The Presence of Minimal Residual Disease Pre- and Post-Transplant Is Predictive of Outcome in Recipients of Allogeneic Transplantation for Acute Myeloid Leukaemia.

Abstract 2601

Poster Board II-577

Minimal residual Disease (MRD) monitoring is known to be of importance in guiding management in patients with acute myeloid leukaemia (AML) treated with chemotherapy. Many subtypes of AML do not, however, have a molecular marker available for MRD monitoring and in this setting the use of multiparametric flow cytometry (MFC) has been evaluated.

Few studies have investigated the impact of MRD in the setting of allogeneic transplantation for AML. We speculate that, as with chemotherapy, the outcome in MRD positive patients will be inferior to MRD negative patients due to an increase in disease relapse. In order to investigate this hypothesis we studied 74 patients who underwent allogeneic transplantation for AML in a single centre from 2004–2008. The median age was 47.1 years (range: 20.8–70.1). The overall survival at 2 years was 48% with a median follow-up of 1.96 years. Conditioning was myeloablative in 37 patients and reduced intensity in 37 patients. 46 patients had T cell depletion as part of the conditioning. 48 patients had an unrelated donor and in 26 a sibling donor was used. MRD analysis by MFC (sensitivity 0.04%)in the bone marrow was performed pre-transplant and on day 28 and 100 post-transplant.

Using conventional criteria, 47 patients were in CR1, 15 in CR2, 4 in CR > 2 and 8 patients were not in CR at the time of the transplant. Survival at 2 years was 58%, 50%, 25% and 12,5% respectively (p=0.001). We examined the impact of MRD by MFC in patients who were in CR at transplant. Survival at 2 years in patients who were MRD negative (MRD-) pre-transplant was 70% compared to 29% in those who were MRD positive (MRD+) (p=0.003). In multivariate analysis, the most significant factor affecting outcome was being MRD+ pre-transplant (RR of death of 3.04, p=0.011; 95% CI 1.2–7.1). Interestingly, there was no significant difference in survival seen between patients who were MRD+ and those who were not in CR (by conventional tests) pre-transplant (p=0.305).

The MRD analysis at day 28 was not predictive of outcome. However, in patients alive and without relapse at day 100, MRD analysis was predictive of outcome (2 year: MRD-70% vs 27% MRD+; p=0.045). In patients who were MRD- at both time-points, survival at 2 years was 81% compared to 42% in those who were MRD+ at one or both time-point (p=0.040).

The incidence of disease relapse was 32% at 2 years and non-relapse mortality (NRM) was 8% at day 100 and 22% at 1 year. Interestingly, while the pre-transplant stage was predictive of relapse (2 years: those in CR 28% (no difference for CR1 or >CR1) compared to those not in CR 63%, p=0.004), the presence of MRD positivity was not; neither pre-transplant (p=0.47) nor at day 100 (p=0.29). Conversely, the NRM was significantly higher in those who were MRD+ pre-transplant, but not at day 100. The NRM at day 100 was 4% in MRD- patients and 15% in MRD+ patients (p=0.01). In multivariate analysis this conferred a RR of NRM of 2.12 (p=0.014; 95% CI 1.1–3.8) to MRD+ patients.

We conclude that the presence of MRD detected by flow cytometry pre- and day 100 post-transplant is predictive of a worse patient outcome. Interestingly this appears to be predominantly due to an increase in NRM rather than an increase in disease relapse. We speculate that differences in the duration and/or intensity of GVHD prophylaxis may be implicated in this finding and we are currently investigating this hypothesis. A better understanding of these factors will allow us to tailor treatment based on MRD status in an intelligent fashion.