To The Editor

Chronic graft-versus-host disease (cGVHD) is diagnosed in 30%–70% of patients after allogeneic hematopoietic cell transplantation (HCT), and is a leading cause of late morbidity and non-relapse mortality [1]. Safe and effective long-term therapies are necessary to control symptomatic manifestations, and repress the disease activity causing organ damage and disability.

Belumosudil mesylate (formerly ME3208) is a selective inhibitor of Rho-associated coiled-coil kinase 2, and demonstrated good response rates and symptom improvements in patients with cGVHD in the United States (US) [2, 3]. Belumosudil received Japanese regulatory approval (March 2024) for the treatment of cGVHD after HCT in patients with an inadequate response to steroid therapy. This was based on the primary analysis from a Japanese clinical trial (jRCT2011210041), which evaluated belumosudil 200 mg once daily as second or subsequent line of therapy (LOT) in patients with steroid-dependent/resistant cGVHD, and showed that treatment was associated with a best overall response rate (ORR) at 24 weeks after last patient enrollment of 85.7% and was well-tolerated [4]. Herein, we report the efficacy and safety findings at the end of the Japanese clinical trial (26 March 2024), with a median belumosudil treatment duration of 20.9 months.

Full details of the study design, eligibility criteria, and characteristics of the 21 patients with cGVHD included in the analyses have been published [4]. These are summarized, along with the analysis procedures, in the Supplementary Methods.

The best ORR at the end of treatment was the same (85.7%) as at 24 and 48 weeks (Fig. 1a); all were partial responses (PR). The median duration of response (DOR) was noncalculable (Fig. 1b). Among 18 responders, 8 (44.4%) had responses ongoing at data cut-off, 13 (72.2%) had sustained responses for ≥24 weeks and 9 (50.0%) for ≥48 weeks. Patients who discontinued treatment tended to respond relatively late, if at all; conversely, many patients with treatment ongoing responded early (within the first 2–3 months) after belumosudil initiation (Fig. 1c). ORR over time is shown in Supplementary Fig. 1a. Of the patients with lung involvement at baseline (Fig. 1d), one had a complete response (CR) at Week 60, and most patients demonstrated stable lung function after starting belumosudil (Supplementary Fig. 1b). The best ORR exceeded 66.7% in all subgroups except patients with steroid-resistant cGVHD (Supplementary Fig. 2). Overall, 13/21 patients (61.9%) had ≥1 clinically meaningful improvement ( ≥ 7 points decrease from baseline) in the Lee cGVHD symptom scale (LSS) score, and 14/21 (66.7%) had ≥1 corticosteroid dose reduction from baseline (Supplementary Tables 1, 2). Median survival outcomes were non-calculable (Supplementary Fig. 3).

Fig. 1: Selected efficacy evaluations at the end of the Japanese clinical trial.
figure 1

a Best ORR at key time points; b DOR rate (belumosudil responders only); c treatment status by individual patient; d best ORR by organ; e changes in steroid dose and LSS overall scores over time in the adolescent patient. In panel b + indicates censored patients. In panel c the symbols indicating the start and end of response refer to the start and end of periods judged to be CR/PR. Since the end of response could refer to a situation in which one affected organ had worsened, but other organs were still responding (i.e., not an overall evaluation of progressive disease), treatment was continued after the end of response with the aim of maintaining any persisting responses (in order to preserve the treatment benefit in the organ(s) that continue to respond), whilst also ensuring patient safety. Reasons for discontinuation are indicated as follows: a = withdrawal of consent; b = physician’s judgment; c = progression/recurrence of malignancy; d = progression of cGVHD; e = protocol deviation. The adolescent patient is indicated by #14. In panel e the steroid dose is shown as a plane and corresponds to the treatment ongoing in panel c. At baseline, the corticosteroid dose was 10.0 mg/day; this was reduced to 5.0 mg/day on the day following Week 24, 2.5 mg/day on the day following Week 52, 1.5 mg/day on the day following Week 61, and treatment was discontinued on the day following Week 65. The LSS score was improved by 19.0 points from baseline (score 24.8 points) to Week 2 (score 5.8 points), and clinically meaningful improvements (a decrease of ≥7 points from baseline, indicated by the blue dotted line) endured between Weeks 2–84. BL baseline, CI confidence interval, CR complete response, EOT end of treatment, GI gastrointestinal, LR-M lack of response-mixed, NIH National Institutes of Health, PR partial response.

Full size image

The study included one adolescent patient (aged 13 years; patient #14 in Fig. 1c) with steroid-dependent cGVHD, with a best overall response of PR occurring after 2 weeks of treatment. The patient had a CR in the esophagus and upper gastrointestinal tract between Weeks 2–84 and a lung CR between Weeks 60–72 (Supplementary Table 3). This patient had improvements in the LSS score, and was able to discontinue corticosteroids (Fig. 1e), despite having a steroid-dependent condition.

There were no large or unexpected increases in treatment-emergent adverse event (TEAE) frequency or severity over the longer treatment duration, compared with the primary analysis (Supplementary Tables 4, 5). Levels of blood cells were maintained (Supplementary Fig. 4), and no withdrawal symptoms were observed within 28 days after belumosudil discontinuation (Supplementary Table 6).

Overall, belumosudil 200 mg once daily as second or subsequent LOT was an effective long-term treatment option for cGVHD. The majority of study patients responded to treatment (18/21 [85.7%]), and responses were sustained. There were differences between our data and those from the US ROCKstar study [3, 5], including the DOR (non-calculable after a median belumosudil treatment duration of 20.9 months vs. 54 weeks after 14 months) and the estimated failure-free survival rate at 2 years (76% vs. 48%). This may be explained by the fact that our study allowed earlier administration of belumosudil within the treatment sequence (median 1 prior LOT) compared with ROCKstar (median 3 prior LOTs), and outcomes may be prolonged if belumosudil is initiated when patients have less resistant disease.

Our data indicate that long-term treatment may provide additional efficacy improvements compared with the primary analysis [4]. Belumosudil elicited responses in organs including the joints and fascia, skin, and mouth, which are likely to result in quality of life improvements. This inference was supported by clinically meaningful LSS improvements in 61.9% of patients on ≥1 occasion. Long-term belumosudil administration may allow gradual reduction, and potential discontinuation, of corticosteroids. This is of interest, as long-term use of high doses of corticosteroids can be associated with serious side-effects. The disease course of the adolescent patient also indicates that remission of pulmonary symptoms is possible, although it is unclear whether factors such as patient age and duration of cGVHD may play a part. Most patients in our study with lung involvement demonstrated stable lung function after starting belumosudil, and one achieved a pulmonary CR before the data cut-off. These results align with those of a combined analysis of the US belumosudil studies in which percent predicted forced expiratory volume in one second (%FEV1) measurement trajectories indicated maintenance of pulmonary function over time among responders and non-responders [6]. The best lung-specific ORR (among patients with lung involvement at enrollment) was 26% (12/47 patients) in the ROCKstar study [3], and 37% (24/65 patients) in the combined phase II analysis [7]. In a Chinese phase II study in 30 patients with cGVHD who had received ≥1 prior LOT, lung-specific responses were observed in 2/30 patients (15.4%) [8]. Further investigation after long-term belumosudil use is warranted to assess the occurrence of late responses or lengthy durations of stable disease. In terms of safety, this analysis did not identify any long-term tolerability concerns, and there were no rebound phenomena associated with cessation of belumosudil, suggesting that cGVHD treatment can be safely changed.

Belumosudil may be a valuable alternative treatment for patients with cGVHD compared with ruxolitinib or ibrutinib. Indirect comparisons suggest that fewer patients discontinued this study due to TEAEs (all-cause, 9.5%; drug-related, 0%) or cGVHD progression (9.5%) compared with other treatments. In the pivotal phase II study of ibrutinib, discontinuation rates were 33.3% for all-cause AEs and 11.9% for cGVHD progression [9]. In a comparative study in patients with steroid-resistant cGVHD, 17.0% discontinued ruxolitinib due to all-cause AEs and 14.5% due to lack of efficacy, while 4.9% discontinued investigator’s choice of therapy due to AEs and 42.7% due to lack of efficacy [10]. In long-term studies, 5.7% of patients discontinued ruxolitinib due to drug-related AEs [11], while 15.8% discontinued ibrutinib due to all-cause AEs and 10.5% due to cGVHD progression [12].

The main limitations of our study relate to the design (open-label, single arm, small population) which restricts the inferences that can be made. Additionally, the use of belumosudil in the second-line setting is currently only applicable to Japanese and Chinese clinical practice.

In conclusion, our findings demonstrated that long-term treatment with belumosudil as second or subsequent LOT was effective and well-tolerated in Japanese patients aged ≥12 years with steroid-dependent/resistant cGVHD. Taken together, this analysis and previously published data [3, 5, 6, 8] serve to confirm the efficacy of belumosudil for the treatment of cGVHD.