Outcome Of MDS and AML With MDS-Related Changes: Treatment Versus Prognostic Factors

From January, 2004-August, 2012, 367 pts (MDS, n=208; sAML, n=159) consecutively treated (median age 63y) at the University of Leipzig were included. Patients (84%) with marrow blasts >10% received induction chemotherapy (CT; 59%) or azacitidine (AZA; 25%) (after its approval in the EU in January, 2009) with the intention of performing a subsequent allogeneic hematopoietic cell transplantation (HCT) in pts <70y. Up-front HCT was scheduled if blasts were <10% (n=56). Cytogenetics were categorized according to Schanz et al, JCO 2012 for MDS and the WHO classification for sAML. As confounders in the estimation of therapy, host- and disease-related features were investigated in a multistep process. 38% of pts had >2 comorbidities with no difference between MDS and sAML. The sAML group (median blasts 40%) included 69 and 81 pts with previous MDS and MDS-related cytogenetic abnormality respectively. Cytogenetics were poor and very poor in 34% of MDS. Outcome at two years are presented.

Median interval between diagnosis of MDS and therapy was 3.6 months.

Median survival time for sAML was 15 vs 72 months for MDS (p<0.0005). Overall, age was higher (median 68y) and blasts lower (median 13%) in the AZA group compared to CT (62y and 27%) (p<0.0005). Cytogenetics and the comorbidity burden (CB) were comparable. OS with AZA was similar to up-front HCT (68%) and superior to CT (48%) (p=0.01). OS was 50% if HCT was performed after CT (136 pt) compared to CT alone (p=0.01). In the 20% of pts >70y, AZA was given to 52% and CT to 48%. OS was 55% and best with AZA (p=0.01). Median survival times were 30 for AZA/MDS, 27 for AZA/AML, 15 for CT/AML, and 5 months for CT/MDS. Of the 110 pts <70y with MDS, AZA was given to 50 (45%) and CT to 60 (55%). The IPSS, cytogenetics, CB, BM blasts (10% vs 11.5%) were similar in both groups. With a median age of 63y, the AZA/MDS group was older than the CT/MDS group (median age 60y) (p=0.005). OS for both groups was 68%. NRM (16%) and RI (38% vs 34%) were alike. For the 114 pts <70y with sAML (median age 62%; median blasts 44%) treated with CT, OS was 40% and inferior to MDS (AZA/MDS, p=0.007; CT/MDS, p=0.01) due to higher RI (57%) (p=0.008).

Overall, 218 (78%) pts <70y received HCT (after a median of 3 AZA cycles for AZA/MDS). Ferritin, cytogentics, CB, type of donor, and blasts at HCT (median 4%) were comparable in the transplant groups (AZA/MDS, CT/MDS, HCT up-front, CT/sAML). Irrespective of prior therapy (p=0.6), interval between therapy and HCT, and blasts <5 vs >5-<10%, outcome in the MDS groups (OS, 60%, NRM 29%, RI, 32%) was similar. In multivariate analysis, >2 comorbidities and very poor cytogenetics were associated with an inferior OS (p=0.001)and a higher RI (p=0.003). With a median survival time of 11 months for sAML and a RI of 49%, outcome after HCT for sAML was inferior to MDS (p=0.005). In multivariate analysis, blasts <5%, >2 comorbidities were associated with a poor outcome. For MDS/CT and sAML/CT, a complex karyotype (38%) tended to decrease OS (p=0.06) and increase RI (p=0.01) after HCT.

Treatment was able to reduce the significance of most negative host- and disease-specific prognostic factors on outcome in MDS. AZA is superior to CT in elderly patients and equal to CT in younger patients with MDS and seems to have no negative impact on outcome after HCT. Despite the improvement achieved with allogeneic HCT, AML with MDS-related changes remains a distinct clinic-pathologic entity associated with a worse outcome. Genetics rather than marrow blasts are an important determinant of prognosis after treatment including allogeneic HCT.

No relevant conflicts of interest to declare.