Myeloid Leukemia of Down Syndrome: The Results of An International Retrospective Study

Abstract 2718

Children with Down Syndrome (DS) have an increased risk of developing acute myeloid leukemia. Myeloid leukemia of Down Syndrome (ML-DS) has a better prognosis than sporadic pediatric AML, using reduced-intensity treatment protocols as the leukemic cells of these patients are relatively sensitive to chemotherapy. Moreover, ML-DS patients have an increased risk of side-effects, hence there is a delicate balance between anti-leukemic efficacy and treatment related mortality.

So far, no prognostic factors in ML-DS have been identified that allow a risk-stratified approach in future ML-DS treatment protocols. The majority of the ML-DS cases are characterized by additional karyotypic changes in their leukemic cells besides the constitutional trisomy 21 (Forestier et al, Blood 2008), but the potential prognostic impact of these cytogenetic abnormalities is not known. We therefore conducted a large international retrospective study and collected clinical and cytogenetic data of 451 children with ML-DS from 13 collaborative study groups. Patients were eligible between 6 months and up to 5 years of age, and were diagnosed between Jan 1, 1995 and Jan 1, 2005. Patients who were not treated with curative intent were excluded. All karyotypes were centrally reviewed before assigning patients to cytogenetic groups.

The Kaplan-Meier method was used to estimate the 5-years OS and 5-years EFS, and outcome estimates were compared using the log-rank test. Cumulative incidence functions of relapse (CIR) were compared by the Gray test. For multivariate analysis, the Cox proportional-hazard regression model was used.

The median age of all ML-DS patients (n= 451) was 1.8 years (range 0.5 – 5.0) and the median white blood cell count (WBC) was 7.0 × 10⁹/l (range 0.8 – 290). The overall 5-years EFS was 78%, the 5-years OS 79%, the 5-years CIR 8%, and treatment related mortality (TRM) was 11%. The main groups were the patients with a normal karyotype (NK) (n= 103; 29%), trisomy 8 (n= 63; 18%) and ‘other’ patients (n= 187; 53%). Outcome estimates showed large differences across these different cytogenetic groups.

There were no significant differences in age between the NK, trisomy 8, and the other patients. However, trisomy 8 patients did have a significantly lower median WBC (6.3 × 10⁹/l; P²⁵=4.0 – P⁷⁵=10.3) compared to both other groups (7.2 × 10⁹/l; P²⁵=4.6 – P⁷⁵=16.7); P= 0.03). The NK patients had a significantly worse 5-years OS, EFS, and CIR compared to the other patients and the trisomy 8 group (OS 69% vs. 83% vs. 90%; P= 0.002 and 0.006; EFS 68% vs. 81% vs. 88%; P= 0.006 and P= 0.005; and CIR 19% vs. 9% vs. 5%; overall p(Gray)= 0.002). TRM was highest in the NK patients (13%). In addition, patients with a loss of 7q/monosomy 7 (n=16) had a 5-years EFS of 69%, which is remarkably high compared to non-DS AML (39%) (Hasle et al, Blood 2007).

Furthermore, younger patients, age < 3 years showed a significantly better 5-years EFS and CIR than patients aged ≥ 3 years (EFS 79% vs. 65%; P = 0.04; CIR 9% vs. 21%; P = 0.03). OS was not significantly different (80% vs. 69%; P = 0.10). Patients with a WBC < 20 × 10⁹/l had a borderline significantly better 5-years EFS and CIR than patients with a WBC ≥ 20 × 10⁹/l (EFS 80% vs. 70%; P = 0.05, CIR 9% vs. 16%; P = 0.05). OS was not significantly different. Multivariate analysis revealed NK patients (hazard ratio [HR] = 1.82, P = 0.027), WBC ≥ 20 × 10⁹/l (HR = 1.77, P = 0.029) and age > 3 years (HR = 2.37, P = 0.005) as independent predictors for poor EFS. The trisomy 8 group was not independently associated with a favorable outcome, which may be explained by the fact that the WBC of these patients was significantly lower and hence that there is interference between these 2 characteristics.

In conclusion, this study shows independent risk-groups within ML-DS that may benefit from stratified therapy. It has to be mentioned that patients were treated on different treatment protocols, which were sometimes adjusted for DS. As NK ML-DS patients have a higher CIR than TRM, therapy reduction is not an option for this group. To improve the prognosis of these patients, the biological background has to be investigated. As known in non-DS AML normal karyotype patients, mutations in WT1 and FLT3 are associated with poor outcome. FLT3 mutations are not found in ML-DS patients and as WT1 mutations have not yet been well studied in this group, this may be subject of further research.