Abstract
INTRODUCTION: T-cell depleted allogeneic peripheral blood stem cell transplant (TCD PBSCT) using CD34 selection achieves relapse rates comparable to those of unmodified grafts (Pasquini et al., JCO 2012), but disease-related predictors of outcome have not been fully characterized in the TCD setting. We evaluated the prognostic utility of the refined Disease Risk Index (DRI; Armand et al., Blood 2014) in TCD PBSCT.
METHODS: This was a retrospective analysis of patients who underwent first allogeneic HCT with TCD PBSCT for AML, ALL, or MDS at a single center between 1/2000 and 12/2015. Overall survival (OS), relapse-free survival (RFS), and chronic GVHD/relapse-free survival (CRFS) were estimated by the Kaplan-Meier method. Cumulative incidence of relapse, non-relapse mortality (NRM), acute GVHD (aGVHD), and chronic GVHD (cGVHD) were estimated using the cumulative incidence method for competing risks. The univariate association between variables of interest and OS/RFS/CRFS was evaluated using the log-rank test; Cox regression models assessed the adjusted effect of significant covariates on OS and RFS. Given only 1 patient with very high DRI, the high/very high DRI groups were combined. Similarly, given few patients with low DRI, the low/intermediate groups were combined in multivariate analysis.
RESULTS: The analysis comprised a total of 519 patients. Median age was 55 years (range 18-73). There were 302 patients (58%) transplanted for AML, 144 (28%) for MDS, and 73 (14%) for ALL. Seventeen patients had low DRI scores (3%), 431 intermediate (83%), and 71 high/very high (14%). Median follow-up among survivors was 53.1 months (range 4.6-171.0). Two-year estimates for outcomes of interest were OS 62.8% (95% CI 58.5, 66.9), RFS 58.1% (95% CI 53.7, 62.3), and CRFS 54.0% (95% CI 49.5, 58.2). The cumulative incidence of relapse at 2 years was 17.3% (95% CI 14.2, 20.7). There were 0 relapse events in patients with low DRI, whereas intermediate and high/very high DRI scores were associated with a significantly increased incidence of relapse (p < 0.001), with 2 year estimates 14.7% (95% CI 11.5, 18.3) and 37.1% (95% CI 25.8, 48.4), respectively. The cumulative incidence of NRM was 24.6% (95% CI 20.9, 28.4) at 2 years. The cumulative incidence of aGVHD at 100 days was 12.5% (95% CI 9.8, 15.5) for grade 2-4 and 2.5% for grade 3-4 (95% CI 1.4, 4.1); with a cumulative incidence of cGVHD of 4.7% (95% CI 3.1, 6.7) at 1 year. NRM, aGVHD, and cGVHD did not vary with DRI.
In univariate analysis, DRI was associated with significant differences in OS, RFS, and CRFS (Table 1; Figure). Additional factors associated with poorer OS in univariate analysis were HCT-CI score > 0, KPS < 90, donor type (matched unrelated or mismatched vs. matched related donor), and age > the median of 55.3 years; HCT-CI and KPS also correlated with significant differences in RFS. On multivariate analysis (Table 2), high/very high DRI corresponded to significantly greater risk of death (HR 1.72 for OS, [95% CI 1.24, 2.40]) and relapse or death (HR 1.86 for RFS [95% CI 1.35, 2.55]), compared with low/intermediate DRI. Multivariate analysis also showed that KPS < 90 was associated with worse OS and RFS, as did a higher HCT-CI score. Neither age nor donor type was significantly associated with OS in multivariate analysis.
CONCLUSION: In a large cohort of patients undergoing first TCD PBSCT at a single center for acute leukemia or MDS, DRI score significantly correlated with relapse incidence as well as OS, RFS, and CRFS. We have previously shown that the HCT-CI score, which incorporates patients' baseline comorbidities, is also predictive of outcomes after TCD PBSCT. Combining these prognostic tools will serve to better select appropriate patients for TCD PBSCT, a transplant approach currently under investigation in a multicenter phase 3 trial (BMT CTN 1301).
Koehne:Atara Biotherapeutics: Consultancy.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal