Elsevier

Leukemia Research

Volume 39, Issue 11, November 2015, Pages 1146-1153
Leukemia Research

Comparison of three prognostic scoring systems in a series of 146 cases of chronic myelomonocytic leukemia (CMML): MD Anderson prognostic score (MDAPS), CMML-specific prognostic scoring system (CPSS) and Mayo prognostic model. A detailed review of prognostic factors in CMML

https://doi.org/10.1016/j.leukres.2015.05.017Get rights and content

Highlights

  • Extensive and actualized review of the novel defined prognostic factors in CMML.

  • MDAPS, CPSS and Mayo prognostic model are excellent to distinguish low and high-risk patients.

  • Platelets < 100 × 109/L is the independent prognostic factor with the highest weight in our series.

  • CPSS-P was constructed after including the platelet count information to the CPSS.

  • CPSS-P was the index with the higher capability for predicting survival and AML transformation.

Abstract

Although specific prognostic models for chronic myelomonocytic leukemia (CMML) exist, few are based on large series of patients. MD Anderson prognostic score (MDAPS) has been the most useful for CMML risk assessment. Due to recent emergence of CMML-specific prognostic scoring system (CPSS) and Mayo prognostic model, we compared the three scores.

One hundred forty-six CMML patients diagnosed between 1998 and 2014 were retrospectively analyzed. Univariate analysis was performed to assess prognostic impact on overall survival (OS) and leukemia-free survival (LFS) of the variables composing the scores and all items showed prognostic value on OS with the exception of the presence of circulating immature myeloid cells. Regarding LFS, only CPSS variables, bone marrow blast ≥10% and an absolute monocyte count >10 × 109/L had an impact. When the scores were applied, all showed an impact on OS and retained their significance in multivariate analysis. By using ROC curves and C-index, CPSS showed a slightly better predictive value for mortality and leukemia transformation. Variables composing the three indexes were compared in multivariate analysis and only CPSS parameters and platelets < 100 × 109/L retained their significance. Based on these findings, by adding platelet count to CPSS, a new score was implemented (CPSS-P) showing the best risk prediction capability in our series. This study reinforces the validity of the tested scores.

Introduction

Chronic myelomonocytic leukaemia (CMML) is a clonal haematopoetic malignancy with overlapping features of myeloproliferative neoplasms and myelodysplastic syndromes (MDS). It is characterized by persistent monocytosis (monocytes > 1 × 109/L in peripheral blood for at least 3 months), absence of BCR-ABL1 fusion gene, no rearrangement of PDGFRA or PDGFRB and fewer than 20% myeloblasts and promonocytes in the peripheral blood (PB) and bone marrow (BM). Dysplasia involving one or more myeloid lineages is almost always present but the diagnosis of CMML can still be made if the other requirements are met and dysplasia is not present [1].

The French–American–British (FAB) classification distinguished in 1994 two subtypes of CMML according to the leukocyte count in peripheral blood: the myelodysplastic subtype (CMML–MD) with 13 × 109/L leukocytes or less, and the myeloproliferative subtype (CMML–MP) with a leukocyte count above 13 × 109/L [2]. Based on the 2008 WHO classification, CMML is subdivided into CMML-1 (<5% PB blasts and <10% BM blasts) and CMML-2 (5–19% PB blasts and 10–19% BM blasts or when Auer rods are present irrespective of the blast count) [1].

Median survival of patients with CMML is reported to vary from 18 to 24 months and progression to AML occurs in approximately 15–30% of cases [3], [4], [5], [6], [7], [8], [9]. Although several prognostic models have been created to stratify CMML patients into different risk categories, there is a paucity of specific prognostic scores based in large series of patients [8], [9], [10], [11], [12]. In this regard, the MD Anderson prognostic scoring system (MDAPS) published in 2002 was the first score based in a large cohort of 213 CMML patients [13]. This identified four independent adverse prognostic factors: hemoglobin level <12 g/dl, absolute lymphocyte count (ALC) > 2.5 × 109/L, presence of circulating immature myeloid cells (IMC) and bone marrow (BM) blasts  10%. In the last two years four new specific prognostic scores have been developed to stratify risk of CMML patients: CMML-specific prognostic scoring system (CPSS) [14], Groupe Français des Myélodysplasies prognostic score (GFM prognostic score) [15], Mayo prognostic model for WHO-defined CMML [16] and Mayo molecular model (MMM) [17]. The CPSS was based in a cohort of 558 patients and included FAB classification (CMML–MD vs. CMML–MP), WHO classification (CMML-1 vs. CMML-2), red blood cell (RBC) transfusion dependency at diagnosis and the Spanish cytogenetic risk classification [5], [14]. The GFM prognostic score was developed from a series of 312 patients and was the first that included gene mutation status information and clinical parameters, defining a prognostic score based on: age >65 years, leukocytes >  15 × 109/L, anemia (hemoglobin < 10 g/dL in women and <11 g/dL in men), platelets < 100 × 109/L and presence of ASXL1 mutation [15], [16], [17], [18], [19], [20], [21], [22]. The Mayo prognostic model was based in a series of 226 patients and, interestingly, the adverse independent prognostic impact of ASXL1 mutation was not observed. This index identified three risk categories based on: absolute monocyte count (AMC) > 10 × 109/L, presence of circulating immature myeloid cells, hemoglobin < 10 g/dL and platelet count < 100 × 109/L [16]. Finally, after combining the data of 466 patients from the GFM and Mayo clinic group, ASXL1 mutation (nonsense and frameshift) confirmed its independent prognostic value in terms of overall survival but not on leukemic transformation. By the inclusion of ASXL1 mutational status to the Mayo prognostic model, the novel Mayo molecular model (MMM) was created [17].

A series of 146CMML patients was used to assess the usefulness of the CPSS, the MDAPS and the Mayo prognostic model. These three scores, and not those based on clinical and molecular variables, were selected as the most easy-to-apply in normal clinical practice. The prognostic value of each variable that composed every score was tested and a comparison of the three indexes was made. A modified CPSS based on our results was performed in order to get a better risk stratification of patients.

Section snippets

Patients

From January 1997 to August 2013 a retrospective analysis including 146 patients diagnosed with CMML was performed in Hospital Clínic of Barcelona (n = 134) and Hospital Universitari Germans Trias i Pujol (n = 12). CMML was morphologically defined according to the FAB and WHO 2008 criteria.

In the analysis of prognostic factors, variables analyzed were those included in the CPSS [14], MDAPS [13] and Mayo prognostic model [16]: absolute monocyte count (AMC), hemoglobin levels, platelet count,

Baseline characteristics

The main characteristics of 146 patients with WHO-defined CMML are shown in Table 1. The median age was 76 years (range 27–96 years) and 63% were males. One-hundred and twenty-nine (88%) had a CMML-1, 17 (12%) had a CMML-2, one-hundred and two (70%) had a CMML–MD and forty-four (30%) had a CMML–MP. Cytogenetic risk according to the Spanish CMML classification was assessed in one-hundred and twenty-six (86%) patients being low, intermediate, high or not available in 68%, 7%, 12% and 13%,

Discussion

This study, based on a series of 146CMML patients, confirms the prognostic value of the scores assessed: CPSS, MDAPS and Mayo prognostic model.

CMML patients present overlapping features of myeloproliferative neoplasms and MDS adding increasing ambiguity in terms of diagnosis and prognosis. Despite the existence of specific prognostic scoring systems [8], [9], [13], [14], [15], [16], [17], [24], IPSS [11] has been the most used score for the evaluation of CMML, although it is not applicable for

Conflict of interests

None of the authors have any conflicts of interest or disclosures.

Author contributions

XC designed the study, collected data, performed statistical analysis and wrote the paper. MN, RS, NM and DC collected data and reviewed the paper. AP performed statistical analysis and reviewed the paper. NE, BX and JE collected data and reviewed the paper. BN designed the study, analyzed data and reviewed the paper.

Acknowledgements

This work has been supported by the “Emili Letang” Grant (2011) from Hospital Clínic of Barcelona.

References (26)

  • P. Fenaux et al.

    Prognostic factors in adult chronic myelomonocytic leukemia: an analysis of 107 cases

    J. Clin. Oncol.

    (1988)
  • U. Germing et al.

    Risk assessment in chronic myelomonocytic leukemia (CMML)

    Leuk. Lymphoma

    (2004)
  • A. Worsley et al.

    Prognostic features of chronic myelomonocytic leukaemia: a modified Bournemouth score gives the best prediction of survival

    Br. J. Haematol.

    (1988)
  • Cited by (14)

    • Outcomes and molecular profile of oligomonocytic CMML support its consideration as the first stage in the CMML continuum

      2022, Blood Advances
      Citation Excerpt :

      Otherwise, it is important to assess whether the specific CMML prognostic factors5,6,7,8,9,10,11,12 applied to OM-CMML provide better predictive accuracy of outcome than classic MDS prognostic variables,13,14 given that most patients are currently categorized as having MDS according to the World health Organization (WHO) 2017 classification.3,15,16 Although most of the prognostic factors described in MDS are also consistent prognostic determinants in CMML, there are some that are specific to the latter (eg, trisomy 8 as a high-risk cytogenetic abnormality, and blast cell count defined as the sum of blasts and promonocytes).5,6,7,8,9,10,11,12 Of note, leukocytosis is a well-recognized independent prognostic factor in CMML,6,7,10,12 but it is not applicable to OM-CMML prognostic stratification, because the presence of ≥10 × 109/L leukocytes implies having at least 1 × 109/L monocytes and therefore meeting CMML diagnostic criteria.

    • Moving towards a uniform risk stratification system in CMML - How far are we?

      2020, Best Practice and Research: Clinical Haematology
      Citation Excerpt :

      It was developed based on a large cohort (n = 558) of CMML patients compared to the other models, and it was externally validated by 274 patients from the Germany group. Like the MDAPS-M1, minor revision had been proposed for CPSS to include platelets <100 × 109/L (known as CPSS-P) to the initial model after more analysis using a different cohort of patients was done [26]. Notably, all existing CMML models are established using both patients who are eligible and ineligible for allogeneic hematopoietic cell transplantation (HCT), the only potentially curative treatment for CMML at this time.

    • Blastic Transformation in Mexican Population With Chronic Myelomonocytic Leukemia

      2017, Clinical Lymphoma, Myeloma and Leukemia
      Citation Excerpt :

      The factors related to blastic transformation are well known and have been consistently related to shorter LFS. In our study, differences in the cytogenetic abnormalities,18-20 female sex,21 transfusion dependency, and the hemoglobin, platelet, neutrophil, lymphocyte, and absolute monocyte count were not statistically significant and were not predictive of a greater risk of progression to AML.12-15 Of the different prognostic scores used, only the Mayo Clinic prognostic model was able to clearly differentiate among the 3 risk groups.

    View all citing articles on Scopus

    Presented at the 56th annual meeting of the American Society of Hematology, San Francisco, CA, December 2014.

    View full text