Prognostic factors in myelodysplastic syndromes: analysis of five scoring systems

To evaluate the prognostic significance of five scoring systems applied with predictive trends to myelodysplastic syndromes (MDS).

This study was comprised of 226 patients with MDS diagnosed in accordance with the FAB criteria and followed up in our department between January 1975 and April 1992. The following MDS subtypes were found: refractory anaemia (RA), 59 cases; refractory sideroblastic anaemia (RSA), 49 cases; refractory anaemia with excess of blasts (RAEB), 56 cases; RAEB in transformation (RAEB-T) 48 cases; and chronic myelomonocytic leukemia (CMML), 14 cases. The following scoring systems were applied: Mufti's 1985, Varela's 1985, Sanz's 1989, Rubio's 1991 and Aul's 1992. The statistical analysis was performed according to Kaplan-Meier actuarial systems and the log-rank test of survival.

(1) Three groups (A, B and C) can be defined by Mufti's system, with median survival of 54.0, 16.0 and 8.5 months, respectively. The majority of cases (138) were included in group B. Group A did not reach 25 per cent of actuarial survival probability, whereas groups B and C did at 31.1 and 12.2 months, respectively. With regard to the morphologic subtypes, RA and RSA were included in groups A and B, and RAEB, RAEB-T and CMML pertained mostly to group C. Sixty-six cases (33.6 per cent) developed into acute leukemia (AL) corresponding to those last groups. (2) The three groups defined by Varela's system (0-1, 2-5 and 6 or more) had median survival of 91.8, 24 and 13 months, respectively. As in the former system, group 0-1 did not reach 25 per cent actuarial probability, this appearing at 60 and 20 months, respectively, in groups 2-5 and > 6. The distribution of the cytological varieties, RA and RSA among the groups is heterogenous although they were more common within the cases included in groups 0-1. All cases developing AL were included in the groups 2-5 and > 6. (3) The three groups of the system proposed by Sanz (0-1, 2-3 and 4-5) had median survival of, 55.3, 15 and 12.6 months respectively. As in the preceding cases, group 0-1 did not reach the 25 per cent actuarial probability, while this figure appeared at 28.2 months for group 2-3 and at 19.3 months for group 4-5. RA and RSA varieties were included chiefly in group 0-1, while RAEB and RAEB-T appear mostly in groups 2-3 and 4-5. The distribution of the cases and the evolution of AL was heterogeneous according to this system, although they predominate in groups 2-3 and 4-5. (4) Using the Aul's system, three groups A, B and C were defined. The median survival time was 14 months for group C and 24 months for group B. For group A, the median survival was not reached. RA and RSA were exclusive for group A, while RAEB and RAEB-T varieties were outstanding in group C. Regarding the evolution to leukemia the differences observed had no statistical relevance. (5) Three prognostic groups were defined by Rubio's system (namely 0-2.5, 3-5.5, > or = 6) with median survival of 53.3, 16.8 and 10.5 months, respectively. A striking difference was seen when studying the cumulated survival observed, in each of the three percentages considered, between the groups. The different cytological varieties were reasonably distributed with higher incidence of RA and RSA in group I and RAEB, RAEB-T and CMML in group III. This system offers statistical significance when comparing RA with RSA, RAEB with RAEB-T and, obviously RA+RSA with RAEB+RAEB-T+CMML. The evolution into AL also showed statistical significance with respect to the three groups.