Abstract
In the two-step approach to haploidentical hematopoietic stem cell transplantation (HI-HSCT), following conditioning, recipients receive the lymphoid portion of their graft ("DLI") containing a fixed dose of 2 x 10 8/kg CD3+ cells (Step 1 of the transplant). After the DLI, a 2 day period of alloreactivity and haploimmunostorm occurs which resolves with the administration of Cyclophosphamide (CY) for GVHD prophylaxis/rejection prevention. After the completion of CY, a CD34-selected donor product is infused (Step 2 of the transplant). It has been postulated that a rapid reduction in quantity and a functional impairment of alloreactive T cells by CY contributes to the establishment of tolerance between the donor and recipient immune systems. Based on their degree of disease burden at HSCT, a potentially important difference between patients undergoing HI-HSCT is the percentage of graft versus tumor (GVT) versus graft versus host (GVH) reactive T cells that are likely to become activated during the haploimmunostorm, and as a result of this activation, their subsequent inhibition by CY. In the initial two-step trial (2006-2010-initial group), relapsed disease was the major barrier to long-term survival (OS). In the follow-up trial (2013-2016-later group), extra time between the end of conditioning and the DLI was added to allow for a greater reduction in disease burden prior to the introduction of the DLI (Figure). We hypothesized that by allowing more of the malignancy to die off, less GVT specific T-cells in the DLI would become activated prior to CY administration. The result would be the avoidance of suppression of these cells by CY resulting in reduced relapse rates.
We compared the outcomes of patients treated on the initial trial to those treated on the later trial. The number and median age of subjects in the initial and the later groups were 27 vs 40 and 51 vs 52 years. Diagnoses were AML (59% vs 33%), MDS (7% vs 18%), ALL (15% vs 27.5%), Burkitt's (0% vs 5%), NHL (11% vs 10%), and other (8% vs 6.5%). There was no significant differences between the groups in Revised Disease Risk Index on chi square analysis or HCT-CI on Mann-Whitney analysis. All patients received 12 Gy total body irradiation for conditioning; the initial group over 4 days, the later group over 3 days. Patients in both groups received a fixed dose of 2 x 10 8/kg of T cells in the DLI, and tacrolimus and mycophenolate mofetil for additional GVHD prophylaxis. Median CD34 cell dose was 3.6 x 10 6/kg versus 6.0 x 10 6/kg (p=0.01) in the initial versus later groups respectively.
At two years in the initial and later groups, the cumulative incidence of relapse-related mortality was 30% vs 10% (p=0.04), non-relapse mortality was 22% vs 30% (p=0.28), acute GVHD 48% vs 38% (p=0.198), and chronic GVHD 15% vs 20% (p=0.591). Kaplan-Meier probability of OS at 2 years was 48% vs 63% (p=0.431).
The analysis was limited by its retrospective comparison of two groups and small sample size. However, the data supports a potential relapse benefit, without increased toxicity, of the later approach. Improved GVT effects due to the alteration in DLI timing potentially contributed to lower relapse related-mortality, although superior tumor control based on a different radiation strategy should be considered. A prospective, randomized two arm trial comparing the two conditioning approaches is warranted.
No relevant conflicts of interest to declare.
