cochran-Mantel-Haenszel test.¹⁸

DOR was a secondary outcome in the REACH 2 trial and was calculated for all patients who had CR or PR at day 28. The cumulative incidence rates and associated 95% CIs at 1, 2, 6, 18, and 24 months were assessed for each study group and the cumulative incidence curve was provided. Reasons for censoring included not having experienced events or competing risks before or at the data cut-off date. Censoring occurred at the last response assessment before or at the analysis cut-off date.¹⁸

Cumulative steroid dosing up to day 56 was a secondary outcome in the REACH 2 trial. Overall and weekly cumulative steroid doses for each patient up to day 56 or at discontinuation of randomized treatment were summarized. RDI by week was analyzed relative to the starting dose of corticosteroids, and classified as a complete reduction in which patients are tapered off corticosteroids by day 56, a RDI of no more than 50%, and a RDI of more than 50%. The proportion of patients in each aforementioned group with associated 95% CIs were presented by study group. For weeks ending on days 14, 28, 56, 84, and 168, the average weekly corticosteroid dose was also assessed and plotted.¹⁸

OS was a secondary outcome in the REACH 2 trial. The analysis for OS was conducted according to the randomized treatment group, stratified by aGvHD grade (i.e., aGvHD grade II versus III versus IV). A KM plot of OS was presented by treatment group. Medians and KM estimated hazard risk and corresponding 95% CIs (Brookmeyer and Crowley [1982]⁵⁴) at 1, 2, 6, 12, 18, and 24 months were presented. The HR and 95% CIs were estimated from a stratified Cox proportional hazards regression model. Patients were censored at the latest date a patient was known to be alive (on or before the data cut-off date). Patients who crossed over to ruxolitinib from the BAT group were included in the OS analysis for the BAT group.³

The following exploratory sensitivity analyses were planned a priori if the primary analysis is significant and sufficient OS events have occurred:³

• Rank-preserving structural failure time model (Robins and Tsiatis [1991]⁵⁵ and Korhonen et al. [1999]⁵⁶) to estimate the treatment effect, taking into account the switch from the BAT group to the ruxolitinib group.

• Stratified Cox regression model adjusted for prognostic factors, using a stepwise selection process, if appropriate. Goodness-of-fit of the model was examined.

The following exploratory analyses were planned a priori to be performed as necessary if the primary analysis is significant.³

• Relationship between ORR at day 28 and OS will be performed in the ruxolitinib group in the form of a landmark analysis, using day 28 as the landmark time (patients who died before day 28 were excluded). The null hypothesis was that survival after day 28 would not depend on response status at day 28. KM curve and log-rank test were to be used to compare the survival of responders and nonresponders.

EFS was a secondary outcome in the REACH 2 trial. It was analyzed using the same statistical methods as OS. Reasons for censoring included no known experience of any event. Patients were censored at the latest date a patient was known to be alive (on or before the data cut-off date). Patients who crossed over to ruxolitinib from the BAT group were included in the OS analysis for the BAT group. A sensitivity analysis was planned in which aGvHD progression was included as an event.³

FFS was a secondary outcome in the REACH 2 trial. The cumulative incidence of FFS at 1, 2, 6, 12, 18, and 24 months with associated 95% CIs was assessed with onset of chronic GvHD as the competing risk. The cumulative incidence of each of the 3 components (i.e., date of hematologic disease relapse or progression, NRM, and addition of new systemic aGvHD treatment), taking the other 2 components as a competing risk, was also calculated. The cumulative incidence curves were presented for each study group. A sensitivity analyses in which aGvHD progression was considered an event was planned.³

NRM was a secondary outcome in the REACH 2 trial and was analyzed according to randomized treatment group and strata assigned at randomization. The cumulative incidence of NRM at 1, 2, 6, 12, 18, and 24 months was assessed with underlying disease relapse or progression as the competing events. Reasons for censoring included no confirmed death and the experience of the competing event. Censoring occurred at the latest date the patient was known to be alive (on or before the cut-off date). Sensitivity analyses were planned in which the cumulative incidence curve of NRM and estimates at 1, 2, 6, 12, 18, and 24 months with 95% CIs were presented for patients with underlying hematologic malignant disease in each study group.¹⁸

The incidence of malignancy relapse or progression was a secondary outcome in the REACH 2 trial. The cumulative incidence of malignancy relapse or progression at 1, 2, 6, 12, 18, and 24 months with associated 95% CIs was assessed, with underlying hematologic malignant disease, accounting for NRM, as the competing risk. The proportion of patients who had hematologic malignancy relapse or progression and the associated 95% CI at 1, 2, 6, 12, 18, and 24 months were presented by study group for patients with underlying hematologic malignant disease. Odds ratio and 95% Wald confidence limits were also calculated using the stratified Cochran-Mantel-Haenszel test. Reasons for censoring included no known relapse or progression. Censoring occurred at the latest date the patient was known to be alive (on or before the cut-off date).¹⁸

The incidence of cGvHD was a secondary outcome in the REACH 2 trial. The cumulative incidence of cGvHD at 1, 2, 6, 12, 18, and 24 months with associated 95% CIs was assessed with deaths without prior onset of cGvHD and hematologic disease relapse or progression as the competing risks. Reasons for censoring included no known event or competing risks. Censoring occurred at the latest date of the patient was known to be alive (on or before the cut-off date).¹⁸

BOR response was a secondary outcome in the REACH 2 trial. BOR was summarized using descriptive statistics (N, %) by treatment group and 2-sided exact binomial 95% CIs (Clopper and Pearson [1934]⁵³). Odds ratio and 95% Wald confidence limits were calculated using the stratified Cochran-Mantel-Haenszel test.¹⁸

Resource use was a secondary outcome in the REACH 2 trial. Summary statistics will be provided for each resource component (i.e., time to discharge, readmissions to hospital, duration of readmission to hospital, and proportion of patients with readmission to hospital). The 95% CIs will be presented for the proportion of patients with readmission to hospital, by study group. No further analyses were specified a priori in the statistical analysis plan. Data related to resource use were to be collected for the purpose of the economic evaluation and were analyzed and reported separately.³

REACH 1

OS was a secondary outcome in the REACH 1 trial. The KM plot of OS was presented with its 95% CI; 95% CI was estimated using the method of Brookmeyer and Crowley (1982).⁵⁴ Survival rate at month 6 was estimated. Reasons for censoring included still being alive or being lost to follow-up at the data cut-off date, and occurred at the last date a patient was known to be alive. Additional analyses were performed, as reported in the Clinical Study Report, for OS by response status, deaths by response status, and deaths while still receiving ruxolitinib or within 30 days of the last ruxolitinib dose; however, these analyses were not pre-specified in the statistical analysis plan of the REACH 1 trial.³

FFS was a secondary outcome in the REACH 1 trial. The KM plot of FFS was presented with its 95% CI; 95% CI was estimated using the method of Brookmeyer and Crowley (1982).⁵⁴ Survival rate at month 6 was estimated. Reasons for censoring included still being alive, being lost to follow-up, having no relapse or progression of the underlying malignancy, requiring no additional therapy for aGvHD, and having no demonstrated signs or symptoms of cGvHD.³

ORR at days 14, 56, and 100 was a secondary outcome in the REACH 1 trial. Statistical methods were the same as those used for the primary outcome of ORR at day 28.³

DOR at month 3 was a secondary outcome in the REACH 1 trial and assessed when all patients completed the day 84 visit. Statistical methods were the same as those used for the key secondary outcome of DOR at month 6.³

NRM was a secondary end point in the REACH 1 trial. The cumulative incidence of NRM at 6, 9, 12, and 24 months was assessed with relapse-related mortality as the competing event. Reasons for censoring included no confirmed death. Censoring occurred at the latest date the patient was known to be alive. Per the statistical analysis plan, the 95% CI for NRM was estimated using the exact method for binomial distribution. However, according the Clinical Study Report, changes to the planned analyses included using Marubini and Valsecchi's method (1995)⁵⁷ and the delta method (Hosmer et al. [2008])⁵⁸ with log-log transformation, respectively, for the calculation of cumulative incidence rates and 95% CIs.^3,19

Relapse rate was a secondary end point in the REACH 1 trial. The 95% CI for the relapse rate was estimated using the exact method for binomial distribution.³

Relapse-related mortality rate was a secondary end point in the REACH 1 trial. The 95% CI was estimated using the exact method for binomial distribution.³

Exploratory End Points

REACH 1

Average and cumulative steroid dosing on days 28, 56, 100, and 180 were exploratory end points in the REACH 1 trial. However, according the Clinical Study Report, changes to the planned analyses included replacing cumulative steroid dosing with a summary of the average and relative (to initial dose) corticosteroid dose at selected time points during ruxolitinib treatment.¹⁹

The incidence of cGvHD was an exploratory outcome and the proportion of patients with cGvHD was calculated.

Subgroup Analyses

REACH 2

For each subgroup, the point estimate and 2-sided exact binominal 95% CI (Clopper and Pearson [1934]⁵³) were calculated. The odds ratio was calculated with 95% CI, using a logistic regression model with covariates (i.e., treatment and stratification factors). A forest plot was presented. No formal statistical tests of hypotheses were conducted. Subgroup analyses were planned a priori in the statistical analyses plan. Planned subgroup analyses were to be conducted for the primary end point if statistically significant results were observed:¹⁸

• age (12 years to younger than 18 years, 18 to 65 years, older than 65 years)

• sex

• race

• Europe plus Australia and Canada, Asia excluding Japan, Japan

• aGvHD grade (grade II, III, IV)

• source of grafts (related, unrelated)

• criteria for SR-aGvHD (progression after at least 3 days, failure to achieve a response after 7 days, flare failure during taper)

• prior aGvHD therapy (steroids ± CNIs, steroids ± other systemic aGvHD treatment, steroids ± CNIs ± other systemic aGvHD treatment).

The following subgroups, planned a priori in the statistical analysis plan, aligned with the subgroups pre-specified in the protocol for this CADTH review: overall aGvHD grade, organ involvement for aGvHD, disease underlying aGvHD, age, criteria for SR-aGvHD, prior aGvHD therapy, and stem cell source. Only subgroups identified in the CADTH review protocol are reported in the efficacy section.¹⁸

REACH 1

Additional supportive analyses for the primary end point were performed, as reported in the Clinical Study Report; however, none of the following supportive analyses were pre-specified in the statistical analysis plan of the REACH 1 trial¹⁹:

• day 28 ORR by baseline aGvHD grade

• day 28 ORR by baseline SR subcategory

• day 28 ORR by use of immunosuppressive medications

• day 28 ORR by use of CNIs

• day 28 ORR by average reported daily ruxolitinib dose from day 1 to day 28

• day 28 ORR by age, sex, race, and baseline GvHD organ involvement (liver, upper GI, lower GI, and skin).

Multiplicity

REACH 2

Apart from the pre-specified hierarchical testing of ORR at day 28 and durable ORR at day 56, no adjustments for multiplicity were performed to control the type I error rate.

REACH 1

Not applicable.

Amendments

REACH 2

The protocol of the REACH 2 trial was amended twice (Amendment 1 on May 31, 2017, and Amendment 2 on June 21, 2018). Amendment 1 included changes to exclusion criterion 5, other eligibility criteria, and administrative changes. Amendment 2 included changes to allow for more flexibility in the tapering of corticosteroids, CNIs, and ruxolitinib, and to allow for the complete taper to occur beyond week 24, if required. Amendment 2 also implemented a post-trial-access commitment by Novartis, in which patients who meet certain protocol treatment discontinuation criteria or who were still receiving ruxolitinib at their end of study (approximately 2 years from randomization) and were judged by the investigator to be deriving clinical benefit from ruxolitinib were given the option to continue ruxolitinib outside the study. Further, to align with the clinical management of adolescents and to increase their enrolment in the study, other systemic medications for aGvHD prophylaxis could be continued after randomization for all patients. It was noted in the protocol that the impact of this change on overall patient homogeneity was judged to be limited. The secondary end point, BOR, was added to align with aGvHD publications. A data monitoring committee was added to uphold blinding of members of the study steering committee during their review of pooled safety data. In the decision to add a data monitoring committee, no efficacy or safety data from the study were considered.³

REACH 1

The protocol of the REACH 1 trial was amended twice (Amendment 1 on September 12, 2016, and Amendment 2 on October 4, 2016). Both amendments occurred before the first patient was enrolled in December 2016. Amendment 1 included changes to eligibility criteria, secondary end points, GvHD staging and grading criteria, the starting dose of ruxolitinib, and the ruxolitinib dose in participants with liver GvHD, and added an interim analysis and the ability of investigators to taper ruxolitinib. Amendment 1 addressed regulatory feedback. Key changes included:¹⁹

• adding the assessment of DOR as a key secondary end point

• changing the starting dose to 5 mg twice daily and providing guidance to investigators for escalating the dose to 10 mg twice daily after 3 days if hematologic parameters were stable and no treatment-related toxicity was observed

• changing GvHD staging and grading criteria to Mount Sinai Acute GvHD Consortium guidelines

• expanding eligibility to participants 12 years and older, clarifying requirements for patients defined as SR, expanding eligibility to participants with ANC ≥ 0.5 × 10⁹/L, removing the requirement for platelet engraftment, removing the Eastern Cooperative Oncology Group performance status requirement, clarifying the definition of severe organ dysfunction, and expanding eligibility with respect to prior use of JAK inhibitors

• adding language to reflect a planned interim analysis for futility once 35 participants completed the day 28 visit

• adding toxicity management guidelines to provide guidance on the appropriate management of bilirubin elevations in participants with and without liver GvHD, given the potential for pre-existing cytopenia and liver function test abnormalities

• removing a statement about reducing the total daily dose of ruxolitinib by approximately 50% when administering ruxolitinib with strong CYP3A4 inhibitors or dual moderate inhibitors of CYP2C9 and CYP3A4 (e.g., fluconazole). As azole antifungal support is routinely administered to GvHD patients as a part of their prophylaxis regimen, and as available retrospective ruxolitinib data indicated that starting doses of 5 mg to 10 mg twice daily with concomitant azole use did not result in significant changes in safety, reducing the starting dose of ruxolitinib in patients receiving concurrent azole therapy as is done in other indications was not required in this study.

Amendment 2 included changes to secondary end points, dose modifications of ruxolitinib in participants with liver GvHD, the ability of investigators to taper ruxolitinib, and guidance on corticosteroid tapering. Amendment 2 addressed regulatory feedback received after Amendment 1. Key changes included:¹⁹

• changing the key secondary end point from 3-month DOR to 6-month DOR, and 3-month DOR was added as a secondary end point

• modifying toxicity management guidelines to provide guidance on the appropriate management of bilirubin elevations in participants with and without liver GvHD based on the upper limit of the normal range instead of total bilirubin concentration, given the potential for different ranges for adolescents

• revising instructions for tapering ruxolitinib to permit tapering after day 180 for participants achieving CR or VGPR; allowing investigators to continue treating participants with CNIs and other GvHD prophylaxis medications as appropriate; allowing investigators to initiate an earlier taper of ruxolitinib, with sponsor approval; and allowing participants who have completely tapered off ruxolitinib to restart treatment when GvHD symptoms reappear

• deleting specific guidance on corticosteroid tapering; corticosteroids were to be tapered per institutional guidelines at a rate commensurate with the resolution of GvHD manifestations.

Table 23: Statistical Analysis of Efficacy End Points in REACH 2 (Outcomes Are Presented in Order of Priority as Identified in the CADTH Review Protocol)

aGvHD = acute graft-versus-host disease; BMT = bone marrow transplant; BOR = best overall response; cGvHD = chronic graft-vs.-host disease; CI = confidence interval; DOR = duration of response; EFS = event-free survival; EQ-5D-5L = 5-level EQ-5D; FACT-BMT = Functional Assessment of Cancer Therapy-Bone Marrow Transplant; FFS = failure-free survival; HR = hazard ratio; KM = Kaplan–Meier; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival; RDI = relative dose intensity; SD = standard deviation; SR = steroid refractory; VAS = visual analogue scale; vs. = versus.

^aORR at crossover is defined as the proportion of crossover patients with a CR or PR at crossover day 28, according to the standard criteria of Harris et al. (2016).¹⁶ Response is relative to the last assessment of aGvHD before or at the start of crossover treatment (ruxolitinib).

^bDurable ORR at crossover day 56 was defined as the proportion of all crossover patients who achieve a CR or PR at crossover day 28 and maintained a CR or PR at crossover day 56. Response was assessed relative to the last assessment of aGvHD before or at the start of crossover treatment (ruxolitinib).³

Sources: Statistical analysis plan (REACH 2),³ Study Protocol (REACH 2).³

Table 24: Statistical Analysis of Efficacy End Points in REACH 1 (Outcomes Are Presented in Order of Priority as Identified in the CADTH Review Protocol)

aGvHD = acute graft-versus-host disease; BOR = best overall response; cGvHD = chronic graft-vs.-host disease; CI = confidence interval; CNI = calcineurin inhibitor; CR = complete response; DOR = duration of response; FFS = failure-free survival; GI = gastrointestinal; GvHD = graft-vs.-host disease KM = Kaplan–Meier; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival; SR = steroid refractory; PR = partial response; VGPR = very good partial response; vs. = versus.

^aParticipants who had progressive GvHD after 3 days of primary treatment, GvHD that had not improved after 7 days of primary treatment, previously started steroid therapy at a lower dose but developed new GvHD in another organ system, and who could not tolerate a steroid taper.

Sources: Statistical analysis plan (REACH 1),³ Clinical Study Report (REACH 1).¹⁹

Analysis Populations

REACH 2

At the time of the primary analysis, the primary end point and all secondary end points were analyzed using the full analysis set, as defined in Table 25. In supportive analyses at the primary analysis, ORR at day 28 was analyzed using the per-protocol set, as defined in Table 25. The crossover analysis set included patients randomized to the BAT group who crossed over to ruxolitinib treatment between day 28 and week 24. This set was used for all analyses for the crossover patients. Analyses of safety were performed using the safety population.

REACH 1

At the final analysis, all efficacy analyses were analyzed using the efficacy-evaluable population, as defined in Table 25. Analysis of safety was performed using the safety-evaluable population.

Table 25: Analysis Populations in the REACH 2 and REACH 1 Trials

aGvHD = acute graft-versus-host disease; BAT = best available therapy; CNI = calcineurin inhibitor; ITT = intention-to-treat; SR = steroid refractory.

Sources: Clinical Study Report (REACH 2),¹⁸ Clinical Study Report (REACH 1).¹⁹

Results

Patient Disposition

REACH 2

Details of the patient disposition in the REACH 2 trial are summarized in Table 26. A total of 620 patients were screened and, of those, 49.8% (N = 309) of patients were randomized to receive ruxolitinib (n = 154) or BAT (n = 155). Reasons for not being randomized included not meeting the inclusion criteria (n = 296), death (n = 7), patient chose not to participate (n = 4), and other reasons (n = 3). The 302 (97.7%) randomized patients who were treated included 152 of 154 (98.7%) patients in the ruxolitinib group and 150 of 154 (96.8%) patients in the BAT group. As of the primary analysis (July 25, 2019, data cut-off date), 111 (72.1%) patients in the ruxolitinib group and 132 (85.2%) patients in the BAT group had discontinued treatment. The main reasons for discontinuation of the assigned treatment were (ruxolitinib versus BAT) lack of efficacy (20.8% versus 43.9%), AEs (16.9% versus 3.2%), and death (16.2% versus 14.2%). At the primary analysis, 12 (7.8%) patients in the ruxolitinib group and 6 (3.9%) patients in the BAT group were still on randomized treatment.¹⁸

As of the secondary analysis (January 6, 2020, data cut-off date), 116 (75.3%) patients in the ruxolitinib group and 134 (86.5%) patients in the BAT group had discontinued treatment. The main reasons for discontinuation of the assigned treatment were (ruxolitinib versus BAT) lack of efficacy (20.8% versus 44.5%), AEs (17.5% versus 3.9%), and death (16.2% versus 13.5%). At the secondary analysis, 3 (1.9%) patients in the ruxolitinib group and 0 patients in the BAT group were still on randomized treatment.¹⁸

At the final analysis, no patient was on ongoing treatment, and 22.7% and 12.9% of patients in the ruxolitinib and BAT groups, respectively, had completed the treatment period (see Table 26).

Table 26: Patient Disposition in REACH 2, Full Analysis Set (Data Cut-Off Dates of July 25, 2019, January 6, 2020, and April 23, 2021)