Comparison of scoring systems in primary myelodysplastic syndromes

To compare the prognostic value of scoring systems in primary myelodysplastic syndromes (pMDS), four clinicohematological systems (Rennes, Bournemouth, Düsseldorf, Pavia) and the histopathological Hannover Scoring System were applied to 415 MDS patients from the Bone Marrow Registry of Hannover Medical School. According to the FAB classification, 180 patients (43%) were diagnosed as RA, 33 (8%) as RARS, 99 (24%) as RAEB, 36 (9%) as RAEBt, and 48 (12%) as CMMOL; 19 patients (4%) were not further classified (MDS.UC). All scoring systems revealed three or four groups of patients with significantly different survival times. The ranges and standard deviations in these groups were similar but high in all scoring systems. A good differentiation between short-term survivors (< 1 year survival time) and intermediate-term survivors (1-4 years) was possible with all tested scoring systems, but the differentiation between intermediate- and long-term survivors (> 4 years) was not distinctive enough. The problem of risk assessment in the single patient is furthermore elucidated by the values of specificity and sensitivity, which were relatively low in all scoring systems tested. Best results were yielded by the Bournemouth Score for long-term survivors and the Düsseldorf and Hannover Score for intermediate- and short-term survivors. Multivariate analysis of all parameters used in the scoring systems showed the highest negative impact on survival for increase of myeloblasts, anemia, myelofibrosis, high age of the patient, and abnormal localization of immature precursors (ALIPs). Therefore, the histopathological Hannover Scoring System is not only equal to clinicohematological risk assessment in pMDS, but also includes important independent prognostic parameters. Risk assessment for the individual low-risk MDS patient using only initial parameters may be rendered impossible due to the biological nature of pMDS. Therefore, sequential analysis is needed to elucidate random events which alter the prognosis.