Donor lymphocyte infusions for the treatment of minimal residual disease in acute leukemia

To the editor:

Hematologic relapse after an allogeneic hemopoietic stem cell transplantation (HSCT) in patients with acute leukemia is associated with a poor outcome, despite further cell therapy in the form of donor lymphocyte infusions (DLIs).¹  This has been taken as evidence that acute leukemia is less sensitive to the so-called graft-versus-leukemia (GVL) effect, especially if compared with chronic myeloid leukemia (CML), in which DLI alone can induce durable long-term molecular remissions.²  However, several reports suggest a strong protection exerted by chronic graft-versus-host disease (GVHD) in acute lymphoid leukemia (ALL) and acute myeloid leukemia (AML),^3,4  and low-dose immunosuppression in the first days after transplantation also reduces the risk of relapse.^5,6  Perhaps DLI would be effective in patients with ALL and AML if given when the tumor burden is low. Attempts to monitor minimal residual disease (MRD) in acute leukemia after HSCT have been reported and seem to predict hematologic relapse.^7,8  Some of these patients were given immune intervention to prevent hematologic relapse.⁹ 

In this study we wished to assess (1) the predictive value of MRD on hematologic relapse after transplantation in patients with AML/ALL and (2) whether cellular therapy with DLI would protect against leukemia relapse. We studied 80 patients with ALL (n = 44) or AML (n = 36) undergoing an allogeneic HSCT. MRD was evaluated monthly on bone marrow samples using a qualitative nested polymerase chain reaction (PCR) for IgH VDJ, as previously described,¹⁰  and T-cell receptor (TCR) gene rearrangement for T-ALL. Real-time PCR for Wilms tumor 1 (WT1) expression was used in AML.¹¹  MRD was considered positive in AML when WT1 copy numbers every 10⁴ copies of Abl were more than 180. Molecular positivity was defined as a positive PCR assay in the presence of a marrow sample in hematologic remission (blast count less than 5%).

The cumulative incidence of MRD positivity was 45%, with a median interval from transplantation to first MRD positivity of 120 days, and from transplantation to hematologic relapse of 203 days. Hematologic relapse was significantly higher in MRD⁺ patients (36%) compared with MRD⁻ patients (16%; P = .03). Patients were analyzed according to whether they were always MRD⁻ (n = 44), MRD⁺ receiving DLI (MRD⁺DLI⁺; n = 17) or MRD⁺ not receiving DLI (MRD⁺DLI⁻; n = 19). Reasons for not giving DLI were presence or development of GVHD after cessation of immunosuppressive therapy (n = 7), donor unavailable (n = 8), early relapse (n = 2), or other (n = 2). Hematologic relapse was 16% in MRD⁻ patients, 6% in MRD⁺DLI⁺ patients, and 63% in MRD⁺DLI⁻ patients (P < .001; Figure 1); the actuarial 3-year survival in these 3 groups was 78%, 80%, and 26%, respectively (P = .001; Figure 1). Mortality due to acute GVHD following DLI was 12%. In multivariate Cox analysis, the MRD group predicted relapse (P < .001) and survival (P = .01), together with disease phase and chronic GVHD. In MRD⁺ patients, DLI protected against relapse (P = .003) and improved survival (P = .01).

In conclusion, we confirm that MRD detected after transplantation is a significant predictor of relapse. Treatment of MRD with DLIs appears to protect against leukemia relapse, although caution with DLI dosing needs to be used because of the potential risk of GVHD.