A New Prognostic Disease Specific Model To Predict Survival after Intensive Antileukemic Treatment for Young Patients with Poor-Risk MDS and AML: Results of the CRIANT and AML-10 Studies Conducted by the EORTC/GIMEMA/SAKK/HOVON/EBMT Groups.

The use of intensive antileukemic treatment is less widely accepted in high-risk MDS pts compared to de novo AML, due to the reported inferior results. It is questionable whether the poorer outcome reflects an intrinsic property of the involved stem cell or a higher frequency of poor prognostic factors. The purpose of this analysis is to identify disease-specific prognostic factors for outcome of young (aged <56 years) MDS and AML pts. This analysis combines the data of 591 pts in the AML-10 study and 203 pts with high-risk MDS or secondary AML (sAML) in the CRIANT study. Both groups received identical remission-induction therapy (idarubicin, cytarabine and etoposide), followed by one consolidation course with intermediate dose of cytarabine and idarubicin (IDIA). In both studies post-consolidation therapy consisted of alloSCT if an HLA-identical sibling donor was available. The remaining pts received autoSCT (ASCT) in AML-10 or were randomized between ASCT and a 2nd consolidation course in CRIANT. The CR rate was 68% (AML-10) vs. 59% (CRIANT) (p=0.02). The 4-year survival rates were 35% vs. 33% (p=0.80). DFS at 4 years was 43% (AML-10) vs. 35% (CRIANT) (p=0.18). For overall survival (OS) in both studies, study was not of importance (HR=1.09, p=0.45), but the following variables showed independent prognostic value: cytogenetic risk group (the HR for poor vs intermediate risk was 1.68, 95% CI 1.24–2.27, p=0.0008), white blood count (WBC) ≥ 100 x 10⁹/l (HR=2.02, 95% CI 1.53–2.68, p<0.0001), age 46–55 yrs (HR=1.39, 95% CI 1.16–1.67, p=0.0004) and performance status (PS) (HR=1.32, 95% CI 1.17–1.49, p <0.0001). For DFS, the following factors were of an independent prognostic importance: cytogenetics (p<0.0001), age 46–55 (HR=1.23, p=0.05), WBC >100 (HR=1.67, p=0.02) and donor availability (HR=0.77, p=0.04). Some variables were of prognostic value for OS in only one of the studies: in the CRIANT study number of cytopenias (3 vs 0–2) and AHD >6 months appeared of prognostic importance for OS, wherease FAB subtype M2/M4 and cytogenetics inv(16)/t(8;21) were prognostic in AML-10. Therefore a specific prognostic score for OS was established for each study, AML-10 (based on cytogenetics, PS, FAB, WBC and age) and CRIANT (based on cytogenetics, nr of cytopenias, age, AHD and WBC). The AML-10 study distinguished 5 groups with an estimated 4-year survival rate of 69%, 40%, 45%, 26% and 17%, resp. The prognostic value of this score has been validated on patients treated in the AML-10 study with mitoxantrone instead of idarubicin: the 4-year survival were 76%, 46%, 41%, 33% and 18%, resp. The CRIANT study distinguished 5 groups with a 4-year survival rates of 72%, 44%, 39%, 12% and 0%, resp. In conclusion: the prognostic scores identify a group of 26% AML and 42% MDS pts, with a 4-year survival less than 20%. Apparently current treatment modalities are unsatisfactory for these poor-risk pts and novel treatment strategies should be offered to these pts in the context of clinical trials. Our finding that different variables are of prognostic importance in MDS/sAML and de novo AML pts supports the hypothesis that these are intrinsically different disorders. The CRIANT-derived score is a valuable alternative for the IPSS in intensively treated high-risk MDS pts.