- Email: [email protected]
Clinical features of prognostic significance in myelodysplastic patients with normal karyotype at high risk of transformation
Leukemia Research 29 (2005) 33–39
Clinical features of prognostic significance in myelodysplastic patients with normal karyotype at high risk of transformation Massimo Breccia∗ , Marco Mancini, Mauro Nanni, Gianna Maria D’Elia, Ida Carmosino, Roberto Latagliata, Chiara Sarlo, Franco Mandelli, Giuliana Alimena Dipartimento di Biotecnologie Cellulari ed Ematologia, Università La Sapienza, Via Benevento 6, 00161 Rome, Italy Received 3 May 2004; accepted 10 May 2004 Available online 31 July 2004
Abstract The International Prognostic Scoring System (IPSS) for myelodysplastic syndromes (MDS) has defined patients with a normal karyotype as a good risk cytogenetic subgroup, but nevertheless a fraction of these patients has a poor outcome similar to that of high risk patients. We retrospectively analysed our series of myelodysplastic patients with normal karyotype observed in a period of 11 years, with the aim of identifying clinical features of possible prognostic significance within this subgroup of patients. Multivariate analysis showed that among clinical scoring systems, the Bournemouth score appears the best prognostic indicator for risk of leukemic transformation, and platelet count <100 × 109 /l−1 , presence of haemorrhagic symptoms at time of diagnosis and morphologic FAB classification are the main prognostic factors for prediction of survival. In the absence of genetic abnormalities as detected by conventional cytogenetics or even the more sensitive molecular techniques in MDS, clinical variables could be of help in identifying patients with different prognosis, suitable for risk adapted therapeutic strategies. © 2004 Elsevier Ltd. All rights reserved. Keywords: Myelodysplastic syndromes; Normal karyotype; Prognostic factors
1. Introduction Myelodysplastic syndromes (MDS) are a group of heterogeneous clonal haematopoietic stem cell disorders characterized by dysfunctional haematopoiesis leading to bone marrow failure and high risk of progression to acute leukaemia. Through multivariate analyses of large series of patients, several studies have shown that main prognostic factors in these disorders are percentage of bone marrow blasts, number and degree of cytopenias, age >65 years and presence of chromosomal abnormalities [1–3]. Based on these features, different scoring systems for prediction of outcome and risk of transformation to acute myeloid leukaemia (AML) have been created during the last two decades [4–9], the most used being the International Prognostic Scoring System (IPSS) that has defined four prognostic subgroups of MDS patients [10]. ∗ Corresponding author. Tel.: +39 06 85 79 51; fax: +39 06 44 24 19 84. E-mail address: [email protected] (M. Breccia).
0145-2126/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.leukres.2004.05.010
The IPSS evidences a normal karyotype at diagnosis as detected by conventional cytogenetics, in 40–65% of all MDS patients included in low risk group [10]. Nevertheless, this category of patients is very heterogeneous for the wide variability of clinical parameters and the outcome is often unpredictable. In present study, we retrospectively evaluated 98 consecutive MDS patients with normal karyotype at CCA and analysed clinical features as detected at time of diagnosis, in order to establish their possible prognostic impact on survival and risk of evolution to acute leukaemia.
2. Patients and methods 2.1. Patients Between January 1986 and December 1997, 98 patients were diagnosed at our Institution as having a myelodysplastic syndrome with a normal karyotype; this figure represents 58% of the whole MDS patients diagnosed over the same period of time. We included in the study only patients without
34
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
history of previous cytotoxic, chemo- or radiation exposure, and excluded patients with secondary or therapy-related MDS, according to WHO criteria [11]. Morphological examinations were performed using standard methods on peripheral blood and bone marrow Wright–Giemsa-stained films. Iron staining and non-specific esterase were used for the diagnosis of refractory anemia with ringed sideroblasts and chronic myelomonocytic leukaemia. Classification was performed according to FAB criteria [4]; we used FAB and not WHO revisited criteria because in the analysis were included patients with blasts >20% (RAEB-t), that in the past were considered MDS and not AML. Bone marrow aspirate was obtained in all cases at diagnosis and dysplastic features were recognized as previously indicated [4]. Cytogenetic analyses were conducted at the time of diagnosis by GTG banding technique on bone marrow cells obtained from short-term unstimolated cultures (24–48 h). Chromosomal abnormalities were identified by the International System for Human Cytogenetic Nomenclature (1995). At least 20 metaphases were analysed in all cases. Treatment of MDS phase was based on only supportive therapy in all patients. Only five patients received AML-like therapy after evolution. 2.2. Analysis of prognostic factors The following characteristics at time of diagnosis were analysed to establish the association with survival and risk of transformation to acute leukaemia: age, sex, dysplastic lineage involvement, haemoglobin, white blood cell (WBC), platelet and neutrophils counts, percentage of blast cells in the bone marrow, presence of transfusional requirement, presence of hemorrhagic symptoms and/or infections. The prognostic impacts of FAB classification and of three scoring systems were also evaluated. The first two systems, the Bournemouth and the Spanish scoring, are based only on clinical parameters while the International Prognostic Scoring Systems introduces also the karyotype [7,10,12]. According to this last system the stratification of patients in low, intermediate-1, intermediate-2 and high risk was only applied on the basis of clinical features, because for the normal karyotype the attributed score value is 0, i.e. considered as of “good prognosis”. 2.3. Statistical analysis Univariate analysis for each prognostic variable on overall survival and time to disease evolution was estimated according to the Kaplan–Meier method. The terminal event was considered the death attributable to cancer or non-cancer causes. The statistical significance of the differences in survival distribution among the prognostic groups was evaluated by the log-rank test. The Cox proportional hazards model was applied to the multivariate survival analysis.
The prognostic variables for overall survival and time to leukemic evolution included: sex, age, FAB classification, Bournemouth score, Sanz score, IPSS score, lineages of bone marrow involvement, transfusions requirement, infectious and/or hemorrhagic complications, hemoglobin levels, WBC, neutrophils and platelets counts at diagnosis. P-values <0.05 were regarded as statistically significant in two tailed tests. SPSS software (version 10.00, SPSS, Chicago) was used for statistical analysis [14–16]. 3. Results 3.1. Clinical data There were 59 males (60%) and 39 females (40%). Median age was 63.8 years (range 23–82.8). According to FAB classification there were 29 patients with RA (29%), 11 patients with RARS (11.2%), 26 patients with RAEB (26.5%), 8 patients with RAEB-t (8%) and 24 patients with CMML (24.4%). In the peripheral blood, 47 patients had haemoglobin values <10 g/dl, 36 patients had neutrophil counts <2.5 × 109 l−1 , 31 patients had platelet counts <100 × 109 l−1 . Morphological analysis revealed a bone marrow percentages of blasts <5% in 40 patients (41%), 5–10% in 38 patients (38.7%), 11–20% in 12 patients (12%) and >20% in 8 patients (8%). In 51 patients (52%) there was a trilinear bone marrow dysplasia. Sixty patients (61.2%) required transfusional support; 44 patients (44.8%) had an infection and 13 patients (13.2%) presented one or more hemorrhagic events at the time of diagnosis. Median survival of all 98 patients was 28.9 months. These data compared to a median survival of 21.7 months in the whole MDS population, observed in the same period. By the end of follow-up (60 months), 41/98 patients (41.8%) have died, acute leukaemia being the main cause of death, with a transformation rate of 32% (13 patients); in 4 patients (9.8%) death was not related to the myelodysplastic disease and was mainly due to cardiovascular failure or second neoplasias. 3.2. Analysis of prognostic factors for survival and leukemic transformation The results of the univariate analysis of prognostic factors for survival are summarized in Table 1. Male sex (P = 0.03), age >65 years (P = 0.03), FAB category with high percentage of blasts (P = 0.0002) and presence of trilineage involvement (P = 0.03) were strictly associated with a shorter survival. At univariate analysis, also significant were haemorrhagic symptoms at diagnosis (P = 0.002), haemoglobin level <10 g/dl (P = 0.03),
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
platelet count <100 × 109 l−1 (P = 0.003) and WBC count <3.5 × 109 l−1 (P = 0.02). The subdivision of patients in risk subgroups according to different scoring systems (Bournemouth, Spanish, IPSS) was significative for survival only when applying the Bournemouth score (P = 0.006). The considered factors were entered to multivariTable 1 Univariate Analysis of overal survival
35
Table 2 Multivariate cox regression analysis of overall survival Relative risk of death
95% confidential interval
P-value
Sex Male Female
1.450 1
0.678–1.675 –
0.334 (n.s.)
Age <65 ≥65
0.869 1
0.698–1.260 –
0.053 (n.s.)
FAB classification RA (and RARS) RAEB RAEB-t CMML
0.364 0.370 0.741 1
0.206–0.643 0.187–0.733 0.323–1.699 –
0.004 0.001 0.003 0.896 (n.s.)
Median survival
95% confidential interval
P-value
Sex Male Female
2.89 5.15
2.13–3.65 2.84–7.46
0.03
Age <65 ≥65
5.15 2.89
4.12–6.18 2.00–3.78
0.03
Bournemouth score 0–1 ≥2
0.715 1
0.380–1.348 –
0.300 (n.s.)
FAB classification RA (and RARS) RAEB RAEB-t CMML
5.12 4.63 4.33 2.21
3.2–7.04 1.80–7.46 1.44–7.22 1.56–2.86
0.0002
BM involvement Mono-linear Multi-linear
0.737 1
0.439–1.237 –
0.248 (n.s.)
1.084–4.035 –
0.02
Bournemouth score 0–1 ≥2
Hemorrhagic complications Yes 2.094 No 1
4.54 2.74
3.21–5.87 1.81–3.67
0.006
0.890 1
0.703–1.513 –
0.454 (n.s.)
Spanish score 0–1 ≥2
Hb levels ≥10 g/dl <10 g/dl
4.08 3.22
2.53–5.63 2.00–4.44
0.1103 (n.s.)
1.514 1
1.011–2.292 –
0.05
IPSS score 0 0.5–1 1.5–2 >2
PLTs levels <100.000 l−1 ≥100.000 l−1
2.75 2.46 2.1 1.6
1.81–4.87 0.95–3.83 0.70–3.04 0.70–2.04
WBC ≥3.500 l−1 <3.500 l−1
0.789 1
0.609–1.511 –
0.650 (n.s.)
BM involvement Mono-linear Trilinear
5.16 3.31
1.44–8.88 2.58–4.04
0.03
Transfusions No Yes
4.54 3.08
2.22–6.86 1.96–4.20
0.0665 (n.s.)
Infections No Yes
3.71 3.31
2.33–5.09 1.22–5.40
0.8746 (n.s.)
Hemorrhagic complications No 4.08 Yes 2.09
2.87–5.29 1.24–2.94
0.002
Hb levels ≥10 g/dl <10 g/dl
4.08 3.08
1.66–6.50 1.51–3.65
0.03
PLTs levels ≥100.000 l−1 <100.000 l−1
4.09 2.41
3.06–5.52 1.80–3.02
0.003
Neutrophils ≥2.500 l−1 <2.500 l−1
5.16 3.83
4.15–6.17 1.97–5.69
0.4146 (n.s.)
WBC ≥3.500 l−1 <3.500 l−1
6.92 3.31
3.24–10.60 1.85–4.77
0.02
0.0826 (n.s.)
ate analysis (Table 2) and platelet values <100 × 109 l−1 (P = 0.05), haemorrhagic symptoms at diagnosis (P = 0.02) and FAB classification (P = 0.004) appeared to be the main prognostic factors associated with shorter survival (Figs. 1 and 2). Table 3 shows the results of univariate analysis identifying prognostic factors for risk of evolution: trilineage bone marrow involvement (P = 0.03), haemoglobin level <10 g/dl (P = 0.04) and intermediate and high risk according to Bournemouth score (P = 0.0009), calculated at the time of diagnosis, were of statistical significance. At the multivariate analysis (Table 4) only patients with a score value >2 at Bournemouth score (P = 0.02) had a significantly high relative risk of evolution to AML (Fig. 3).
4. Discussion The most frequent cytogenetic pattern in MDS is represented by normal karyotype. In fact 40–60% of de novo MDS and 20% of secondary or therapy-related MDS present a normal karyotype at the time of diagnosis [13,17–19].
36
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
Fig. 1. Analysis of survival based on presence of haemorrhages at presentation.
In our series, the frequency of normal karyotype was 58%. This cytogenetic finding has been considered by the IPSS as a good risk prognostic variable, with an attributed score value of 0; however patients with normal karyotype are heterogeneous in their disease clinical course and final prognosis. In our analysis, only 35 of 98 patients were categorized by IPSS as low risk, whereas 42 patients were identified
as intermediate-1 risk, 18 patients as intermediate-2 and 3 patients as high risk. In our study, there is a small difference in the distribution of patients by this scoring system among the groups that was not significative for survival and for risk of transformation. Cytogenetic analysis is crucial in MDS patients but available techniques all have some limitations: in fact, CCA is mostly unable to precisely identify complex aberrations,
Fig. 2. Analysis of survival for platelet counts.
M. Breccia et al. / Leukemia Research 29 (2005) 33–39 Table 3 Univariate analysis of leukemic evolution
37
Table 4 Multivariate cox regression analysis of leukemic evolution
Median survival (months)
95% confidential interval
P-value
Sex Male Female
2.74 3.19
1.45–4.03 1.30–5.08
Age <65 ≥65
3.34 2.42
FAB classification RA (and RARS) RAEB RAEB-t CMML
95% confidential interval
P-value
0.397–0.925 –
0.02
0.7832 (n.s.)
Bournemouth score 0–1 0.606 ≥2 1
0.7091 (n.s.)
0.705 1
0.445–1.118 –
0.137 (n.s.)
1.91–4.77 1.24–3.60
BM involvement Mono-linear Multi-linear
1.40–4.38 0.87–7.98 0.00–3.72 0.48–4.76
0.3088 (n.s.)
0.897 1
0.769–1.497 –
0.454 (n.s.)
2.89 3.82 1.59 2.62
Hb levels ≥10 g/dl <10 g/dl
Bourrnemouth score 0–1 ≥2
4.08 1.60
2.53–5.63 1.22–1.98
0.0009
Spanish score 0–1 ≥2
3.19 2.39
1.81–4.57 0.87–3.91
0.4709 (n.s.)
IPSS score 0 0.5–1 1.5–2 >2
2.75 2.46 2.1 1.6
1.81–4.87 0.95–3.83 0.70–3.04 0.70–2.04
BM involvement Mono-linear Tri-linear
4.93 2.15
1.25–8.61 1.31–2.99
0.03
Transfusions No Yes
2.74 2.89
1.06–4.42 1.65–4.13
0.5071 (n.s.)
Infections No Yes
3.82 2.93
2.10–5.54 1.30–3.48
0.5151 (n.s.)
Hemorrhagic complications No 2.39 Yes 3.19
1.03–3.75 1.80–4.58
0.8095 (n.s.)
Hb levels ≥10 g/dl <10 g/dl
3.34 2.30
1.02–5.66 0.92–3.68
0.04
PLTs levels ≥100.000 l−1 <100.000 l−1
3.45 1.60
2.09–4.81 0.86–2.34
0.0842 (n.s.)
Neutrophils ≥2.500 l−1 <2.500 l−1
4.08 2.74
1.88–6.28 1.71–3.77
0.7016 (n.s.)
WBC ≥3.500 l−1 <3.500 l−1
4.04 2.42
2.37–5.71 1.23–3.61
0.0890 (n.s.)
0.6 (n.s.)
hidden translocations or insertions, or cryptic mutations, and fluorescence in situ hybridization (FISH) is limited to two or three chromosomes and requires some prior knowledge of chromosomal origin to select the appropriate probes [20].
Relative risk of evolution
Recently, several publications have focused on the importance of the more sensitive molecular cytogenetic investigations in patients displaying a normal karyotype at the CCA, but reached opposite conclusions. In 2001, Rigolin et al. [21] reported on the detection of minor clones by FISH in 18/101 patients (17.8%) with primary MDS and apparently normal karyotype at CCA, which were able to identify a subgroup within these patients with an inferior prognosis. In 2002, Ketterling et al. [22], in a study on 32 MDS patients with normal karyotype at CCA reached an opposite conclusion. In fact, they detected a minor clone in only 1 of the 32 cases in which FISH analysis was applied and suggested for a scant utility of this technique at diagnosis of MDS disease. Bernasconi et al. [23], found occult defects in about 15% of CCA normal MDS patients and indicated the clinical relevance of these findings as they caused a change in the IPSS categorization and could identify a fraction of patients with apparently normal karyotype but unfavourable clinical outcome. Moreover, an abnormal FISH pattern significantly correlated with percentage of bone marrow blasts and IPSS. In the absence of FISH analysis, not always routinely applied in MDS patients, we analysed some clinical features as detected at diagnosis in patients with normal karyotype at CCA. We found that platelets <100 × 109 l−1 , presence of haemorrhagic symptoms and RAEB or CMML FAB categories were strictly associated with a shorter survival. Furthermore, the inclusion of patients with intermediate/high risk categories according to Bournemouth score correlated with an higher risk of evolution into AML. Therefore, we suggest that in the lack of routinely applicable FISH analysis in all MDS patients with normal karyotype at CCA, evaluation of clinical parameters could be of help to select patients in whom molecular investigations should be more appropriate to detect minor abnormal clones or submicroscopic lesions, and consequently choice of risk adapted therapeutic strategies.
38
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
Fig. 3. Relative risk for evolution to AML according to Bournemouth score.
References [1] Tricot G, Vlietinck R, Boogaerts MA, Hendrickx B, De Wolf-Peeters C, Van den Berghe H, et al. Prognostic factors in the myelodysplastic syndromes: importance of initial data on peripheral blood counts, bone marrow cytology, trephine biopsy and chromosomal analysis. Br J Haematol 1985;60:19–32. [2] Goasguen JE, Garand R, Bizet M, Bremond JL, Gardais J, Callat MP, et al. Prognostic factors of myelodysplastic syndromes—a simplified 3-D scoring system. Leuk Res 1990;14:255–62. [3] Sanz GF, Sanz MA, Greenberg PL. Prognostic factors and scoring systems in myelodysplastic syndromes. Haematologica 1998;83:358– 68. [4] Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 1982;51:189–99. [5] Mufti GJ, Stevens JR, Oscier DG, Hamblin TJ, Machin D. Myelodysplastic syndromes: a scoring system with prognostic significance. Br J Haematol 1985;59:425–33. [6] Coiffier B, Adeleine P, Gentilhomme O, Felman P, Treille-Ritouet D, Bryon PA. Myelodysplastic syndromes. A multiparametric study of prognostic factors in 336 patients. Cancer 1987;60:3029–32. [7] Sanz GF, Sanz MA, Vallespi T, Canizo MC, Torrabadella M, Garcia S, et al. Two regression models and a scoring system for predicting survival and planning treatment in myelodysplastic syndromes: a multivariate analysis of prognostic factors in 370 patients. Blood 1989;74:395–408. [8] Aul C, Gattermann N, Heyll A, Germing U, Derigs G, Schneider W. Primary myelodysplastic syndromes: analysis of prognostic factors in 235 patients and proposals for an improved scoring system. Leukemia 1992;6:52–9. [9] Morel P, Hebbar M, Lai JL, Duhamel A, Preudhomme C, Wattel E, et al. Cytogenetic analysis has strong independent prognostic value in de novo myelodysplastic syndromes and can be incorporated in a new scoring system: a report on 408 cases. Leukemia 1993;7:1315– 23.
[10] Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 1997;89:2079–88. [11] Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, Vardiman J, et al. WHO classification of neoplastic diseases of the hematopoietic and lymphoid tissues: report of the clinical advisory committee meeting. J Clin Oncol 1999;17:3835–49. [12] Vallespi T, Imbert M, Mecucci C, Preudhomme C, Fenaux P. Diagnosis, classification, and cytogenetics of myelodysplastic syndromes. Haematologica 1998;83:258–75. [13] Mufti GJ. Chromosomal deletions in the myelodysplastic syndrome. Leuk Res 1992;16:35–41. [14] Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. [15] Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br J Cancer 1977;35:1–39. [16] Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187–220. [17] Fenaux P, Morel P, Lai JL. Cytogenetics of myelodysplastic syndromes. Sem Hematol 1996;33:127–38. [18] Mauritzson N, Johansson B, Albin M, Billstrom R, Ahlgren T, Mikoczy Z, et al. A single-center population-based consecutive series of 1500 cytogenetically investigated adult hematological malignancies: karyotypic features in relation to morphology, age and gender. Eur J Haematol 1999;62:95–102. [19] Sole F, Espinet B, Sanz GF, Cervera J, Calasanz MJ, Luno E, et al. Incidence, characterization and prognostic significance of chromosomal abnormalities in 640 patients with primary myelodysplastic syndromes. Grupo Cooperativo Espanol de Citogenetica Hematologica. Br J Haematol 2000;108:346–56. [20] de Greef GE, Hagemeijer A. Molecular and cytogenetic abnormalities in acute myeloid leukaemia and myelodysplastic syndromes. Baillieres Clin Haematol 1996;9:1–18. [21] Rigolin GM, Bigoni R, Milani R, Cavazzini F, Roberti MG, Bardi A, et al. Clinical importance of interphase cytogenetics detecting occult chromosome lesions in myelodysplastic syndromes with normal karyotype. Leukemia 2001;15:1841–7.
M. Breccia et al. / Leukemia Research 29 (2005) 33–39 [22] Ketterling RP, Wyatt WA, VanWier SA, Law M, Hodnefield JM, Hanson CA, et al. Primary myelodysplastic syndrome with normal cytogenetics: utility of ‘FISH panel testing’ and M-FISH. Leuk Res 2002;26:235–40.
39
[23] Bernasconi P, Cavigliano PM, Boni M, Calatroni S, Klersy C, Giardini I, et al. Is FISH a relevant prognostic tool in myelodysplastic syndromes with a normal chromosome pattern on conventional cytogenetics? A study on 57 patients. Leukemia 2003;17:2107–12.
Clinical features of prognostic significance in myelodysplastic patients with normal karyotype at high risk of transformation Massimo Breccia∗ , Marco Mancini, Mauro Nanni, Gianna Maria D’Elia, Ida Carmosino, Roberto Latagliata, Chiara Sarlo, Franco Mandelli, Giuliana Alimena Dipartimento di Biotecnologie Cellulari ed Ematologia, Università La Sapienza, Via Benevento 6, 00161 Rome, Italy Received 3 May 2004; accepted 10 May 2004 Available online 31 July 2004
Abstract The International Prognostic Scoring System (IPSS) for myelodysplastic syndromes (MDS) has defined patients with a normal karyotype as a good risk cytogenetic subgroup, but nevertheless a fraction of these patients has a poor outcome similar to that of high risk patients. We retrospectively analysed our series of myelodysplastic patients with normal karyotype observed in a period of 11 years, with the aim of identifying clinical features of possible prognostic significance within this subgroup of patients. Multivariate analysis showed that among clinical scoring systems, the Bournemouth score appears the best prognostic indicator for risk of leukemic transformation, and platelet count <100 × 109 /l−1 , presence of haemorrhagic symptoms at time of diagnosis and morphologic FAB classification are the main prognostic factors for prediction of survival. In the absence of genetic abnormalities as detected by conventional cytogenetics or even the more sensitive molecular techniques in MDS, clinical variables could be of help in identifying patients with different prognosis, suitable for risk adapted therapeutic strategies. © 2004 Elsevier Ltd. All rights reserved. Keywords: Myelodysplastic syndromes; Normal karyotype; Prognostic factors
1. Introduction Myelodysplastic syndromes (MDS) are a group of heterogeneous clonal haematopoietic stem cell disorders characterized by dysfunctional haematopoiesis leading to bone marrow failure and high risk of progression to acute leukaemia. Through multivariate analyses of large series of patients, several studies have shown that main prognostic factors in these disorders are percentage of bone marrow blasts, number and degree of cytopenias, age >65 years and presence of chromosomal abnormalities [1–3]. Based on these features, different scoring systems for prediction of outcome and risk of transformation to acute myeloid leukaemia (AML) have been created during the last two decades [4–9], the most used being the International Prognostic Scoring System (IPSS) that has defined four prognostic subgroups of MDS patients [10]. ∗ Corresponding author. Tel.: +39 06 85 79 51; fax: +39 06 44 24 19 84. E-mail address: [email protected] (M. Breccia).
0145-2126/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.leukres.2004.05.010
The IPSS evidences a normal karyotype at diagnosis as detected by conventional cytogenetics, in 40–65% of all MDS patients included in low risk group [10]. Nevertheless, this category of patients is very heterogeneous for the wide variability of clinical parameters and the outcome is often unpredictable. In present study, we retrospectively evaluated 98 consecutive MDS patients with normal karyotype at CCA and analysed clinical features as detected at time of diagnosis, in order to establish their possible prognostic impact on survival and risk of evolution to acute leukaemia.
2. Patients and methods 2.1. Patients Between January 1986 and December 1997, 98 patients were diagnosed at our Institution as having a myelodysplastic syndrome with a normal karyotype; this figure represents 58% of the whole MDS patients diagnosed over the same period of time. We included in the study only patients without
34
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
history of previous cytotoxic, chemo- or radiation exposure, and excluded patients with secondary or therapy-related MDS, according to WHO criteria [11]. Morphological examinations were performed using standard methods on peripheral blood and bone marrow Wright–Giemsa-stained films. Iron staining and non-specific esterase were used for the diagnosis of refractory anemia with ringed sideroblasts and chronic myelomonocytic leukaemia. Classification was performed according to FAB criteria [4]; we used FAB and not WHO revisited criteria because in the analysis were included patients with blasts >20% (RAEB-t), that in the past were considered MDS and not AML. Bone marrow aspirate was obtained in all cases at diagnosis and dysplastic features were recognized as previously indicated [4]. Cytogenetic analyses were conducted at the time of diagnosis by GTG banding technique on bone marrow cells obtained from short-term unstimolated cultures (24–48 h). Chromosomal abnormalities were identified by the International System for Human Cytogenetic Nomenclature (1995). At least 20 metaphases were analysed in all cases. Treatment of MDS phase was based on only supportive therapy in all patients. Only five patients received AML-like therapy after evolution. 2.2. Analysis of prognostic factors The following characteristics at time of diagnosis were analysed to establish the association with survival and risk of transformation to acute leukaemia: age, sex, dysplastic lineage involvement, haemoglobin, white blood cell (WBC), platelet and neutrophils counts, percentage of blast cells in the bone marrow, presence of transfusional requirement, presence of hemorrhagic symptoms and/or infections. The prognostic impacts of FAB classification and of three scoring systems were also evaluated. The first two systems, the Bournemouth and the Spanish scoring, are based only on clinical parameters while the International Prognostic Scoring Systems introduces also the karyotype [7,10,12]. According to this last system the stratification of patients in low, intermediate-1, intermediate-2 and high risk was only applied on the basis of clinical features, because for the normal karyotype the attributed score value is 0, i.e. considered as of “good prognosis”. 2.3. Statistical analysis Univariate analysis for each prognostic variable on overall survival and time to disease evolution was estimated according to the Kaplan–Meier method. The terminal event was considered the death attributable to cancer or non-cancer causes. The statistical significance of the differences in survival distribution among the prognostic groups was evaluated by the log-rank test. The Cox proportional hazards model was applied to the multivariate survival analysis.
The prognostic variables for overall survival and time to leukemic evolution included: sex, age, FAB classification, Bournemouth score, Sanz score, IPSS score, lineages of bone marrow involvement, transfusions requirement, infectious and/or hemorrhagic complications, hemoglobin levels, WBC, neutrophils and platelets counts at diagnosis. P-values <0.05 were regarded as statistically significant in two tailed tests. SPSS software (version 10.00, SPSS, Chicago) was used for statistical analysis [14–16]. 3. Results 3.1. Clinical data There were 59 males (60%) and 39 females (40%). Median age was 63.8 years (range 23–82.8). According to FAB classification there were 29 patients with RA (29%), 11 patients with RARS (11.2%), 26 patients with RAEB (26.5%), 8 patients with RAEB-t (8%) and 24 patients with CMML (24.4%). In the peripheral blood, 47 patients had haemoglobin values <10 g/dl, 36 patients had neutrophil counts <2.5 × 109 l−1 , 31 patients had platelet counts <100 × 109 l−1 . Morphological analysis revealed a bone marrow percentages of blasts <5% in 40 patients (41%), 5–10% in 38 patients (38.7%), 11–20% in 12 patients (12%) and >20% in 8 patients (8%). In 51 patients (52%) there was a trilinear bone marrow dysplasia. Sixty patients (61.2%) required transfusional support; 44 patients (44.8%) had an infection and 13 patients (13.2%) presented one or more hemorrhagic events at the time of diagnosis. Median survival of all 98 patients was 28.9 months. These data compared to a median survival of 21.7 months in the whole MDS population, observed in the same period. By the end of follow-up (60 months), 41/98 patients (41.8%) have died, acute leukaemia being the main cause of death, with a transformation rate of 32% (13 patients); in 4 patients (9.8%) death was not related to the myelodysplastic disease and was mainly due to cardiovascular failure or second neoplasias. 3.2. Analysis of prognostic factors for survival and leukemic transformation The results of the univariate analysis of prognostic factors for survival are summarized in Table 1. Male sex (P = 0.03), age >65 years (P = 0.03), FAB category with high percentage of blasts (P = 0.0002) and presence of trilineage involvement (P = 0.03) were strictly associated with a shorter survival. At univariate analysis, also significant were haemorrhagic symptoms at diagnosis (P = 0.002), haemoglobin level <10 g/dl (P = 0.03),
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
platelet count <100 × 109 l−1 (P = 0.003) and WBC count <3.5 × 109 l−1 (P = 0.02). The subdivision of patients in risk subgroups according to different scoring systems (Bournemouth, Spanish, IPSS) was significative for survival only when applying the Bournemouth score (P = 0.006). The considered factors were entered to multivariTable 1 Univariate Analysis of overal survival
35
Table 2 Multivariate cox regression analysis of overall survival Relative risk of death
95% confidential interval
P-value
Sex Male Female
1.450 1
0.678–1.675 –
0.334 (n.s.)
Age <65 ≥65
0.869 1
0.698–1.260 –
0.053 (n.s.)
FAB classification RA (and RARS) RAEB RAEB-t CMML
0.364 0.370 0.741 1
0.206–0.643 0.187–0.733 0.323–1.699 –
0.004 0.001 0.003 0.896 (n.s.)
Median survival
95% confidential interval
P-value
Sex Male Female
2.89 5.15
2.13–3.65 2.84–7.46
0.03
Age <65 ≥65
5.15 2.89
4.12–6.18 2.00–3.78
0.03
Bournemouth score 0–1 ≥2
0.715 1
0.380–1.348 –
0.300 (n.s.)
FAB classification RA (and RARS) RAEB RAEB-t CMML
5.12 4.63 4.33 2.21
3.2–7.04 1.80–7.46 1.44–7.22 1.56–2.86
0.0002
BM involvement Mono-linear Multi-linear
0.737 1
0.439–1.237 –
0.248 (n.s.)
1.084–4.035 –
0.02
Bournemouth score 0–1 ≥2
Hemorrhagic complications Yes 2.094 No 1
4.54 2.74
3.21–5.87 1.81–3.67
0.006
0.890 1
0.703–1.513 –
0.454 (n.s.)
Spanish score 0–1 ≥2
Hb levels ≥10 g/dl <10 g/dl
4.08 3.22
2.53–5.63 2.00–4.44
0.1103 (n.s.)
1.514 1
1.011–2.292 –
0.05
IPSS score 0 0.5–1 1.5–2 >2
PLTs levels <100.000 l−1 ≥100.000 l−1
2.75 2.46 2.1 1.6
1.81–4.87 0.95–3.83 0.70–3.04 0.70–2.04
WBC ≥3.500 l−1 <3.500 l−1
0.789 1
0.609–1.511 –
0.650 (n.s.)
BM involvement Mono-linear Trilinear
5.16 3.31
1.44–8.88 2.58–4.04
0.03
Transfusions No Yes
4.54 3.08
2.22–6.86 1.96–4.20
0.0665 (n.s.)
Infections No Yes
3.71 3.31
2.33–5.09 1.22–5.40
0.8746 (n.s.)
Hemorrhagic complications No 4.08 Yes 2.09
2.87–5.29 1.24–2.94
0.002
Hb levels ≥10 g/dl <10 g/dl
4.08 3.08
1.66–6.50 1.51–3.65
0.03
PLTs levels ≥100.000 l−1 <100.000 l−1
4.09 2.41
3.06–5.52 1.80–3.02
0.003
Neutrophils ≥2.500 l−1 <2.500 l−1
5.16 3.83
4.15–6.17 1.97–5.69
0.4146 (n.s.)
WBC ≥3.500 l−1 <3.500 l−1
6.92 3.31
3.24–10.60 1.85–4.77
0.02
0.0826 (n.s.)
ate analysis (Table 2) and platelet values <100 × 109 l−1 (P = 0.05), haemorrhagic symptoms at diagnosis (P = 0.02) and FAB classification (P = 0.004) appeared to be the main prognostic factors associated with shorter survival (Figs. 1 and 2). Table 3 shows the results of univariate analysis identifying prognostic factors for risk of evolution: trilineage bone marrow involvement (P = 0.03), haemoglobin level <10 g/dl (P = 0.04) and intermediate and high risk according to Bournemouth score (P = 0.0009), calculated at the time of diagnosis, were of statistical significance. At the multivariate analysis (Table 4) only patients with a score value >2 at Bournemouth score (P = 0.02) had a significantly high relative risk of evolution to AML (Fig. 3).
4. Discussion The most frequent cytogenetic pattern in MDS is represented by normal karyotype. In fact 40–60% of de novo MDS and 20% of secondary or therapy-related MDS present a normal karyotype at the time of diagnosis [13,17–19].
36
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
Fig. 1. Analysis of survival based on presence of haemorrhages at presentation.
In our series, the frequency of normal karyotype was 58%. This cytogenetic finding has been considered by the IPSS as a good risk prognostic variable, with an attributed score value of 0; however patients with normal karyotype are heterogeneous in their disease clinical course and final prognosis. In our analysis, only 35 of 98 patients were categorized by IPSS as low risk, whereas 42 patients were identified
as intermediate-1 risk, 18 patients as intermediate-2 and 3 patients as high risk. In our study, there is a small difference in the distribution of patients by this scoring system among the groups that was not significative for survival and for risk of transformation. Cytogenetic analysis is crucial in MDS patients but available techniques all have some limitations: in fact, CCA is mostly unable to precisely identify complex aberrations,
Fig. 2. Analysis of survival for platelet counts.
M. Breccia et al. / Leukemia Research 29 (2005) 33–39 Table 3 Univariate analysis of leukemic evolution
37
Table 4 Multivariate cox regression analysis of leukemic evolution
Median survival (months)
95% confidential interval
P-value
Sex Male Female
2.74 3.19
1.45–4.03 1.30–5.08
Age <65 ≥65
3.34 2.42
FAB classification RA (and RARS) RAEB RAEB-t CMML
95% confidential interval
P-value
0.397–0.925 –
0.02
0.7832 (n.s.)
Bournemouth score 0–1 0.606 ≥2 1
0.7091 (n.s.)
0.705 1
0.445–1.118 –
0.137 (n.s.)
1.91–4.77 1.24–3.60
BM involvement Mono-linear Multi-linear
1.40–4.38 0.87–7.98 0.00–3.72 0.48–4.76
0.3088 (n.s.)
0.897 1
0.769–1.497 –
0.454 (n.s.)
2.89 3.82 1.59 2.62
Hb levels ≥10 g/dl <10 g/dl
Bourrnemouth score 0–1 ≥2
4.08 1.60
2.53–5.63 1.22–1.98
0.0009
Spanish score 0–1 ≥2
3.19 2.39
1.81–4.57 0.87–3.91
0.4709 (n.s.)
IPSS score 0 0.5–1 1.5–2 >2
2.75 2.46 2.1 1.6
1.81–4.87 0.95–3.83 0.70–3.04 0.70–2.04
BM involvement Mono-linear Tri-linear
4.93 2.15
1.25–8.61 1.31–2.99
0.03
Transfusions No Yes
2.74 2.89
1.06–4.42 1.65–4.13
0.5071 (n.s.)
Infections No Yes
3.82 2.93
2.10–5.54 1.30–3.48
0.5151 (n.s.)
Hemorrhagic complications No 2.39 Yes 3.19
1.03–3.75 1.80–4.58
0.8095 (n.s.)
Hb levels ≥10 g/dl <10 g/dl
3.34 2.30
1.02–5.66 0.92–3.68
0.04
PLTs levels ≥100.000 l−1 <100.000 l−1
3.45 1.60
2.09–4.81 0.86–2.34
0.0842 (n.s.)
Neutrophils ≥2.500 l−1 <2.500 l−1
4.08 2.74
1.88–6.28 1.71–3.77
0.7016 (n.s.)
WBC ≥3.500 l−1 <3.500 l−1
4.04 2.42
2.37–5.71 1.23–3.61
0.0890 (n.s.)
0.6 (n.s.)
hidden translocations or insertions, or cryptic mutations, and fluorescence in situ hybridization (FISH) is limited to two or three chromosomes and requires some prior knowledge of chromosomal origin to select the appropriate probes [20].
Relative risk of evolution
Recently, several publications have focused on the importance of the more sensitive molecular cytogenetic investigations in patients displaying a normal karyotype at the CCA, but reached opposite conclusions. In 2001, Rigolin et al. [21] reported on the detection of minor clones by FISH in 18/101 patients (17.8%) with primary MDS and apparently normal karyotype at CCA, which were able to identify a subgroup within these patients with an inferior prognosis. In 2002, Ketterling et al. [22], in a study on 32 MDS patients with normal karyotype at CCA reached an opposite conclusion. In fact, they detected a minor clone in only 1 of the 32 cases in which FISH analysis was applied and suggested for a scant utility of this technique at diagnosis of MDS disease. Bernasconi et al. [23], found occult defects in about 15% of CCA normal MDS patients and indicated the clinical relevance of these findings as they caused a change in the IPSS categorization and could identify a fraction of patients with apparently normal karyotype but unfavourable clinical outcome. Moreover, an abnormal FISH pattern significantly correlated with percentage of bone marrow blasts and IPSS. In the absence of FISH analysis, not always routinely applied in MDS patients, we analysed some clinical features as detected at diagnosis in patients with normal karyotype at CCA. We found that platelets <100 × 109 l−1 , presence of haemorrhagic symptoms and RAEB or CMML FAB categories were strictly associated with a shorter survival. Furthermore, the inclusion of patients with intermediate/high risk categories according to Bournemouth score correlated with an higher risk of evolution into AML. Therefore, we suggest that in the lack of routinely applicable FISH analysis in all MDS patients with normal karyotype at CCA, evaluation of clinical parameters could be of help to select patients in whom molecular investigations should be more appropriate to detect minor abnormal clones or submicroscopic lesions, and consequently choice of risk adapted therapeutic strategies.
38
M. Breccia et al. / Leukemia Research 29 (2005) 33–39
Fig. 3. Relative risk for evolution to AML according to Bournemouth score.
References [1] Tricot G, Vlietinck R, Boogaerts MA, Hendrickx B, De Wolf-Peeters C, Van den Berghe H, et al. Prognostic factors in the myelodysplastic syndromes: importance of initial data on peripheral blood counts, bone marrow cytology, trephine biopsy and chromosomal analysis. Br J Haematol 1985;60:19–32. [2] Goasguen JE, Garand R, Bizet M, Bremond JL, Gardais J, Callat MP, et al. Prognostic factors of myelodysplastic syndromes—a simplified 3-D scoring system. Leuk Res 1990;14:255–62. [3] Sanz GF, Sanz MA, Greenberg PL. Prognostic factors and scoring systems in myelodysplastic syndromes. Haematologica 1998;83:358– 68. [4] Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 1982;51:189–99. [5] Mufti GJ, Stevens JR, Oscier DG, Hamblin TJ, Machin D. Myelodysplastic syndromes: a scoring system with prognostic significance. Br J Haematol 1985;59:425–33. [6] Coiffier B, Adeleine P, Gentilhomme O, Felman P, Treille-Ritouet D, Bryon PA. Myelodysplastic syndromes. A multiparametric study of prognostic factors in 336 patients. Cancer 1987;60:3029–32. [7] Sanz GF, Sanz MA, Vallespi T, Canizo MC, Torrabadella M, Garcia S, et al. Two regression models and a scoring system for predicting survival and planning treatment in myelodysplastic syndromes: a multivariate analysis of prognostic factors in 370 patients. Blood 1989;74:395–408. [8] Aul C, Gattermann N, Heyll A, Germing U, Derigs G, Schneider W. Primary myelodysplastic syndromes: analysis of prognostic factors in 235 patients and proposals for an improved scoring system. Leukemia 1992;6:52–9. [9] Morel P, Hebbar M, Lai JL, Duhamel A, Preudhomme C, Wattel E, et al. Cytogenetic analysis has strong independent prognostic value in de novo myelodysplastic syndromes and can be incorporated in a new scoring system: a report on 408 cases. Leukemia 1993;7:1315– 23.
[10] Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 1997;89:2079–88. [11] Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, Vardiman J, et al. WHO classification of neoplastic diseases of the hematopoietic and lymphoid tissues: report of the clinical advisory committee meeting. J Clin Oncol 1999;17:3835–49. [12] Vallespi T, Imbert M, Mecucci C, Preudhomme C, Fenaux P. Diagnosis, classification, and cytogenetics of myelodysplastic syndromes. Haematologica 1998;83:258–75. [13] Mufti GJ. Chromosomal deletions in the myelodysplastic syndrome. Leuk Res 1992;16:35–41. [14] Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. [15] Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br J Cancer 1977;35:1–39. [16] Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187–220. [17] Fenaux P, Morel P, Lai JL. Cytogenetics of myelodysplastic syndromes. Sem Hematol 1996;33:127–38. [18] Mauritzson N, Johansson B, Albin M, Billstrom R, Ahlgren T, Mikoczy Z, et al. A single-center population-based consecutive series of 1500 cytogenetically investigated adult hematological malignancies: karyotypic features in relation to morphology, age and gender. Eur J Haematol 1999;62:95–102. [19] Sole F, Espinet B, Sanz GF, Cervera J, Calasanz MJ, Luno E, et al. Incidence, characterization and prognostic significance of chromosomal abnormalities in 640 patients with primary myelodysplastic syndromes. Grupo Cooperativo Espanol de Citogenetica Hematologica. Br J Haematol 2000;108:346–56. [20] de Greef GE, Hagemeijer A. Molecular and cytogenetic abnormalities in acute myeloid leukaemia and myelodysplastic syndromes. Baillieres Clin Haematol 1996;9:1–18. [21] Rigolin GM, Bigoni R, Milani R, Cavazzini F, Roberti MG, Bardi A, et al. Clinical importance of interphase cytogenetics detecting occult chromosome lesions in myelodysplastic syndromes with normal karyotype. Leukemia 2001;15:1841–7.
M. Breccia et al. / Leukemia Research 29 (2005) 33–39 [22] Ketterling RP, Wyatt WA, VanWier SA, Law M, Hodnefield JM, Hanson CA, et al. Primary myelodysplastic syndrome with normal cytogenetics: utility of ‘FISH panel testing’ and M-FISH. Leuk Res 2002;26:235–40.
39
[23] Bernasconi P, Cavigliano PM, Boni M, Calatroni S, Klersy C, Giardini I, et al. Is FISH a relevant prognostic tool in myelodysplastic syndromes with a normal chromosome pattern on conventional cytogenetics? A study on 57 patients. Leukemia 2003;17:2107–12.