Acute Leukemia

Having a higher blast percentage in circulation than bone marrow: clinical implications in myelodysplastic syndrome and acute lymphoid and myeloid leukemias

Abstract

Determining the percentage of peripheral blood (PB) and bone marrow (BM) blasts is important for diagnosing and classifying acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS). Although most patients with acute leukemia or MDS have a higher percentage of BM blasts than PB blasts, the relative proportion is reversed in some patients. We explored the clinical relevance of this phenomenon in MDS (n=446), AML (n=1314), and acute lymphoblastic leukemia (ALL) (n=385). Among patients with MDS or ALL, but not AML, having a higher blast percentage in PB than in BM was associated with significantly shorter survival. In multivariate analyses, these associations were independent of other relevant predictors, including cytogenetic status. Our findings suggest that MDS and ALL patients who have a higher percentage of PB blasts than BM blasts have more aggressive disease. These data also suggest that MDS classification schemes should take into account the percentage of blasts in PB differently from the percentage of blasts in BM.

Introduction

Blast percentage plays a central role in the diagnosis and classification of acute leukemias and myelodysplastic syndromes (MDS). The French–American–British (FAB) classification requires a blast percentage of at least 30% in bone marrow (BM) or peripheral blood (PB) for the diagnosis of acute myeloid leukemia (AML),1, 2 and also requires specific blast percentages to subclassify MDS into refractory anemia with excess blasts (RAEB) and RAEB in transformation (RAEBt): patients with <5% PB or 5–20% BM blasts are considered to have RAEB, whereas those with 5% PB or 21–29% BM blasts have RAEB in transformation (RAEBt).3 In contrast, the World Health Organization (WHO) classification does not include the RAEBt category, as it decreases the cutoff limit for the diagnosis of AML from 30 to 20% BM or PB blasts.4 Thus, patients with up to 19% PB or BM blasts are considered to have RAEB, which is subdivided as RAEB-1 (<5% PB blasts and 5–9% BM blasts) and RAEB-2 (5–19% PB blasts or 10–19% BM blasts).3

The percentage of PB and BM blasts is not as important for the diagnosis of acute lymphoblastic leukemia (ALL), because the presence of any clonal blast population is diagnostic. However, post-therapy PB blast percentage is an important prognostic index that reflects the outcome in ALL.5

The diagnosis of leukemia and MDS is based on BM blasts because, in most cases, the percentage of blasts is higher in BM than in PB. In a small proportion of patients with acute leukemia known as ‘peripheral leukemia,’ however, there is no diagnostic increase in BM blast percentage; the diagnosis is based on the presence of at least 20% PB blasts. Although it is believed that homing molecules play a role in determining why some blasts tend to circulate in PB and some do not, the clinical relevance of circulating blasts is not well studied. In this retrospective study, we investigated the clinical and prognostic significance of having a higher percentage of blasts in PB than in BM in patients with MDS, AML, ALL, or CMML.

Patients and methods

Patients and laboratory data

Previously untreated patients admitted to the adult leukemia Department at MD Anderson Cancer Center with a diagnosis of MDS, AML, or ALL from 1992 to 2002 were selected for this study. Only patients treated at MD Anderson were included in the analysis. The diagnosis was based on morphologic, cytochemical, and immunophenotypic studies. Cytogenetic analyses and molecular diagnostic studies were also performed.

Disease status was classified according to the FAB system.1, 2 When abnormal cells were detected in PB, manual differential counting of leukocytes was performed. PB smears obtained from all patients with leukemia were evaluated for the presence of blasts, irrespective of whether the automated differential leukocyte count was flagged. BM blast counts were based on a total of 500 cells obtained from four different BM aspirate smears and/or touch imprints, whenever possible. Total leukocyte, platelet, and absolute lymphocyte counts and hemoglobin, β-2 microglobulin (B2M), LDH levels were assessed with standard techniques. Cytogenetic and molecular diagnostic studies were performed on most of the patients.

Statistical analyses

The association between the clinical characteristics of the patients and the relative percentage of blasts in BM and PB was studied. In each case, performance status was classified as 0 or 1 vs 2, 3, or 4. Antecedent hematologic disease (AHD) classified as positive or negative. Cytogenetic findings in AML and MDS were categorized as favorable (t(8;21), t(15;17) or inv16), unfavorable (−5, −7 or 11q abnormalities), or intermediate (others). Similarly, cytogenetic status in ALL patients was classified as favorable (hyperdiploidy), unfavorable (t(8,14) or t(9;22)), or intermediate (others) groups.

Spearman's rank correlation coefficient was used to assess correlations between variables. The χ2 test was used to investigate dependence between two categorical variables. The Wilcoxon rank sum (or Kruskal–Wallis) test was used to compare groups of continuous independent variables. The Kaplan–Meier product-limit method was used to estimate the survival distribution. The two-sided log-rank test was used to test the association between single categorical variables and survival. Martingale residual plots were used to investigate the association of continuous factors with overall survival. Multivariate analysis was performed using the Cox proportional hazards regression model to determine the associations of two markers with survival after adjusting for the roles of other factors. P-values <0.05 were considered statistically significant; all P-values presented are two-sided. Statistical analyses were performed using SAS (version 8.2, SAS Institute Inc., Cary, NC, USA) and S-PLUS 6 (Insightful, Seattle, WA, USA) software programs.

Results

Correlation of blast percentages with laboratory characteristics

The characteristics of the 2201 consecutive patients who met the inclusion criteria and were included in the study are shown in Table 1. All patients were treated with regimens containing ara-C and/or idarubicin, with or without fludarabine and topotecan. The minor differences in therapy did not result in a significant difference in outcome among the therapy arms (Cox regression analysis; P=0.31 for AML, 0.81 for MDS, and 0.08 for ALL).

Table 1 Patient characteristics

In all the four disease groups, most patients had a higher blast percentage in BM than in PB. However, 341 (26%) of the 1314 patients with AML, 58 (13%) of the 446 patients with MDS, 35 (9%) of the 385 patients with ALL, and 2 (3%) of the 58 patients with CMML had a higher blast percentage in PB than in BM. As a result of the small number of CMML patients with a higher percentage of blasts in PB than in BM, this group was not included in subsequent analyses.

Table 2 shows the correlations of laboratory variables with PB and BM blast percentages, and indicates which levels varied according to the relative blast percentages in PB and BM. In AML patients, PB blast percentage correlated significantly with white blood cell (WBC) count, absolute lymphocyte count, and lactate dehydrogenase (LDH) and B2M levels. MDS and AML patients with a higher percentage of blasts in PB than in BM had significantly higher WBC counts, absolute lymphocyte counts, and LDH and B2M levels. Among ALL patients, those with higher blast percentages in PB than in BM also had significantly higher absolute lymphocyte and WBC counts.

Table 2 Laboratory variables in AML, MDS, and ALL: correlation with blast percentage in PB and BM and association with relative percentage of blasts (PB>BM blast % vs BM>PB blast %)

PB blasts and survival

In the initial univariate survival analyses we studied most of the known relevant factors, including cytogenetic status, performance status, AHD, age, LDH level, platelet count, B2M level, absolute lymphocyte count, and blast percentage in PB alone and BM alone, as well as the presence of a higher blast percentage in PB than in BM. Only significant associations are listed in Table 3. In MDS patients, overall survival was associated with cytogenetic status, performance status, AHD, age, LDH level, platelet count, International Prognostic Scoring System (IPSS) and B2M level. MDS patients who had a higher percentage of blasts in PB than in BM had significantly shorter survival (Table 3; Figure 1). WBC count was not associated with survival in MDS patients. Although the percentage of blasts in PB was associated with survival, the percentage of blasts in BM, when considered alone, was not: median overall survival was 9.7 months for patients with <10% BM blasts and 10.1 months for those with 10–20% BM blasts.

Table 3 Univariate survival analysis in MDS, ALL, and AML patientsa
Figure 1
figure1

Survival in patients with MDS according to the relative percentage of blasts in PB and BM. Survival rates were significantly lower in patients with a higher percentage of blasts in PB than in BM. (Number of dead/total number of patients is shown.)

In patients with ALL, survival was associated with cytogenetic status, performance status, WBC count, absolute lymphocyte count, age, platelet count, and B2M level (Table 3). Similar to MDS patients, ALL patients with a higher percentage of blasts in PB than in BM also had significantly shorter survival (Table 3; Figure 2). However, neither the PB nor the BM blast percentage alone was associated with survival (Table 3).

Figure 2
figure2

Survival in patients with ALL according to the relative percentage of blasts in PB and BM. Survival rates were significantly lower in patients with a higher percentage of blasts in PB than in BM. (Number of dead/total number of patients is shown.)

In the AML group, overall survival was associated with cytogenetic status, performance status, AHD, age, LDH level, platelet count, and B2M level (Table 3). The presence of a higher percentage of blasts in PB than in BM was not associated with survival, despite the fact that the both the PB and the BM blast percentage, when considered individually, were (Table 3; Figure 3).

Figure 3
figure3

Survival in patients with AML according to the relative percentage of blasts in PB and BM. Survival rates did not differ significantly between patients with a higher percentage of blasts in PB than in BM and those with a higher percentage of blasts in BM than in PB. (Number of dead/total number of patients is shown.)

In a multivariate model incorporating all factors found to be significant in univariate analysis, cytogenetic status and the presence of a higher blast percentage in PB than in BM were retained as independent prognostic indicators in MDS (Table 4). Furthermore, multivariate analysis incorporating IPSS classification and having higher blasts in PB than BM showed that both were independent of each other in predicting survival. In ALL, only the presence of a higher percentage of blasts in PB than in BM was independently associated with survival. In AML, the blast percentages in PB and BM alone were not independent prognostic factors.

Table 4 Multivariate analysis of factors associated with shorter survival in MDS and ALL patients

Discussion

Our results suggest that having a higher blast percentage in PB than in BM is associated with more aggressive disease in MDS and ALL, but not in AML. Importantly, this association was independent of other key prognostics factors, including cytogenetic status. The exact biological explanation for the relative increase in PB blast percentage in some patients with MDS or acute leukemia is not known.

The presence of higher blast percentages in PB than in BM was most common in AML (26%), followed by MDS (13%) and ALL (9%). Very few CMML patients (3%) had higher blast percentages in PB than in BM, most likely because of the ability of leukemic cells to differentiate in chronic leukemia. Recent studies showed that the presence of an increased number of myeloid colony-forming units in the PB of patients with MDS is associated with decreased survival and a tendency toward transformation into AML.6, 7

Little is known about the exact factors that control the mobilization of BM blasts into PB, although adhesion molecules, chemokines, and angiogenic factors are believed to play a role. We previously showed that angiogenin levels are high in patients with AML or advanced MDS.8 Aref et al9 found increased soluble serum hepatocyte growth factor and serum vascular endothelial growth factor levels in patients with AML and extramedullary involvement; increased serum hepatocyte growth factor levels were associated with worse clinical outcome in that study. Most likely, these and other factors interact with each other in a complex fashion to determine which blasts will circulate and which will stay in BM.10, 11, 12, 13

Irrespective of the cause, an increased tendency of blasts to circulate, when defined as the presence of higher blast percentages in PB than in BM, has prognostic value in ALL and MDS. The independent prognostic value of this phenomenon highlights the importance of exploring the causes behind it, and suggests that investigating its underlying causes may help in devising therapeutic strategies. It is possible that blasts in patients with relatively high PB blast percentages are more capable of invading extramedullary tissues, which may protect them from chemotherapy.

Our study raises an important question related to the relevance of considering the PB and BM blast percentage differently in the classification of MDS. As mentioned earlier, the WHO classification considers MDS patients with up to 19% PB blasts to have RAEB (RAEB-1 or RAEB-2), whereas the FAB classification considers patients with 5% PB blasts to have more aggressive disease (RAEB-T).14 Although the percentage of blasts in PB is prognostically important irrespective of BM percentage, having a higher percentage of blasts in PB than in BM is particularly important (Table 3). MDS patients who have a higher percentage of blasts in PB than in BM are also likely to have >4% blasts in PB (74% of our patients).

The reason that the relative blast percentage in PB and BM lacked clinical relevance in AML, even though the PB and BM blast percentages alone were each prognostic (Table 3), remains unclear. It is possible that the higher percentage of PB blasts reflects a greater tumor mass and is not necessarily related to homing. However, a significant number of AML patients have massive infiltration of BM by blasts, but lack circulating blasts. Investigation of the characteristics of PB cells in relation to BM blasts may not only allow us to better understand the biology of circulating blasts, but may also help in developing therapeutic strategies.

References

  1. 1

    Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposals for the classification of acute leukaemias. French–American–British (FAB) cooperative group. Br J Haematol 1976; 33: 451–458.

  2. 2

    Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French–American–British Cooperative Group. Ann Intern Med 1985; 103: 620–625.

  3. 3

    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: 189–199.

  4. 4

    Brunning RD, Bennett JM, Flandrin G, Flandrin G, Matutes E, Head D et al. Myelodysplastic syndromes: introduction. In: Jaffe ES, Harris NL, Stein H, Vardiman JW (eds). World Health Organization Classification of Tumors: Pathology and Genetics of Tumors of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press, 2001, pp 63–67.

  5. 5

    Griffin TC, Shuster JJ, Buchanan GR, Murphy SB, Camitta BM, Amylon MD . Slow disappearance of peripheral blood blasts is an adverse prognostic factor in childhood T cell acute lymphoblastic leukemia: a Pediatric Oncology Group study. Leukemia 2000; 14: 792–795.

  6. 6

    Vehmeyer K, Haase D, Alves F . Increased peripheral stem cell pool in MDS: an indicator of disease progression? Leuk Res 2001; 25: 955–959.

  7. 7

    Berer A, Jager E, Sagaster V, Streubel B, Wizamal F, Sperr WR et al. Circulating myeloid colony-forming cells predict survival in myelodysplastic syndromes. Ann Hematol 2003; 82: 271–277.

  8. 8

    Verstovsek S, Kantarjian H, Aguayo A, Manshouri T, Freireich E, Keating M et al. Significance of angiogenin plasma concentrations in patients with acute myeloid leukaemia and advanced myelodysplastic syndrome. Br J Haematol 2001; 114: 290–295.

  9. 9

    Aref S, Mabed M, Sakrana M, Goda M, El-Sherbiny M . Soluble hepatocyte growth factor (sHGF) and vascular endothelial growth factor (sVEGF) in adult acute myeloid leukemia: relationship to disease characteristics. Hematology 2002; 7: 273–279.

  10. 10

    Recher C, Ysebaert L, Beyne-Rauzy O, Mansat-De Mas V, Ruidavets JB, Cariven P et al. Expression of focal adhesion kinase in acute myeloid leukemia is associated with enhanced blast migration, increased cellularity, and poor prognosis. Cancer Res 2004; 64: 3191–3197.

  11. 11

    Thomas X, Anglaret B, Bailly M, Maritaz O, Magaud JP, Archimbaud E . Differential adhesiveness between blood and marrow leukemic cells having similar pattern of VLA adhesion molecule expression. Leuk Res 1998; 22: 953–960.

  12. 12

    Bradstock KF, Gottlieb DJ . Interaction of acute leukemia cells with the bone marrow microenvironment: implications for control of minimal residual disease. Lymphoma 1995; 18: 1–16.

  13. 13

    Voermans C, van Heese WP, de Jong I, Gerritsen WR, van Der Schoot CE . Migratory behavior of leukemic cells from acute myeloid leukemia patients. Leukemia 2002; 16: 650–657.

  14. 14

    Greenberg P, Anderson J, de Witte T, Estey E, Fenaux P, Gupta P et al. Problematic WHO reclassification of myelodysplastic syndromes. Members of the International MDS Study Group. J Clin Oncol 2000; 18: 3447–3452.

Download references

Acknowledgements

We thank Jeff Radcliff for his excellent help in preparing this article.

Author information

Correspondence to M Albitar.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Amin, H., Yang, Y., Shen, Y. et al. Having a higher blast percentage in circulation than bone marrow: clinical implications in myelodysplastic syndrome and acute lymphoid and myeloid leukemias. Leukemia 19, 1567–1572 (2005). https://doi.org/10.1038/sj.leu.2403876

Download citation

Keywords

  • blasts
  • acute leukemia
  • myelodysplastic syndrome
  • survival

Further reading