Ruxolitinib (RUX) Vs Best Available Therapy (BAT) in Patients with Steroid-Refractory/Steroid-Dependent Chronic Graft-Vs-Host Disease (cGVHD): Primary Findings from the Phase 3, Randomized REACH3 Study

BACKGROUND

Standard first-line treatment of cGVHD includes systemic corticosteroids; however, about 50% of patients (pts) are steroid refractory or dependent (SR/D) and require additional treatment. The best second-line therapy option has not yet been defined.

RUX, an oral JAK1/2 inhibitor, was superior to BAT in SR acute GVHD (aGVHD) in a phase 3 study (REACH2). Here we present the primary analysis of the REACH3 study (NCT03112603), a phase 3, open-label, randomized study evaluating RUX vs BAT in pts with SR/D cGVHD.

METHODS

Eligible pts were ≥ 12 years old, had received allogeneic hematopoietic cell transplant, and had developed moderate or severe SR/D cGVHD. Pts transitioning from aGVHD to cGVHD without tapering steroids were excluded. Pts treated with JAK inhibitors for aGVHD were allowed if they achieved complete response (CR) or partial response (PR) and had been off JAK inhibitor treatment for ≥ 8 weeks prior to cycle 1 day 1. Those treated with ≥ 2 prior lines of systemic therapy for cGVHD in addition to corticosteroids ± calcineurin inhibitors (CNI) were ineligible.

Pts were randomized (1:1) to RUX 10 mg bid or investigator-selected BAT (10 options) and were treated for 6 cycles (cycle = 28 days). Pts continued receiving their regimen of corticosteroids ± CNI. Viral prophylaxis and antibiotics were allowed as needed for infection prevention and treatment. Addition or initiation of a new BAT was allowed only after lack of response, intolerable toxicity, or cGVHD flare and was considered treatment failure. Crossover from BAT to RUX was allowed on or after cycle 7 day 1 (C7D1) in pts who did not achieve or maintain CR/PR, developed toxicity to BAT, or had a cGVHD flare.

The primary endpoint was overall response rate (ORR) at C7D1. ORR was defined as the proportion of pts achieving CR or PR, per NIH consensus criteria. Key secondary endpoints were failure-free survival (FFS; defined as time to the earliest of recurrence of underlying disease, start of new systemic treatment for cGVHD, or death) and improvement in symptoms based on change in the modified Lee symptom score (mLSS; 0 [no symptoms] to 100 [worst symptoms]) at C7D1. An mLSS responder was defined as having achieved a ≥ 7-point reduction from baseline in the total symptom score. The primary and key secondary endpoints were α-controlled via an overall hierarchical testing procedure.

RESULTS

A total of 329 pts were randomized to RUX (n = 165) or BAT (n = 164). Baseline characteristics were balanced in the 2 arms; 61% were male, and median age was 49 years (range, 12-76 years). Twelve pts were < 18 years old. Overall, 48% and 52% of pts had moderate and severe cGVHD, respectively. At data cutoff (May 8, 2020), 125 pts (38%) were on randomized treatment (RUX, 50%; BAT, 26%). Eighty-two (50%) and 122 (74%) pts discontinued RUX and BAT, respectively; 61 (37%) had crossed over to RUX. Reasons for discontinuation included lack of efficacy (15% RUX vs 43% BAT), adverse events (AEs; 17% vs 5%), and relapse (5% vs 4%).

The study met its primary endpoint with the efficacy boundary crossed at interim analysis (ORR, P = 0.0003). At C7D1, ORR was significantly higher in the RUX arm vs BAT (50% vs 26%; odds ratio, 2.99; P < 0.0001^a); the CR rate was higher with RUX (7% vs 3%; Table).

Both key secondary endpoints showed superiority of RUX vs BAT. FFS was significantly longer for RUX-treated pts (median FFS, not reached vs 5.7 months; HR, 0.370 [95% CI, 0.268-0.510]; P < 0.0001; Figure), and the mLSS responder rate was higher (24% vs 11%; odds ratio, 2.62; P = 0.0011). Overall, 31 RUX (19%) and 27 BAT (16%) pts died; the main cause of death was cGVHD (RUX, n = 22 [13%]; BAT, n = 13 [8%], including 2 deaths after crossover to RUX).

Rates of AEs up to C7D1 were comparable in the 2 arms (RUX, 98% [grade ≥ 3, 57%]; BAT, 92% [grade ≥ 3, 58%]. The most common AEs (≥ 15%) in the RUX vs BAT arms were anemia (29% vs 13%), hypertension (16% vs 13%), pyrexia (16% vs 9%), and ALT increase (15% vs 4%). Infections of any type occurred in 64% of RUX and 56% of BAT pts (19% vs 18% grade 3, grading based on Cordonnier et al 2006), and included fungal (12% vs 6%), viral (34% vs 29%), and bacterial (28% vs 26%) infections.

CONCLUSIONS

This is the first successful randomized phase 3 trial of RUX in pts with SR/D cGVHD. RUX demonstrated superior efficacy vs BAT, measured by a higher ORR, longer FFS, and greater symptom improvement. RUX was effective for moderate or severe SR/D cGVHD and its safety profile is consistent with that expected for this drug and this population.