New system for assessing the prognosis of refractory anemia patients

Transcription

1 Leukemia (1999) 13, Stockton Press All rights reserved /99 $ New system for assessing the prognosis of refractory anemia patients A Matsuda 1, I Jinnai 2, F Yagasaki 1, S Kusumoto 1, I Murohashi 1, M Bessho 1, K Hirashima 1, S Honda 3, M Minamihisamatsu 4, K Fuchigami 2, T Matsuo 2, K Kuriyama 2 and M Tomonaga 2 1 First Department of Internal Medicine, Saitama Medical School, Saitama; 2 Department of Hematology, Atomic Bomb Disease Institute, Nagasaki University School of Medicine, Nagasaki; 3 Department of Radiation Epidemiology, Atomic Bomb Disease Institute, Nagasaki University School of Medicine, Nagasaki; 4 Division of Radiobiology and Biodosimetry, National Institute of Radiological Sciences, Chiba, Japan Refractory anemia (RA) is a very heterogeneous disease regarding biological and clinical features. The International Prognostic Scoring System (IPSS) was useful for assessing the prognosis in the whole group of 219 myelodysplastic syndrome (MDS) patients. However, the IPSS was not sufficient in 132 RA patients. To predict survival and freedom from acute myeloid leukemia (AML) evolution, we investigated individual prognostic factors based on the clinical parameters (age, gender, morphologic features, cytopenias and cytogenetics) of 132 RA patients using univariate and multivariate analyses. Based on the results, we devised a new system for assessing the prognosis of RA patients. In our system, RA patients with pseudo- Pelger-Huët anomalies 3% were classified as high risk (12 patients); of patients without pseudo-pelger Huët anomalies 3%, those with intermediate/poor karyotype according to IPSS, Hb 6 g/dl or mmgk 10% were classified as intermediate risk (57 patients); and those without high or intermediate risk were classified as low risk (67 patients). In our system, the analyses of both survival times and leukemia-free survival times revealed significant differences among the three groups (P ). Keywords: MDS; RA; prognosis; pseudo-pelger Huët anomalies; micromegakaryocyte Introduction Myelodysplastic syndromes (MDS) are acquired clonal stem cell disorders characterized by ineffective hematopoiesis with myelodysplasia 1 and are associated with a high risk of progression to acute leukemias. 2 MDS are very heterogeneous diseases regarding their morphology, clinical features and survival. 3 Refractory anemia (RA) is generally classified as a lowrisk group in MDS. 4 Therefore, for the treatment of RA patients, hematopoietic growth factors, nonspecific differentiation inducing agents, steroid and/or supportive therapy are generally chosen. 4 However, RA patients show a wide spectrum of clinical and cytogenetic features, and some RA patients with a poor prognosis need more intensive treatment. To help identify high-risk patients, several scoring systems have been proposed, such as the Bournemouth score (BS), 5 Rochester score, 6 Japanese score (JS) 7 and International Prognostic Scoring System (IPSS). 8 These scoring systems appear to be useful in patients with MDS, which consists of five subgroups, but the usefulness of these scoring systems in the single category of RA patients is not clear. In our previous study, we proposed the new designations RA with severe dysplasia (RASD) and RA with minimal dysplasia (RAminiD). According to our criteria, RASD is considered present if a bone marrow (BM) examination shows pseudo-pelger Huët anomalies of mature neutrophils 3% Correspondence: A Matsuda, First Department of Internal Medicine, Saitama Medical School, 38 Morohongo, Moroyama, Iruma-gun, Saitama, , Japan; Fax: Received 12 January 1999; accepted 13 July 1999 and/or mmgk of megakaryocytes 10% in RA patients. RAminiD is defined as RA cases other than RASD. We observed that the outcome of the RASD patients was unfavorable. 9 Allogeneic bone marrow transplantation (BMT) therapy currently offers the only potentially curative treatment for MDS patients. 3 However, this choice for RA patients is usually difficult because of varied individual prognoses. Therefore, it is necessary to clarify the individual prognoses of RA patients, for it would be possible to devise better therapeutic strategies for RA patients including the option of allogeneic BMT. In the present study, we devised a new system for a more efficient discrimination of RA cases, by employing parameters not only of cytopenias and cytogenetics, but also of morphology such as pseudo-pelger Huët anomalies and mmgk. Patients and methods Patients A total of 219 consecutive patients with the diagnosis of primary RA, RA with ringed sideroblasts (RARS), RA with excess of blasts (RAEB) and RAEB in transformation (RAEB-t) were included in this retrospective analysis. Patients were diagnosed at the Saitama Medical School Hospital, Nagasaki University Hospital or some affiliated hospitals between January 1976 and October Of the patients, 132 were diagnosed as having RA, 15 as RARS, 41 as RAEB and 31 as RAEB-t. Morphological analysis Hematological examinations were performed using standard methods (peripheral blood (PB) and bone marrow (BM) Wright Giemsa-stained films). In our study, the light microscopic examinations were performed excluding the period of infection. For the French American British (FAB) classification 10 and IPSS, PB and BM differential counts were performed on 100 and 500 cells, respectively. For evaluation of mmgk and pseudo-pelger Huët anomalies, a minimum of 25 megakaryocytes and 200 mature neutrophils in BM was examined in each patient. In addition, two especially distinct dysplastic changes were chosen, and their morphological abnormalities (pseudo-pelger Huët anomalies and mmgk) were evaluated. In this study, pseudo-pelger Huët anomalies were defined as hypo-segmented mature neutrophils (Figure 1). mmgk were defined as mono- or bi-nucleated megakaryocytes with a size equal to or 1.5 times smaller than promyelocytes (Figures 2, 3).

2 Prognostic system in RA 1728 Figure 1 do not. Photomicrographs of bone marrow smears (Wright-Giemsa stain, 1000). (a e) show pseudo-pelger Huët anomalies, but (f h) affiliated hospital until death due to any cause or until the last patient contact, as of November Leukemic transformation was measured from diagnosis. The clinical variables are shown in Table 1. The analyses of survival curve and leukemic transformation were done using the log rank test. A univariate analysis of each clinical variable was analyzed using the Kaplan Meier method and log rank test. A multivariate analysis of clinical variables was performed using proportional hazards model regression analysis. We used stepwise procedure for the purpose of selecting variables for the model. The significance level for entering an explanatory variable into the model and for removing from the model was Figure 2 Photomicrographs of bone marrow smears (Wright- Giemsa stain, 1000). Mononuclear micromegakaryocytes are shown in (a d). (e) shows a mononuclear megakaryocyte the size of which is 1.5 times larger than a promyelocyte. Scoring system The IPSS score of each patient was calculated based on the hematologic data obtained at the time of each patient s initial visit to our hospitals. Cytogenetic analysis Cytogenetic analyses were performed with a trypsin-giemsa banding technique on BM cells from aspirates. Cells from short-term unstimulated cultures ( 48 h) were examined. Chromosomal abnormalities were identified by the International System for Human Cytogenetic Nomenclature. 11 Statistical analysis The Kaplan Meier method was used to estimate the probability of overall survival and leukemic transformation. Survival was measured from the date of diagnosis at Saitama Medical School Hospital, Nagasaki University Hospital or the Results Follow-up periods ranged from 1 to 250 months (median 27 months) in all 219 MDS patients. Follow-up periods of the 132 RA patients ranged from 1 to 212 months (median 46 months). All of the 219 MDS patients were reviewed according to the IPSS criteria. Fifty-eight cases (RA, RARS, RAEB, RAEB-t) could not be classified because cytogenetic data were not obtained. After reclassification, they were composed of 26 Low, 83 Int-1, 23 Int-2 and 29 High patients. In RA patients, the reclassification according to IPSS was 22 Low, 73 Int-1 and Int-2 patients. The analysis of survival curves revealed a significant difference between the RA/RARS patients vs the RAEB/RAEB-t patients (Figure 4a). The IPSS were useful for assessing the prognosis, but the analysis of survival curves did not reveal a significant difference between Low vs Int-1 (Figure 4b). In the RA patients only, IPSS were not sufficient for assessing the prognosis (Figure 4c). In the univariate analysis of RA patients, age 60 years old (P ), pseudo-pelger Huët anomalies of mature neutrophils 3% (P ), mmgk of megakaryocytes 10% (P = ), hemoglobin concentration 6 g/dl (P = ) and chromosome (intermediate/poor karyotype according to IPSS) (P = ) were significantly correlated with overall survival. And pseudo-pelger Huët anomalies of mature neutrophils 3% (P ), mmgk of megakaryocytes 10% (P = ) and chromosome (poor karyotype according to IPSS) (P = ) were significantly correlated with leukemia-

3 Prognostic system in RA 1729 Figure 3 Photomicrographs of bone marrow smears (Wright-Giemsa stain, 1000). Binuclear micromegakaryocytes are shown in (a d), but (e) is not a micromegakaryocyte. Figure 4 Cumulative survival of MDS patients. (a) FAB classification subgroups (RA/RARS vs RAEB/RAEB-t: P ); (b) IPSS classification subgroup (Low/INT-1 vs INT-2/High: P , Low vs INT-1: P = , INT-1 vs INT-2: P , INT-2: vs High: P = ). Cumulative survival of RA patients. (c) IPSS classification subgroup (P = ).

4 1730 Table 1 Prognostic system in RA Univariate analysis of overall and leukemia-free survival in refractory anemia patients Variable No. of Percentile of Overall survival Percentile of Leukemia-free patients survival (months) P value AML (months) survival 25% P value 75% 50% Age 60 y.o NR Gender Male NR NR Female NR ANC /l NR /l NR NR Hb 10 g/dl NR g/dl NR NR Hb 6 g/dl NR g/dl NR NR 6 g/dl NR g/dl 56 NR NR NR 6 g/dl g/dl NR PLT /l NR /l Cytopenias accoding to IPSS 0/ NR NR / NR mmgk 10% % NR 10% % NR NR 10% % Pelger 3% % NR 3% % NR 3% NR % No. of abnormal lineages NR NR NR NR NR NR NR NR NR Continued

5 Prognostic system in RA Table 1 Continued 1731 Variable No. of Percentile of Overall survival Percentile of Leukemia-free patients survival (months) P value AML (months) survival 25% P value 75% 50% Chromosome Good NR Intermediate NR NR Poor 8 27 NR 22 Good NR NR Intermediate 8 23 NR NR Poor Good Intermediate NR Poor 5 5 NR 4 Chromosome Good NR Intermediate/Poor NR 37 Good NR NR Intermediate/Poor NR 37 Good Intermediate/Poor NR IPSS LOW NR NR INT NR INT NR 22 LOW NR NR INT NR INT LOW NR NR INT INT NR 4 Pelger, pseudo-pelger Huët anomalies; Hb, hemoglobin level; mmgk, micromegakaryocytes; PLT, platelet count; ANC, absolute neutrophil count; BM, bone marrow; IPSS, International Prognostic Scoring System; NR, not reached. No. of abnormal lineages: number of Hb 6 g/dl, pelger 3% or mmgk 10%. Good: normal, del(5q) only, del(20q) only, Y only. Intermediate: +8, single miscellaneous, double abnormalities. Poor: complex (ie, 3 anomalies) or chromosome 7 abnormalities. Table 2 Multivariate analysis by stepwise procedure of overall and leukemia-free survival in refractory anemia patients No. of Variable P value patients Overall survival including chromosome data 103 Pelger 3% Age chromosome (Intermediate/Poor) Overall survival not including chromosome data 132 Pelger 3% Age Hb 6 g/dl Leukemia-free survival including chromosome data 103 Pelger 3% chromosome (Intermediate/Poor) Leukemia free-survival not including chromosome data 132 mmgk 10% Pelger 3% The variables were selected by the stepwise procedure. The significance level for entering or removing a variable was Abbreviations are explained in Table 1.

6 Prognostic system in RA 1732 Figure 5 New system for assessing the prognosis of RA patients. Pelger, pseudo-pelger Huët anomalies; Hb, hemoglobin level; mmgk, micromegakaryocytes. Figure 6 (a) Cumulative survival of RA patients according to our new system subgroups: P (b) Leukemia-free survival of RA patients according to our new system subgroups: P Table 3 Causes of death in RA patients Type Bleeding Infection Acute leukemia Heart failure Solid malignancy Unknown Total (%) (%) (%) (%) (%) (%) (%) High risk 2 (22) 4 (44) 3 (33) 0 (0) 0 (0) 0 (0) 9 (100) Intermediate risk 6 (25) 8 (33) 5 (21) 3 (13) 0 (0) 2 (8) 24 (100) Low risk 1 (11) 4 (44) 2 (22) 0 (0) 2 (22) 0 (0) 9 (100) Table 4 Cytogenetic findings in MDS patients FAB No. of Cytogenetic subgroups (IPSS) (%) studied patients Good Intermediate Poor RA (78) 15 (15) 8 (8) RARS 10 6 (60) 1 (10) 3 (30) RAEB 25 7 (28) 8 (32) 10 (40) RAEB-t (56) 2 (11) 6 (33) Abbreviations are explained in Table 1. free survival (Table 1). The analysis of survival curves did not reveal a significant difference between the RA patients with poor karyotype and those with intermediate karyotype according to IPSS. To evaluate the effect of prognostic factors on overall survival and freedom from AML evolution, we used a multivariate proportional hazards model regression analysis. The clinical variables of age 60 years, gender, absolute neutrophil count (ANC) /l, hemoglobin (Hb) concentration 6 g/dl, platelet count /l, mmgk 10%, pseudo-pelger Huët anomalies 3% and chromosome data were included in the mode. An analysis not including chromosome data was also done on the assumption that those data are not always available. For survival, the significant variables

7 were pseudo-pelger Huët anomalies 3% (P ), age 60 years (P = ) and chromosome (intermediate/poor karyotype according to IPSS) (P = ) in the analysis including chromosome data, and pseudo-pelger Huët anomalies 3% (P ), age 60 years (P = ) and Hb concentration 6 g/dl (P = ) in that not including chromosome data (Table 2). For freedom from AML evolution, pseudo-pelger Huët anomalies 3% (P = ) and karyotype (intermediate/poor karyotype according to IPSS) (P = ) were significant in the analysis including chromosome data, and mmgk 10% (P = ) and pseudo- Pelger Huët anomalies 3% (P = ) in that not including chromosome data (Table 2). Based on these findings, we devised a new system for assessing the prognosis of RA patients (Figure 5). Our new system is also useful for assessing the prognosis of RA patients whose chromosome data are not available. In our system, RA patients with pseudo-pelger Huët anomalies 3% were classified as high risk. Of the patients without pseudo-pelger Huët anomalies 3%, those with intermediate/poor karyotype according to IPSS were classified as intermediate risk. Of the patients with good karyotype according to IPSS or whose data of cytogenetic analysis were not available, those with Hb 6 g/dl or mmgk 10% were also classified as intermediate risk. Those without high or intermediate risk were classified as low risk. In this system, the analysis of survival curves revealed significant differences among the three groups; the intermediate risk group (57 cases) had lower survival probabilities than those of the low risk group (63 cases), and higher probabilities than those of the high risk group (12 cases) (P ) (Figure 6a). Concerning the causes of death in the RA patients, none of the high risk patients died due to non-hematological causes (Table 3). Thirty-three of 120 RA patients who were initially classified as having intermediate or low risk according to our system died. In the follow-up period, 28 of these 33 patients became red blood cell (RBC) transfusion dependent. Fourteen of 33 patients initially showed a Hb concentration 6 g/dl, and 11 of 19 patients that initially showed a Hb concentration 6 g/dl subsequently fell to 6 g/dl. Seven of the 33 nonsurviving intermediate/low risk patients showed pseudo- Pelger Huët anomalies 3% in subsequent BM aspirations, and four of these patients developed acute leukemia. The leukemia-free survival (LFS) time calculation revealed significant differences among the three groups in our system; the intermediate risk patients had lower survival probabilities than those of the low risk group, and higher probabilities than those of the high risk group (P ) (Figure 6b). Discussion There have been several reports of scoring systems for predicting the prognosis of patients with MDS. Most of them have adopted cytopenias as a prognostic factor. However, Cunningham et al 12 reported that neutropenia and thrombocytopenia were not associated with infective and bleeding episodes. Our study also showed that the initial degrees of neutropenia and thrombocytopenia were not related to survival. These results suggest that the risk of bleeding and infection in RA patients may be associated with not only cytopenias but also dysfunctions of neutrophil and platelet, and the dysfunctions may be a more significant risk factor than cytopenias. It has been reported that the incidence of chromosomal Prognostic system in RA aberrations varies among FAB types of MDS, and is lower in RA and RARS than in RAEB, RAEB-t, and CMML. 13 In addition, the distribution of the types of chromosomal abnormalities is different among the subgroups according to FAB classification. Chromosomal abnormalities associated with good prognosis such as isolated 5q or 20q (no adverse point in cytogenetic scoring of IPSS), predominate in RA and RARS whereas others (associated with poor prognosis), such as 7/7q and complex abnormalities, predominate in RAEB and RAEB-t. 13 In fact, our present series revealed similar results (Table 4). In our study, the chromosomal findings were a significant prognostic factor according to the multivariate analysis. However, the analysis of survival curves did not reveal a significant difference between the RA patients with poor karyotype and those with intermediate karyotype according to the IPSS. This result may be due to the low frequency of RA patients with poor karyotype. However, there is no denying the possibility that the impact of cytogenetic analysis on prognosis might be less apparent in RA patients only compared to MDS as a whole. The Rochester score 6 differs in MDS patients as it is based on the degree of granulocytic and megakaryocytic dysplasia. The clinical use of this score is difficult because of varied individual interpretation of dysplastic changes in hematopoietic cells. 4 Gattermann et al 14 suggested that, in the setting of RARS, patients with dysplastic granulocytic and/or megakaryocytic elements were at higher risk of developing acute leukemia than were those who lacked these findings. Rosati et al 15 indicated that refractory cytopenia with multilineage dysplasia (RCMD) is a distinct subset of MDS, with an unfavorable clinical outcome. We also believe that the degree of myelodysplasia reflects the prognosis in MDS patients including those with RA and RARS. However, the individual interpretation of myelodysplasia is variable. In addition, there are low frequencies of dysplasia in the BM of healthy subjects 16 and patients with liver disease 17 or viral infection, 18 for example, and myelodysplastic features except pseudo-pelger Huët anomalies, complete degranular neutrophils and mmgk are not able to distinguish MDS patients from healthy subjects. 19 We reported that a high concordance rate between the observers was found regarding pseudo-pelger Huët anomalies and mmgk, and that they were correlated with the length of survival. 9 There are several recent reports of a relation between the abnormality of some genes and myelodysplasia. Soenen et al 20 reported the very strong relation in MDS between 17p deletion, the presence of pseudo-pelger Huët anomalies, and p53 mutation. In our study, none of the RA patients had 17p deletion, and an analysis on p53 mutations was not performed. Bouscary et al 21 reported that the c-mpl expression was significantly correlated with poor survival and the presence of dysmegakaryopoiesis in MDS patients. Our study revealed myelodysplasia in RA patients had a prognostic impact. It appears that genetic or some other abnormality(ies) which cause myelodysplasia might relate to prognosis. It is important to clarify the mechanism of this result biologically. The median age of RA patients in Saitama Medical School Hospital and Nagasaki University Hospital were 56 and 57 years, respectively. It has been reported that the median age of Japanese MDS patients is 60 years, about 10 years younger than that in most reports of Western MDS patients. 22 The median age of the MDS patients in Thailand and the mean age of those in Central Africa were reported to be and 57 years, 24 respectively. On the other hand, in the IPSS study, 8 the median age of the MDS patients was 69 years. There may be a different age distribution in Asian/African and Western 1733

8 1734 Prognostic system in RA MDS patients. The IPSS study and our present investigation showed that age ( 60) was a significant poor prognostic factor. Our RA patients demonstrated favorable outcomes compared to those of the IPSS study. We consider this result to be associated with a difference between the median age of Japanese and that of Western MDS patients. In this retrospective study, consecutive examinations at regular periods were not performed in every RA patient. Therefore, an optimal follow-up study of our RA patients was not done regarding prognosis. However, most of the 33 nonsurviving RA patients without pseudo-pelger Huët anomalies 3% in initial BM aspirations became red blood cell transfusion dependent. Seven of these 33 non-surviving patients had pseudo-pelger Huët anomalies 3% in subsequent BM aspirations, and four of these seven patients developed acute leukemia. These findings show that RBC transfusion dependence and pseudo-pelger Huët anomalies 3% in the followup periods may be related to poor prognosis. A future planned consecutive study will clarify this issue. Our system is useful for assessing the prognosis of RA patients. Our data indicate that high risk patients in our system have particularly poor prognoses. We propose that our new system can be used to devise better therapeutic strategies for RA patients including the option of allogeneic BMT. References 1 Goasguen JE, Bennett JM. Classification and morphologic features of the myelodysplastic syndromes. Semin Oncol 1992; 19: Ganser A, Hoelzer D. Clinical course of myelodysplastic syndromes. Hematol Oncol Clin N Am 1992; 6: Koeffler HP. Introduction: myelodysplastic syndromes. Semin Hematol 1996; 33: Hofmann WK, Ottmann OG, Ganser A, Hoelzer D. Myelodysplastic syndromes: clinical features. Semin Hematol 1996; 33: Mufti GJ, Stevens JR, Oscier DG, Hamblin TJ, Machin D. Myelodysplastic syndromes: a scoring system with prognostic significance. Br J Haematol 1985; 59: Varela BL, Chuang C, Woll JE, Bennet JM. Modifications in the classification of primary myelodysplastic syndromes: the addition of a scoring system. Haematol Oncol 1985; 3: Toyama K, Ohyashiki K, Yoshida Y, Abe T, Asano S, Hirai H, Hirashima K, Hotta T, Kuramoto A, Kuriya S, Miyazaki T, Kakishita E, Mizoguchi H, Okada M, Shirakawa S, Takaku F, Tomonaga M, Uchino H, Yasunaga K, Nomura T. Clinical implications of chromosomal abnormalities in 401 patients with myelodysplastic syndromes: a multicentric study in Japan. Leukemia 1993; 7: Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G, Sanz M, Vallespi T, Hamblin T, Oscier D, Ohyashiki K, Toyama K, Aul C, Mufti G, Bennett J. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 1997; 89: Matsuda A, Jinnai I, Yagasaki F, Kusumoto S, Minamihisamatsu M, Honda S, Murohashi I, Bessho M, Hirashima K. Refractory anemia with severe dysplasia: clinical significance of morphological features in refractory anemia. Leukemia 1998; 12: Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnic HR, Sultan C. Proposal for the classification of the myelodysplastic syndromes. Br J Haematol 1982; 51: ISCN (1995). Mitelman F (ed). An International System for Human Cytogenetic Nomenclature. Karger: Basel, Cunningham I, MacCallum SJ, Nicholls MD, Byth K, Hewson JW, Arnold B, Motum PI, Mulligan SP, Crane GG. The myelodysplastic syndromes: an analysis of prognostic factors in 226 cases from a single institution. Br J Haematol 1995; 90: Fenaux P, Morel P, Lai JL. Cytogenetics of myelodysplastic syndromes. Semin Hematol 1996; 33: Gattermann N, Aul C, Schneider W. Two types of acquired sideroblastic anemia (AISA). Br J Haematol 1990; 74: Rosati S, Mick R, Xu F, Stonys E, Le Beau MM, Larson R, Vardiman JW. Refractory cytopenia with multilineage dysplasia: further characterization of an unclassifiable myelodysplastic syndrome. Leukemia 1996; 10: Bain BJ. The bone marrow aspirate of healthy subjects. Br J Haematol 1996; 94: Hadnagy C, Laszlo GA. Acquired dyserythropoiesis in liver disease. Br J Haematol 1991; 78: Karcher DS, Frost AR. The bone marrow in human immunodeficiency virus (HIV)-related disease. Morphology and clinical correlation. Am J Clin Pathol 1991; 95: Kuriyama K, Tomonaga M, Matsuo T, Ginnai I, Ichimaru M. Diagnostic significance of detecting pseudo-pelger Huët anomalies and micro-megakaryocytes in myelodysplastic syndrome. Br J Haematol 1986; 63: Soenen V, Preudhomme C, Roumier C, Daudignon A, Lai JL, Fenaux P. 17p deletion in acute myeloid leukemia and myelodysplastic syndrome. Analysis of breakpoints and deleted segments by fluorescence in situ. Blood 1998; 91: Bouscary D, Preudhomme C, Ribrag V, Melle J, Viguie F, Picard F, Guesnu M, Fenaux P, Gisselbrecht S, Dreyfus F. Prognostic value of c-mpl expression in myelodysplastic syndromes. Leukemia 1995; 9: Oguma S, Yoshida Y, Uchino H, Maekawa T, Nomura T, Mizoguchi H, the Refractory Anemia Study Group of the Ministry of Health anf Welfare. Clinical characteristics of Japanese patients with primary myelodysplastic syndromes: a co-operative study based on 838 cases. Leukemia Res 1995; 19: Intragumtornchai T, Prayoonwiwat W, Swasdikul D, Suwanwela N, Chaimongkol B, Jootar S, Chansung K, Chancharunee S, Leelasiri A, Yoshida Y. Myelodysplastic syndrome in Thailand: a retrospective pathologic and clinical analysis of 117 cases. Leukemia Res 1997; 21 (Suppl. 1): S5. 24 Mukiibi JM, Paul B. Myelodysplastic syndrome (MDS) in Central Africans. Trop Geogr Med 1994; 46: