CADTH Reimbursement Review

Ruxolitinib (Jakavi)

Sponsor: Novartis Pharmaceutical Canada Inc.

Therapeutic area: 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.

Introduction

Hematopoietic stem cell transplants provide stem cells to patients whose bone marrow has been destroyed by disease, chemotherapy, or radiation.¹ The 2 main types of stem cell transplant are autologous and allogeneic transplants. Although an allogeneic stem cell transplant (alloSCT) has curative potential, there is a risk 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 graft-versus-host disease (GvHD).¹ GvHD is a multi-system disorder in which donor-derived immune cells initiate an adverse immune reaction to the transplant recipient’s tissues, cells, and organs, leading to tissue damage, organ failure, or death.² GvHD has been shown to be the leading cause of morbidity and nonrelapse mortality (NRM) in patients after alloSCT, affecting up to 70% of patients who receive stem cell transplants.^2,3 Between 2008 and 2019, an estimated 13,033 transplants were performed in Canada, of which 5,672 were alloSCTs.⁴

Chronic GvHD (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.⁵ cGvHD commonly involves the skin, mouth, and liver, and less frequently involves the eye, lung, gastrointestinal (GI) tract, joint/fascia, and genital tract.⁶ Signs and symptoms can include lichen planus-like lesions or full sclerosis, muscle pain or joint fasciitis, vulvo-vaginitis, bronchiolitis obliterans, Sjogren’s syndrome, chronic immunodeficiency, primary biliary cirrhosis, wasting syndrome, immune cytopenias, and chronic immunodeficiency, and can cause damage to the GI tract and liver.⁷

Patients with mild cGvHD, whose disease is limited to the skin are commonly treated with topical therapies (e.g., corticosteroids or calcineurin inhibitors [CNIs]). Patients with moderate-to-severe cGvHD commonly receive systemic treatment with corticosteroids as the standard first-line therapy.² Systemic corticosteroids may be administered alone or in combination with additional immune-suppressants, such as CNIs. First-line therapies have challenges, including immunosuppression that may lead to an increase in malignancy relapse, and other side effects, such as infections, myopathy, cataracts, hyperglycemia, decline in bone mass, and avascular necrosis.⁷ In patients who experience clinical improvements, tapering of corticosteroids is recommended to reduce the risk of infection and other toxicities.² About 50% to 60% of patients do not respond to first-line systemic corticosteroid treatments, have inadequate control of disease that requires the addition of another systemic therapy, or fail to taper corticosteroids. Currently, there is no consensus on standard second-line therapies for patients with steroid-refractory cGvHD (SR-cGvHD).² According to the clinical experts consulted by CADTH, available second-line options in Canada include extracorporeal photopheresis (ECP), mycophenolate mofetil (MMF), etanercept, low-dose methotrexate (MTX), infliximab, mechanistic target of rapamycin (mTOR) inhibitor (e.g., sirolimus), imatinib, rituximab, ibrutinib, low-dose interleukin (IL)-2, pulsed cyclophosphamide, and, in rare cases, pentostatin. In the absence of proven treatment options, there is interprovincial variation in standard practices and access to therapies. The clinical experts expressed several challenges with currently available therapies in this heavily pre-treated target population, including low rates of partial responses (PRs) in patients with cGvHD (estimated at 30%; complete responses [CRs] are uncommon) and refractoriness or intolerance to currently available therapies. According to the clinical experts, responses in this patient population enable the tapering of steroids to mitigate long-term side effects (e.g., osteoporosis, avascular necrosis, hypertension, and diabetes mellitus with end-organ damage) and the risk of opportunistic infection. It was emphasized by the clinical experts that infectious complications are a leading cause of nonrelapse mortality in SR-cGvHD.

Adolescents 12 years and older comprise approximately 5% of the cGvHD population. Adolescents and adults are managed in similar ways in clinical practice. Although the long-term disease outcome of cGvHD appears to be slightly more favourable in children than in adults, the 2 groups suffer from similar complications and have poor long-term outcomes.²

There is consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve health-related quality of life (HRQoL), reduce symptoms of cGvHD, enhance a patient’s performance status, and improve overall survival (OS). They highlight the need for a convenient oral route of administration to improve adherence and reduce the need for hospital-based or ambulatory centre resource use.

Ruxolitinib is a Janus kinase (JAK) inhibitor undergoing review by Health Canada, through the expedited pathway (Orbis), for the treatment of GvHD in patients 12 years and older who have an inadequate response to corticosteroids or other systemic therapies. The sponsor’s requested reimbursement criteria for ruxolitinib are per the Health Canada indication under review. This CADTH review focuses on cGvHD. Ruxolitinib is concurrently being reviewed by CADTH in the acute GvHD (aGvHD) setting for the treatment of aGvHD 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 who are resistant to hydroxyurea or who are intolerant of hydroxyurea, according to the modified European LeukemiaNet criteria used in the RESPONSE trial, and have a good performance status. Ruxolitinib is available as 5 mg, 10 mg, 15 mg, and 20 mg tablets. The recommended dose is 10 mg administered orally twice daily. The product monograph states that the tapering of ruxolitinib may be considered in patients who have responded and have discontinued corticosteroids. It is recommended that ruxolitinib be tapered with a dose reduction to 50% every 2 months; in the event that signs or symptoms of GvHD reoccur during or after the taper, re-escalation of ruxolitinib should be considered.⁸

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

Stakeholder Perspectives

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

Patient Input

Eight patient groups — Lymphoma Canada, Lymphoma and Leukemia Society of Canada, Chronic Lymphocytic Leukemia Canada, Myeloma Canada, Aplastic Anemia and Myelodysplasia Association of Canada, Canadian MPN Research Foundation and the Chronic Myelogenous Leukemia Network, the Chronic Myeloid Leukemia Network, and Cell Therapy Transplant Canada (CTTC) — co-created a joint patient input for this review. The input was based on an online survey and responses from a total of 68 participants were included in the patient input. Sixty patients reported having received a stem cell transplant (SCT), 6 reported not having received an SCT, and 2 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% developed aGvHD, 24% developed cGvHD, and 62% developed 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%). To manage their GvHD, respondents reported requiring numerous medical consultations, hospital stays, and nights away from home. Respondents indicated that a range of GvHD symptoms significantly affected ’their daily activities and had detrimental effects on ’their quality of life. Respondents highlighted interruption of life goals and accomplishments (career, school), difficulty sleeping, impact on mental health (stress, anxiety, worry, and concentration problems), and financial issues. Respondents commonly reported experiencing symptoms such as 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 difficulties, 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 received treatment in the outpatient setting (rather than requiring an overnight hospital stay), having access to treatment locally (rather than requiring extensive amount of travel), treatment being covered by insurance or drug plans, and the treatment being recommended by health care professionals were perceived to be very important by respondents. Respondents who had direct experience with ruxolitinib indicated that, overall, the drug was effective, improved their quality of life, had tolerable side effects, they would take it again if recommended by their physician, and they 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-cGvHD, except ibrutinib, which has been authorized since 2017 for the treatment of adults with SR-cGvHD but has not undergone review by CADTH and is not publicly reimbursed; it is available through private drug insurance only. According to the clinical experts consulted by CADTH, available second-line options in Canada include ECP, MMF, etanercept, low-dose MTX, infliximab, mTOR inhibitor (e.g., sirolimus), imatinib, rituximab, ibrutinib, low-dose IL-2, pulsed cyclophosphamide, and, in rare cases, pentostatin. 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 cGvHD, enhance a patient’s performance status, and improve OS. The clinical experts highlighted the need for a convenient oral route of administration to improve adherence and reduce the need for hospital-based or ambulatory centre resource use. Ruxolitinib can be used, they explained, as an add-on to an ’immunosuppressive regimen of corticosteroids with or without CNI in patients aged 12 years and older with moderate or severe SR-cGvHD, per the REACH3 trial. They agreed that ruxolitinib, as a therapy for SR-cGvHD, will likely shift the current treatment paradigm. The clinical experts consulted by CADTH agreed that patients who meet the inclusion and exclusion criteria of the REACH3 trial should be eligible for ruxolitinib therapy. They noted that potential subgroups most in need of ruxolitinib therapy are patients with glucocorticoid-refractory (as opposed to glucocorticoid-dependent) cGvHD and those with bronchiolitis obliterans. Potential patient subgroups least likely to benefit from ruxolitinib include patients with isolated lichenoid cGvHD, who may preferentially be treated with ECP or subcutaneous low-dose IL-2, rather than ruxolitinib. Patients with strictly autoimmune cGvHD manifestations, such as immune thrombocytopenia (as well as immune hemolytic anemia, immune glomerulonephritis, and myasthenia gravis), may preferentially be treated with rituximab. The clinical experts consulted by CADTH felt that it would be reasonable to generalize results from the REACH3 trial to patients who have received 2 or more systemic treatments for cGvHD in addition to corticosteroids with or without CNI, as well as to patients with an Eastern Cooperative Oncology Group (ECOG) performance status score of 3 or a Karnofsky performance status (KPS) or Lansky performance status (LPS) scores below 60%, if performance status is related to cGvHD and its symptoms. Furthermore, it was agreed 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 mild cGvHD.

In the opinion of the clinical experts consulted by CADTH, an accurate assessment of response in patients with cGvHD is based on the National Institutes of Health (NIH) consensus criteria, as was used in the REACH3 trial. Response to treatment is usually assessed every 2 to 4 weeks, depending on the severity of cGvHD. Weekly assessment may be required initially. The clinical experts indicated that the most clinically meaningful responses to treatment include overall response (complete or partial), improvements in HRQoL and performance status, reduction in cGvHD symptoms (frequency and severity), stability of disease (no deterioration), and improvements in OS, as well as the ability to reduce the dose of immunosuppression drugs and/or corticosteroids without a flare in cGvHD signs and symptoms that leads to the initiation of another drug for cGvHD.

In the opinion of the clinical experts consulted by CADTH, treatment with ruxolitinib should be discontinued if a patient experiences cGvHD disease progression, relapse of underlying hematologic malignancy, or intolerable toxicity (e.g., anemia, thrombocytopenia, or neurologic toxicity that cannot be managed with a drug interruption and/or dose reduction). Tapering ruxolitinib in responders may be considered after 24 weeks of therapy.

Clinician Group Input

Two clinician group inputs were provided: 1 from CTTC (based on input from 8 clinicians), and 1 from Ontario Health (Cancer Care Ontario [CCO]) Complex Malignant Hematology (based on input from 2 clinicians). Overall, the views of the clinician groups were consistent with those of the clinical experts consulted by CADTH, indicating that, based on the evidence from the REACH3 trial, it is anticipated that ruxolitinib will become the dominant first-line therapy for SR-cGvHD. The outcomes assessed in the REACH3 trial were judged to be applicable to Canadian clinical practice and reflective of clinically meaningful responses. It was noted by both inputs that ruxolitinib is not as immunosuppressive as other available therapies. The clinicians from Ontario Health CCO noted the drawbacks of currently available therapies, such as 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. It was highlighted by the input from CTTC that a Health Canada–approved and provincially funded therapy for SR-cGvHD would be an important step forward in the current target setting, with existing therapies offering low response rates and high rates of toxicity. According to input from CTTC, experience with ruxolitinib (accessible through a compassionate access program) and real-world effectiveness appear to be similar to findings observed in the REACH3 trial, with 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 REACH3 trial and other clinical considerations to provide responses to Provincial Advisory Groups drug program implementation questions. Refer to Table 4 for more details.

Pivotal Studies and Protocol-Selected Studies

Description of Studies

The REACH3 trial is a an ongoing, international, multi-centre, open-label, randomized phase III trial comparing ruxolitinib (10 mg, oral, twice daily) with the investigator’s choice of best available therapy (BAT) in patients 12 years and older with moderate or severe SR-cGvHD. Patients continued to receive their systemic immunosuppressive regimen of corticosteroids with or without CNI, per the standard of care. A total of 329 patients were randomized in a 1:1 ratio to receive ruxolitinib or BAT. Randomization was stratified by cGvHD severity, per 2014 NIH consensus criteria (Jagasia et al., 2015)⁹ (moderate versus severe) (refer to Table 38 in Appendix 3 for NIH consensus criteria for cGvHD staging). The primary outcome was overall response rate (ORR) on cycle 7 day 1, and key secondary outcomes included failure-free survival (FFS) and modified Lee Symptom Scale score. Additional secondary outcomes were ORR at cycle 4 day 1, HRQoL, symptom severity, duration of response (DoR), best overall response (BOR), OS, NRM, incidence of malignancy relapse or progression (MR), steroid dosing, resource use, and safety.

The REACH3 trial enrolled patients aged 12 years and older who had undergone alloSCT, had evidence of myeloid and platelet engraftment (absolute neutrophil count > 1,000/mm³ and platelet count > 25,000/mm³), and were diagnosed with moderate or severe cGvHD, which was determined to be corticosteroid-refractory per NIH consensus criteria (Martin et al. 2015)¹⁰ (refer to Table 6). Patients with 2 or more systemic treatments (BAT) for cGvHD in addition to corticosteroids with or without CNI, impaired renal or GI function, or liver disease associated or not associated with GvHD were excluded (refer to Table 6). The baseline demographic characteristics of patients, underling disease history, and GvHD history at baseline are summarized in Tables 7, Table 8, and Table 9, respectively (refer to Table 36 in Appendix 3 for more detailed baseline demographic characteristics). The mean ages of the ruxolitinib and BAT groups, respectively, were 45.9 years (SD = 15.68) and 47.2 years (SD = 16.17). The ruxolitinib group had a lower proportion (ruxolitinib versus BAT) of female patients (33.0% versus 43.9%), of patients meeting the corticosteroid-refractory A criteria (37.6% versus 44.5%), and of patients who had received only steroids as prior systemic cGvHD, SR-GvHD, or cGvHD therapy (42.4% versus 49.4%), and had a higher proportion of patients with prior aGvHD of grade II (32.1% versus 26.2%) and of patients meeting the corticosteroid-refractory B criteria (35.2% versus 25.6%). The majority of patients (ruxolitinib versus BAT) had severe SR-cGvHD (58.0% versus 54.9%) and met corticosteroid-refractory A (37.6% versus 44.5%) or B (35.2% versus 25.6%) criteria, and most patients had received either only steroids (42.4% versus 49.4%) or steroids plus CNI (41.2% versus 42.1%) as previous systemic cGvHD or SR-cGvHD therapy. Malignant leukemia/ myelodysplastic syndrome was the most common underlying disease (ruxolitinib versus BAT) (73.3% versus 74.4%), and the mean time from transplant to cGvHD diagnosis and from initial diagnosis to randomization were similar across groups (247.0 versus 230.0 days and 3.90 versus 3.52 years, respectively).

This CADTH review is based on the data cut-off date of May 8, 2020. Release of the final study results is expected after completion of the study, which is expected to be in the third or fourth quarter of 2022.¹¹ A fixed-sequence hierarchical testing procedure was applied to the primary and 2 key secondary end points, which included the interim analysis (when 196 patients [60% of the targeted 324 patients] completed cycle 7 day 1 visit or discontinued earlier; July 9, 2019 data cut-off date) and the primary analysis (all 329 patients completed the cycle 7 day 1 visit or discontinued earlier; May 8, 2020 data cut-off date).¹²

Efficacy Results

The key efficacy results from the REACH3 trial are summarized in Table 2. For the primary analysis (May 8, 2020 data cut-off date), median FFS was not reached (95% confidence interval [CI], 18.6 to not estimable [NE]) in the ruxolitinib group, and was 5.7 (95% CI, 5.6 to 6.5) months in the BAT group, with a stratified hazard ratio (HR) of 0.370 (95% CI, 0.268 to 0.510) in favour of ruxolitinib. FFS was not formally tested in the primary analyses, given that results reached statistical significance in the interim analysis.

For the primary analysis (May 8, 2020), the proportion of patients who achieved an overall response at cycle 7 day 1 was higher in the ruxolitinib group than in the BAT group. ORR at cycle 7 day 1 was achieved by 49.7% (95% CI, 41.8 to 57.6) of patients in the ruxolitinib group and by 25.6% (95% CI, 19.1 to 33.0) of patients in the BAT group, with a stratified odds ratio of 2.99 (95% CI, 1.86 to 4.80). ORR at cycle 7 day 1 was not formally tested in the primary analyses, given that results reached statistical significance in the interim analysis. The proportion of patients with CR and PR was 6.7% (n = 11) and 43.0% (n = 71), respectively, in the ruxolitinib group, and 3.0% (n = 5) and 22.6% (n = 37), respectively, in the BAT group. The ORR cycle 7 day 1 supportive analysis, which used the per-protocol analysis set, showed results consistent with ORR results for the full analysis set. The treatment effect on ORR at cycle 7 day 1 was consistent with the primary analysis across patient subgroups, except for the “prior cGvHD therapy with steroid + CNI + other systemic therapy” and “prior cGvHD therapy with steroid + other systemic therapy” subgroups. Of note, the sample sizes of these subgroups were small (fewer than 20 patients in each study group). Several other subgroups (i.e., 12 to < 18 years, > 65 years, criteria C) had relatively small sample sizes (< 50 patients in either group). The wide CIs in subgroups reflected uncertainty in the effect estimates.¹²

In the primary analysis, results for the modified Lee Symptom Scale suggested that the rate of responders (responders include patients who achieved an improvement of ≥ 7 points on the total symptom score [TSS] from baseline) up to cycle 7 day 1 was higher in the ruxolitinib group (24.2% [95% CI, 17.9 to 31.5]) than in the BAT group (11% [95% CI, 6.6 to 16.8]), with an odds ratio of 2.62 (95% CI, 1.42 to 4. 82). Improvement in the TSS response was formally tested in the primary analysis, as the result in the interim analysis did not reject the null hypothesis.

At the May 8, 2020 data cut-off date, the proportion of patients that had achieved BOR (CR or PR at any time point up to and including cycle 7 day 1, and before the start of change in or addition of systemic therapy for cGvHD) was 76.4% (95% CI, 69.1 to 82.6) in the ruxolitinib group and 60.4% (95% CI, 52.4 to 67.9) in the BAT group, with an odds ratio of 2.17 (95% CI, 1.34 to 3.52). Among the patients who achieved a BOR, median DoR was not reached (95% CI, 20.2 to NE) in the ruxolitinib group and was 6.2 (95% CI, 4.7 to 13.1) months in the BAT group.

As of the May 8, 2020 data cut-off date, 58 deaths occurred across both study groups. The median duration of follow-up for OS was 57.3 weeks in all patients, 56.6 weeks in the ruxolitinib group, and 57.9 weeks in the BAT group. Median OS was not reached (95% CI, NE to NE) in the 2 study groups, with a stratified HR of 1.86 (95% CI, 0.648 to 1.820).

During the day 166 to day 168 time interval (end of cycle 6), a similar number of patients in both study groups achieved a reduction in corticosteroid dose of at least 50% (normalized for body weight) from baseline (ruxolitinib group: 84 of 118 patients [71.2%]; BAT group: 80 of 115 patients [69.6%]). The reduction in steroid dose in the ruxolitinib group was slightly (but consistently) higher than in the BAT group. The number of patients with no steroids during the day 155 to day 168 time interval was 37 (31.4%) in the ruxolitinib group and 32 (27.8%) in the BAT group.¹²

Harms Results

The key harms outcomes reported in the primary analysis (May 8, 2020 data cut-off date) of the REACH3 trial are summarized in Table 2. Safety data in Table 2 are summarized up to cycle 7 day 1. Overall, safety data for the main treatment period were consistent with safety data up to cycle 7 day 1, and can be found in Table 30.

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

BAT = best available therapy; BOR = best overall response; CI = confidence interval; CR = complete response; DoR = duration of response; FFS = failure-free survival; HR = hazard ratio; NE = nonevaluable; PR = partial response; OS = overall survival; TEAE = treatment-emergent adverse event.

^aHR obtained from stratified Cox model using cGvHD severity at randomization as strata.

^bP value is nominal.

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

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

^eOther: Patient with additional systemic therapies along with CR/PR per investigator assessment.

^fSubjects with change in or addition of new systemic cGvHD treatment are counted as nonresponders irrespective of the TSS value.

^gOne-sided P value, odds ratio, and 95% CI are calculated using stratified Cochran-Mantel-Haenszel test.

Source: Clinical Study Report.¹²

Up to cycle 7 day 1, the percentage of patients reporting at least 1 treatment-emergent adverse event (TEAE) was 97.6% in the ruxolitinib group and 91.8% in the BAT group. The most commonly reported TEAEs in the ruxolitinib group (ruxolitinib versus BAT) were anemia (29.1% versus 12.7%), pyrexia (15.8% versus 9.5%), increase in alanine aminotransferase (15.2% versus 4.4%), hypertension (15.8% versus 12.7%), and increase in blood creatine (13.9% versus 4.4%).¹² The most commonly reported TEAEs in the BAT group (ruxolitinib versus BAT) were diarrhea (10.3% versus 13.3%), anemia (29.1% versus 12.7%), hypertension (15.8% versus 12.7%), pneumonia (10.9% versus 12.7%), and nausea (9.1% versus 10.1%). The percentage of patients reporting grade 3 or higher TEAEs up to cycle 7 day 1 was similar across the study groups (57.0% of patients in the ruxolitinib group and 57.6% in the BAT group). The most commonly reported grade 3 or higher TEAE (ruxolitinib versus BAT) was anemia in the ruxolitinib group (12.7% versus 7.6%) and pneumonia in the BAT group (8.5% versus 9.5%). Other commonly reported grade 3 or greater TEAEs that occurred across the 2 treatment groups up to cycle 7 day 1 (ruxolitinib versus BAT) included neutropenia (8.5% versus 3.8%), thrombocytopenia (10.3% versus 5.7%), increase in gamma-glutamyl transferase (6.7% versus 1.9%), and hypertension (4.8% versus 7.0%).

Up to cycle 7 day 1, the percentage of patients reporting serious TEAEs was 33.3% in the ruxolitinib group and 36.7% in the BAT group. The most commonly reported serious TEAE in the 2 study groups was pneumonia, occurring in 7.9% of patients in the ruxolitinib group and 8.2% of patients in the BAT group. Other commonly reported serious TEAEs occurring across both treatment groups (ruxolitinib versus BAT) included pyrexia (4.8% versus 1.9%), febrile neutropenia (1.8% versus 1.3%), pulmonary embolism (1.2% versus 1.9%), and acute kidney injury (1.2% versus 1.9%).

Up to cycle 7 day 1, the percentage of patients discontinuing study treatment due to TEAEs was 16.4% in the ruxolitinib group and 7.0% in the BAT group. The most commonly reported TEAEs leading to treatment discontinuation in the 2 study groups was pneumonia (4.8% in the ruxolitinib group and 1.3% in the BAT group), followed by anemia (0.6% in the ruxolitinib group and 0.6% in the BAT group). Pneumothorax occurred in 1.2% of patients in the ruxolitinib group and in no patients in the BAT group.

Up to cycle 7 day 1, there were 13 (7.9%) on-treatment deaths in the ruxolitinib group and 9 (5.7%) in the BAT group. The most common cause of on-treatment death up to cycle 7 day 1 was the study indication (cGvHD and/or complications attributed to treatment for cGvHD) in 12 (7.3%) and 6 (3.8%) patients in the ruxolitinib and BAT groups, respectively. One patient in the ruxolitinib group died more than 30 days after the last dose due to general physical health deterioration. In the BAT group, 1 patient each died from pneumonia, sepsis, and systemic infection.

Up to cycle 7 day 1, the most commonly reported infections in the ruxolitinib and BAT groups were infections excluding tuberculosis (62.4% and 58.2%, respectively), other infections (48.5% and 47.5%, respectively), pneumonia (19.4% and 17.1%, respectively), opportunistic infections (11.5% and 12.0%, respectively), urinary tract infections (8.5% and 6.3%, respectively), cytomegalovirus (CMV) infection disease (9.1% and 10.8%, respectively), and sepsis and septic shock (2.4% and 6.3%, respectively).

Up to cycle 7 day 1, there were 31 (18.8%) patients in the ruxolitinib group and 23 (14.6%) patients in the BAT group who experienced lipid abnormality events of any grade. The most commonly reported lipid abnormalities in the ruxolitinib and BAT groups, respectively, were hypertriglyceridemia (9.7% and 8.2%), increase in blood cholesterol (7.3% and 4.4%), hypercholesterolemia (5.5% and 1.3%), and hyperlipidemia (2.4% and 2.5%).

Up to cycle 7 day 1, there were 16 (9.7%) patients in the ruxolitinib group and 17 (10.8%) patients in the BAT group who experienced renal and urinary disorders of any grade. The most commonly reported renal and urinary disorders in the ruxolitinib and BAT groups were acute kidney injury (2.4% and 3.8%, respectively), renal failure (1.2% and 1.3%, respectively), hematuria (1.2% and 1.9%, respectively), and proteinuria (0.6% and 1.3%, respectively).

Up to cycle 7 day 1, the most commonly reported cytopenia events of any grade in the ruxolitinib and BAT groups were erythropenia (29.7% and 12.7%, respectively), leukopenia (18.8% and 13.9%, respectively), thrombocytopenia (21.2% and 14.6%, respectively), and other cytopenias (1.2% and 1.3%, respectively).

Up to cycle 7 day 1, the most commonly reported bleeding events of any grade in the ruxolitinib and BAT groups were hemorrhage (11.5% and 14.6%, respectively), hemorrhage events (6.7% and 10.1%, respectively), bruising (4.2% and 2.5%, respectively), and GI bleeding (1.2% and 3.2%, respectively).

Critical Appraisal

The REACH3 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 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 [e.g., modified Lee Symptom Scale]) may be particular at risk of bias because of the open-label design. Furthermore, the underlying complexity of cGvHD has been acknowledged to be a key challenge in the design and analysis of clinical trials in the current target setting, and may contribute to subjective inter-physician variation in response assessments. To mitigate the impact of this bias, the investigators used standardized criteria (i.e., cGvHD disease evaluation and response-assessment criteria for all organs were in accordance with NIH consensus criteria [Lee 2015])¹³ 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., sex, race, SR criteria B met), they were unlikely to influence clinical outcomes, according to the clinical experts consulted by CADTH. Patients in the BAT group who did not achieve responses were allowed to add or initiate a new systemic therapy up to cycle 7 day 1 without having to discontinue the initial study treatment; however, in the ruxolitinib group, treatment was discontinued if patients changed or added a systemic therapy. This design feature may have biased the reporting of adverse events (AEs) leading to treatment discontinuation against the ruxolitinib group. The clinical experts consulted by CADTH noted that changing or initiating new systemic cGvHD therapies is reflective of clinical practice. It was felt by the clinical experts, that changes to the BAT treatment up the cycle 7 day 1 were unlikely to affect OS results, given the similar efficacy and similar responses achieved with various systemic therapies. Modified Lee Symptom Scale scores were measured up to cycle 7 day 1 (cycle length = 28 days), which may not represent an accurate picture of a patient’s experience with ruxolitinib over a prolonged period of time. However, the assessment time frame coincided with the assessment of the primary outcome, ORR at cycle 7 day 1, and the clinical experts consulted by CADTH agreed that changes in symptom severity would be apparent during the first 6 cycles after the start of treatment. Given several important limitations, including the noninferential analyses, the significant decline in patients available to provide assessment over time, and the open-label design of the trial, the ability to interpret results for the European Quality of Life 5-Five Dimensions 5-Levels (EQ-5D-5L) and the Functional Assessment of Cancer Therapy (FACT)-Bone Marrow Transplantation (BMT) scores is limited.

It was noted that few patients in the trial were younger than 18 years of age. However, the clinical experts supported generalizing the study results to adolescents younger than 18 years of age, as the management of these patients is similar to that of adults in clinical practice, the safety profile of ruxolitinib in these patients was acceptable and similar to the overall safety set, and there is no biologic rational to assume that outcomes with ruxolitinib would be different between adults and adolescents with SR-cGvHD. It was agreed by the clinical experts that the NIH consensus criteria used in the trial for cGvHD disease and response assessment, as well as the tarping schedule for treatments applied in the trial, were, overall, reflective of Canadian clinical practice. The proportions of patients with cGvHD disease staging of mild, moderate, and severe, as well as the proportions of patients meeting the SR-cGvHD criteria (A versus B versus C) were reflective of patients seen in clinical practice.

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 other relevant evidence section included:

• 1 additional relevant study (Moiseev et al., 2020¹⁴) in the sponsor’s submission to CADTH that reported results for ruxolitinib in adults and children with SR-cGvHD

• a brief summary of methods and results of post hoc analyses of the REACH3 trial that were applied in the submitted pharmacoeconomic model

• Table 33 of ongoing trials.

Moiseev et al. (2020) Study

Description of the Study

The article by Moiseev et al. (2020)¹⁴ was a prospective, single-centre, open-label study conducted in Russia that included 75 patients with either acute (n = 32) or chronic (n = 43) SR-GvHD. The study sample included both adults and children, with half of the sample comprised of children (53% in the acute and 39% in the chronic GvHD groups). 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. Study participants received ruxolitinib at a starting dose of 10 mg twice a day for adults, 10 mg twice a day for children weighing more than 40 kg, and 0.15 mg/kg twice a day for children weighing less than 40 kg. 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 with joint statement criteria by Martin et al. (2009)¹⁵ and NIH criteria by Lee et al. (2015),¹³ 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. The 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 (P = 0.31 for aGvHD; P = 0.35 for cGvHD) and survival (P = 0.44 for aGvHD; P = 0.12 for cGvHD).

Harms Results

The most common AE was hematological 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 cGvHD patients (n = 43), the generalizability of these results may be limited. Moreover, as this trial was conducted in Russia, there may be limitations in the generalizability of these findings to the Canadian context.

Relevance for CADTH Review

In the REACH3 trial, the number of patients 12 to 18 years represented a small proportion of the study sample (3.6%). In the study by Moiseev et al. (2020),¹⁴ approximately 50% of the study sample included children younger than 18 years. Hence, this additional study supplements the evidence for ruxolitinib in patients younger than 18 years.

Post hoc Analyses of the REACH3 Trial²

Several post hoc analyses of the REACH3 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 ORR, duration of treatment from randomization, duration of treatment by response at and from cycle 7 day 1, resource use by study group and response at cycle 7 day 1, and weekly dosing. The CADTH review team was unable to rigorously evaluate 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 CADTH review team’s ability to interpret results of such analyses is significantly limited.

Conclusions

One phase III, open-label, multi-centre RCT (REACH3) provided evidence of the efficacy and safety of ruxolitinib compared with the investigator’s choice of BAT in patients 12 years and older with moderate or severe SR-cGvHD. Compared with BAT, patients who were treated with ruxolitinib showed statistically significant improvements in ORR at cycle 7 day 1, the primary end point, and in FFS and the modified Lee Symptom Scale, key secondary outcomes. The improvements in ORR, FFS, and the modified Lee Symptom Scale of the magnitude observed in the REACH3 trial were considered clinically meaningful by the clinical experts consulted by CADTH. Other secondary outcomes, BOR, DoR, and steroid use, were also supportive of the observed ORR cycle 7 day 1 benefit with ruxolitinib. The open-label design of the trial and reliance on local investigators’ assessments of trial outcomes may have introduced a bias that is difficult to quantify. The results of HRQoL measures — EQ-5D-5L and FACT-BMT — remain uncertain because of several important limitations. The actual degree of OS benefit with ruxolitinib was unknown at the time of the primary analysis because of OS data immaturity; median OS not been reached in either study group. According to the clinical experts consulted by CADTH, no unexpected safety concerns were observed with ruxolitinib.

Introduction

Disease Background

Hematopoietic SCTs provide stem cells to patients whose bone marrow has been destroyed by disease, chemotherapy, or radiation.¹ The 2 main types of stem cell transplant are autologous and allogeneic transplants. alloSCTs use stem cells from either a matched related or unrelated donor; whereas autologous stem cell transplants use stem cells from the same patient who will get the transplant.¹ Between 2006 and 2014, nearly 1 million SCTs were performed worldwide, of which approximately 40% were alloSCTs.² alloSCT can be used for the treatment of malignant and nonmalignant hematologic diseases.² According to a Canadian population-based cohort study, there were 547 alloSCTs in Ontario between 2012 and 2015.¹⁶ Between 2008 and 2019, an estimated 13,033 transplants were performed in Canada, of which 5,672 were alloSCTs; in 2018 and 2019, 2,843 transplants were recorded across Canada.⁴ Most transplants were conducted in patients with plasma cell disorders (30.28%), followed by non-Hodgkin’s lymphoma (20.07%), and acute myeloid leukemia (15.93%). Although alloSCT has curative potential, there is a risk 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.¹

GvHD is a multi-system disorder in which donor-derived immune cells initiate an adverse immune reaction to the transplant recipient’s tissues, cells, or organs, leading to tissue damage, organ failure, or death.² GvHD is the leading cause of morbidity and NRM in patients after alloSCT, affecting up to 70% of patients who receive stem cell transplants.^2,3 GvHD is estimated to be responsible for 21% to 31% and 31% to 40% of post-alloSCT deaths in patients who received a transplant from a human leukocyte antigen (HLA)–matched sibling and from an unrelated donor, respectively.²

GvHD has a multitude of syndromes defined by clinical manifestations, according to NIH consensus criteria. GvHD is typically classified as acute or chronic, based on a set of distinct clinical symptoms rather than the time of onset (i.e., before or after day 100 of transplantation, as was used previously).^17,18 Overlap syndrome may also occur, in which diagnostic or distinctive features of acute GvHD and chronic GvHD appear together.⁵ cGvHD typically occurs 100 days or more after alloSCT, and can last a few months or a lifetime and can affect almost any part of the body.⁵ aGvHD typically occurs in the first 100 days after alloSCT and often affects the skin, liver, and intestines.⁵

Chronic GvHD occurs in 35% to 70% of patients who undergo alloSCT.¹⁹ Most patients will experience the onset of cGvHD in the first year after alloSCT; however, about 5% to 10% of patients may not develop signs and symptoms until later.⁶ Although 1 of the major risk factors for the development of cGvHD is previous aGvHD, approximately 30% of cGvHD occurs de novo, without any previous aGvHD.^2,20 Additional risk factors for the development of cGvHD include alloSCT from HLA-nonidentical or unrelated donors, advanced age of the patient, grafting with growth factor mobilized blood cells, and the use of a female donor for male recipients.^2,21

Approximately 50% of patients with cGvHD are cured within 7 years after the start of systemic immunosuppression, about 10% of patients require ongoing systemic treatment for life, and roughly 40% of patients relapse or die within 7 years during prolonged immunosuppressive treatment.²

Chronic GvHD can be classified as mild, moderate, or severe disease, according to established NIH criteria for cGvHD.⁹ In 1 prospective cohort study that enrolled 911 patients from 13 centres, 19%, 53%, and 28% of patients had mild, moderate, and severe cGvHD disease, respectively, at onset.²² Mild cGvHD has been associated with better disease outcomes than severe disease, which has higher treatment-related mortality and lower survival rates.^2,23 Patients with mild, moderate, and severe cGvHD had a 2-year OS of 97%, 86%, and 62%, respectively.^2,23

Clinical manifestations of cGvHD are variable, resembling autoimmune or other immunologic disorders. Signs and symptoms can include lichen planus-like lesions or full sclerosis, muscle pain or joint fasciitis, vulvo-vaginitis, bronchiolitis obliterans, Sjogren’s syndrome, chronic immunodeficiency, primary biliary cirrhosis, wasting syndrome, immune cytopenias, and chronic immunodeficiency, in addition to damage of the GI tract and liver.^2,7 Chronic GvHD significantly negatively affects patients’ quality of life.² cGvHD commonly involves the skin, mouth, and liver, with less frequent involvement of eye, lung, GI tract, joint and fascia, and genital tract.⁶

Because of the lack of diagnostic biomarkers for cGvHD and the challenge of obtaining pathologic samples, cGvHD evaluations in clinical practice often rely on clinician reporting and patient interviews. In an effort to standardize reporting, NIH recommendations on diagnostic, severity scoring, and disease response criteria for cGvHD (i.e., NIH Consensus Development Projects on Criteria for Clinical Trials in Chronic GvHD) were first proposed in 2004 and updated in 2014.⁶

Eight organs or sites (skin, mouth, eyes, GI tract, liver, lungs, joint and fascia, and genital tract) are included in the calculation of severity of GvHD, according to the NIH criteria.⁹ Criteria included in the disease severity score encompass both the number of organs or sites involved and the severity score within each affected organ.⁹ Table 38 in Appendix 3 describes the calculation of the cGvHD severity scoring into mild, moderate, or severe disease, per NIH criteria.⁹ Briefly, mild disease involves 2 or fewer organs with a score of no more than 1 and no lung involvement; moderate disease involves 3 or more organs with a score of 1, any organ with a score of 2, or lung with a score of 1; and severe disease involves any organ with a score of 3 or lung with a score of 2, and the existence of substantial organ damage.

The diagnosis of cGvHD requires either at least 1 diagnostic manifestation or, if a diagnostic feature is not present, at least 1 distinctive manifestation confirmed by histologic, radiologic, or laboratory evidence of GvHD from any site.⁹ Patients with cGvHD have many concurrent medical signs and symptoms that can resemble cGvHD but are unrelated to cGvHD (e.g., skin rashes due to drug toxicity or infection, skin erythema from sun exposure, infectious diarrhea, or poor pulmonary function tests that predate transplantation). If cGvHD is suspected because of certain symptoms (e.g., papulosquamous lesions, oral ulcers, onycholysis, or dry gritty eyes), confirmatory tests for the diagnosis of cGvHD are required, as other non-cGvHD causes could account for similar symptoms. Confirmatory tests may include tissue biopsies (e.g., skin, mouth, lung, liver, GI, genital), organ-specific testing (e.g., pulmonary function tests, Schirmer tests), imaging (e.g., a barium swallow showing an esophageal ring), or evaluation by a specialist (e.g., ophthalmologist or gynecologist) to confirm cGvHD. As there are no accepted diagnostic biomarkers for cGvHD, and pathologic samples may be difficult to obtain, cGvHD evaluations in clinical practice are often based on clinical examinations and patient interviews. Although the NIH diagnostic criteria were established for use in clinical trials to ensure that the characterization of cGvHD is the same in all study participants, some patients in clinical practice may not meet the NIH diagnostic criteria but may nevertheless require systemic immunosuppression to manage symptoms and prevent organ damage.⁶

Adolescents 12 years of age and older comprise approximately 5% of the cGvHD population. Adolescents and adults are managed in similar way in clinical practice. Although the long-term disease outcomes of cGvHD appear to be slightly more favourable in children than in adults, both groups suffer from similar complications and have poor long-term outcomes.² The clinical experts consulted by CADTH agreed that there is a significant unmet need in children with cGvHD.

Standards of Therapy

Currently available treatment options initially involve corticosteroids with lymphopenic and anti-inflammatory properties, which have been the first-line treatment for cGvHD for the past 3 decades.⁷ Patients with mild cGvHD whose disease is limited to the skin are commonly treated with topical therapies (e.g., corticosteroids or CNIs). Patients with moderate-to-severe cGvHD commonly receive systemic treatment with corticosteroids as the standard first-line therapy.² Systemic corticosteroids can be administered alone or in combination with other immunosuppressants, such as CNIs. Median duration of first-line immunosuppressive therapy is estimated to be 2 to 3 years.² First-line therapies have challenges, including immunosuppression that can lead to an increase in malignancy relapse and other side effects, such as infections, myopathy, cataracts, hyperglycemia, decline in bone mass, and avascular necrosis, which have all been associated with extended corticosteroid use.⁷ In patients with clinical improvements, tapering of corticosteroids is recommended to reduce the risk of infection and other toxicities.² Among patients who respond to first-line systemic corticosteroids, responses are durable in 20% to 40% patients; the remaining patients (i.e., about 50% to 60%) do not respond or have inadequate disease control and require the addition of another systemic therapy, or fail to taper corticosteroids.² Patients with cGvHD who do not respond to steroids or are unable to taper SR-cGvHD have a poor 5-year survival rate of 50% to 70%.² Second-line drugs are commonly added to first-line therapy for patients who fail to respond to corticosteroids or do not tolerate or fail to taper corticosteroids.² Given the myelosuppression and risk of infection, patients with SR-cGvHD are also treated with standard alloSCT supportive care, which includes anti-infective medications and transfusions, along with GvHD prophylactic treatment with corticosteroids or CNIs.

Currently, there is no consensus on standard second-line therapies for patients with SR-cGvHD. Ibrutinib is the only Health Canada–approved second-line therapy for cGvHD; it is indicated for the treatment of patients with steroid-dependent or refractory cGvHD.^2,24 However, ibrutinib has not been reviewed by CADTH and is currently not publicly reimbursed in Canada for the current target indication; it is available only through private drug insurance. Evidence from patients with SR-cGvHD has mostly been obtained from retrospective, single-arm, phase II studies. Small numbers of enrolled patients, heterogenous patient populations, and the lack of standardized end points make comparisons of data across studies challenging.² In the absence of sufficient evidence to guide second-line treatment selection, factors that influence the treatment choice include the experience of the treating physician, types of aGvHD prophylaxis used, and the risk of potential toxicities and worsening pre-existing comorbidities.²

According to the clinical experts consulted by CADTH, available second-line options in Canada include ECP, MMF, etanercept, low-dose MTX, infliximab, mTOR inhibitor (e.g., sirolimus), imatinib, rituximab, ibrutinib, low-dose IL-2, pulsed cyclophosphamide, and, in rare cases, pentostatin. The clinical experts noted that available therapies, except ibrutinib, are currently used off-label. In the absence of proven treatment options, there is interprovince variation in standard practices and access to therapies. The clinical experts expressed several challenges with currently available therapies in this heavily pre-treated target population, including low rates of PRs (estimated to be 30%; CRs are uncommon in cGvHD), and refractoriness or intolerance to currently available therapies. Response rates observed in the literature with the second-line drugs vary significantly and are a challenge to interpret because of heterogeneity across study designs, end points, and populations, and small sample sizes.² 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, avascular necrosis, hypertension, and diabetes mellitus with end-organ damage) and the risk of opportunistic infections. It was emphasized by the clinical experts that infectious complications are a leading cause of NRM in SR-cGvHD.

There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve HRQoL, reduce disease symptoms of cGvHD, enhance a patient’s performance status, and improve OS. They highlighted the need for a convenient oral route of administration to achieve high adherence and reduce the need for hospital-based or ambulatory centre resource use.

Drug

Ruxolitinib is a JAK inhibitor that mediates the signalling of a number of cytokines and growth factors, which are important for hematopoiesis and immune function.⁸ Ruxolitinib binds to and inhibits protein tyrosine kinases JAK 1 and 2, which can lead to a reduction in inflammation and an inhibition of cellular proliferation.²⁵

Ruxolitinib is undergoing review by Health Canada through the expedited pathway (Orbis) for the treatment of GvHD in patients 12 years of age and older who have an inadequate response to corticosteroids or other systemic therapies. The sponsor’s requested reimbursement criteria for ruxolitinib are per the Health Canada indication under review. This CADTH review focuses on cGvHD. Ruxolitinib is concurrently being reviewed by CADTH for the treatment of aGvHD in patients aged 12 years and older who have an inadequate response to corticosteroids or other systemic therapies. Ruxolitinib has 2 Health Canada–approved indications: 1 for the treatment of splenomegaly and/or its associated symptoms in adults with primary myelofibrosis (also known as chronic idiopathic myelofibrosis), post-polycythemia vera myelofibrosis, or post-essential thrombocythemia myelofibrosis; and 1 for the control of hematocrit in adults with polycythemia vera resistant to or intolerant of a cytoreductive drug. Ruxolitinib received a positive conditional CADTH recommendation in March 2016 for the treatment of patients with polycythemia vera who are resistant to hydroxyurea or who are intolerant of hydroxyurea, according to modified European LeukemiaNet criteria used in the RESPONSE trial, and have a good performance status. Ibrutinib is the only Health Canada–approved second-line therapy for cGvHD²; it is indicated for the treatment of patients with steroid-dependent or refractory cGvHD.²⁴ However, ibrutinib has not been reviewed by CADTH and is currently not publicly reimbursed in Canada for the current target indication; it is available only through private drug insurance.

After being granted priority review with orphan product designation, the FDA approved ruxolitinib in September 2021 for cGvHD to be used after failure of 1 or 2 lines of systemic therapy in adults and pediatric patients aged 12 years and older based on evidence from the phase III REACH3 trial.²⁶ In May 2019, after granting priority review, the FDA approved ruxolitinib for SR-aGvHD in adults and pediatric patients aged 12 years and older, based on evidence from the phase II REACH1 trial.²⁷ In addition to the GvHD setting, the FDA has approved ruxolitinib for intermediate or high-risk myelofibrosis, including primary myelofibrosis, post-polycythemia vera myelofibrosis, and post-essential thrombocythemia myelofibrosis in adults, and for polycythemia vera 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 adult and pediatric patients aged 12 years and older with cGvHD after failure of at least 2 prior lines of systemic therapy based on evidence from the phase II KD025-213 trial.²⁸ In August 2017, the FDA approved ibrutinib for the treatment of adult patients with cGvHD after failure of 1 or more lines of systemic therapy, based on evidence from a phase II trial. Ibrutinib was the first therapy for the treatment of cGvHD approved by the FDA.²⁹ The European Medicines Agency has approved ruxolitinib for the treatment of disease-related splenomegaly or symptoms in adult patients with primary myelofibrosis (also known as chronic idiopathic myelofibrosis), post-polycythemia vera myelofibrosis, or post-essential thrombocythemia myelofibrosis, and for the treatment of adult patients with polycythemia vera who are resistant to or intolerant of hydroxyurea.

Ruxolitinib is available as 5 mg, 10 mg, 15 mg and 20 mg tablets. The recommended dose is 10 mg administered orally twice daily. The product monograph states that tapering of ruxolitinib may be considered in patients with a response and after having discontinued corticosteroids. It is recommended to taper ruxolitinib by reducing the dose to 50% every 2 months; in the event that signs or symptoms of GvHD reoccur during or after the taper, re-escalation of ruxolitinib should be considered.⁸ Table 3 summarizes key characteristics of ruxolitinib.

Table 3: Key Characteristics of Ruxolitinib

cGvHD = chronic graft-versus-host disease; JAK = Janus kinase.

^aHealth Canada–approved indication.

Source: Product monograph.⁸

Stakeholder Perspectives

Patient Group Input

The information in this section is a summary of input provided by the patient groups that responded to CADTH’s call for patient input for the purpose of this review. The full patient input received has been included in Appendix 5.

Eight patient groups — Lymphoma Canada, Lymphoma and Leukemia Society of Canada, Chronic Lymphocytic Leukemia Canada, Myeloma Canada, Aplastic Anemia and Myelodysplasia Association of Canada, Canadian MPN Research Foundation and the Chronic Myelogenous Leukemia Network, the Chronic Myeloid Leukemia Network, and Cell Therapy Transplant Canada (CTTC) — co-created a joint patient input for this review. The input was based on an online survey and responses from a total of 68 participants were included in the patient input. Sixty patients reported having received a stem cell transplant (SCT), 6 reported not having received an SCT, and 2 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% developed aGvHD, 24% developed cGvHD, and 62% developed 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%). To manage their GvHD, respondents reported requiring numerous medical consultations, hospital stays, and nights away from home. Respondents indicated that a range of GvHD symptoms significantly affected ’their daily activities and had detrimental effects on ’their quality of life. Respondents highlighted interruption of life goals and accomplishments (career, school), difficulty sleeping, impact on mental health (stress, anxiety, worry, and concentration problems), and financial issues. Respondents commonly reported experiencing symptoms such as 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 difficulties, 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 received treatment in the outpatient setting (rather than requiring an overnight hospital stay), having access to treatment locally (rather than requiring extensive amount of travel), treatment being covered by insurance or drug plans, and the treatment being recommended by health care professionals were perceived to be very important by respondents. Respondents who had direct experience with ruxolitinib indicated that, overall, the drug was effective, improved their quality of life, had tolerable side effects, they would take it again if recommended by their physician, and they would recommend it to other patients.

Clinician Input

Input From 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 cGvHD.

Unmet Needs

The clinical experts consulted by CADTH noted that there are currently no Health Canada–authorized standard-care regimens specific for patients with SR-cGvHD, except ibrutinib, which has been authorized since 2017 for the treatment of adults with SR-cGvHD but has not undergone review by CADTH, is not publicly reimbursed in Canada, and is available through private drug insurance only. The clinical experts noted that in the absence of proven treatment options, there is interprovincial variation in standard practices and access to therapies. Available therapies are currently used off-label (except ibrutinib) and, due to the lack of large-scale, RCTs, no consensus could be reached on the choice of optimal second-line therapy for SR-cGvHD. There was consensus among the clinical experts that there is an unmet need for effective therapies with acceptable toxicity profiles that improve HRQoL, reduce disease symptoms of cGvHD, enhance patient’s performance status, and improve OS. They highlighted the need for a convenient oral route of administration to improve adherence and reduce the need for hospital-based or ambulatory centre resource use. The clinical experts expressed several challenges with currently available therapies in this heavily pre-treated target population, including low rates of PRs (estimated to be 30%; CRs are uncommon in cGvHD) and refractoriness or intolerance to currently available therapies. 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, avascular necrosis, hypertension, and diabetes mellitus with end-organ damage) and the risk of opportunistic infections. It was emphasized by the clinical experts that infectious complications are a leading cause of NRM in SR-cGvHD.

Place in Therapy

Ruxolitinib is to be used as an add-on to the immunosuppressive regimen of corticosteroids with or without CNI in patients 12 years and older with moderate or severe SR-cGvHD, per the REACH3 trial, the clinical experts explained. They agreed that ruxolitinib, as a therapy for SR-cGvHD, would likely shift the current treatment paradigm. The clinical experts noted that ruxolitinib, which works as a JAK inhibitor and blocks the JAK/STAT pathway and its mechanism of action, is novel in the context of other immunosuppressives used in the management of cGvHD, potentially acting additively or in synergy with other therapies. The clinical experts highlighted the way ruxolitinib differentially affects distinct T-cell subsets, which helps prevent post-transplant relapse. The REACH3 trial excluded patients who received 2 or more systemic treatments for cGvHD in addition to corticosteroids with or without CNI for cGvHD, patients with overlap syndrome, patients with ECOG performance status scores of 3 or KPS or LPS scores below 60%, and patients with mild, as opposed to moderate or severe, cGvHD. The clinical experts consulted by CADTH felt that it would be reasonable to generalize the REACH3 trial results to patients who received 2 or more systemic treatments for cGvHD in addition to corticosteroids with or without CNI. The clinical experts noted that the benefit of treatment with ruxolitinib was observed in patients after crossover (potential third and fourth liners) in the REACH3 trial. As well, given the acceptable safety profile of ruxolitinib, it was felt by the clinical experts that it would be reasonable to offer ruxolitinib to patients with ECOG performance status scores of 3 or KPS or LPS scores below 60% in patients whose performance status may be related to cGvHD and its symptoms. Furthermore, it was agreed 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 mild cGvHD.

Patient Population

Overall, the clinical experts consulted by CADTH agreed that patients who meet the inclusion and exclusion criteria of the REACH3 trial should be eligible for ruxolitinib therapy. Although it was agreed that there is currently insufficient evidence to guide a recommendation on which patient subgroups would most likely show a response to ruxolitinib, the clinical experts identified the following potential subgroups as being most in need of ruxolitinib therapy: patients with glucocorticoid-refractory as opposed to glucocorticoid-dependent cGvHD; and patients with bronchiolitis obliterans. Patient subgroups that would potentially benefit the least from ruxolitinib could include patients with isolated lichenoid cGvHD who may preferentially be treated with ECP or subcutaneous low-dose IL-2 rather than ruxolitinib. Patients with strictly autoimmune cGvHD manifestations, such as immune thrombocytopenia (as well as immune hemolytic anemia, immune glomerulonephritis, and myasthenia gravis) may preferentially be treated with rituximab. Patients with thrombocytopenia and anemic cGvHD or active uncontrolled infections 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 noted that patients would be identified as possible candidates for ruxolitinib treatment if they are considered to have SR-cGvHD. The clinical experts stated that cGvHD is a complex clinical diagnosis that is made on a daily basis by experienced SCT physicians. cGvHD is associated with a typical constellation of symptoms and signs, although tissue diagnosis (biopsy) confirmation or sequential pulmonary function tests may be pursued in patients with diagnostic suspicion (patients with moderate-to-severe cGvHD are generally not asymptomatic). It was noted that SCT programs are highly specialized, multi-disciplinary, and well supported by relevant subspecialists. The clinical experts agreed that a potential for misdiagnosis exists, given the underlying complexity of cGvHD, the nonspecific presentation of cGvHD, and the fact that a differential diagnosis must be considered. However, misdiagnosis is minimized in Canada because patients who have undergone an alloSCT receive follow-up care in specialized clinics with expertise in the diagnosis and management of cGvHD.

Assessing Response to Treatment

In the opinion of the clinical experts consulted by CADTH, an accurate assessment of response in patients with cGvHD is based on the NIH consensus criteria, as was used in the REACH3 trial. Given the underlying complexity of cGvHD evaluation, it was noted by the experts that formal assessment of treatment response can be accomplished but is a challenge to undertake outside the structure of a clinical trial. Clinical assessments in Canadian clinical practice to evaluate the response to treatment include objective improvement in the target organs involved (e.g., oral or genital lichenoid mucosal changes; improvement in liver function tests; objective improvement in keratitis on slit lamp or corneal staining; objective improvement in skin rash and/or joint flexibility; improvement in weight; reduction in proteinuria; subjective improvement in appetite; improvement of nausea and/or diarrhea; improvement in eye, mouth, or vaginal pain; improvement in dyspnea and/or wheeze; and overall performance status).

The clinical experts noted that response to treatment is usually assessed every 2 to 4 weeks, depending on the severity of cGvHD. Weekly assessment may be required initially.

The clinical experts indicated that the most clinically meaningful responses to treatment include overall response (CR or PR), improvements in HRQoL and performance status, reduction in cGvHD symptoms (frequency and/or severity), stability of disease (no deterioration), improved OS, as well as the ability to reduce the dose of immunosuppression and/or corticosteroids without flare of cGvHD signs and symptoms and the need to start another drug for cGvHD.

Discontinuing Treatment

In the opinion of the clinical experts consulted by CADTH, treatment with ruxolitinib should be discontinued if a patient experiences cGvHD disease progression, relapse of underlying hematologic malignancy, or experiences intolerable toxicity (e.g., anemia and thrombocytopenia or neurologic toxicity that cannot be managed with drug interruption and/or dose reduction). Tapering ruxolitinib in responders may be considered after 24 weeks of therapy.

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 undertaken by providers who are familiar with GvHD. Occasionally, patients with severe multi-system cGvHD require admission to the hospital, and treatments, including steroids and ruxolitinib, will be given on an appropriate inpatient service.

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 complete clinician inputs received are included in Appendix 6.

The 2 clinician group inputs were provided by CTTC (based on input from 8 clinicians) and from Ontario Health (CCO) Complex Malignant Hematology (based on input from 2 clinicians). The views of the clinician groups are overall consistent with the clinical experts consulted by CADTH, indicating that, based on evidence from the REACH3 trial, it is anticipated that ruxolitinib will become the dominant first-line therapy for SR-cGvHD. The outcomes assessed in the REACH3 trial were judged to be applicable to Canadian clinical practice and reflective of clinically meaningful responses. It was noted by both inputs that ruxolitinib is not as immunosuppressive as other available therapies. The clinicians from Ontario Health CCO 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. It was highlighted by the input from CTTC that a Health Canada–approved and provincially funded therapy for SR-cGvHD would be an important step forward in the current target setting, with existing therapies offering low response rates and high rates of toxicity. According to the input from CTTC, the experience with ruxolitinib (accessible through a compassionate access program) and real-world effectiveness appear similar to what was observed in the REACH3 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 affect 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

AEs = adverse events; BAT = best available therapy; cGvHD = chronic GvHD; CNI = calcineurin inhibitor; ECP = extracorporeal photopheresis; IL-2 = interleukin-2; NIH = National Institutes of Health; SCT = stem cell transplant.

Clinical Evidence

The clinical evidence included in the review of ruxolitinib is presented in 3 sections. The first section, the systematic review, includes the pivotal study 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 evidence that was considered to address important gaps in the evidence presented 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 cGvHD 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 to patients, clinicians, and drug plans.

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

Table 5: Inclusion Criteria for the Systematic Review

AEs = adverse event; aGvHD = acute graft-vs.-host disease; BOR = best overall response; cGvHD = chronic graft-versus-host disease; DoR = duration of response; ECP = extracorporeal photopheresis; FFS = failure-free survival; HRQoL = health-related quality of life; MMF = mycophenolate mofetil; MR = malignancy relapse or progression; mTOR = mammalian target of rapamycin; MTX = methotrexate; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival; RCTs = randomized controlled trials; SAEs = serious adverse events; vs. = versus; WDAEs = withdrawal due to adverse events.

^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.³⁰

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. The following 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 the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to 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 March 23, 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 checklist.³¹ 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. Refer to Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the sponsor 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 dealing with GvHD was run in MEDLINE All (1946–-) on graft versus host disease. 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.

Description of Study

The REACH3 trial is an ongoing, international, multi-centre, open-label, randomized phase III trial comparing ruxolitinib (10 mg, oral twice daily) with investigator’s choice of BAT in patients 12 years and older with moderate or severe SR-cGvHD. Patients continued to receive their systemic immunosuppressive regimen of corticosteroids with or without CNI, per standard of care. The diagnosis and disease staging of cGvHD was based on NIH criteria (Jagasia et al. 2015)⁹ (Table 38 in Appendix 3). This CADTH review is based on the data cut-off date of May 8, 2020. Final study results are expected to be released after completion of the study (estimated to be in the third or fourth quarter of 2022).¹¹ The REACH3 trial was sponsored by Novartis (all countries except US) and Incyte-Corporation (US).¹²

The primary end point was the ORR at the cycle 7 day 1 visit. Patients in this international trial were randomized at 149 sites across 28 countries, which are listed in Table 6. The majority of sites were in Europe and North America (Canada and US); 6 patients were randomized at the 3 Canadian sites in the REACH3 trial.¹²

A total of 329 patients were randomized (using an interactive voice-response system) in a 1:1 ratio to receive ruxolitinib or BAT. Enrolment occurred between July 11, 2017 and November 18, 2019. Randomization was stratified by cGvHD severity, per 2014 NIH consensus criteria (moderate versus severe).¹²

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of Included Study

AEs = adverse events; aGvHD = acute graft-versus-host disease; alloSCT = allogeneic stem cell transplant; BAT = best available therapy; BOR = best overall response; cGvHD = chronic graft-versus-host disease; CNI = calcineurin inhibitor; CR = complete response; DLI = donor lymphocyte infusion; DoR = duration of response; ECOG PS = Eastern Cooperative Oncology Group performance status; ECP = extracorporeal photopheresis; EQ-5D-5L = EuroQol 5-Dimensions 5-Levels; FACT-BMT = Functional Assessment of Cancer Therapy – Bone Marrow Transplant; FFS = failure-free survival; GI = gastrointestinal; JAK = Janus kinase; KPS = Karnofsky performance score; LPS = Lansky performance score; MMF = mycophenolate mofetil; mTOR = mechanistic (formerly mammalian) target of rapamycin; MTX = methotrexate; NIH = National Institutes of Health; NRM = nonrelapse mortality; ORR = overall response rate; PR = partial response; RCT = randomized controlled trial; SR-GvHD = steroid-refractory graft-versus-host disease.

^aChoice of 10 commonly used therapies (ECP, low-dose MTX, MMF, mTOR inhibitors [everolimus or sirolimus], infliximab, rituximab, pentostatin, imatinib, or ibrutinib). Investigational drugs only available via clinical trials, including JAK inhibitors and combinations of different treatments, were not permitted in the BAT group. No other BAT was allowed. Concomitant use of CNI and steroids was allowed. If any medication in the BAT list was used as prophylaxis for underlying malignancy relapse, it was discontinued before randomization and entered into the electronic case report form. For patients randomized to either the ruxolitinib or BAT treatment group, rituximab could be administered after randomization for the treatment of Epstein-Barr virus. The Epstein-Barr virus infection was captured either in the medical history or AE electronic case report form. If any medication in the BAT list was used as prophylaxis for cGvHD before study entry, it was allowed to be continued after randomization; however, it was captured on the concomitant medication electronic case report form.

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

^cUse of growth factor supplementation and transfusion support was allowed during the trial; however, transfusion to reach a minimum platelet count for inclusion was not allowed during screening and at baseline.

^dModerate cGvHD: At least 1 organ (not lung) with a score of 2 or 3, or more than 1 organ each with a score of 1, or a lung score of 1; severe cGvHD: at least 1 organ with a score of 3, or a lung score of 2 or 3.

^eChanges or additions of a new systemic therapy in the BAT group due to documented lack of response or toxicity was allowed in the first 6 cycles, but was considered a treatment failure. At cycle 7 day 1, or later after randomization, patients randomized to BAT who did not achieve or maintain a CR PR, or who developed toxicity to BAT, were allowed to cross over from BAT to ruxolitinib.²

^fInfections were considered to be controlled if appropriate therapy was instituted and, at the time of screening, no signs of infection progression were present.

^gOne additional report was included: the Clinical Study Report.¹²

Sources: Zeiser et al. (2021),³³ Clinical Study Report,¹² sponsor’s response to additional information request.¹¹

The study consisted of 4 main periods: the screening period (28-day duration); main treatment period (day 1 until end of treatment; i.e., the primary efficacy period [cycle 1 to end of cycle 6] plus the extension period [cycle 7 to cycle 39]); long-term survival follow-up (end of treatment to 39 cycles on study treatment)¹²; and safety follow-up (Figure 2). The end of the study was to occur when all patients completed 39 cycles, discontinued from the study, or died.³³

The primary efficacy assessments were performed on cycle 7 day 1. During the extension period, cross over between treatment groups was allowed on cycle 7 day 1 and thereafter. Patients in the BAT group could cross over to the ruxolitinib group if they experienced disease progression, mixed response, unchanged response, toxicity to BAT, or a cGvHD flare.¹² Patients in the BAT group who achieved a CR or PR on cycle 7 day 1 were not permitted to cross over to the ruxolitinib group until disease progression, mixed response, or occurrence of a toxicity to BAT.¹²

During the randomized treatment period, cGvHD disease assessments were planned to occur every week in cycle 1 (days 1, 8, 15, 22) (± 3 days) and every 4 weeks from cycle 2 day 1 until cycle 7 day 1 (± 7 days). Visits after cycle 7 day 1 were at cycle 9 day 1 (± 7 days) and every 12 weeks thereafter until cycle 39 (± 14 days) or end of treatment, whichever occurred first. Unscheduled visits could be performed as necessary.³³

Figure 2: REACH3 Trial Study Design

BAT = best available therapy; cGvHD = chronic graft-versus-host disease; EOT = end of treatment; ICF = informed consent form.

Source: Clinical Study Report¹²

Populations

Inclusion and Exclusion Criteria

The key inclusion and exclusion criteria used in REACH3 trial are described in Table 6. Briefly, the trial enrolled male or female patients 12 years and older who had undergone alloSCT, had evidence of myeloid and platelet engraftment (absolute neutrophil count > 1,000/mm³ and platelet count > 25,000/mm³), and were diagnosed with moderate or severe cGvHD that was determined to be corticosteroid-refractory, per NIH consensus criteria¹⁰ (refer to Table 6). At screening, patients had to have an ECOG performance status score of 0 to 2 or a KPS or LPS score of 60% to 100%. Patients with 2 or more systemic treatments (BAT) for cGvHD in addition to corticosteroids with or without CNI, impaired renal, GI function, or liver disease associated or not associated with GvHD were excluded (refer to Table 6).

Baseline Characteristics

Demographic characteristics, underlying disease history, and GvHD history at baseline are summarized in Tables 7, Table 8, and Table 9, respectively (refer to Table 36 in Appendix 3 for detailed baseline demographic characteristics). The mean ages for the ruxolitinib and BAT groups, respectively, were 45.9 (SD = 15.68) and 47.2 (SD = 16.17) years. The ruxolitinib group had a lower proportion (ruxolitinib versus BAT) of female patients (33.0% versus 43.9%), White patients (70.3% versus 80.5%), patients meeting the corticosteroid-refractory A criteria (37.6% versus 44.5%), and patients who only received steroid as prior systemic cGvHD or SR-cGvHD therapy (42.4% versus 49.4%), and a higher proportion of Asian patients (20.0% versus 12.8%), patients with unknown Center for International Blood and Marrow Transplant Research risk assessment (24.2% versus 17.7%), patients with prior aGvHD of grade II (32.1% versus 26.2%), and patients meeting the corticosteroid-refractory B criteria (35.2% versus 25.6%). Patients in the ruxolitinib group had a slightly shorter time from transplant to cGvHD diagnosis (mean days = 371.44 versus 404.53) and from aGvHD diagnosis to randomization (mean days = 578.76 versus 631.05), and a slightly longer time from initial cGvHD to diagnosis of corticosteroid-refractory disease (mean days = 200.84 versus 186.87) and from aGvHD diagnosis to resolution (mean days = 143.12 versus 105.78). The majority of patients had an ECOG performance status score of 1 (55.8% versus 50.0%).

The majority of patients (ruxolitinib versus BAT) had severe SR-cGvHD (58% versus 54.9%), met corticosteroid-refractory A (37.6% versus 44.5%) or B (35.2% versus 25.6%) criteria, and most patients had received either only steroid (42.4% versus 49.4%) or steroid plus CNI (41.2% versus 42.1%) as prior systemic cGvHD or SR-cGvHD therapy. Myelodysplastic syndrome was the most common underlying disease (ruxolitinib versus BAT) (73.3% versus 74.4%), and the mean time from transplant to cGvHD diagnosis, as well as the mean time from initial diagnosis to randomization, were similar across groups (mean days = 247.0 versus 230.0 and mean years = 3.90 versus 3.52, respectively).

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

BAT = best available therapy; ECOG = Eastern Cooperative Oncology Group; SD = standard deviation.

^aBaseline characteristics are only given for patients with a valid baseline record.¹¹

Source: Clinical Study Report.¹²

Table 8: Summary of Underlying Disease History at Baseline, Full Analysis Set

BAT = best available therapy; CIBMTR = Center for International Blood and Marrow Transplant Research; SD = standard deviation.

Note: All results are as documented on the pretransfusion disease history electronic case report form.

Source: Clinical Study Report.¹²

Table 9: Summary of GvHD History at Baseline, Full Analysis Set

aGvHD = acute graft-versus-host disease; BAT = best available therapy; cGvHD = chronic graft-versus-host disease; SD = standard deviation; SR = steroid refractory; SR-aGvHD = steroid-refractory acute graft-versus-host disease; SR-cGvHD = steroid-refractory chronic graft-versus-host disease.

Note: Analyses of “time to variables” are given for patients with available start and end dates only. Prior treatment for cGvHD, as documented in the prior medication data; topical and local treatments not counted.

^aData were reported by investigator.

^bBaseline characteristics are only given for patients with a valid baseline record.¹¹

Source: Clinical Study Report.¹²

The great majority of patients received a systemic treatment for cGvHD or SR-cGvHD before enrolment into the trial, and the proportions and types of therapies were balanced across groups, as summarized in Table 10. The most commonly received prior systemic treatments (ruxolitinib versus BAT) were glucocorticoids (98.2% versus 95.7%; including mainly prednisone, prednisolone, and methylprednisolone), followed by corticosteroids (plain) (95.8% versus 93.9%; including mainly prednisone, prednisolone, and methylprednisolone) and corticosteroids (acting locally) (78.2% versus 82.9%; including mainly prednisone and prednisolone). The most frequently received CNIs (received by 31.5% and 31.7% of patients in the ruxolitinib and BAT groups, respectively) included cyclosporine, tacrolimus, and tacrolimus monohydrate. The majority of patients in the REACH3 trial received either a steroid only or a steroid plus CNI as prior systemic cGvHD or SR-cGvHD therapy (Table 11). The percentages of patients who received a steroid plus CNI plus other systemic therapy were 6.1% and 2.4% in the ruxolitinib and BAT groups, respectively. The percentages of patients who received a steroid plus other systemic therapy were 8.5% and 5.5% in the ruxolitinib and BAT groups, respectively.

Table 10: Summary of Types of Prior Systemic Treatment for cGvHD or SR-cGvHD by ATC Class, Full Analysis Set

ATC = anatomical therapeutic chemical; BAT = best available therapy; cGvHD = chronic graft-versus-host disease; CNIs = calcineurin inhibitors; SR-cGvHD = steroid-refractory chronic graft-versus-host disease.

Source: Clinical Study Report.¹²

Table 11: Prior Systemic cGvHD or SR-cGvHD Therapy, Full Analysis Set

BAT = best available therapy; CNI = calcineurin inhibitor; cGvHD = chronic graft-versus-host disease; SR-cGvHD = steroid-refractory chronic graft-versus-host disease.

Source: Clinical Study Report.¹²

Approximately 1-third of patients (32.5%) received treatment as cGvHD prophylaxis before enrolment into the trial, and the proportions and types of therapies were balanced across groups, as summarized in Table 12. Prior prophylaxis was defined as all systemic treatments that started before the diagnosis of cGvHD and were received by patients as prophylaxis for GvHD. The most commonly received prior prophylactic treatments (ruxolitinib versus BAT) were CNIs (23.6% versus 24.4%; including mainly ciclosporin), followed by other ophthalmologicals (15.8% versus 18.3%; including mainly ciclosporin), and corticosteroids (plain) and glucocorticoids (11.5% versus 8.5%, each).

Table 12: Summary of Prior cGvHD Prophylactic Therapies, Full Analysis Set

ATC = anatomical therapeutic chemical; BAT = best available therapy; cGvHD = chronic graft-versus-host disease; CNIs = calcineurin inhibitors.

Source: Clinical Study Report.¹²

Interventions

Details on response assessment and instructions on treatment during the primary efficacy and extension periods are depicted in Figure 3 and Figure 4. Patients were entered into the long-term survival period if they had permanently discontinued the study treatment before completing 3 years on study (during either the main treatment or crossover periods) for reasons other than achieving a CR or PR. The main treatment period ended for patients who entered the long-term survival follow-up period, as well as for those who started ruxolitinib treatment after crossover from the BAT group.¹² Patients who tapered off the study treatment and all immunosuppressive therapy due to achieving a CR or PR continued to be followed in the main treatment period for safety and efficacy assessments, per the assigned visit schedule.³³ During the long-term survival follow-up period, patients were followed approximately every 3 months for 3 years. A 30-day safety follow-up assessment was completed after the last dose of ruxolitinib or BAT for patients who permanently discontinued the study treatment for reasons other than a CR or PR.

Patients randomized in the REACH3 trial were allocated to either the ruxolitinib group or the BAT group. Treatments and doses are described in Table 13. Standard of care was received by both study groups, and included systemic immunosuppressive regimens of corticosteroids with or without CNI (cyclosporine or tacrolimus), per standard of care recommended by the investigator. Routine administration of antibiotics, anti-infectives, and immunizations as prophylactic therapies was allowed, per institutional guidelines. Supportive care measures (e.g., use of anti-emetics and anti-motility drugs for diarrhea management) were allowed at the investigator’s discretion.² Prophylactic therapies (i.e., antibiotics, anti-infectives, and immunizations) started before enrolment in the REACH3 trial could be continued after randomization, per institutional guidelines.

Tapering of Therapies

Tapering of corticosteroids, CNI, and ruxolitinib was done in 2 steps: first, systemic corticosteroids were tapered upon documented CR or PR, then CNI or ruxolitinib were tapered.³³

Tapering guidelines for corticosteroids are outlined in Table 37 in Appendix 3. Taper of corticosteroids was to be initiated approximately 2 weeks after the achievement of a CR. If a flare were to occur during the taper, treatment was to continue for at least 3 months before resuming the taper.³³

Tapering guidelines for CNI indicated that, starting at cycle 7 day 1, 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 following a taper schedule, per institutional practice.³³

Taper guidelines for ruxolitinib indicated that, starting at cycle 7 day 1, taper could be initiated 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 cGvHD response (i.e., no worsening of cGvHD signs and 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.

Treatment of cGvHD Recurrence

Recurrence was defined as the return of cGvHD symptoms after tapering off all immunosuppressive therapy and study treatment due to response.³³ If worsening of cGvHD symptoms occurred, patients were allowed to resume treatment for cGvHD, per the investigator’s discretion. Patients in the BAT group who experienced cGvHD recurrence were allowed to restart a single drug BAT (± corticosteroids ± CNI). Restart of treatment was handled the same as the addition or initiation of new systemic treatment for statistical analyses.³³ Addition or initiation of a second BAT systemic treatment was not allowed. Upon second recurrence, patients initially assigned to the BAT group had to either crossover to the ruxolitinib group or discontinue study treatment and enter the long-term survival follow-up.³³

Treatment of cGvHD Flare

A cGvHD flare was defined as any increase in symptoms during taper of any immunosuppressive therapy for cGvHD after an initial response (CR or PR). The dose of corticosteroids could be re-escalated at the investigator’s discretion and was not considered a treatment failure as long as there was no change in or addition of another systemic treatment. If addition or initiation of a new systemic therapy was required because it was not possible to taper corticosteroids below methylprednisolone 1 mg/kg per day (or equivalent < 1.25 mg/kg per day of prednisone) for a minimum 7 days, or to re-escalate corticosteroids to methylprednisolone > 2 mg/kg/day (or equivalent > 2.5 mg/kg per day of prednisone) patients were considered to have experienced treatment failure. Refer to Table 13 for instructions on treatment of a cGvHD flare in the BAT group. For patients in the ruxolitinib group who experienced a cGvHD flare during ruxolitinib taper after cycle 7 day 1, the dose could be increased to the prior dose level (maximum 10 mg twice daily), their response monitored, and ruxolitinib taper restarted if they achieved a response within 28 days. If no response was achieved within the 28 days, or more than 1 flare was observed, patients were considered to have experienced cGvHD flare failure, and further treatment with ruxolitinib was allowed, per the investigator’s judgment.³³

In the event of cGvHD that developed as a result of an abrupt ruxolitinib interruption due to worsening cytopenia, the patient’s corticosteroid dose was to be maintained or increased to > 0.4 mg/kg per day methylprednisolone (or equivalent prednisone to > 0.5 mg/kg per day) for a minimum 7 days.³³

Figure 3: Overall cGvHD Response Assessment Versus Baseline (Pre-Cycle 7 Day 1)

cGvHD = chronic graft-versus-host disease; CR = complete response, PC = partial response.

Note: Any of the following therapies represents new or additional systemic therapy for cGvHD:

Any new CNI therapy being initiated as “treatment for cGvHD” or “treatment for SR-cGvHD” after the baseline, as recorded either on the dosage administration record or on prior and concomitant medication, and never received before or at baseline.

Any other systemic treatments (excluding CNIs and steroids) being initiated as “treatment for cGvHD” or “treatment for SR-cGvHD” after baseline and documented in the prior and concomitant medication, and never received before or at baseline.

New or additional BAT at or after the start of the initial BAT, as reported on the dosage administration record electronic case report forms.

Treatment with ruxolitinib after crossover from BAT, as reported on the dosage administration record electronic case report forms.

Source: Clinical Study Report.¹²

Figure 4: Overall cGvHD Response Assessment Versus Baseline (From Cycle 7 Day 1)

BAT = best available therapy; cGvHD = chronic graft-versus-host disease; CR = complete response, PC = partial response.

Source: Clinical Study Report¹²

Dose Modifications

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. The dose of BAT and corticosteroids with or without CNI was modified per the investigator’s discretion, institutional guidelines, or the product label. Once a dose or schedule modification occurred, titration back up to the original dose or schedule was permitted. If the dose interruption for ruxolitinib or BAT exceeded 21 days, the study treatment had to be discontinued.

Concomitant Medication

Concomitant treatment included all interventions (therapeutic treatments, supportive care and procedures) that were received by patients at the same time as the study treatment or before study entry and continued after initiation of the study intervention.¹² Standard cGvHD prophylaxis and treatment medications started before initiation of the study intervention — including systemic corticosteroids, CNI (cyclosporine or tacrolimus), and topical or inhaled corticosteroids — were allowed to be continued, per institutional guidelines.¹² It was recommended that patients who received CNI before enrolment remain on the same CNI, as needed, while they received the study intervention.² Routine administration of antibiotics, anti-infectives, and immunizations as prophylactic therapies was allowed, per institutional guidelines.¹² Supportive care measures (e.g., use of anti-emetics and anti-motility drugs for diarrhea management) were allowed at the investigator’s discretion. Prohibited concomitant mediation included fluconazole at daily doses higher than 200 mg, nonsteroidal anti-inflammatory drugs, and medications that were expected to reduce platelet function or have an adverse effect on blood coagulation.³³

Change in or addition of new systemic immunosuppressive therapy after randomization and up to completion of cycle 7 day 1 was allowed and was recorded as concomitant medication.²

Outcomes

A list of end points identified in the CADTH review protocol that were assessed in the clinical trials and included in this review is provided in Table 14. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Table 14: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

AEs = adverse events; BOR = best overall response; DoR = duration of response; EQ-5D-5L = EuroQol 5-Dimensions 5-Levels; FACT-BMT = Functional Assessment of Cancer Therapy – Bone Marrow Transplant; FFS = failure-free survival; HRQoL = health-related quality of life; MR = malignancy relapse or progression; PGIC = Patient Global Impression of Change; PGIS = Patient Global Impression of Severity; ORR = overall response rate; OS = overall survival.

Note: One treatment cycle = 28 days.

Failure-Free Survival

FFS was a key secondary outcome in the REACH3 trial, and was defined as relapse, recurrence, or death due to underlying disease; NRM; or the addition or initiation of another systemic therapy for cGvHD. Disease relapse or recurrence for the analyses of FFS was determined according to local institutional practices.³³

ORR on Cycle 7 Day 1

ORR on cycle 7 day 1 was the primary outcome of the REACH3 trial. ORR on cycle 7 day 1 was defined as the proportion of patients who had achieved a CR or PR, according to NIH consensus criteria.¹³ (Response was assessed relative to the disease evaluation of cGvHD at baseline, based on the NIH criteria.¹³ For the crossover patients, baseline was defined as the last cGvHD evaluation before the initiation of ruxolitinib). cGvHD disease evaluation and response assessments to the study treatment were assessed by the treating physician at every visit during the treatment period, according to NIH consensus criteria. The cGvHD disease evaluation captured individual organ symptoms, including eyes, mouth, lungs, GI, liver, skin, joints, and fascia. A high-level summary of the cGvHD disease evaluation, as well as response-assessment criteria used in the REACH3 trial, per NIH consensus guidelines, is provided Table 15. The efficacy assessments for ORR were based on a composite assessment of all organs at each post-baseline assessment visit to categorize a patient’s response to treatment as CR, PR, mixed response, progression, or unchanged response (refer to Table 16 for details).³³

CR was defined as complete resolution of all signs and symptoms of cGvHD in all evaluable organs without the initiation or addition of any new systemic therapy (refer to Table 16).³³

PR was defined as an improvement in at least 1 organ (e.g., improvement of 1 or more points on a 4 to 7 point scale, or improvement of 2 or more points on a 10- to 12-point scale) without progression in other organs or sites and without the initiation or addition of new systemic therapies (refer to Table 16).³³

Lack of response was defined as unchanged, mixed response, or progression (refer to Table 16 for criteria of unchanged, mixed response, and progression).

In addition, patients were not considered to be responders at cycle 7 day 1 if they experienced any of the following events before the cycle 7 day 1 visit: missing overall cGvHD response assessment at cycle 7 day 1; no CR or PR at cycle 7 day 1; or addition of or start of new systemic therapy for cGvHD.³³

Table 15: Summary of cGvHD Evaluation and Response-Assessment Criteria for All Organs at Post-Baseline cGvHD Response Assessments in the REACH3 Trial, According to NIH Consensus Criteria¹³

ALT = alanine aminotransferase; BSA = body surface area; cGvHD = chronic graft-versus-host disease; eCRF = electronic case report form; GI = gastrointestinal; NIH = National Institutes of Health; FEV₁ = forced expiratory volume in the first second; GI = gastrointestinal; OMRS = sum of scores for erythema, lichenoid, and ulcers; P-ROM = photographic range of motion.

Source: Protocol.³³

Table 16: Post-Baseline Overall Response Evaluation Based on All Organs

CR = compete response; PR = partial response; GI = gastrointestinal.

^aAt least 1 organ must be involved at baseline. Organ-specific responses vs. baseline status.

Source: Protocol.³³

ORR at Cycle 4 Day 1 (at End of Cycle 3)

ORR at cycle 4 day 1 was a secondary outcome of the REACH3 trial. The same response evaluation and assessment criteria were applied as for the primary outcome (for details, refer to previous section on ORR on cycle 7 day 1).

Health-Related Quality of Life

The HRQoL outcomes measured in the REACH3 trial included the FACT-BMT and EQ-5D-5L instruments as secondary outcomes. The Response Criteria Working Group within the NIH Consensus Development Project on Criteria for Clinical Trials in cGvHD recommends the FACT-BMT as a cGvHD nonspecific supportive measure to assess quality of life in adults.¹³ The EQ-5D-5L is frequently used for economic evaluations of health care and is recommended to be used by the National Institute of Health and Care Excellence as part of its Health Technology Assessments.³³ 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 to be calculated according to the user’s guide issued by the EuroQol Group.³⁴ According to the study’s protocol scores for each scale (mean, SDs, median, minimum, and maximum) and changes from baseline to each visit were measured and summarized descriptively. The FACT-BMT and EQ-5D-5L questionnaires were assessed every 4 weeks from baseline until cycle 7 day 1 (± 7 days). Visits after cycle 7 day 1 were at cycle 9 day 1 (± 7 days) and every 12 weeks thereafter until end of treatment.³³ During the crossover period (after completion of all assessments at cycle 7 day 1), disease assessments were planned to occur in same frequency as during the randomized treatment period.³³ 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 across 23 items and the bone marrow transplant subscale, which assesses specific BMT-related concerns across 23 items.² The higher the score, the better the quality of life.^35,36

Construct validity and responsiveness of the total score of the FACT-BMT instrument were evaluated in a systematic review, which included studies that reported patient-reported outcome measures in patients with aGvHD or cGvHD³⁷ (the FACT-BMT was included in 23 studies with a total of 5,071 patients). While the instrument appeared to have construct validity, responsiveness to change was not demonstrated. Estimates for MIDs in the literature were not found for the FACT-BMT in patients with GvHD.

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).³⁴ 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 ceiling effects, as compared to the EQ-5D-3L.³⁴ 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 an individual’s value or judgment of his or her current health state. The EQ-5D VAS scores are not used to create utility scores but provide complementary information to the EQ-5D-5L index score.³⁴

Construct validity of the total score of the EQ-5D-5L instrument was evaluated in a systematic review, which included studies that reported on patient-reported outcome measures in patients with aGvHD or cGvHD³⁷ (the EQ-5D-5L was included in 1 study with a total of 67 patients). The instrument appeared to have construct validity in patients with cGvHD. Estimates for MIDs in the literature were not found for the EQ-5D-5L in patients with GvHD.

Symptom Severity

The symptom severity outcomes measured in the trial included the modified Lee Symptom Scale and the Patient Global Impression of Change (PGIC) and Patient Global Impression of Severity (PGIS) instruments. The modified Lee Symptom Scale was a key secondary outcome, whereas the PGIC and PGIS instruments were exploratory end points. The Response Criteria Working Group within the NIH Consensus Development Project on Criteria for Clinical Trials in cGvHD recommends the Lee Symptom Scale as a GvHD-specific core measure for the assessment of quality of life in adults with cGvHD. It was reported in the sponsor’s submission that the PGIC and the PGIS questionnaires were administered to support the data collected with the modified Lee Symptom Scale; to help define a clinically meaningful score change in the modified Lee Symptom Scale (i.e., to confirm a responder definition or threshold for the modified Lee Symptom Scale total score).³³

The Lee Symptom Scale is a self-administered 30-item scale with 7 subscales (skin, eye, mouth, lung, nutrition, energy, and psychological status). Respondents are asked to report on symptom “bother” over the previous month on a 5-point Likert sale. The summary and subscale scores range from 0 to 100, with a higher score indicating worse symptoms.^33,38 The REACH3 trial included a modified version of the Lee Symptom Scale that focuses on the outcome of symptom severity, rather than symptom bother, and shortens the recall period from 1 month to 1 week (e.g., “Please let us know how severe any of the following problems have been in the past week”).³³ The objective of the modified Lee Symptom Scale analysis was to assess the improvement of symptoms based on the TSS. A responder was defined as having achieved a clinically relevant reduction from baseline in the TSS.³³ The modified Lee Symptom Scale was assessed every 4 weeks from baseline until cycle 7 day 1 (± 7 days). The modified Lee Symptom Scale was not assessed during the crossover period.³³

No studies evaluating the psychometric properties of the modified Lee Symptom Scale that was used in the REACH3 trial were found. It was reported in the sponsor’s submission that the modifications to the current Lee Symptom Scale will be validated following FDA guidelines using data collected from cognitive interviews and the REACH3 trial to support the validation of the modified Lee Symptom Scale.³³

No studies reporting a MID for the modified Lee Symptom Scale, which was used in the REACH3 trial, were found. According to the Lee Symptom Scale development study by Lee et al. (2002),³⁹ a difference of 6 to 7 points in the summary score is considered to be a clinically meaningful difference for symptoms in patients with cGvHD, based on a distribution-based method (i.e., 0.5 times the standard deviation of the baseline responses). A study by Teh et al. (2020)⁴⁰ that assessed a modified version of the Lee Symptom Scale (including a 28-item scale, rather than the 30-item scale used in the REACH3 trial) in 68 patients with cGvHD suggested that, based on the distribution methods, a 5- to 6-point difference (half a standard deviation) was considered clinically meaningful. However, a distribution-based approach to estimate the MID internally from the trial data is not an established method; rather, a triangulation with an anchor-based approach would be required.

The PGIC is a single-item measure of overall change in cGvHD symptoms since initiation of the study medication.³³ Respondents were asked to rate the change in the cGvHD symptoms since starting the study medication on a 7-point response scale that ranged from “very much better” to “very much worse.”³³ The PGIS is a single-item measure of overall cGvHD symptom severity in the previous week. Respondents were asked to rate the severity of their cGvHD symptoms over the past week on a 5-point scale that ranged from “no symptoms” to “very severe symptoms.” The results of the PGIC and PGIS were planned a priori to be used as anchors in the analyses designed to confirm a responder definition for the modified Lee Symptom Scale score. The PGIC and PGIS questionnaires were assessed every 4 weeks from baseline until cycle 7 day 1 (± 7 days). The PGIC and PGIS instruments were not assessed during the crossover period.³³

No studies evaluating the psychometric properties of the PGIC or PGIS scales were provided in the sponsor’s submission to CADTH for patients with GvHD. A MID for the PGIC or PGIS scales was neither identified by the sponsor nor found in the literature for patients with GvHD.

Duration of Response

DoR, a secondary outcome in the REACH3 trial, was assessed in responders (BOR = CR or PR) only and was defined as the time from first response to cGvHD progression, death, or the date of a change in or addition of systemic therapies for cGvHD.³³ Clinical response for the analysis of DoR was determined based on NIH consensus criteria¹³ and assessed by local investigators’ assessments.¹²

Best Overall Response

BOR was a secondary outcome in the REACH3 trial and was defined as the proportion of patients who achieved an overall response (CR or PR) at any time point up to and including cycle 7 day 1 and before the start of a change in or addition of systemic therapy for cGvHD. Clinical response for the analyses of BOR was based on local investigators’ overall response assessments.¹²

Overall Survival

OS was a secondary outcome in the REACH3 trial and was defined as the time from date of randomization to date of death from any cause.¹²

Nonrelapse Mortality

Nonrelapse mortality was a secondary outcome in the REACH3 trial and was defined as the time from date of randomization to date of death not preceded by underlying disease relapse or recurrence.¹²

Incidence of Malignancy Relapse or Progression

The incidence of MR was a secondary outcome in the REACH3 trial and was defined as the time from date of randomization to date of relapse or recurrence of the underlying malignant disease.¹² Assessment of MR was based on local institutional practices.²

Proportion of Patients With a Reduction of at Least 50% in Daily Corticosteroid Dose at Cycle 7 Day 1

The proportion of patients with a reduction of at least 50% in daily corticosteroid dose at cycle 7 day 1 was a secondary outcome in the REACH3 trial and was determined by systemic corticosteroid use and dose during the day 155 to day 168 interval (end of cycle 6).¹²

Proportion of Patients Successfully Tapered off All Corticosteroids at Cycle 7 Day 1

The proportion of patients successfully tapered off all corticosteroids at cycle 7 day 1 was a secondary outcome in the REACH3 trial and was determined by systemic corticosteroid use and dose during the day 155 to day 168 interval (end of cycle 6).¹²

Resource Use

The assessment of resource use was a secondary outcome in the REACH3 trial. Health care resource use was recorded in electronic case report forms (eCRFs). Resource-use data collected included the duration and frequency of hospitalization (from baseline up to end of study); frequency of emergency room visits (from baseline up to end of study); frequency of additional outpatient office visits to general practitioners, specialists, and urgent care (from baseline up to end of study); and the frequency of concomitant treatments for SR-cGvHD symptom assessment.^12,33

Safety

Safety was a secondary outcome in the REACH3 trial and was defined as the appearance of (or worsening of any pre-existing) undesirable sign(s), symptom(s), or medical condition(s) that occur after a patient’s signed informed consent has been obtained.³³ AEs reported for the first time or worsening of a pre-existing event after informed consent were recorded as AEs in the eCRFs. 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., a hematologic abnormality that required transfusion or hematological stem cell support), or required changes in the study drug. Components of study end points (i.e., worsening of study indication [cGvHD], including occurrence of a cGvHD flare, relapse or recurrence of underlying disease [including fatal outcomes]) were not reported as a serious adverse event (SAE) and were reported on other eCRFs but not AE eCRFs.³³

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

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

The REACH3 trial included the following parameters for the analysis of AEs: AEs by system organ class and preferred term, summarized by relationship (all AEs and AEs related to study treatment), seriousness (SAEs and non-SAEs), and AEs leading to treatment discontinuation, leading to dose interruption or adjustment, and leading to fatal outcome.³³

The REACH3 trial monitored the following parameters: physical exams, changes in vital signs from baseline, routine serum chemistry, hematology results, and coagulation profile.³³

Statistical Analysis

Details related to statical analyses of efficacy end points are summarized in Table 17.

A fixed-sequence hierarchical testing procedure was applied for the primary and the 2 key secondary end points: the interim analysis (when 196 patients [60% of the targeted 324 patients] completed their cycle 7 day 1 visit or discontinued earlier than that; July 9, 2019 data cut-off date) and the primary analysis (all 329 patients completed cycle 7 day 1 visit or discontinued earlier than that; May 8, 2020 data cut-off date).¹² The trial is still ongoing and a planned final analysis is expected to be conducted once all patients have completed the study (estimated completion date between the third and fourth quarter of 2022).¹² No formal statistical testing is planned a priori for the final analysis.

Sample Size Determination

The trial sample size of 324 patients was determined based on the primary end point of ORR at cycle 7 day 1. Assuming an ORR of 0.66 in the ruxolitinib group, with an observed odds ratio of at least 1.68, the study would have 90% power to demonstrate a statistically significant difference at a 1-sided alpha of 0.025.^12,33 With regard to the assumed ORR of 0.66, the sponsor’s submission referenced 1 study, by Olivieri et al. (2015),⁴¹ that reported an ORR of 0.66 (95% CI, 0.62 to 0.70) across 9 systemic treatments of SR-cGvHD from 82 nonrandomized studies. Olivieri et al. (2015)⁴¹ suggested that adherence to the NIH consensus criteria in the assessment of ORR would lead to lower response rates. Therefore, it was anticipated that the application of the NIH consensus criteria to assess ORR in the REACH3 trial could lead to a lower ORR.³³

For the analysis of the key secondary outcome, FFS, 324 patients was considered sufficient. If at the time of the FFS analysis the 6-month FFS rate for the ruxolitinib and BAT study groups was 0.69 and 0.56, respectively, the study would have 90% power to demonstrate a statically significant difference in FFS at a 1-sided alpha of 0.025. With regard to the assumed FFS rate of 0.56 for the BAT study group, the sponsor’s submission referenced a study by Inamoto et al. (2014)⁴² that provided FFS rates based on a nonrandomized cohort of 400 patients who received systemic treatment for moderate or severe cGvHD.⁴² No rationale was provided in the sponsor’s submission for the assumption of an FFS of 0.69 for the ruxolitinib group.

Interim Analysis

An interim analysis of efficacy and safety based on 194 patients (approximately 60% of targeted total sample size of 324) who completed the cycle 7 day 1 visit or discontinued the study earlier than that, for whom data of assessments were available, was planned with data cut-off date of July 9, 2019. The data monitoring committee (DMC) was responsible for reviewing the efficacy and safety data from this analysis. The interim analysis was added as part of protocol amendment 1 on December 20, 2017 to ensure timely assessment of and patient access to the investigational drug, despite protracted enrolment to the study because additional studies had been initiated in the first-line cGvHD setting.³³ An overall hierarchical testing procedure was applied, which combined the interim analysis (when 196 patients completed cycle 7 day 1 visit or discontinued earlier than that) and the primary analysis (all 329 randomized patients completed cycle 7 day 1 visit or discontinued earlier than that)¹² to control the family-wise error rate at the pre-specified overall significance level of alpha of 0.025 for the primary and key secondary end points.³³ The efficacy stopping bound and the respective alpha to be spent at the interim analysis was calculated based on actual information fraction (i.e., patients included in the full analysis set at the interim analyses, divided by 324) using the pre-specified alpha spending function. The interim analysis was conducted with efficacy data from 196 patients (60.5% of the targeted 324 patients) at a resulting alpha of 0.01176. For the primary analyses, 329 patients were available and the targeted alpha to retest the hypotheses (if not rejected at the interim analysis) was 0.01858.¹²

The interim analysis was conducted by an external vendor (not involved with the conduct of the study).^12,33 The DMC reviewed the unblinded results of the interim analyses based on a data cut-off date of July 9, 2019 in October 2019. The DMC was to inform Novartis only if “all 3 end points are positive” when providing their recommendations to prevent any impact on the further conduct of the study. The DMC recommended that the study continue as planned.³³

The trial is still ongoing and a planned final analysis is expected to be conducted once all patients have completed the study (estimated completion date is between the third and fourth quarter of 2022).¹² No formal statistical testing is planned a priori for the final analysis.

Primary Outcome

The primary outcome in the REACH3 trial was ORR on cycle 7 day 1. A brief overview of statistical methods used for the primary outcome is provided in Table 17.

Table 17: Statistical Analysis of Efficacy End Points (Outcomes Are Presented in Order of Priority as Identified in the CADTH Review Protocol)

BMT = bone marrow transplantation; BOR = best overall response; cGvHD = chronic graft-versus-host disease; CI = confidence interval; CR = complete response; DoR = duration of response; EQ-5D-5L = EuroQol 5-Dimensions 5-Levels; 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; PR = partial response; SD = standard deviation; TSS = total symptom score; VAS = visual analogue scale.

^aThe regulatory recommendations for determining additional benefit to patients with SR-GvHD differed between US and the rest of the world (the Committee for Medicinal Products for Human Use/Pharmaceuticals and Medical Devices Agency agreed that failure-free survival could be a meaningful measure of clinical benefit, the FDA recommended improvement in patient-reported outcomes).³³

^bORR was evaluated based on the full analysis set using local investigators’ overall response assessed at the cycle 7 day 1 visit according to the NIH consensus criteria.¹³

Source: Statistical analysis plan,³³ study protocol,³³ Clinical Study Report.¹²

To compare ORR between the 2 study groups, a Cochrane-Mantel-Haenszel Chi-square test, stratified by the randomization factor (i.e., cGvHD moderate versus severe) was used at a 1-sided 2.5% level of significance.³³ ORR was also summarized using descriptive statistics (N, %) and 2-sided exact binomial 95% CIs.⁴³ P value, odds ratio, and 95% Wald confidence limits were also calculated using the stratified Cochran-Mantel-Haenszel test. Patients with missing assessments or patients initiating or adding a new systemic therapy before cycle 7 day 1 were considered nonresponders.³³ No data imputation was applied.³³

As noted previously, an efficacy and safety interim analysis was added as part of Amendment 1, with a group-sequential methodology in which the stopping boundary and the respective alpha to be spent was estimated based on actual information fraction (number of patients included in the full analysis set at the interim analysis, divided by 324) using the pre-specified alpha spending function.

Several supportive analyses for ORR on cycle 7 day 1 were planned as follows:

• A detailed description of response rates (CR, PR, unchanged response, mixed response, and progression) at cycle 7 day 1 by study group

• ORR at cycle 7 day 1 assessed with the same analysis as for the primary efficacy analysis using all patients in the per-protocol set (if the per-protocol set differs from the full analysis set)

• A detailed description of the organ-specific response for all organs at cycle 7 day 1.

Key Secondary Outcomes

The REACH3 trial had 2 key secondary outcomes: FFS and the modified Lee Symptom Scale. The sponsor’s submission noted that in all regions in the world except the US, FFS was used as the first key secondary end point and was tested before the modified Lee Symptom Scale. However, the FDA recommended improvement in patient-reported outcomes to support the primary outcome, so patient-reported outcomes were tested as first key secondary end point, before FFS, to support the FDA recommendation. This CADTH review will focus on the results presented for all regions outside the US.³³ As noted previously (refer to Interim Analysis section), an overall hierarchical testing procedure was applied, which combined the interim analysis and the primary analysis, to control the family-wise error rate at the pre-specified overall significance level of an alpha of 0.025 for the primary and key secondary end points. Only hypotheses not rejected at the interim analysis were tested again at the primary analysis, applying the remaining alpha. Refer to Figure 5 for the overall hierarchical testing strategy, encompassing interim and primary analysis for all regions except the US.³³

Figure 5: Overall Hierarchical Testing Strategy Used for All Regions Except the US

FFS = failure-free survival; mLSS = modified Lee Symptom Scale score; ORR = overall response rate.

Note: Final analysis = primary analysis.

Source: Statistical analysis plan.³³

FFS was 1 of the key secondary outcomes in the REACH3 trial. In the event that the null hypothesis for ORR on cycle 7 day 1 was rejected at the 1-sided 2.5% level of significance, then FFS was formally tested for statistical significance. The P value from the stratified log-rank test at an overall 1-sided 2.5% level of significance was the confirmatory P value. Reasons for censoring included no experience of relapse, recurrence of underlying disease, or death due to underlying disease; NRM; or the addition or initiation of another systemic therapy for cGvHD. FFS was censored at the latest date of contact (on or before the cut-off date). No sensitivity analyses were planned a priori for FFS.³³

The following supportive analyses for FFS were conducted:

• A Kaplan–Meier (KM) plot of FFS was presented by treatment group. Medians and KM estimated probabilities at 3, 6, 12, 18, and 24 months with corresponding 95% CIs⁴⁴ were presented for each group. The HR and 95% CIs were estimated from a stratified Cox proportional hazards regression model.

• The cumulative incidence curve of each of the 3 FFS components (the other 3 components were considered a competing risk) was reported with estimates at 3, 6, 12, 18, and 24 months and associated 95% CIs.¹²

Data were not imputed when missing values, censoring, or discontinuations occurred.

The modified Lee Symptom Scale was 1 of the key secondary outcomes in the REACH3 trial. In the event that the null hypothesis for the ORR on cycle 7 day 1 and for FFS were rejected, then the modified Lee Symptom Scale was formally tested for statistical significance. Each patient providing responses to the Lee Symptom Scale was categorized as either a responder (i.e., achieved a reduction of 7 or more points on the TSS) or a nonresponder (i.e., did not achieve a reduction of 7 or more points on the TSS, data were missing or insufficient to calculate total symptom baseline score, or additional systemic therapy was used for cGvHD).¹² The main analysis of the modified Lee Symptom Scale aimed to assess the treatment-effect difference between the study groups on responder rates at cycle 7 day 1 in the ruxolitinib and BAT groups, respectively. To compare the responder rates between the study groups, the Cochrane-Mantel-Haenszel Chi-square test, stratified by the randomization stratification factor (i.e., cGvHD moderate versus severe) was applied at the 1-sided 2.5% level of significance. Additionally, the TSS and its 95% CI were presented by study group.³³

No imputation was performed if the total or subscale scores were missing at a visit.³³

Other Secondary Outcomes

All analyses for other secondary end points were noncomparative in nature and were not included in formal hypothesis testing.

BOR was a secondary end point in the REACH3 trial. BOR and its 95% CI was reported for each study group. P value, odds ratio, and 95% Wald confidence limits were also calculated using the stratified Cochran-Mantel-Haenszel test.³³

ORR at cycle 4 day 1 was a secondary outcome in the REACH3 trial. ORR at cycle 4 day 1, and its 95% CI, was reported for each study group. P value, odds ratio, and 95% Wald confidence limits were also calculated using the stratified Cochran-Mantel-Haenszel test.³³

DoR was a secondary outcome in the REACH3 trial. DoR was calculated for all patients with BOR. A KM plot of DoR was presented by treatment group. Medians and KM estimated hazard risk at 6, 12, 18, and 24 months with corresponding 95% CIs⁴⁴ were presented for each group. Patients with no event (i.e., cGvHD progression, additional systemic cGvHD therapy, or death) were censored at the last contact date.¹¹

OS was a secondary outcome in the REACH3 trial. Analyses for OS were conducted according to the randomized treatment group, stratified by cGvHD grade (i.e., cGvHD moderate versus severe). A KM plot of OS was presented by treatment group. Medians and KM estimated hazard risk and corresponding 95% CIs (Brookmeyer and Crowley 1982)⁴⁴ at 3, 6, 12, 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 a before the data cut-off date).³³

NRM was a secondary outcome in the REACH3 trial. The cumulative incidence of NRM and associated probabilities at 1, 2, 6, 18, and 24 months was assessed with underlying disease relapse or recurrence as the competing events.¹² Reasons for censoring included “not known to have died” or “experiencing the competing event.” Censoring occurred at the latest date the patient was event-free (on or before the cut-off date).³³

Incidence of MR was a secondary outcome in the REACH3 trial. The cumulative incidence of MR at 3, 6, 12, 18, and 24 months with associated 95% CIs were assessed for patients with underlying hematologic malignant disease, accounting for NRM as the competing risk. The proportion of patients who had hematologic malignancy relapse or recurrence and the associated 95% CI at 3, 6, 12, 18 and 24 months were presented by study group for patients with underlying hematologic malignant disease. Reasons for censoring included “not known to have relapse/recurrence” or “death without relapse/recurrence.” Censoring occurred at the latest date the patient was event-free (on or before the cut-off date).³³

The proportion of patients with a reduction of at least 50% in daily corticosteroid dose at cycle 7 day 1 and the proportion of patients who successfully tapered off all corticosteroids at cycle 7 day 1 were secondary outcomes in the REACH3 trial. Analyses assessed systemic corticosteroid use and dose by time intervals and the proportion of patients having successfully tapered off all systemic corticosteroids at cycle 7 day 1.³³

The tabulation of resource-use data was planned for the on-treatment period, and all data were to be listed. No further analyses were specified a priori in the statistical analysis plan. It was noted that data related to resource use were to be collected for the purpose of the economic evaluation and were analyzed and reported separately.²

Subgroup Analyses

For each subgroup, the point estimate and 95% CI were calculated. No formal statistical tests of hypotheses were conducted. Subgroup analyses were planned a priori in the statistical analyses plan for groups of patients listed below. Subgroup analyses were planned for the primary end point if statistically significant results were observed.³³

• Age group (12 to 65 years)

• Sex

• Race

• Regions: Europe (including Australia and Canada), US, Asia excluding Japan, Japan

• Chronic GvHD severity (moderate versus severe)

• Source of grafts

• Criteria for SR-cGvHD: lack of response or disease progression after administration of minimum prednisone 1 mg/kg per day for at least 1 week; disease persistence without improvement despite continued treatment with prednisone at more than 0.5 mg/kg per day or 1 mg/kg per every other day for at least 4 weeks; OR increase the prednisolone dose to more than 0.25 mg/kg per day after 2 unsuccessful attempts to taper the dose

• Prior cGvHD therapy (corticosteroid only versus corticosteroid ± CNI).

The following subgroups, planned a priori in the statistical analysis plan, aligned with the subgroups pre-specified in the protocol for this CADTH review: chronic GvHD severity, age, criteria for SR-cGvHD, and prior cGvHD therapy. Only subgroups identified in the CADTH review protocol are reported in the Efficacy section of this report.

Multiplicity

Apart from the pre-specified hierarchical testing of ORR on cycle 7 day 1, FFS, and modified Lee Symptom Scale, no adjustments for multiplicity on other outcomes were performed to control the type I error rate.

Amendments

The protocol of the REACH3 trial was amended once (Amendment 1: December 20, 2017). Amendment 1 included changes to inclusion and exclusion criteria and added ibrutinib to the protocol-defined list of BAT options. Ibrutinib had received regulatory approval in the US and Canada for the treatment of cGvHD since the time of the initial protocol. Additionally, a DMC, as an independent review group, was added based on health authority feedback to review safety data on a regular basis. Furthermore, an interim analysis was added to ensure timely assessment of and access for patients to the investigational drug, despite protracted enrolment to the study because additional studies had been initiated in the first-line cGvHD setting. The interim analysis allowed assessment of efficacy based on fewer patients but maintained the control of the error rates in the study. The decision to add an interim analysis was not influenced by efficacy or safety data from the study. The changes applied in Amendment 1 were not expected to affect the interpretability of the study or induce a cohort effect; when Amendment 1 was implemented, 15 patients had been enrolled in the REACH3 trial, and none of these patients had reached the 6-month visit used for the primary end point.³³

Analysis Populations

All efficacy data available at the time of the primary analysis were analyzed using the full analysis set, as defined in Table 18. The results from the interim analysis for the primary and key secondary outcomes were based on the first 196 patients in the full analysis set who completed cycle 7 day 1 visit. In supportive analyses at the primary analysis, ORR at cycle 7 day 1 was analyzed using the per-protocol set, as defined in Table 18. The crossover analysis set included patients randomized to the BAT group who crossed over to ruxolitinib treatment after the cycle 7 day 1 efficacy assessment. This set was used for all analyses for the crossover patients. Analyses of safety were performed using the safety population.

Table 18: Analysis Populations in the REACH3 Trial

aGvHD = acute graft-versus-host disease; BAT = best available therapy; cGvHD = chronic graft-versus-host disease; CNI = calcineurin inhibitor.

Source: Clinical Study Report.¹²

Results

Patient Disposition

Details of the patient disposition in the REACH3 trial are summarized in Table 19. A total of 404 patients were screened, and of those, 81.4% (N = 329) of patients were randomized to receive ruxolitinib (n = 165) or BAT (n = 164). Reasons for not being randomized included not completing screening (n = 74), failing screening (n = 72), or patient/guardian decision (n = 2), and 1 patient who was misrandomized was excluded.¹² A total of 323 (98.2%) randomized patients were treated: 165 of 165 (100%) in the ruxolitinib group and 158 of 164 (96.3%) in the BAT group. In the ruxolitinib group, 82 (49.7%) patients had discontinued treatment as of the primary analysis data cut-off date of May 8, 2020. In the BAT group, 122 (74.4%) patients had discontinued BAT treatment as of the primary analysis. The main reasons for discontinuing the assigned treatment were AEs (17.0% in ruxolitinib group versus 4.9% in the BAT group) and lack of efficacy (14.5% in the ruxolitinib group versus 42.7% in the BAT group). At the primary analysis, 83 (50.3%) patients in the ruxolitinib group and 42 (25.6%) patients in the BAT group were still on randomized treatment.

Table 19: Patient Disposition — End of Randomized Treatment (May 8, 2020 Data Cut-Off Date)

AEs = adverse events; BAT = best available therapy.

^aOngoing treatment and/or assessments during the main treatment period at cut-off date of May 8, 2020.

^bBased on the first 196 patients who completed cycle 7 day 1 visit (data cut-off date of July 9, 2019).

Source: Clinical Study Report.¹²

About one-third of patients (n = 61; 37.2%) randomized to the BAT group crossed over to ruxolitinib treatment at the end of cycle 6. At the time of the primary analysis, no patient had completed the crossover treatment period with ruxolitinib and 46 of 61 (75.4%) patients were still receiving ruxolitinib (refer to Table 20). Of the patients who crossed over to ruxolitinib, 15 (24.6%) discontinued the crossover treatment period. Reasons for discontinuing ruxolitinib treatment included lack of efficacy (n = 4) and physician or patient/guardian decision (n = 3), and 1 patient died. Eleven of 15 (18.0%) patients who discontinued, entered survival follow-up.¹²

Table 20: Patient Disposition — Crossover Period, Crossover Analysis Set (May 8, 2020 Data Cut-Off Date)

AEs = adverse events.

^aOngoing treatment at cut-off date of May 8, 2020.

Source: Clinical Study Report.¹²

Protocol deviations were reported for 104 (63.0%) patients randomized to the ruxolitinib group and 109 (66.5%) patients randomized to the BAT group. Patients may have had more than 1 violation. The overall type and frequency of protocol deviations appeared balanced between the treatment groups. The most commonly reported protocol deviations (ruxolitinib versus BAT) were “other deviations” (32.7% versus 37.2%), which included Lee Symptom Scale score not collected (13.9% versus 17.1%), hepatitis B virus and/or hepatitis C virus viral load testing value missing at cycle 1 day 1 (9.7% versus 7.9%), and deviations caused by the COVID pandemic, such as visits conducted outside the study site (6.1% versus 10.4%) and changes in the drug-supply method (5.5% versus 5.5%). Inclusion criteria deviations were observed in 18.2% of patients in the ruxolitinib group and 20.1% of patients in the BAT group, including, most commonly, SR-cGvHD criteria not met (13.3% versus 14.0%); exclusion criteria deviations were observed in 13.7% versus 13.4% of patients in the ruxolitinib and BAT groups, respectively, including, most commonly, the absence of active hepatitis C virus infection not confirmed or hepatitis C virus viral load test not performed (5.5% versus 6.7%). Deviations associated with treatment deviations and prohibited drugs occurred in 19.4% versus 13.4% and 10.9% versus 15.9% of patients in the ruxolitinib and BAT groups, respectively.¹²

Exposure to Study Treatments

Exposure to ruxolitinib and BAT as of the primary analysis is summarized in Table 21. Duration of exposure accounts for the time from first dose of randomized treatment until discontinuation of ruxolitinib or the latest BAT treatment in the BAT treatment group during the main treatment period. The median duration of treatment with ruxolitinib in the main treatment period (including the primary efficacy period [cycle 1 to the end of cycle 6] and the extension period [cycle 7 to cycle 39]; the main treatment period ended when a patient entered long-term survival or started treatment with ruxolitinib after crossover) was nearly twice as long as the treatment duration with BAT (41.3 weeks ([range = 0.7 to 127.3] and 24.1 weeks [range = 0.6 to 108.4] in the ruxolitinib and BAT groups, respectively). The duration of median follow-up from the first dose of randomized treatment until the end of the main treatment period (including patients who achieved a CR and stopped study treatment but continued regular assessments) was 42.6 weeks and 25.1 weeks in the ruxolitinib and BAT groups, respectively. The median duration of treatment up to cycle 7 day 1 was similar in the ruxolitinib and BAT groups (25.6 weeks [range = 0.7 to 25.6] and 24.0 weeks [range = 0.6 to 25.6], respectively). The median duration of exposure to the initial treatment was almost twice as long for the ruxolitinib group (41.3 weeks) as for BAT group (22.3 weeks). The duration of exposure to the initial treatment was defined as the time from start of initial treatment to when the initial randomized treatment was changed or a new systemic cGvHD treatment was added.¹²

Table 21: Duration of Follow-Up and Duration of Exposure to Randomized Treatment, Safety Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; cGvHD = chronic graft-versus-host disease.

Note: Subject-time is the sum of each subject’s follow-up/exposure in months.

^aDuration from cycle 1 day 1 to the end of the main treatment period (e.g., when entering long-term survival follow-up period or crossover from BAT to ruxolitinib).

^bDuration from cycle 1 day 1 of ruxolitinib or BAT treatments up to the data cut-off during the main treatment period.

^cDuration from cycle 1 day 1 of ruxolitinib or BAT treatments up to cycle 7 day 1; “up to cycle 7 day 1” refers to data up to the upper bound of the cycle 7 day 1 visit window.

^dRefers to the first initial treatment (with ruxolitinib or the initial BAT) until a change in or addition of new systemic cGvHD treatment or the end of initial cGvHD treatment on randomized treatment.

Source: Clinical Study Report¹²

In the crossover set, the median duration of exposure to ruxolitinib was 42.3 weeks (range = 2.6 to 98.7). The duration of exposure to ruxolitinib was longer than 24 weeks for 44 or 61 (72.1%) of patients, and 46 patients had not completed ruxolitinib treatment at the time of the analysis.

The median dose intensity and median relative dose intensity of ruxolitinib up to cycle 7 day 1 was high, indicating high treatment adherence. The median dose intensity for ruxolitinib was 19.6 mg/day (mean = 17.5 mg/day; SD = 3.72; minimum to maximum range = 4.8 to 20.5) up to cycle 7 day 1. The median relative dose intensity for ruxolitinib was 97.8% (mean = 87.4%; SD = 18.61; minimum to maximum range = 23.9 to 102.4) up to cycle 7 day 1. During the main treatment period, the median dose intensity for ruxolitinib was 18.6 mg/day (mean = 16.7 mg/day; SD = 3.98; minimum to maximum range = 4.9 to 20.5); the median relative dose intensity for ruxolitinib was 93.1% (mean = 83.4%; SD = 19.92; minimum to maximum range = 24.3 to 102.4).¹²

For 61 patients during the crossover period, the median dose intensity for ruxolitinib was 20.0 mg/day (mean = 17.8 mg/day; SD = 3.53; minimum to maximum range = 6.8 to 20.5); the median relative dose intensity for ruxolitinib was 99.8% (mean = 88.8%; SD = 17.67; minimum to maximum range = 33.8 to 100.0).¹²

Initial BAT treatment and the number of lines of BAT treatments started before cycle 7 day 1 are summarized in Table 22. The majority of patients (80.4%) in the BAT group received 1 BAT treatment before cycle 1 day 1; 17.7% of patients received 2 BAT treatments, and 1.9% of patients received more than 2 lines of BAT treatments. Most patients (34.8%) received ECP as initial BAT treatment, followed by MMF (22.2%), and ibrutinib (17.1%).¹²

The median duration of exposure to ECP was 24.8 weeks (range = 1.4 to 100.4) and the mean was 29.4 weeks. Treatment exposure to ECP was longer than 24 weeks for 33 (20.9%) patients. The median treatment exposure to MMF was 24.1 weeks (range = 2.1 to 108.4) and the mean was 30.2 weeks. Treatment exposure to MMF lasted longer than 24 weeks for 21 (13.3%) patients. The median treatment exposure with ibrutinib was 22.9 weeks (range = 0.6 to 78.1) and the mean was 23.4 weeks. Treatment exposure to ibrutinib was longer than 24 weeks for 12 (7.6%) patients. The median treatment exposure to low-dose MTX was 16.3 weeks (range = 2.0 to 101.4) and the mean was 25.5 weeks. Treatment exposure to low-dose MTX lasted longer than 24 weeks for 5 (3.2%) patients.¹²

Table 22: Initial BAT Treatment and Number of Lines of BAT Treatment Started Before Cycle 7 Day 1, Safety Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; ECP = extracorporeal photopheresis; MMF = mycophenolate mofetil; MTX = methotrexate.

Source: Clinical Study Report.¹²

Dose Modifications (Interruption, Reduction)

Patients who received ruxolitinib and experienced an AE considered to be related to ruxolitinib required dose modifications or interruptions as specified in the protocol. Patients in the BAT group did not have protocol-specified criteria for dose adjustment or interruptions, but were allowed to follow the investigator’s discretion, institutional guidelines, or product label.¹² According to the sponsor, given the varying nature of different BATs and the absence of a standardized dose for the treatment of cGvHD, dose interruptions, reductions, and changes were recorded but not analyzed.¹¹

As of the primary analysis (May 8, 2020 data cut-off date) during the main treatment period, 123 (74.5%) patients in the ruxolitinib group of the safety population required at least 1 dose reduction or interruption, 69.1% required a dose change, and 43.0% had at least 1 dose interruption. The main reasons given for dose changes or interruptions were AEs (47.3%), “per protocol” (25.5%), “physician decision” (23.0%), and dose tapering (13.9%).¹²

During the crossover treatment period, 32 (52.5%) patients required at least 1 dose reduction or interruption of ruxolitinib, 47.5% had a dose change, and 31.1% patients had a dose interruption. The main reasons given for dose changes or interruptions were AEs (31.1%) and physician decision (19.7%). Six patients experienced a dose re-escalation, per protocol.¹²

Concomitant Medication

The administration of concomitant immunosuppressive medications (starting on or after the start of the study treatment and no more than 0 days after the end of treatment) as of the May 8, 2020 data cut-off date was generally similar across treatment groups and reported for almost all patients (100.0% and 99.4% of patients in the ruxolitinib and BAT groups, respectively).¹² In addition to corticosteroids and CNIs, medications for the treatment of infections, GI symptoms, and sedation were among the most commonly used. Table 23 displays the most common concomitant immunosuppressive medications by anatomical therapeutic chemical (ATC) class (≥ 25% in either treatment group) and preferred term (≥ 8% in either study group).

The most frequently reported immunosuppressive concomitant medications by preferred term (ruxolitinib versus BAT) included bactrim (66.1% versus 62.0%), aciclovir (60.0% versus 50.6%), ursodeoxycholic acid (40.0% versus 34.8%), posaconazole (29.1% versus 34.2%), paracetamol (32.7% versus 27.2%), and colecalciferol (19.4% versus 29.1%).¹² CNIs were received by 27.3% of patients in the ruxolitinib group and by 19.0% of patients in the BAT group; cyclosporin (17.6% and 12.0% in respective groups) and tacrolimus (10.9% and 8.2% in respective groups) were the most commonly received CNI medications.¹²

Table 23: Concomitant Immunosuppressive Medications by ATC Class (≥ 25% in Either Study Group) and Preferred Term^a (≥ 8% in Either Study Group), Safety Set (May 8, 2020 Data Cut-Off Date)

ATC = anatomical therapeutic chemical; CNIs = calcineurin inhibitors; NOS = National Osteoporosis Society.

^aReported only for selected ATC classes.

Source: Clinical Study Report.¹²

Efficacy

Only efficacy outcomes and analyses of subgroups identified in the CADTH review protocol are reported in the following sections. Refer to Appendix 3 for detailed efficacy data.

Failure-Free Survival

The FFS results of the REACH3 trial for the ruxolitinib and the BAT groups are summarized in Table 24 (primary analysis) and Table 25 (interim analysis). As of the interim analysis (July 9, 2019, data cut-off date), the median FFS was not reached (95% CI, 11.9 to NE) in the ruxolitinib group and was 5.6 (95% CI, 4.5 to 5.8) months in the BAT group, with a stratified HR of 0.315 (95% CI, 0.205 to 0.486) in favour ruxolitinib (Table 25). The KM curves are depicted in Figure 6. For the US testing sequence, the results for FFS at the interim analysis did not reject the null hypothesis, but were met at the primary analysis. This CADTH review will focus on the results presented for all regions outside the US.¹²

As of the primary analysis (May 8, 2020 data cut-off date), median FFS was still not reached (95% CI, 18.6 to NE) in the ruxolitinib group and was 5.7 (95% CI, 5.6 to 6.5) months in the BAT group, with a stratified HR of 0.370 (95% CI, 0.268 to 0.510) in favour of ruxolitinib (Table 24). FFS was not formally tested at the primary analyses, given that results reached statistical significance at the interim analysis. The KM curves are depicted in Figure 7. The curves diverged, with a drop around 6 months noted for the BAT group, which was associated with the crossover of patients in the BAT group to ruxolitinib. The probabilities of patients surviving failure-free at 6, 12, and 24 months were 74.89 (95% CI, 67.48 to 80.85), 64.00 (95% CI, 55.78 to 71.09), and 58.94 (95% CI, 49.80 to 66.98), respectively, in the ruxolitinib group, and 44.46 (95% CI, 36.46 to 52.14), 29.62 (95% CI, 22.34 to 37.23), and 20.28 (95% CI, 9.33 to 34.19), respectively, in the BAT group.¹²

An analysis of cumulative incidences was conducted to assess the contribution to FFS of each of the 3 FFS components (i.e., relapse, nonrelapse mortality, or adding additional systemic cGvHD therapy; the 2 others were considered competing events). The results of this analysis suggested that the overall risk of the 3 events was lower in the ruxolitinib group than in the BAT group. For both study groups, the leading event related to FFS was change in systemic treatment.¹¹ The probabilities of patients experiencing treatment changes after randomization and up to completion of cycle 7 day 1 were 13.46% in the ruxolitinib group and 48.51% in the BAT group.¹¹ Figure 19 and Figure 20 in Appendix 3 show the cumulative incidence of FFS for each study group.¹²

Table 24: Summary of Efficacy End Points, REACH3, Full Analysis Set (May 8, 2020 Data Cut-Off Date) (Outcomes Are Presented in Order of Priority as Identified in the CADTH Review Protocol)

BAT = best available therapy; BOR = best overall response; cGvHD = chronic graft-versus-host disease; CI = confidence interval; CR = complete response; DoR = duration of response; FFS = failure-free survival; KM = Kaplan–Meier; HR = hazard ratio; MR = malignancy relapse or recurrence ; NE = nonevaluable; NRM = nonrelapse mortality; ORR = overall response rate; OS = overall survival; PR = partial response; TSS = total symptom score.

^aHR obtained from stratified Cox model using cGvHD severity at randomization as strata.

^bP value is nominal.

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

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

^eOther: Patient with additional systemic therapies along with CR/PR, per investigator assessment.

^fOne-sided P value, odds ratio, and 95% CI are calculated using stratified Cochran-Mantel-Haenszel test.

^gSubjects with change in or addition of new systemic cGvHD treatment are counted as nonresponders irrespective of the TSS value.

Source: Clinical Study Report.¹²

Table 25: Kaplan–Meier Estimate of Failure-Free Survival, Full Analysis Set (Interim Analysis)

BAT = best available therapy; CI = confidence interval; FFS = failure-free survival; HR = hazard ratio; KM = Kaplan–Meier.

Source: Clinical Study Report¹²

Figure 6: Kaplan–Meier Estimates of Failure-Free Survival, Full Analysis Set (Interim Analysis)

BAT = best available therapy; BID = twice daily; CI = confidence interval; FFS = failure-free survival; KM = Kaplan–Meier; RUX = ruxolitinib.

Source: sponsor’s response to additional information request.¹¹

Figure 7: Kaplan–Meier Estimates of Failure-Free Survival, Full Analysis Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; BID = twice daily; CI = confidence interval; FFS = failure-free survival; KM = Kaplan–Meier; RUX = ruxolitinib.

Note: P value is obtained from the log-rank test.

Source: Clinical Study Report.¹²

ORR at Cycle 7 Day 1

The ORR at cycle 7 day 1 (assessed by local investigator review, according to the NIH response criteria) for the REACH3 trial are summarized in Table 26 (interim analysis) and Table 24 (primary analysis). As of the interim analysis (July 9, 2019 data cut-off date), the REACH3 trial met its primary objective. The proportion of patients who achieved an overall response at cycle 7 day 1 was 50.5% (n = 49) (95% CI, 40.2 to 60.8) in the ruxolitinib group and 26.3% (n = 26) (95% CI, 17.9 to 36.1) in the BAT group, with a stratified odds ratio of 2.98 (95% CI, 1.62 to 5.48) (Table 26). For the US testing sequence, the results for ORR at cycle 7 day 1 were also met at the interim analysis. This CADTH review will focus on the results presented for all regions outside the US.¹²

As of the primary analysis (May 8, 2020) the proportion of patients that achieved an overall response at cycle 7 day 1 remained higher in the ruxolitinib group than in the BAT group. ORR at cycle 7 day 1 was achieved by 49.7% (n = 82) (95% CI, 41.8 to 57.6) of patients in the ruxolitinib group and 25.6% (n = 42) (95% CI, 19.1 to 33.0) of patients in the BAT group, with a stratified odds ratio of 2.99 (95% CI, 1.86 to 4.80) (Table 24). ORR at cycle 7 day 1 was not formally tested at the primary analyses, given that results reached statistical significance at the interim analysis. The proportion of patients with CR and PR was 6.7% (n = 11) and 43.0% (n = 71), respectively, in the ruxolitinib group, and 3.0% (n = 5) and 22.6% (n = 37), respectively, in the BAT group. The ORR cycle 7 day 1 supportive analysis using the per-protocol analysis set showed consistent results with the ORR results from the full analysis set. ORR at cycle 7 day 1 was achieved by 55.2% (n = 53) (95% CI, 44.7 to 65.4) of patients in the ruxolitinib group and 23.9% (n = 22) (95% CI, 15.6 to 33.9) of patients in the BAT group, with a stratified odds ratio of 4.08 (95% CI, 2.16 to 7.69). Additional supportive analyses were conducted to present the organ-specific response for all organs at cycle 7 day 1 (refer to Table 39 in Appendix 3).¹²

Table 26: Overall Response Rate at Cycle 7 Day 1, Full Analysis Set (Interim Analysis)

BAT = best available therapy; CI = confidence interval; CR = complete response; ORR = overall response rate; PR = partial response.

Source: Clinical Study Report.¹²

The ORR cycle 7 day 1 results by subgroups of interest, as specified in the protocol for this CADTH review, are summarized in Table 27. The treatment effect on ORR at cycle 7 day 1 was consistent with the primary analysis across patient subgroups, except for the subgroups of “prior cGvHD therapy with steroid + CNI + other systemic therapy” and “prior cGvHD therapy with steroid + other systemic therapy.” Of note, the sample sizes of these subgroups were small (fewer than 20 patients in each study group). Several other subgroups (i.e., “12 ≤ 18 years,” “> 65 years,” “criteria C”) had relatively small sample sizes (< 50 patients in each group). The wide CIs in subgroups reflected uncertainty in the effect estimates.¹²

Table 27: Subgroup Results for ORR at Cycle 7 Day 1, Full Analysis Set (March 8, 2020 Data Cut-Off Date)

BAT = best available therapy; cGvHD = chronic graft-versus-host disease; CI = confidence interval; CNI = calcineurin inhibitor; ORR = overall response rate; SR-cGvHD = steroid-refractory chronic graft-versus-host disease.

Note: Odds ratio and 95% CI are calculated using stratified Cochran-Mantel-Haenszel test.

Criteria for SR-cGvHD:

A. Lack of response or disease progression after administration of minimum prednisone 1 mg/kg per day for at least 1 week

B. Disease persistence without improvement despite continued treatment with prednisone at > 0.5 mg/kg per day or 1 mg/kg per every other day for at least 4 weeks

C. Increase prednisone dose to > 0.25 mg/kg per day after 2 unsuccessful attempts to taper the dose.

Source: Clinical Study Report.¹²

ORR at Cycle 4 Day 1

The ORR at cycle 4 day 1 (assessed by local investigator review, according to the NIH response criteria) results for the REACH3 trial are summarized in Table 24. As of the primary analysis (May 8, 2020), the proportion of patients that achieved an overall response at cycle 7 day 1 was 54.5% (n = 90) (95% CI, 46.6 to 62.3) in the ruxolitinib group and 31.1% (n = 51) (95% CI, 24.1 to 38.8) in the BAT group, with stratified odds ratio of 2.77 (95% CI, 1.75 to 4.39).

Health-Related Quality of Life

FACT-BMT

The completion rates declined over time in both study groups. At cycle 7 day 1, cycle 9 day 1, and cycle 12 day 1, 60.0%, 52.1%, and 35.8% of patients, respectively, were still available for completion in the ruxolitinib group, whereas the number of patients available for completion declined more rapidly in the BAT group, with less than 57.3% of patients available at cycle 7 day 1. After cycle 9 day 1, there were only 28.7% of patients available for completion in the BAT group, reflecting patients on BAT crossing over to receive ruxolitinib.¹¹

A definition of what constituted a clinically meaningful change from baseline in the current target population was not provided. Baseline scores were similar between study groups. Whereas the observed scores for the FACT-BMT trial outcome index from baseline to cycle 7 day 1 appeared to gradually increase (suggesting better HRQoL) in the ruxolitinib group, they appeared to remain constant in the BAT group; the mean change from baseline to cycle 7 day 1 was 4.11 (SD = 11.762) in the ruxolitinib group and –0.19 (SD = 11.742) in the BAT group. Similarly, whereas the observed scores from the FACT-BMT total score appeared to gradually increase in the ruxolitinib group, they appeared to be maintained in the BAT group; the mean change from baseline to cycle 7 day 1 was 3.76 (SD = 15.028) in the ruxolitinib group and 0.66 (SD = 16.816) in the BAT group.

EQ-5D-5L

The completion rates declined over time in both study groups. At cycle 7 day 1, cycle 9 day 1, and cycle 12 day 1, 60.0%, 50.9%, and 34.8% of patients, respectively, were still available for completion in the ruxolitinib group, whereas the number of patients available for completion declined more rapidly in the BAT group, with less than 56.7% of patients available at cycle 7 day 1. After cycle 9 day 1, there were only 26.8% of patients available for completion in the BAT group, reflecting patients on BAT crossing over to receive ruxolitinib.¹¹

A definition of what constituted a clinically meaningful change from baseline in the current target population was not provided. Baseline scores were similar between study groups. Overall observed scores from baseline to cycle 7 day 1 appeared variable, with overall small changes from baseline in both study groups. The mean change from baseline to cycle 7 day 1 of the EQ-5D-5L score was similar between study groups, at 0.07 (SD = 0.233) in the ruxolitinib group and 0.00 (SD = 0.226) in the BAT group.

Symptom Severity

Modified Lee Symptom Scale

The completion rates declined over time in both study groups. At cycle 7 day 1, 55.8% of patients were still available for completion in the ruxolitinib group, whereas the percentage of patients available for completion in the BAT group was 53.0% at cycle 7 day 1.¹¹

The rate of responders (responders included patients who achieved an improvement of ≥ 7 points on the TSS from baseline) from baseline to cycle 7 day 1 was higher in the ruxolitinib group (24.2% [95% CI, 17.9 to 31.5]) than in the BAT group (11% [95% CI, 6.6 to 16.8]), with an odds ratio of 2.62 (95% CI, 1.42 to 4. 82) (Table 24). The improvement in the TSS response was formally tested at the primary analysis, as the result at the interim analysis (Table 28) did not reject the null hypothesis. For the US testing sequence, statistical significance of the modified Lee Symptom Scale was also met at the primary analysis. This CADTH review will focus on the results presented for all region outside the US.¹²

Table 28: Responders at Cycle 7 Day 1 Based on the TSS From the Modified Lee Symptom Scale, Full Analysis Set (Interim Analysis)

BAT = best available therapy; cGvHD = chronic graft-versus-host disease; NA = not applicable; TSS = total symptom score.

Source: Clinical Study Report.¹²

Patient Global Impression of Change

The completion rates declined over time in both study groups. At cycle 7 day 1, 50.9% of patients were still available for completion in the ruxolitinib group, whereas the percentage of patients available for completion in the BAT group was 46.3% at cycle 7 day 1.¹¹

A definition of what constituted a clinically meaningful change from baseline in the current target population was not provided. Overall, the proportion of patients that responded “no change” was lower at cycle 7 day 1 than at cycle 2 day 1 in the ruxolitinib group (from 20.9% to 7.1%) and in the BAT group (from 28.6% to 22.4%); the reduction appeared more markedly in the ruxolitinib group. In addition, the proportion of patients that reported feeling “very much better” or “moderately better” was consistently higher in the ruxolitinib group than in the BAT group at every assessment point; at cycle 7 day 1, the proportions of patients that rated feeling “very much better” and “moderately better,” respectively, were 22.6% and 40.5% in the ruxolitinib group, and 15.8% and 23.7%, respectively, in the BAT group.

Patient Global Impression of Severity

The completion rates declined over time in both study groups. At cycle 7 day 1, 50.9% of patients were still available for completion in the ruxolitinib group, whereas the percentage of patients available for completion in the BAT group was 45.7% at cycle 7 day 1.¹¹

A definition of what constituted a clinically meaningful change from baseline in the current target population was not provided. The proportion of patients that responded as having “moderate,” “severe,” and “very severe” symptoms reduced from cycle 1 day 1 to cycle 7 day 1 in both groups; the reduction appeared markedly better in the ruxolitinib group (“moderate symptoms” from 39.8% to 21.4%; “severe symptoms” from 16.3% to 6%, “very severe symptoms” from 4.1% to 0%) than in the BAT group (“moderate symptoms” from 43.3% to 30.7%; “severe symptoms” from 21.6% to 10.7; “very severe symptoms” from 5.2% to 4.0%). The proportion of patients that responses as having “no” or “mild” symptoms increased from cycle 1 day 1 to cycle 7 day 1 in both groups; the increase was more pronounced in the ruxolitinib group (“no symptoms” from 15.3% to 28.6%; “mild symptoms” from 24.5% to 44.0%) than in the BAT group’ (“no symptoms” from 6.2% to 17.3%, “mild symptoms” from 23.7% to 37.3%).

Duration of Response

The DoR results (assessed by local investigator review, according to the NIH response criteria) for the REACH3 trial are summarized in Table 24. At the May 8, 2020 data cut-off date, among the patients who achieved a BOR (CR or PR at any time point up to and including cycle 7 day 1 and before the start of and change in or addition of systemic therapy for cGvHD), median DoR was not reached (95% CI, 20.2 to NE) in the ruxolitinib group and was 6.2 (95% CI, 4.7 to 13.1) months in the BAT group. The KM curves of DoR are depicted in Figure 8. The probabilities of maintaining a BOR for at least 6, 12, and 24 months were higher in the ruxolitinib group (76.58% [95% CI, 67.87 to 83.22]; 68.48% [95% CI, 58.94 to 76.26]; and 59.97% [95% CI, 47.58 to 70.33], respectively) than in the BAT group (52.11% [95% CI, 41.78 to 61.45]); 40.33% [95% CI, 30.28 to 50.15]; and 29.33% [95% CI, 15.32 to 44.84], respectively).¹²

Figure 8: Kaplan–Meier Curve of Duration of Response, Full Analysis Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; BID = twice daily; DoR = duration of response; KM = Kaplan–Meier; RUX = ruxolitinib.

Source: Clinical Study Report.¹²

Best Overall Response

The BOR results for the REACH3 trial are summarized in Table 24. At the May 8, 2020 data cut-off data, the proportion of patients that had achieved BOR in the ruxolitinib group was 76.4% (95% CI, 69.1 to 82.6) and in the BAT group was 60.4% (95% CI, 52.4 to 67.9), with an odds ratio of 2.17 (95% CI, 1.34 to 3.52). At cycle 7 day 1, a lower proportion of patients experienced an unchanged response in the ruxolitinib group than in the BAT group (16.4% versus 20.1%), and progression of cGvHD did not occur in the ruxolitinib group but occurred in 3.0% of patients in the BAT group.¹²

In the crossover analysis set, the proportion of patients who achieved BOR was 78.7% (95% CI, 66.3 to 88.1) (Table 29).¹²

Table 29: Best Overall Response During Crossover Treatment With Ruxolitinib, Crossover Set (May 8, 2020 Data Cut-Off Date)

CI = confidence interval; CR = complete response; NA = not applicable; ORR = overall response rate; PR = partial response.

Note: The 95% CI for the response rate was calculated using Clopper Pearson exact method.

Source: Clinical Study Report.¹²

Overall Survival

The OS results of the REACH3 trial for the ruxolitinib and the BAT groups are summarized in Table 24. As of the May 8, 2020 data cut-off date, 58 deaths occurred across both study groups. The median duration of follow-up for OS was 57.3 weeks for all patients, and was 56.6 weeks and 57.9 weeks in the ruxolitinib and BAT groups, respectively.¹¹ Median OS was not reached (95% CI, NE to NE) in either study group, with a stratified HR of 1.86 (95% CI, 0.648 to 1.820). The KM curves are depicted in Figure 9. The survival curves lay close to each other (were nearly equal) at the time of the analysis. The survival probabilities of patients surviving to 12 and 24 months were 81.37 (95% CI, 74.08 to 86.79) and 73.98 (95% CI, 61.77 to 82.82), respectively, in the ruxolitinib group, and 83.77 (95% CI, 76.51 to 88.95) and 75.24 (955 CI, 62.42 to 84.23), respectively, in the BAT group.¹²

Figure 9: Kaplan–Meier Curves of Overall Survival, Full Analysis Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; BID = twice daily; CI = confidence interval; KM = Kaplan–Meier; OS = overall survival; RUX = ruxolitinib.

Source: Clinical Study Report¹²

Nonrelapse Mortality

The results for nonrelapse mortality in the REACH3 trial are summarized in Table 24; event rates by time interval are shown in Table 40 in Appendix 3. As of the May 8, 2020 data cut-off date, the number of patients who experienced nonrelapse mortality was 27 in the ruxolitinib group and 22 in the BAT group. The number of patients who were censored was high; 129 and 134 patients in the ruxolitinib and BAT groups, respectively. The 12- and 18-month probabilities of nonrelapse mortality were 17.10% (95% CI, 11.54 to 23.59) and 17.10% (95% CI, 11.54 to 23.59), respectively, in the ruxolitinib group, and 14.13% (95% CI, 8.93, 20.48) and 15.12% (95% CI, 9.66 to 21.71), respectively, in the BAT group, suggesting similar event rates over time. The competing risk (hematological disease relapse or progression) was low in both study groups (9/165 and 8/164 for the ruxolitinib and BAT groups, respectively). The cumulative incidence of nonrelapse mortality is depicted in Figure 10.¹²

Figure 10: Cumulative Incidence of Nonrelapse Mortality by Treatment, Full Analysis Set (May 8, 2020 Data Cut-Off Date)

BID = twice daily; NRM = nonrelapse mortality.

Source: Clinical Study Report.¹²

Malignancy Relapse or Recurrence

The results for MR in the REACH3 trial are summarized in Table 24; event rates by time interval are shown in Table 41 in Appendix 3. At baseline, there were 156 and 160 patients in the ruxolitinib and BAT groups, respectively, who had underlying malignant disease. As of the May 8, 2020 data cut-off date, the number of patients who had events of MR was 9 in the ruxolitinib group and 8 in the BAT group. The number of patients who were censored was high (121 and 130 patients in the ruxolitinib and BAT groups, respectively). The 6- and 12-month probabilities of MR were low, with 4.94% (95% CI, 2.16 to 9.45) and 6.96% (95% CI, 3.34 to 12.38), respectively, in the ruxolitinib group, and 5.80% (95% CI, 2.68 to 10.65) and 5.80% (95% CI, 2.68 to 10.65), respectively, in the BAT group, suggesting similar event rates over time. The competing risk (underlying hematologic malignant disease) was relatively low in both study groups (26/156 and 22/160 for the ruxolitinib and BAT groups, respectively). The cumulative incidence of malignant relapse or recurrence is depicted in Figure 11.¹²

Figure 11: Cumulative Incidence of Malignant Relapse or Recurrence by Treatment, Full Analysis Set (May 8, 2020 Data Cut-Off Date)

BID = twice daily; MR = malignant relapse or recurrence.

Source: Clinical Study Report¹²

Proportion of Patients With a Reduction of at Least 50% in Daily Corticosteroid Dose at Cycle 7 Day 1 (Day 155 to Day 168 Interval [End of Cycle 6])

The results for the proportion of patients with a reduction of at a least 50% in daily corticosteroid dose at cycle 7 day 1 in the REACH3 trial are summarized in Table 24, and steroid exposure by time interval up the cycle 7 day 1 is described in Table 42 in Appendix 3. During the day 166 to day 168 interval (end of cycle 6), a similar number of patients in the study groups achieved a reduction of at least 50% in corticosteroid dose (normalized to body weight) from baseline (84 of 118 patients, or 71.2%, in the ruxolitinib group; 80 of 115 patients, or 69.6%, in the BAT group). The average biweekly steroid dose up to cycle 7 day 1 for both study groups is depicted in Figure 12.¹²

The average dose of systemic steroid was gradually reduced from cycle 1 day 1 up to cycle 7 day 1 (day 168) (Figure 12).

The reduction in steroid dose was slightly (but consistently) higher in the ruxolitinib group than in the BAT group. The dose intensity for systemic steroids in both study groups up to cycle 7 day 1 is described in Table 42 in Appendix 3.¹²

Proportion of Patients Successfully Tapered off All Corticosteroids at Cycle 7 Day 1 (Day 155 to Day 168 Interval)

The results for the proportion of patients successfully tapered off corticosteroids at cycle 7 day 1 in the REACH3 trial are summarized in Table 24, and steroid exposure by time interval up the cycle 7 day 1 is described in Table 42 in Appendix 3. The number of patients with no steroids during the day 155 to day 168 interval was 37 (31.4%) in the ruxolitinib group and 32 (27.8%) in the BAT group.¹²

Figure 12: Average Biweekly Weight-Standardized Steroid Up to Cycle 7 Day 1, Safety Set (May 8, 2020 Data Cut-Off Date)

BAT = best available therapy; RUX = ruxolitinib.

Source: Clinical Study Report.¹²

Resource Use

No results were available for the resource use outcome. Upon request, the sponsor noted that data for this outcome will be available in the final Clinical Study Report.¹¹

Harms

Only harms identified in the review protocol are reported here. Refer to Table 30 for detailed harms data. Safety data in Table 30 are summarized separately for the main treatment period until data cut-off (May 8, 2020) and up to cycle 7 day 1. The main treatment period (i.e., the on-randomized treatment period) refers to the randomized treatment period up to the data cut-off date, excluding crossover treatment and survival follow-up.¹²