Myelodysplastic syndrome (MDS) constitutes a heterogeneous group of clonal stem cell disorders characterized by ineffective hematopoiesis and an increased propensity for transformation to acute leukemia. For hematologists who treat patients with MDS, many new treatment modalities have emerged that can go beyond supportive care and change the natural course of the disease.1–9 However, because the treatment can carry substantial risks, the choice of therapy always should be adapted to the risk profile of the patient's disease. Cytogenetics has an established role in the diagnosis and assessment of prognosis for patients with MDS and is emerging as an important factor in treatment selection and monitoring response to therapy. The International Prognostic Scoring System (IPSS)10 has placed cytogenetic abnormalities into 3 risk categories: ‘good,’ which includes a normal karyotype, deletion of 5q [del(5q)], del(20q), and loss of chromosome (−Y); ‘poor,’ which includes a complex karyotype (>3 abnormalities) as well as del(7q) and −7 present either as a single anomaly or in combination with other anomalies; and ‘intermediate,’ which includes all other abnormalities.
Although some chromosomal anomalies in MDS, such as del(5q), −7, del(7q), gain of chromosome 8 (+8) and del(20q), have been characterized fairly well, the significance of many cytogenetic abnormalities has not been defined well, primarily because of their low frequencies. According to the 1997 IPSS classification,10 all of these miscellaneous, single chromosomal aberrations are grouped together into the intermediate cytogenetic group. Recently, several reports generated from large patient series from the European/Australian groups have suggested that some of these less common cytogenetic abnormalities may play a very significant role in the clinical outcome of patients with MDS.11–14 Therefore, those authors have proposed a reclassification of the IPSS cytogenetic risk groups to incorporate these new findings. However, in those reports, although some findings on the significance of certain infrequent cytogenetic aberrations, such as del(12p), which harbors a good prognosis, are concordant among the studies, some discrepancies remain, such as the prognostic significance of del(11q), rearrangements of (3)(q21q26), del(17p), and +21.11–14 It is noteworthy that some of those studies included both primary MDS and therapy-related MDS, whereas some studies did not uniformly follow the World Health Organization (WHO) criteria to classify diseases. We believe that a study conducted on large cohorts of patients with primary MDS from the US, in which disease is defined strictly by the current WHO criteria and patients are treated according to the same algorithm, would provide valuable information for further clarification.
In the current article, we report the cytogenetic findings from a series of 1029 patients who had a diagnosis of primary MDS classified according to the WHO criteria. We analyze the prognostic value of the cytogenetic aberrations, focusing on the single cytogenetic abnormalities currently categorized under the IPSS intermediate and poor risk groups.
MATERIALS AND METHODS
We retrieved clinical and laboratory data on 1029 consecutive patients with MDS who had cytogenetic studies and were diagnosed at Rush University Medical Center (Rush) or University of Massachusetts Memorial Medical Center (UMass) over an 11-year period (1995-2006). Bone marrow biopsy materials were reviewed by the hematopathologists at Rush or UMass. In the majority of patients, the disease classification was based originally on the French-American-British (FAB) criteria15 and then was reclassified according to the WHO proposals16 based on morphology, laboratory characteristics, and cytogenetics. Patients with myeloproliferative diseases (MPD), chronic myelomonocytic leukemia (CMML), atypical chronic myeloid leukemia, MDS/MPD-unclassifiable, or refractory anemia (RA) with excess blasts in transformation (RAEB-t) were excluded. Patients with an ambiguous diagnosis of MDS or with therapy-related MDS (who previously had received chemotherapy/radiation for primary malignancies) also were excluded from the final analysis. The prognostic score in patients with MDS was calculated by using the IPSS (Table 1).10
Table 1. Clinicopathologic Characteristics of the Patients
|Total no. of patients||1029 (100)|
|Median age [range], y||67 [19–92]|
|Age group, y|
| <45||49 (4.9)|
| 45-65||330 (32)|
| ≥65||650 (63.1)|
| Men||670 (65.1)|
| Women||359 (33.9)|
|Hemoglobin, g/dL, n = 905|
| >10||497 (54.9)|
| ≤10||408 (45.1)|
|Absolute neutrophil count, ×109/L, n = 906|
| >1.8||427 (47.1)|
| ≤1.8||479 (52.9)|
|Platelet count, ×109/L, n = 904|
| >100||446 (49.2)|
| ≤100||458 (50.8)|
| None||141 (15.6)|
| 1||343 (37.8)|
| 2||271 (29.8)|
| 3||152 (16.8)|
|Bone marrow blasts, %|
| <5||695 (67.6)|
| 5-10||208 (20.2)|
| 11-19||126 (12.2)|
| Del(5q) syndrome||47 (4.6)|
| RA||185 (18)|
| RARS||86 (8.3)|
| RCMD and RCMD-RS||331 (32.2)|
| RAEB-1||203 (19.7)|
| RAEB-2||131 (12.7)|
| MDS-U||46 (4.5)|
| Low||295 (28.7)|
| Intermediate-1||468 (45.5)|
| Intermediate-2||194 (18.8)|
| High||72 (7)|
All patients were treated by the same group of hematologists at both institutions. Treatment modalities included transfusion of blood products; administration of growth factors; administration of Food and Drug Administration-approved drugs for MDS, such as azacitidine, decitabine, and lenalidomide; and experimental protocols using a variety of agents, such as thalidomide, arsenic trioxide, lintizumab, and tipifarnib. In our cohort, most patients were treated with experimental drugs, and only approximately 40 patients received lenalidomide, 45 patients received decitabine, and 50 patients received azacitidine. The treatment modalities could not be unpacked easily, because many patients received more than 1 therapeutic drug in the course of their disease. None of the patients either received immune suppressive therapy, such as antithymocyte globulin and cyclosporine, or underwent bone marrow transplantation.
Patient survival was calculated from the date of MDS diagnosis until death from any cause or until the last patient follow-up. The information on acute myeloid leukemia (AML) transformation was incomplete, because approximately 30% to 40% of patients also were under the care of local hematologists and were transferred to different healthcare facilities when they developed AML. Table 2 lists the AML transformation information that was available in our files; however, these data may be underestimates; thus, this information was not used for the outcome analysis. Survival data were obtained on 998 patients (97%).
Table 2. Clinical Characteristics of Cytogenetic Abnormalities, Median Survival, and Acute Myelogenous Leukemia Progression Rate
|Good, n = 682||67 [19-92]||454:228||64*||54 (7.9)||157||78||212||149||39|
|Normal, n = 571||66 [19-92]||394:177||68||45 (7.9)||145||72||193||126||35|
|Del(5q), n = 61||68 [43-88]||19:42||43||6 (9.8)|| || || ||14|| |
|Del(20q), n = 28||68 [47-85]||19:9||50||2 (7.1)||6||1||13||6||2|
|−Y, n = 22||76 [54-88]||22:0||39||1 (4.5)||6||5||6||3||2|
|Intermediate, n = 143||68 [26-90]||91:52||31*||16 (11.1)||23||8||52||58||2|
|Inv(3); t(3;3), n = 3||59 [54-67]||2:1||25||1 (33.3)|| || ||2||1|| |
|+8, n = 38||69 [36-87]||28:10||19||9 (23.7)||7||6||8||16||1|
|?Del(9q), n = 2||64, 72||2:0||12, 26||0||1|| || ||1|| |
|+11, n = 4||81 [75-90]||2:2||13.5||1 (25)||1|| || ||3|| |
|Del(11q), n = 5||71 [58-81]||2:3||53||0|| || ||4||1|| |
|Del(12p), n = 3||70 [66-73]||3:0||69||0|| || ||2||1|| |
|I(14)(q10), n = 2||78, 84||2:0||16, 17||0|| || || ||2|| |
|+15, n = 2||70, 79||1:1||23, 31||0||1|| ||1||1|| |
|I(17q), n = 2||45, 66||2:0||9, 11||0|| || || ||2|| |
|+19, n = 4||74 [64-85]||3:1||23.5||0||1||1||2|| || |
|+21, n = 2||75, 78||1:1||8, 33||0|| || || ||2|| |
|Other single abnormalities, n = 21||61 [26-90]||10:11||33||3 (14.3)||5||1||7||7||1|
|2 Anomalies n = 55||67 [40-82]||23:22||32||2 (4.2)||7|| ||27||21|| |
|Poor, n = 204||66 [25-86]||125:79||12*||37 (18)||5|| ||67||127||5|
|Del(7q), n = 9||63 [26-84]||7:2||26||1 (11.1)|| || ||4||5|| |
|−7, n = 8||68 [37-82]||5:3||8.5||1 (12.5)|| || ||2||6|| |
|Der(1;7)(q10;p10), n = 6||59 [55-64]||5:1||45.5||0|| || ||2||1||3|
|Complex, n = 181||66 [25-86]||108:73||11||35 (19.3)||5|| ||59||115||2|
All patients who were included in the final analysis had representative bone marrow trephine biopsy specimens and smears readily available for evaluation. The Perls reaction for iron was performed on bone marrow aspirates. For the diagnosis of morphologic dysplasia in bone marrow samples, features of dyserythropoiesis, dysgranulopoiesis, and dysmegakaryopoiesis had to be present in ≥10% of cells from the respective lineage. Unilineage dysplasia was defined by dysplasia that involved ≥10% of cells from a single cell lineage, whereas multilineage dysplasia involved ≥10% of cells from ≥2 lineages. The bone marrow differential was based a 500-cell count from multiple fields of the smears.
Conventional Giemsa banding analysis was performed on 24-hour and 48-hour bone marrow cultures. At least 20 metaphases were examined. Criteria defined by the International System for Human CytogeneticNomenclature were used to describe abnormal clones.17
A karyotype was considered complex when ≥3 cytogenetic abnormalities were identified within the same clone. When ≥2 clones with ≥2 aberrations were present, patients were categorized into the complex aberration group; whereas patients who had 2 cytogenetically independent, single-anomaly clones were categorized into the 2-aberrations group.
Patient survival was estimated by using he Kaplan-Meier method from the date of MDS diagnosis until death from any cause or until the last patient follow-up. Survival curves were compared statistically using the log-rank test. Differences between 2 groups were considered statistically significant if P values were <.05 in a 2-tailed test.
In the current series, which consisted of 1029 patients with de novo MDS, clonal cytogenetic abnormalities were identified in 458 patients (44.5%), similar the rate reported by others.11–14, 18 Of the patients who had with abnormal cytogenetics, single abnormalities comprised 48.5% (222 of 458 patients), including 50 patients who had poorly defined, infrequent, single chromosomal aberrations that currently are categorized in the intermediate cytogenetic risk group according to IPSS criteria. The large number of patients with primary MDS who were included in this study allowed us to define each of the cytogenetic categories more carefully.
The good cytogenetic risk group comprised the majority of the cases (682 patients; 66.3%) and had a median OS of 64 months. Although the median OS within this group ranged from 39 months (−Y) to 68 months (normal karyotype), the difference was not statistically significant. Our results corroborate previously published data on the cytogenetic abnormalities that are included in this category [del(5q), del(20q) and −Y]. Loss of chromosome Y can represent a normal age-related process or an MDS clone.19 However, even when it represents an abnormal clone, it is uncertain whether the loss of chromosome Y is related to disease pathogenesis.
The intermediate cytogenetic risk group was comprised of 143 patients (13.9%) and represented a cytogenetically diverse population that encompassed patients with 1 or 2 structural and/or numeric abnormalities. The median OS for this group was 31 months (range, 10–69 months) (Table 2). Although our findings overall endorse the 1997 IPSS categorization in defining the intermediate cytogenetic risk group, the median OS for patients with various cytogenetic abnormalities within this group ranged from 10 months to 69 months, with a significant deviation from the mean of 31 months, suggesting different prognoses and potential reclassification of some of the patients into a different risk group.
Del(12p), which is a recurrent cytogenetic abnormality that may carry a good prognosis,11–14 occurred at a very low frequency in our series, with only 3 cases identified, including 1 patient with RAEB-1 and 2 patients with RCMD. All 3 patients with del(12p) had an indolent clinical course; 2 patients were alive at the time of the last follow-up (105 months and 16 months), and 1 patient died at 33 months after the initial diagnosis. It has been reported in some studies that deletion of 11q23 in MDS harbors a good prognosis,11–13, 18 but it is classified as intermediate risk by others.14, 20 In our series, in total, 5 patients who had del(11)(q23) as a single cytogenetic abnormality were identified, including 1 patient with RCMD, 3 patients with RCMD-RS, and 1 patient with RAEB-1. Although these patients presented with severe anemia and leucopenia with or without thrombocytopenia, all 5 patients had an indolent clinical course, and the median OS for this group was 53 months, which was similar to the reports by Bernasconi et al. and Sole et al.11, 18 Although our numbers were small, they suggest that del(12p) and del(11)(q23) may belong to the good cytogenetic group. Rearrangement involving 3q reportedly indicates an intermediate prognosis in some studies13, 14, 18 but a poor prognosis in others.11, 12 In our series, 3 patients (2 with RCMD and 1 with RAEB-1) had 3q rearrangement [t(3;3)(q21;q26) or inv(3)(q21q26)] as a sole abnormality, and all 3 died at 8 months, 25 months, and 30 months after their initial diagnosis, suggesting an aggressive behavior. Trisomy 19, which is a characteristic abnormality in de novo myeloid malignancies,21 was observed in 4 patients who had a median OS of 24 months, similar to what was reported by Haase et al.14 On the basis of these data, both patients with 3q rearrangement and patients with +19 are likely to have been assigned appropriately to the intermediate cytogenetic risk group. Some other cytogenetic aberrations within the intermediate group for which a different prognostic indication has been suggested were extremely rare in our series. Isochromosome 17q reportedly harbors a poor prognosis,13, 22, 23 and it has been suggested that it should be placed in the poor prognostic category. We encountered only 2 patients who had i(17q) as a single chromosomal abnormality. Both patients presented with RAEB-1, had a very aggressive clinical course, and died at 9 months and 11 months after the initial diagnosis. Other infrequent single chromosomal abnormalities that may be potential candidates for reclassification, such as +21 (2 patients), del(13q) (1 patient), and +15q (2 patients), were extremely rare in our series. It is noteworthy that we identified 4 patients who had +11 as the sole cytogenetic anomaly, the significance of which has not been reported previously. These patients presented with RA (1 patient), RAEB-1 (1 patient), and RAEB-2 (2 patients). All of these patients died within 21 months after the diagnosis, indicating an aggressive clinical course.
Another intriguing finding within the intermediate cytogenetic group was a heterogeneous clinical course and outcome in patients who had +8 as the sole abnormality, which is considered to confer an intermediate risk. The median OS for this entire group was 19 months (range, 1-123 months). Within the intermediate cytogenetic risk category, the patients who died within 10 months and the patients who survived for >5 years mainly were those who had +8 (data not shown). Furthermore, no significant difference in OS was observed between patients who had +8 as a single abnormality and patients who had +8 associated with another cytogenetic abnormality. Despite the finding that +8 is 1 of the most common myeloid abnormalities and is present in 11% of patients with MDS, its clinical impact on the course of the disease largely remains unknown. For example, the reported transformation rate to AML in patients who have MDS with +8 varies greatly from 8% to 62%.12, 24–26 Recently, Paulsson et al. detected cryptic cytogenetic abnormalities, such as del(7)(p14p14) and del(12)(p13), in 4 of 10 patients with AML and MDS who had +8 as their sole cytogenetic abnormality by using high-resolution, genome-wide, array-based comparative genomic hybridization.27 In addition, somatic point mutations of leukemia-associated genes, eg, the CCAAT/enhancer binding protein α gene CEBPA, the fms-related tyrosine kinase 3 gene FLT3, the Kirsten rat sarcoma viral oncogene homolog KRAS, the neuroblastoma ras viral oncogene homolog NRAS, and the runt-related transcription factor 1 gene RUNX1, have been detected in patients who have MDS with +8.28–30
Within the poor cytogenetic risk category defined by the current IPSS criteria, our 8 patients who had del(7q) as a sole cytogenetic abnormality had a superior median OS compared the patients who had −7 or a complex karyotype (26 months, 8.5 months, and 11 months, respectively; P < .05). This observation is in agreement with the reports by Bernasconi et al. and Haase et al.11, 12, 14 and supports the proposal to reclassify interstitial deletion of 7q tothe intermediate cytogenetic risk categories. Der(1;7)(q10;p10), a derivative chromosome which is the result of an unbalanced translocation leading to +1q and −7q when it is present with 2 normal chromosomes 1 and 1 normal chromosome 7, was reported first in 3 patients who had myelofibrosis and myeloid metaplasia.31 Later case reports indicated that der(1;7)(q10;p10) often is associated with therapy-related MDS/AML32–34 but is rare in de novo cases35 and appears to carry an adverse prognosis. In a recent, multi-institutional study that included 64 patients who had primary and therapy-related MDS with der(1;7)(q10;p10), this abnormality appeared to carry a better clinical outcome with a median OS of 23 months.36 However, that study included both patients with der(1;7)(q10;p10) as an isolated abnormality and patients in whom it was present along with other cytogenetic abnormalities. In our series, we identified 6 patients with primary MDS who had der(1;7)(q10;p10), including 1 patient with RAEB-1, 2 patients with RCMD, and 3 patients with MDS-U. These patients had a median OS of 45.5 months, which was significantly superior to the OS of patients who had poor-risk cytogenetics (P = .01) (Table 1). One possible explanation for this finding is the beneficial effect of +1q, which, as a single abnormality, reportedly carries a good prognosis in patients with MDS.11, 14, 18 It is noteworthy that 5 of 6 patients (83%) who had der(1;7)(q10;p10) had hypocellular bone marrow (mean cellularity, 20%; range, 10%–40%). We have demonstrated that hypocellularity in MDS confers a favorable outcome that is independent of other risk factors, including cytogenetics.37 Although the underlying pathogenesis is not known, der(1;7)(q10;p10) appears to carry at least an intermediate prognosis if not a good prognosis. This observation is in keeping with our other finding that patients who had del(7q) as a single abnormality had a better OS than patients the poor category who had either −7 or a complex karyotype (26 months, 8.5 months, and 11 months, respectively; P < .05) (Table 2). It is noteworthy that some of the new therapeutic/experimental drugs may alter the natural course of these patients and may have a positive impact on OS.38–40 It is likely that the adverse biology of MDS can be affected for the better by certain drugs, and it is possible that cytogenetic IPSS risk categorization may be affected by it.
In summary, cytogenetics in MDS plays an important role in disease prognosis and risk stratification for treatment. Our study, which was conducted on a large cohort of patients with primary MDS from the US, confirmed that most of the cytogenetic groups could be defined according to the current IPSS. However, our data also indicated that some single cytogenetic abnormalities, such as del(7q) and der(1;7)(p10;q10), may be excluded from the original IPSS poor-risk category and reclassified into a different category, such as intermediate-risk group. Del(12p) and del(11)(q23), which currently are included in the intermediate-risk group, may belong to the good-risk group; patients who have rearrangement of 3q and +19 are likely to have been assigned appropriately to the intermediate-risk group; and i(17q) and +11 may carry a worse prognosis than their current assignment as intermediate risk. Overall, in patients with primary MDS, the single chromosomal abnormalities other than +8 in the intermediate category are very infrequent, and a multi-institutional study is needed to better define their prognostic significance.