Accumulation of classical monocytes defines a subgroup of MDS that frequently evolves into CMML

To the editor:

Even though the diagnosis criteria of chronic myelomonocytic leukemia (CMML) have been recently revised by the World Health Organization (WHO),¹  recognition of this disease can be challenging. We demonstrated recently that a percentage of classical monocytes CD14⁺⁺CD16⁻ (MO1) ≥94% of total monocytes, as measured by flow cytometry, could rapidly and efficiently distinguish a CMML from a reactive monocytosis with a specificity of 95.1% and a sensitivity of 91.9%.²  The association between MO1 accumulation and CMML was subsequently validated³  and proposed as an additional diagnostic modality in CMML.^4,5  A relative monocytosis, defined as a percentage of peripheral blood monocytosis ≥10%, has been described in a subgroup of myelodysplastic syndrome (MDS) likely to evolve into genuine CMML.^6-9  Here, we compare the monocyte subset repartition in CMML and in MDS.

Between January 2015 and March 2017, we analyzed 158 CMML patients and 84 MDS patients for whom the diagnosis was made according to the 2008 WHO classification,¹⁰  following local ethical committee’s rules. Of the 158 patients with CMML, 152 (96%) fulfilled the latest 2016 WHO criteria that include both a persistent peripheral blood monocytosis ≥1 × 10⁹/L and monocytes accounting for ≥10% of the white blood cell (WBC) differential count.¹  These patients were subdivided into CMML-0 (62), CMML-1 (66), and CMML-2 (24) (Figure 1A; Table 1).^1,11  All peripheral blood parameters but platelet count were similar between these 3 groups (Figure 1B-E). CMML-2 patients displayed a significantly deeper thrombocytopenia (104 ± 90 × 10⁹/L) in comparison with CMML-1 (132 ± 93 × 10⁹/L; P < .05) and CMML-0 (204 ± 143 × 10⁹/L; P < .05). In the 124 cases in which the CMML Prognostic Scoring System (CPSS)¹²  could be tested, we did not notice any difference between CMML-0 and CMML-1, both mainly in the low and intermediate-1 categories, whereas 67% of CMML-2 were in the intermediate-2 and high groups. A small fraction of CMML-0 (19%) and CMML-1 (30%) were proliferative,¹  leading us to test also the Revised International Prognostic Scoring System (IPSS-R),^13,14  with most CMML-0 in the very low and low-risk categories, whereas most CMML-1 were in the low and intermediate groups (Table 1).

In the peripheral blood, the absolute count as well as the percentage of monocytes was similar in the 3 CMML groups. In the bone marrow, the percentage of mature and immature monocytes,¹⁵  excluding monoblasts and promonocytes,^1,10  was ≥5% in 89% of CMML patients and significantly higher in CMML-1 compared with CMML-0 (9% ± 5% vs 12% ± 7%; P < .05; Figure 1F). These results support the categorization of CMML into 3 groups, including CMML-0 with <2% peripheral blood and <5% medullary blasts.

MO1 accumulation ≥94% was identified in 141 of the 152 CMML (Figure 1A,G), indicating a sensitivity of 92.8% in accordance with our previous results.²  This sensitivity increased with CMML subtype, reaching 100% in CMML-2. The 11 CMML cases, according to the 2016 WHO criteria, displaying MO1 < 94%, included 7 CMML-0 and 4 CMML-1. Six of them, all with CMML-0, showed a monocyte subset repartition similar to that observed in healthy donors, notably including the presence of MO3 (CD14^lowCD16⁺) cells (Figure 1H). Molecular analyses and follow-up are then necessary to support or exclude CMML diagnosis in these cases. In the 5 other patients, including the 4 CMML-1 cases, a characteristic, easily recognized, “bulbous” aspect of monocyte subset repartition was observed (Figure 1I) due to the disappearance of MO3 subset combined with the increase of intermediate monocytes MO2 (CD14⁺⁺CD16⁺) subset. Each of these patients demonstrated an associated inflammatory state (eg, one of them had a typical ankylosis spondylitis), whereas the others had an elevated C-reactive protein (23.3 ± 21.0 mg/L). Accordingly, inflammation was previously shown to provoke an accumulation of MO2 subset.^16-18  For example, in a woman with a typical monocyte subset repartition at CMML diagnosis (Figure 1J), the occurrence of an auricular chondritis was associated with an increase in MO2 subset, generating a “bulbous” aspect of the flow image (Figure 1K), which disappeared with the resolution of inflammation after corticotherapy (Figure 1L). Therefore, interpretation of monocyte subset repartition at CMML diagnosis must remain cautious in case of a “bulbous” profile and should take into account an associated inflammatory situation.

In 6 of the 158 patients with a CMML according to 2008 WHO criteria, WBC differential count did not display ≥10% of monocytes. All these patients had a proliferative disease with hyperleukocytosis (35.6 ± 25.6 × 10⁹/L), fulfilled the other WHO 2016 criteria, displayed a marrow monocytosis ≥5% (8% ± 6%), showed an MO1 accumulation (97.4% ± 1.5%) by flow analysis, and demonstrated a molecular profile compatible with a CMML diagnosis. These observations suggest that a WBC count showing ≥10% of monocytes may not be an absolute criterion for CMML diagnosis.

Among the 152 patients with a CMML diagnosis according to WHO 2016 criteria, 15 had evolved from a preexisting MDS with excess of blasts (MDS-EB, N = 7), multilineage lineage dysplasia (MDS-MLD, N = 4), single lineage dysplasia (N = 2), ring sideroblasts and multilineage dysplasia (N = 1), and isolated del(5q) (N = 1). At MDS diagnosis, their peripheral blood monocyte count was 0.7 ± 0.1 × 10⁹/L. A marrow monocytosis ≥5% was detected in 12 out of 14 cases in which bone marrow could be reevaluated. In one of these MDS patients, monocyte subset analysis was available at MDS diagnosis and showed an MO1 accumulation. These observations suggested that MDS with marrow monocytosis, peripheral blood monocyte count neighboring the threshold of monocytosis, and MO1 accumulation could evolve in genuine CMML. To explore this hypothesis, we prospectively analyzed monocyte subset repartition in the blood of 84 MDS patients at diagnosis, including 26 MDS-MLD, 6 MDS–single lineage dysplasia, 12 MDS–ring sideroblasts, 28 MDS-EB1, 10 MDS-EB2, 1 MDS with isolated del(5q), and 1 MDS unclassifiable. MO1 accumulation ≥94% was detected in 29 of them (35%) (Figure 1A), validating a recently reported observation.³  Compared with other MDS, these “CMML-like” MDS displayed a higher WBC number (6.1 ± 3.9 vs 4.4 ± 2.0 × 10⁹/L; P = .05), a higher absolute monocyte count (0.6 ± 0.2 vs 0.4 ± 0.2 × 10⁹/L; P < .05), and a higher fraction of monocytes in the bone marrow (5.8 ± 3.8 vs 3.7 ± 3.5 × 10⁹/L; P < .05). Unlike the recent report of Lee Moffitt’s group,³  MDS without MO1 accumulation was not associated with poor cytogenetic risk or higher R-IPSS in this series.

Follow-up of 44 MDS patients with 2 distinct complete blood counts (CBCs) available showed that a monocyte count ≥1 × 10⁹/L with a monocyte fraction ≥10% of WBC was detected significantly more often in those with a CMML-like phenotype at diagnosis than in other MDS patients (P < .01; Figure 1M). In less than 1 year, 7 out of 16 MDS patients with a “CMML-like” phenotype at diagnosis evolved into overt CMML.

We also evaluated the relevance of the 1 × 10⁹/L threshold for monocytosis^1,10,19  by analyzing CBC from 192 healthy adults (median age: 51; range: 19-99). Median value of blood monocytes in this healthy population was 0.49 × 10⁹/L (Figure 1N). Mathematical extrapolation in relation to age showed an increase of monocytosis with years, reaching 0.6 × 10⁹/L at 100 years of age, suggesting that the French-American-British¹⁹ /WHO^1,10  threshold is probably overestimated (Figure 1O). Further studies with larger sample size would be useful to redefine accurately the monocytosis threshold. Of note, a category of “oligomonocytic” CMML with a monocyte fraction ≥10% but absolute monocyte count between 0.5 and 1 × 10⁹/L was recently described, with a subset of these patients eventually developing genuine CMML.²⁰ 

Overall, we suggest that “CMML-like” MDS could be an entity that is likely to evolve into genuine CMML and that an MO1 accumulation ≥94% should be considered to be included as a major criterion for both CMML and CMML-like MDS diagnosis.