Management of Acute Leukemia Relapse After Allogeneic Hematopoietic Stem Cell Transplantation: Comparison of Different Salvage Strategies

The prognosis of patients with acute leukemia (AL) relapsed after allogeneic stem cell transplantation (allo-SCT) is namely poor. The management relies on implementation of the Graf versus Leukemia (GvL) effect either by immunosuppression withdrawal or administration of Donor Lymphocyte Infusions (DLI); these measures are particularly effective if tumor burden is limited. In case of higher tumor burden, instead, other strategies can be exploited, such as cytoreductive chemotherapy (CT) and/or further allo-SCT with a different conditioning regimen or the use of a more mismatched donor.

Analysis of variables associated with outcome and response to different salvage strategies in pts with AL relapsed after allo-SCT.

we performed a retrospective analysis of a cohort of 64 consecutive pts with AL transplanted at the San Raffaele Scientific Institute and relapsed during follow up. Chi-square analysis was used for the association between variables; Kaplan Meyer curves and Long-rank test were used for survival analyses.

Between 09/2003 and 10/2010, 64 pts relapsed after receiving allo-SCT (median time to relapse: 6 months, range: 1–38); 54/64 pts (84%) had overt hematological relapse, 10/64 pts (16%) MRD positivity and/or Mixed Chimerism (MC). Therapeutic options applied were DLI or second allo-SCT, preceded or not by reinduction CT.

Ten pts with MRD positivity and/or MC received DLI as first choice strategy. Among them, 4 had a sustained CR (40%) and 6 progressed to overt relapse; in 5 cases an allo-SCT was performed, and 1 pt was rescued.

Fifty-four pts presented with overt relapse: 14 received DLI as first line salvage therapy; 2/14 (14%) obtained durable CR; an allo-SCT was attempted in 5/12 pts who progressed, with success in one. Fourteen pts went directly to a further allo-SCT; among them, 1 (7%) obtained durable CR. The remaining 26 pts received allo-SCT after re-induction CT; 5 (19%) obtained sustained CR.

A further allo-SCT, given as first-choice strategy or used as a rescue treatment after DLI failure, was followed by a longer OS compared to DLI only (median OS 328 vs. 187 days after relapse, p=0.03). Survival, in particular, was longer in pts receiving DLI or CT as “debulking” strategy before undergoing further allo-SCT (DLI only v.s. SCT only vs. SCT + debulking: Median OS 187 vs. 159 vs. 351 days after relapse, p=0.07).

Of note, all the long-surviving patients changed donor for the second SCT (alive 31% vs. 0% p=0.02)

Taken together, 24 pts were given DLI as fist choice strategy, and received a median of 2 DLI infusions (range 1–4), with a median dose of 1×10^6 CD3/Kg (range: 1×10^5–1×10^8 CD3/Kg); 5/24 pts (21%) developed acute GvHD (max grade III in a single case).

In univariate analysis, variables with a tendency to be associated with response to DLI were: early disease at transplant (p=0.07), Reduced Intensity Conditioning (p=0.06) and the use of Rapamycin for the GvHD prophylaxis (p=0.1). Variables with a tendency to association of longer OS after DLI were MRD positivity and/or MC (p=0.01), response to first DLI (p=0.07) and time to relapse longer than 6 months (p=0.1).

DLI doses associated with the best response (50%) were 1×10^7 CD3/Kg for MRD and MUD and 1×10^6 CD3/Kg for MMRD respectively. Of note, no association was observed between DLI doses and incidence of GVHD.

DLI is a reasonable option to treat pts with MRD positivity or MC after allo-SCT. When CR is not achieved after the first DLI, anyway, the switch to a more aggressive approach such as a further allo-SCT should be considered. The association of DLI with low-dose CT or other immune-modulating agent should be further exploited. For pts in over relapse, a further allo-SCT, if feasible, offers the best opportunity of prolonging survival. A debulking strategy such as CT or DLI should be considered according to the characteristics of the pts and the disease.

Bonini:MolMed: Consultancy.