Fax: (507) 266–4972
Prognostic diversity among cytogenetic abnormalities in myelofibrosis with myeloid metaplasia†
Article first published online: 15 AUG 2005
Copyright © 2005 American Cancer Society
Volume 104, Issue 8, pages 1656–1660, 15 October 2005
How to Cite
Tefferi, A., Dingli, D., Li, C.-Y. and Dewald, G. W. (2005), Prognostic diversity among cytogenetic abnormalities in myelofibrosis with myeloid metaplasia. Cancer, 104: 1656–1660. doi: 10.1002/cncr.21358
Presented as an abstract at the 2004 American Society of Hematology Meeting, San Diego, California, December 4–7, 2004.
- Issue published online: 3 OCT 2005
- Article first published online: 15 AUG 2005
- Manuscript Accepted: 16 MAY 2005
- Manuscript Revised: 30 APR 2005
- Manuscript Received: 8 FEB 2005
- cytogenetic abnormalities;
- myeloid metaplasia;
- leukemic transformation;
Approximately 30–50% of patients with myelofibrosis with myeloid metaplasia (MMM) demonstrate detectable cytogenetic abnormalities, the prognostic value of which has not been completely defined by previous retrospective studies. The current prospective study addresses this issue in the context of currently accepted independent prognostic variables.
The current study is a single institution study in which patients with MMM were accrued between January 2000 and August 2001 and followed in a prospective fashion. All study patients underwent bone marrow examination with cytogenetic studies as well as comprehensive clinical and laboratory evaluation at the time of karyotype analysis.
Among the study cohort of 81 patients (with a median age of 61 years; 54 males), the cytogenetic findings were normal in 44 patients (54%; Group 1). The remaining 37 patients (46%) demonstrated either interstitial deletions involving the long arm of chromosome 13 or 20 (9 patients; Group 2) or other abnormalities (28 patients; Group 3). All study patients were followed prospectively for a minimum of 40 months (range, 40–55 months). Survival from the time of karyotypic analysis was found to be similar between Groups 1 and 2 but was significantly worse in Group 3. Furthermore, none of the patients in Group 2 experienced leukemic transformation, whereas five patients each from the other two groups did. Multivariate analysis identified an unfavorable cytogenetic profile (Group 3), ≥ 1% circulating blasts, a hemoglobin level of <10 g/dL, and constitutional symptoms as adverse prognostic features for overall survival.
Specific cytogenetic lesions in patients with MMM might carry an independent prognostic effect for both survival and the risk of leukemic transformation. Such information should assist in decision making when considering aggressive treatment approaches. Cancer 2005. © 2005 American Cancer Society.
Large retrospective studies in patients with MMM report an overall median survival ranging from 3.5–10 years.1–4 Factors that have been associated with inferior survival include anemia (hemoglobin level <10 g/dL),1–5 constitutional symptoms,2, 4 circulating blasts (≥ 1–3%),1, 2, 4 leukocytosis (> 30,000/μL),1, 4 leukopenia (< 3000/μL),1, 4 age > 60 years,1–6 male gender,1, 4 thrombocytopenia (< 100,000/μL),1, 4 the percentage of circulating immature myeloid cells,5, 6 hepatomegaly,1 weight loss,1 a markedly increased lactate dehydrogenase (LDH) level,2 and cytogenetic findings.1, 7–9 Among these, a hemoglobin level of <10 g/dL, constitutional symptoms, ≥ 1% circulating blasts, and extreme ranges of leukocyte counts were the factors that usually were found to maintain their prognostic value on multivariate analysis.1, 10 Accordingly, these latter parameters have been utilized effectively to construct prognostic scoring systems that are widely used in current practice.1, 10 However, a recent retrospective study of 165 patients with MMM suggested the possibility that the prognostic value of cytogenetic findings in MMM might be refined by stratifying the observed abnormalities as being either “favorable” (sole abnormalities representing interstitial deletions of the long arms of either chromosome 13 or 20) or “unfavorable” (abnormalities other than 13q- and 20q-).8 The current prospective study examines this possibility in the context of currently accepted independent prognostic variables.
MATERIALS AND METHODS
The diagnosis of MMM was based on previously published criteria.11, 12 Both de novo MMM (agnogenic myeloid metaplasia [AMM]) and postpolycythemic (PPMM) and postthrombocythemic (PTMM) myeloid metaplasia were included in the current study. Special attention was given not to include patients with myelodysplastic syndrome associated with myelofibrosis.13 Bone marrow biopsy slides prepared from paraffin-embedded blocks were stained with hematoxylin and eosin and each bone marrow specimen was examined and subsequently graded for both reticulin fibrosis (Grades 1–4) and osteosclerosis (Grades 0–3) according to previously published criteria.14 The circulating CD34+ cells were quantified using direct immunofluorescence flow cytometry according to previously described methods.15 For cytogenetic studies, both direct technique and unstimulated 24-hour culture methods were used to harvest 20 metaphases, whenever possible, in all bone marrow specimens.16 For the purposes of the current communication as well as based on previously published retrospective data,8 we have referred to single abnormalities of either chromosome 13q- or 20q- as “favorable” and all other abnormalities as “unfavorable” cytogenetic profiles.
Descriptive and statistically analyzed data were obtained from the entire cohort of patients at the time of study entry (i.e., the time of cytogenetic evaluation). A variety of univariate techniques were applied, including the Fisher exact test for categoric variables and the Kruskal-Wallis and the Wilcoxon rank-sum test for continuous variables to explore clinical correlates of specific cytogenetic profiles. Survival was calculated using Kaplan-Meier plots taking the interval from the date of karyotype analysis to death or last contact and checked for significance by both the log-rank and Breslow-Gehan tests. A similar procedure was used to compare leukemic transformation rates among different cytogenetic groups. A Cox proportional hazards regression analysis was used to assess the prognostic relevance of key prognostic variables. To avoid the problem with multiple tests of significance, our multivariate model included only those variables that are currently considered to have independent prognostic value in MMM (≥ 1% circulating blasts, a hemoglobin level of <10 g/dL, constitutional symptoms, or a leukocyte count of either >30 or <4 × 109/L).1, 10 All data were analyzed using SAS software (SAS Institute, Inc., Cary, NC).
A total of 81 consecutive patients with MMM (median age of 61 years [range, 26–78 years]; 54 males) underwent bone marrow examination with cytogenetic studies as well as comprehensive clinical and laboratory evaluation at the study institution between January 2000 and August 2001. The study population was selected from 87 successive cases of MMM among whom cytogenetic studies were successfully performed in 81 patients. In the majority of the study patients, the number of metaphases analyzed was at least 20. Clinical and laboratory information at the time of karyotypic analysis is outlined in Table 1. At that point, cytogenetic results were normal in 44 patients (54%) and abnormal in the remaining 37 patients (46%), including 9 patients (11%) with favorable and 28 patients (35%) with unfavorable cytogenetic profiles (Table 2). Comparison of the three cytogenetic groups (normal vs. favorable vs. unfavorable) did not reveal differences with regard to age, gender, or MMM subtype distributions. Similarly, neither the time of diagnosis (new vs. previously diagnosed) nor the duration of disease before study entry were found to differ among the three cytogenetic groups. However, low-risk disease according to the Lille scoring system1 was more than represented by patients with normal cytogenetic results, and high-risk disease was more than represented by patients with an unfavorable cytogenetic profile (P = 0.03 for the latter association). Consistent with this observation, significant statistical associations also were demonstrated with both a hemoglobin level of <10 g/dL (P = 0.03) and a leukocyte count of >30 or <4 × 109/L (P = 0.001). Conversely, specific cytogenetic profiles did not demonstrate any significant statistical associations with either the presence of circulating blasts (P = 0.86) or constitutional symptoms (P = 0.2). It should be remembered that the latter two variables are integral to one of two popular prognostic scoring systems in MMM.10
|Parameter||n = 81|
|Median age (range)||61 ys (26–78 yrs; 54)|
|No. of males|
|Distribution of cases at the time of karyotypic analysis||No. of patients|
|Newly vs. previously diagnosed||16 vs. 65|
|Receiving active cytoreductive therapy||14|
|Previously splenectomy performed||6|
|Red blood cell transfusion dependent||24|
|Hemoglobin level < 10 g/dL||39|
|Presence of circulating blasts||43|
|Presence of constitutional symptoms||23|
|Leukocyte count either > 30 or < 4 × 109/L||27|
|Lille† score of 0/1/2||33/31/17|
|Median interval between the initial diagnosis and karyotype analysis in previously diagnosed cases (n = 65) (range)||43 mos (4–221 mos)|
|Measurements at the time of karyotype analysis||Median and range|
|• Hemoglobin in transfusion-independent cases (g/dL)|
|• Leukocyte count (× 109/L)||10.9 (7–14.9)|
|• Platelet count (× 109/L)||10.1 (1.3–109.5)|
|• % circulating blasts||189 (14–1007)|
|• Peripheral blood CD34 cell count (× 106/L)||1 (0–17)|
|• Spleen size in non-splenectomized patients (cm below LCM)||52.9 (0–5345)|
|Bone marrow histology||No. of patients|
|• Reticulin fibrosis/osteosclerosis|
|Bone marrow cytogenetic findings at the time of study entry||No. of patients|
|• Sole abnormalities of either 13q- or 20q-||44|
|• Cytogenetic abnormalities other than 13q- or 20q-||9|
|Lesion||No of patients|
At the time of last follow-up, all patients had been followed for a minimum of 40 months (range, 40–55 months) from the time of their cytogenetic evaluation at study entry. At the time of last follow-up (November 2004), 37 patients (46%%) had died, including 10 patients from leukemic transformation. Univariate analysis confirmed the adverse prognostic relevance to overall survival of circulating blasts ≥1% (P = 0.0004), a hemoglobin level of <10 g/dL (P = 0.0005), constitutional symptoms (P = 0.004), and a leukocyte count of either >30 or <4 × 109/L (P = 0.02).1, 10 In addition, the presence of cytogenetic abnormalities other than 13q- and 20q- also was found to be associated with inferior survival (P = 0.01) (Fig. 1). On multivariate analysis, all the aforementioned variables with the exception of the leukocyte count retained their significance. When the analysis was restricted to patients with AMM, despite the substantial reduction in sample size (60 cases rather than 81 cases), an unfavorable cytogenetic profile maintained its independent prognostic value when analyzed together with the two strongest prognostic variables (circulating blasts ≥1% and a hemoglobin level <10 g/dL) using multivariate analysis,. Furthermore, during the study period of 40–55 months, none of the 9 patients with a favorable cytogenetic profile (13q-, 20q-) developed a transformation into acute leukemia whereas 5 patients each from the other 2 groups did (P = 0.02 by the Breslow-Gehan test and P = 0.19 by the log-rank test).
The most frequently encountered cytogenetic abnormalities in MMM include 13q-; 20q-; +8; and abnormalities of chromosomes 1, 7, and 9.7–9, 17 In general, the incidence of each individual lesion in MMM is less than 20%; therefore, such lesions are considered secondary rather than primary genetic events.8 Regardless, their occurrence could influence disease progression depending on the nature of the underlying molecular alterations. In this regard, the observations from both the current prospective study and the immediately preceding larger retrospective study8 strongly suggest a prognostically indolent nature for sole abnormalities of either chromosome 13q- or 20q-. This was manifested by an overall survival that was not significantly different from that of patients with normal cytogenetic findings as well as remarkably low rates of leukemic transformation. In the current study, none of the 9 patients with a favorable cytogenetic profile converted to acute leukemia after a prospective follow up period of 40–55 months. Similarly, in the retrospective study we published previously, the rate of leukemic transformation was 0% with 13q- anomalies (19 cases) and 10% with 20q- anomalies (20 cases).8
The significantly lower risk of transformation into acute leukemia that was linked to 13q- and 20q- abnormalities suggests a reduced vulnerability to clonal evolution that may be imparted by genetic alterations linked to the specific structural lesions. Such a contention is consistent with previous observations regarding the molecular characterization of 13q- in bcr/abl-negative chronic myeloid disorders that revealed microdeletions that did not involve the RB1 tumor suppressor gene locus.18 In contrast, the RB1 locus is commonly deleted in the prognostically detrimental occurrence of 13q- in multiple myeloma.19, 20 With regard to 20q- abnormalities, a recent study involving 113 patients with myeloid malignancies that were associated with the particular chromosome abnormality defined commonly deleted regions that were specific to either myeloproliferative disorders or myelodysplastic syndrome/acute myeloid leukemia.21 Further inquiry into the specific genes that are located in the commonly deleted regions of myeloproliferative disorder-associated 20q- might shed light on the molecular explanation for the current clinical observations. Regardless, it should be noted that 20q- abnormalities in MDS also have been associated with both prolonged survival and lower rates of leukemic transformation.22, 23
We believe that the current study provides potential explanation for the discrepancy noted among previous reports regarding the prognostic relevance of cytogenetic findings in MMM.1, 7–9 Some studies suggested that the presence of cytogenetic abnormalities portended inferior survival,7 whereas other studies were not able to make similar observations.10 Obviously, sample size limitations do not allow valid comparisons across studies but we hypothesized that part of the reason for the divergent conclusions stemmed from a differential effect on prognosis linked to specific cytogenetic abnormalities. Accordingly, we were able to demonstrate, in a previous retrospective study, a significant difference in survival in patients with either 13q- or 20q- lesions when compared with those with either trisomy 8 or 12p- lesions.8 The current study supports this view and suggests that the presence of cytogenetic abnormalities other than 13q- and 20q- is an independent indicator of shortened survival in patients with MMM. Whether cytogenetic information should be incorporated into a new prognostic scoring system depends on the validation of the observations from the current study in a much larger group of patients evaluated from the time of the initial diagnosis. However, until that time, the additional information might facilitate decision making when aggressive treatment approaches are suggested.24