Lower levels of certain biomarkers were predictive of treatment response for graft-versus-host disease (GVHD) at day 28, investigators found.
A new report has identified biomarkers that appear to correlate with treatment response in patients with steroid-refractory or dependent acute graft-versus-host disease (aGVHD).
The authors say such prognostic biomarkers could help guide treatment decisions in patients experiencing the transplant complication. The study, which was based on data from the phase 3 REACH2 trial of ruxolitinib (Jakafi), was published in Blood.
They noted that allogeneic hematopoietic stem cell transplantation (allo-HSCT) can cure several hematological conditions, but between 40% to 60% of patients receiving transplants will experience aGVHD within 100 days of the procedure. In such cases, systemic steroids are the standard first-line therapy, but about half of patients will become steroid refractory or dependent.
“Overall survival is reduced in such patients, highlighting the need for effective second-line treatment,” they wrote.
Ruxolitinib, a Janus kinase (JAK) inhibitor, is the only therapy approved by the FDA to treat steroid-refractory aGVHD, the authors noted. In the REACH2 trial, the drug had higher overall response rates and durable response rates compared with best available treatment (BAT).
Still, a subset of patients did not respond, or did not have durable responses, to ruxolitinib. The investigators therefore set out to see whether they could identify biomarkers that were linked with response to therapy. They decided to use data from the REACH2 trial, since clinical markers from before and after treatment were available for participants.
They developed models to identify biomarkers at baseline and at 14 days that were predictive of treatment response at 28 days. In line with the original results, the data showed treatment with ruxolitinib vs BAT was associated with treatment response. Type of conditioning, skin involvement, and age were also associated with treatment response, the authors found.
Of 21 biomarkers included in their final analysis, they found 8 biomarkers whose levels varied significantly between the responder and nonresponder cohorts at both baseline and at 14 days. An additional biomarker, CXCL10 plasma concentration, was significantly different between responders and nonresponders, but only at day 14, not at baseline.
“Lower levels of most aGVHD and immune cell markers at baseline were associated with increased probability of response,” they found. “In the day 14 model, levels of aGVHD markers at day 14, rather than change from baseline, impacted probability of response.”
The investigators said their bias-corrected area under the receiver operating characteristic curve was greater than 0.72 for both models, “indicating a strong monotonic relationship between the fitted and actual probabilities,” they said However, the accuracy and calibration of the day 14 model were greater than the baseline model.
The authors added that the biomarkers appeared to be meaningful for both the ruxolitinib cohort and the BAT cohort.
“Based on the current data, there was no indication of a differential biomarker role between treatment arms,” they said. “[T]he levels of biomarkers in the biomarker groups were prognostic for a response regardless of treatment received.”
The investigators said their results should be evaluated in a larger study. However, they said if their findings are validated, the identified biomarkers could be a useful tool for clinicians treating patients undergoing allo-HSCT.
Socie G, Niederwieser D, von Bubnoff N, et al. Prognostic value of blood biomarkers in steroid-refractory/dependent acute graft-vs-host disease: a REACH2 analysis. Blood. Published online February 24, 2023. doi:10.1182/blood.2022018579