Chronic Graft-Versus-Host Disease Abrogates Standard Risk Factors for Malignancy Relapse after Allogeneic Hematopoietic Stem Cell Transplantation

Abstract

Chronic graft-versus-host disease (cGVHD) reduces relapse risk in patients with hematological malignancies treated with allogeneic hematopoietic stem cell transplant (allo-HSCT). Nevertheless, relapse remains a major barrier to treatment success. Better understanding of factors determining relapse risk in the cGVHD patient population may lead to development of improved strategies for malignancy control. We hypothesized that cGVHD-related factors would contribute to decreased risk of relapse in addition to other well-known transplant related factors. Patients (N=275) were enrolled on the NCI cross-sectional cGVHD natural history study (NCT00092235) and described using NIH criteria for disease severity and organ scoring. Subjects were subsequently followed for malignancy relapse and survival. Potential predictors of relapse were assessed for their association with risk of relapse using Gray s test. Cox proportional hazards modeling was performed to estimate the joint effect of factors on risk of relapse. Seventeen patients experienced relapse at a median follow-up of 85 months. 48-month cumulative incidence of relapse was 5.7% (95% CI 3.3-8.9%). Median progression-free survival was 156 months. Factors associated with increased risk of relapse in multivariable analysis included aggressive malignancy as an indication for transplant (HR 3.93, 95% CI 1.27-9.10), shorter time from transplant to cGVHD evaluation (HR 0.24, 95% CI 0.06-0.88), and higher number of prior lines of systemic immunosuppressive therapy for cGVHD (HR 0.34, 95% CI 0.12-0.98). Interestingly, conditioning intensity, T-cell depletion, HLA-match, female donor to male recipient, blood vs. marrow stem cell source, or malignancy remission status at transplant were not predictive of relapse in univariate or multivariate analyses. These data suggest an important inverse relationship between cGVHD severity and likelihood of relapse post-transplant. Most classical relapse predictors seem to be abrogated by these effects.