Prognostic Factors in Treatment-Related Myelodysplastic Syndromes (t-MDS) and Acute Myeloid Leukemia (t-AML).

In the current WHO classification patients (pts) with therapy-associated MDS are merged into a subgroup of AML, termed “AML with MDS, therapy-related”, regardless of dysplastic features, medullary blast count or cytogenetics. To validate the appropriateness of this approach, we combined the MDS/AML data sets of the Duesseldorf as well as the Goettingen data base. We identified 305 pts with t-MDS or t-AML. There were 138 males and 167 females, median age was 64 (22–90). When classified according to the WHO proposals for primary MDS (pMDS), 38% of the pts had AML (28% blast count >30%, 10% blast count 20–30% (RAEB-T)), 10% had RAEB II, 8% RAEB I, 2% CMML II, 3% CMML I, 6% RA, 1.5% RARS, 20% RCMD, 11% RCMD-RS and 0.5% had 5q- Syndrome. 60% of pts had received chemotherapy only, 12% radiotherapy, and 28% underwent combined radiochemotherapy prior to t-MDS/AML diagnosis. The median latency between treatment and diagnosis was 116 months in pts who received radiotherapy as compared to 81 months in pts treated with chemotherapy alone or in combination with radiation (p= 0.0005). Cytogenetics were available in 192 pts (63%). 58% had an abnormal karyotype, the median number of aberrant chromosomes was 2 (0–12). The most frequent aberrant chromosome was chromosome 5 (18%), followed by 7 (14%), 20 (7%) and 17 (7%). 48% of the pts had a low-risk, 18% an intermediate-risk, and 33% a high-risk karyotype according to IPSS. Median survival of the entire group was 10 months compared to 26 months in pMDS in our registry (p=0.00005). Important prognostic parameters for pMDS, like age, hemoglobin level, platelet count, neutrophile count, as well as classification and scoring systems (FAB, WHO, IPSS) could not show an impact on outcome in t-MDS/AML. Pts with <5% medullary blasts had a median survival of 16 months, with 5–9% of 8 months, with 10–19% of 8 months, with 20–29% of 6 months and pts with a blast count >30% had a median survival of 12months (p=n.s.). When we compared pts with a medullary blast count <5 vs ≥5 months the survival difference (15 vs 7 months) became significant (p=0,013). Pts with an elevated LDH level had a median survival of 7 as compared to 17 months in pts with normal LDH (p=0.0006). The karyotype subgroups (IPSS) showed significant survival differences between each of the groups: low-risk -34 months, intermediate-risk -25 months, high-risk -8 months. In a multivariate analysis only karyotype (p=0,011) and LDH level (p=0,014) remained independent prognostic parameters (blast count: p=0,961). These two variables can be combined in a score: LDH elevated= 1 point, high-risk karyotype= 1 point. This score leads to 3 subgroups (0, 1, or 2 points) with a significantly different prognosis of 42, 15, and 7 months (p=0,01). Our analysis shows that there is no substantial difference in survival between t-MDS and t-AML, justifying the combination of both entities in one AML subgroup. Current morphology based classification systems offer no prognostic information. On the other hand the established karyotype classification according to IPSS remains relevant for prognosis and combining this information with the LDH level leads to a simple score that is able to separate prognostically different risk-groups and thus should be validated further. 0%