Myeloid Neoplasms With Oligomonocytosis Exhibit Heterogenous Pathologic and Genetic Features

Abstract

The criteria used to classify patients with myeloid neoplasms and monocytosis have changed in the fifth edition of the World Health Organization classification (WHO5) and the International Consensus Classification (ICC). Although both classifications have reduced the absolute monocyte count threshold for chronic myelomonocytic leukemia (CMML) to 0.5 × 10⁹/L, the ICC has also introduced new morphologic criteria for CMML. We studied the effect of these changes on a large cohort of myeloid neoplasms with white blood cell count < 13 × 10⁹/L and blasts <20% that were previously classified using the fourth revised edition of the WHO classification of hematologic malignancies (WHO4r). The lower monocyte threshold might reclassify approximately 20% of WHO4r-defined myelodysplastic syndrome as myelodysplastic-type CMML (MD-CMML) in the new classifications. This change led to an 84% increase in MD-CMML by WHO5 criteria. The number of MD-CMML cases in the cohort decreases from 107 by WHO5 criteria to 40 by ICC criteria because of the morphologic criteria. New ICC criteria remove 55% of WHO4r-defined MD-CMML patients from the category, and these and other patients (13% of our cohort) are not classifiable by ICC criteria. ICC-defined CMML enriches for CMML-like genetic signatures, but we find that genetic classification and genetically informed risk models predict outcome better than monocyte counts or morphologic features.

Introduction

In the fourth revised edition of the World Health Organization classification of hematologic malignancies (WHO4r) and in previous classifications, a diagnosis of chronic myelomonocytic leukemia (CMML) required an absolute monocyte count ≥ 1 × 10⁹/L, with monocytes constituting ≥10% of the white blood cell (WBC).^1–3 Recently, the concept of “oligomonocytic” CMML (oCMML) has emerged, defined as myelodysplastic syndrome (MDS) or myelodysplastic/myeloproliferative neoplasm (MDS/MPN) with relative monocytosis (≥10% monocytes) and a monocyte count of between 0.5 and 1 × 10⁹/L; a subset of patients with oCMML eventually develop overt CMML.^4,5 CMML is further subdivided into myelodysplastic CMML (MD-CMML) and myeloproliferative CMML (MP-CMML) subtypes based on whether the WBC count is less than or more than 13 × 10⁹/L, respectively.

Genetic mutations are found in >90% of CMML cases.^6,7 These mutations can be broadly divided into ones that affect DNA methylation (TET2, IDH1/2, and DNMT3A), histone modifications (ASXL1 and EZH2), the spliceosome complex (SRSF2, SF3B1, and U2AF1), and signal transduction molecules in the RAS pathway (NRAS, KRAS, CBL, and PTPN11).^7,8 SRSF2 and TET2 comutations are particularly suggestive of CMML,^9–12 as are multihit TET2 alterations.¹³

Multiple studies have shown that oCMML has a genetic profile similar to traditional CMML, with similar signature mutations occurring at similar frequencies.^4,14,15 Because oCMML has a genetic profile comparable to that of overt CMML and many oCMML cases progress to overt CMML, they now are regarded as a disease in both the fifth edition of the WHO classification (WHO5) and the International Consensus Classification (ICC). Although the WHO5 essentially retains the diagnostic criteria used for CMML in the WHO4r, with the exception of the lower absolute monocyte threshold,¹⁶ the ICC adds additional diagnostic criteria for CMML that were not present in the WHO4r, requiring that CMML marrows exhibit age-adjusted hypercellularity with myeloid pre-dominance and cytopenia.⁶ In addition, the ICC has dropped the explicit requirement that monocytosis is “persistent,” which was present in both the WHO4r and the current WHO5.

With these changes in the diagnostic boundaries, some cases previously classified as MDS would now be reclassified as MD-CMML, and some cases considered to be MD-CMML under WHO4r and WHO5 may not meet ICC criteria. We sought to examine the diagnostic implications of these changes in a large cohort of patients with myeloid neoplasms.

Materials and Methods

We studied bone marrow cases seen at our institution between 2016 and 2021 that were found to have at least 1 mutation in myeloid-directed sequencing studies and that were originally diagnosed as non-MPNs with blasts <20% and with WBC count < 13 × 10⁹/L. We applied WHO4r, WHO5, and ICC diagnoses. The electronic medical record was retrospectively reviewed to identify demographic information and assess hematologic parameters, in particular persistence of monocyte counts (>3 months). Blast counts, bone marrow cellularity, myeloid:erythroid ratio, dysplasia, genetic characteristics including mutational variant allele frequency, and survival characteristics were also extracted. For all cases, aspirate material and bone marrow trephine biopsies were reviewed, and the marrow cellularity was determined based on the trephine core biopsies. We considered marrows as exhibiting hypercellularity if they were >60% cellular in patients aged ≥60 years and, >65% cellular in patients aged between 40 and 60 years and, >70% in patients 40 years and younger based on the recommendations in the ICC book.^17,18 For the purposes of ICC, we also considered marrows with a myeloid:erythroid ratio of > 2:1 to be “myeloid predominant.” We applied the molecular International Prognostic Scoring System (IPSS-M),¹⁹ the clinical/molecular CMML-specific Prognostic Scoring System (CPSS-mol) using sex-specific anemia thresholds,⁹ and the MDS International Working Group molecular taxonomic system (MDS-IWG)²⁰ to the cases.

We also analyzed the publicly available IPSS-M data set of MDS and MDS/MPN patients to assess a subset of our findings in a very large independent cohort: 2412 IPSS-M patients had sufficient monocyte count information to allow for assessment and comparison with our internal cohort.¹⁹ Data regarding patients and specimens were gathered from both the electronic medical record and the laboratory information system and then organized using Microsoft Excel. Descriptive statistics were used for basic demographics. Differences among groups were compared using the Pearson, Kruskal-Wallis, and Wilcoxon tests. Survival data were assessed using the Kaplan-Meier method, and distinctions were determined through the log-rank test. Multivariable analysis was performed using a Cox proportional hazards model. All statistical analyses were performed in R v4.4.1 (R Project for Statistical Computing). The code used to generate the analyses/figures is available from the authors upon reasonable request.

Results

We identified 488 patients with myeloid neoplasms with WBC count < 13 × 10⁹/L and with mutations on sequencing studies. Of these, 59 (12%) were classified as MD-CMML based on WHO4r criteria. A single case with NPM1 mutation and blasts ≥10% was present in the WHO4r-defined MD-CMML cohort and excluded from further analysis because it would be classified as acute myeloid leukemia by both WHO5 and ICC.

Common Oligomonocytosis in WHO4r Defined MDS

Applying the new oligomonocytosis threshold of absolute monocyte count ≥ 0.5 × 10⁹/L and monocyte percentage ≥ 10%, 77 possible additional cases of MD-CMML were identified in the cohort (18% of the original 429 MDS cases). Many of these 77 potential oCMML patients had only transient (<3 months) oligomonocytosis. After review of the medical record, 49 patients in this group had persistent monocytosis (11% of the original 429 MDS cases). For the purposes of this study, we excluded patients with transient monocytosis from further consideration as MD-CMML, bringing the total possible MD-CMML cohort using lower monocyte criteria to 107 patients. Demographic and laboratory characteristics of the cohort are presented in Table 1.

Table 1.

Laboratory and demographic characteristics of the portion of cohort with monocytosis, excluding a single NPM1-mutated case

Characteristic	Monocytosis type				Classification
	Overt monocytosis (n = 58)a	Persistent oligomonocytosis (n = 49)a	Transient oligomonocytosis (n = 28)a	P b	Both ICC and WHO5 CMML (n = 40)a	Only WHO5 CMML (n = 67)a	P b
Age (y)	78 (71–82)	76 (68–81)	71 (64–77)	.039	78 (69–83)	75 (69–81)	.6
Sex				.9			.5
Female	19 (33)	17 (35)	8 (29)		12 (30)	24 (36)
Male	39 (67)	32 (65)	20 (71)		28 (70)	43 (64)
WBC count (×109/L)	7.30 (6.10–9.80)	4.30 (3.20–5.10)	4.65 (3.45–5.50)	<.001	6.25 (4.15–10.20)	5.50 (4.20–7.00)	.088
Hb concentration (g/dL)	10.00 (8.40–12.20)	10.00 (8.70–11.70)	9.10 (8.20–9.85)	.14	8.75 (8.05–11.65)	10.50 (9.10–12.50)	.002
PLT count (×109/L)	147 (61–216)	107 (68–189)	97 (56–256)	.7	101 (46–200)	137 (72–210)	.2
ANC (×109/L)	3.28 (2.20–4.85)	1.91 (1.02–2.66)	1.92 (1.11–2.82)	<.001	2.45 (1.26–4.85)	2.52 (1.74–3.71)	.9
AMC (×109/L)	1.51 (1.10–2.07)	0.75 (0.60–0.87)	0.61 (0.55–0.75)	<.001	1.38 (0.80–2.27)	0.96 (0.74–1.27)	.006
Monocyte %	21 (16–28)	17 (13–21)	14 (11–18)	<.001	21 (16–28)	19 (15–23)	.049
Blast category				.5			.2
Low	43 (75)	39 (80)	20 (77)		27 (68)	55 (83)
5–9	8 (14)	4 (8.2)	1 (3.8)		6 (15)	6 (9.1)
≥10	6 (11)	6 (12)	5 (19)		7 (18)	5 (7.6)

Bold values are P values < .05.

AMC, absolute monocyte count; ANC, absolute neutrophil count; CMML, chronic myelomonocytic leukemia; Hb, hemoglobin; ICC, International Consensus Classification; PLT, platelet; WBC, white blood cell; WHO5, fifth edition of the World Health Organization classification.

^aValues are represented as median (Q1-Q3) and n (%).

^bKruskal-Wallis rank sum test and Pearson χ² test.

We confirmed frequent oligomonocytosis in WHO4r-defined MDS and non-CMML MDS/MPN patients in the external IPSS-M data set. A total of 2412 patients in this data set have sufficient monocyte count information for assessment, although we do not know how many of these patients had persistent monocytosis.¹⁹ As per WHO4r criteria, 257 MD-CMML patients were present in the IPSS-M data set, 11% of the cohort, very similar to the 12% of WHO4r CMML in our internal cohort. Of the remaining 2155 non-CMML MDS and MDS/MPN patients in the IPSS-M cohort, an additional 438 cases of potential new MD-CMML with oligomonocytosis are present, 20% of the total MDS and MDS/MPN cohorts.

Marrow Cellularity and WHO5/ICC Classification Discordance

WHO5

The lower diagnostic monocyte threshold designated by WHO5 increased the cases in our cohort classified as MD-CMML from 58 to 107, an 84% increase.

ICC

Applying the new explicit ICC requirement that CMML exhibits marrow hypercellularity, along with the ICC-endorsed threshold for age-adjusted hypercellularity (>60% for the majority of patients in our cohort),¹⁷ we find that only 40 total patients meet ICC criteria for MD-CMML, with 32 patients who were previously classified as MD-CMML by WHO4r criteria no longer meeting diagnostic requirements. Were a lower, older hypercellularity consensus cutoff used (>40% in patients aged ≥70 years),²¹ then there would be 53 MD-CMML patients in our cohort as per ICC (still excluding 23 WHO4r CMML cases). Cases meeting ICC MD-CMML morphologic criteria had higher absolute monocyte counts (median, 1.38 × 10⁹/L vs 0.96 × 10⁹/L; P = .006) and lower hemoglobin concentrations (median, 8.75 vs 10.5 g/dL; P = .002) than did those only meeting WHO5 criteria (Table 1).

Appropriate classification of cases with persistent monocytosis that do not meet morphologic ICC CMML criteria is difficult because persistent monocytosis ≥ 0.5 × 10⁹/L is an exclusionary criterion for all MDS subtypes in the ICC.⁶ The ICC suggests that cases with monocytosis that do not meet ICC-defined morphologic criteria might be thought of as clonal monocytosis of undetermined significance or clonal cytopenia and monocytosis of undetermined significance (CCMUS), but nearly all cases in our cohort had reported dysplasia, and 11 cases had increased marrow blasts ≥5%, precluding consideration of these diagnoses for most patients. A diagnosis of MDS/MPN, not otherwise specified (NOS), was therefore considered, but ICC criteria for MDS/MPN, NOS, require either a WBC count ≥ 13 × 10⁹/L or platelet count 450 × 10⁹/L, along with cytopenia, leading to only 1 case in our cohort that could be classified as MDS/MPN, NOS, by the ICC. Two cases had absolute monocyte counts ≥ 1 × 10⁹/L and cytopenia but lacked dysplasia, leading to classification as CCMUS by the ICC. Ultimately, 64 cases were essentially unclassifiable myeloid neoplasms with persistent monocytosis by ICC criteria. An overview of the classification of cases by WHO4r, WHO5, and ICC criteria is shown in Figure 1, and a flowchart summarizing the classification criteria across the 3 systems is displayed in Supplementary Figure S1.

Figure 1.

Comparison of individual case classification across WHO4r, ICC, and WHO5. CCMUS, clonal cytopenia and monocytosis of undetermined significance; CMML, chronic myelomonocytic leukemia; EB, excess blasts; ICC, International Consensus Classification; MDS, myelodysplastic syndrome; MPN, myeloproliferative neoplasm; MLD, multilineage dysplasia; NOS, not otherwise specified; RS, ring sideroblasts; SLD, single-lineage dysplasia; WHO4r, fourth revised edition of the World Health Organization classification of hematologic malignancies; WHO5, fifth edition of the World Health Organization classification.

Blast Comparison

Excluding the single case with mutated NPM1, 14 cases of WHO4r-defined MD-CMML exhibited increased blasts ≥5%. When cases with persistent oligomonocytosis were included in the MD-CMML category (as per WHO5), an additional 10 cases with increased blasts were added to the MD-CMML cohort. Using more stringent ICC morphologic criteria, only 4 oligomonocytic cases with increased blasts were considered to be MD-CMML, and 5 former WHO4r-defined MD-CMML cases with increased blasts were excluded from the category.

Genetic Characteristics

Based on previous studies, we considered multihit TET2 alterations, comutation of TET2 and SRSF2, or RAS pathway mutations (KRAS, NRAS, CBL, and PTPN11) together as CMML-like genetic features. We found that persistence of oligomonocytosis was a reasonably effective criterion to distinguish potential oCMML cases from MDS with transient monocytosis because CMML-like genetic alterations were enriched in cases with monocytosis persistence, whereas multihit TP53 mutations were more common in transient monocytosis cases (Table 2). Only 1 of 10 patients with persistent oligomonocytosis and increased blasts had a “CMML-type” mutational pattern (biallelic TET2 alteration); the other cases with increased blasts and oligomonocytosis fell into 7/SETBP1, SRSF2, SF3B1, TP53-complex, and clonal cytopenia of undetermined significance–like MDS-IWG genetic categories.

Table 2.

Genetic comparison of cases by monocytosis persistence and classification across systems

Characteristic	Monocytosis type				Classification
	Overt monocytosis (n = 58)a	Persistent oligomonocytosis (n = 49)a	Transient oligomonocytosis (n = 28)a	P b,c	Both ICC and WHO5 CMML (n = 40)a	Only WHO5 CMML (n = 67)a	P b
CMML likec	35 (60)	23 (47%)	6 (21%)	.002	30 (75)	28 (42)	<.001 d
Multihit TET2	23 (40)	15 (31%)	4 (14%)	.031	16 (40)	22 (33)	.5
TET2 and SRSF2	13 (22)	7 (14%)	3 (11%)	.3	12 (30)	8 (12)	.02
RAS pathway	9 (16)	9 (18%)	2 (7.1%)	.2	11 (28)	7 (10)	.023
ASXL1	20 (34)	9 (18%)	7 (25%)	.8	13 (33)	16 (24)	.3
SF3B1	12 (21)	8 (16%)	7 (25%)	.5	4 (10)	16 (24)	.075
Multihit TP53	2 (3.4)	2 (4.1%)	4 (14%)	.035	2 (5.0)	2 (3.0)	.6
MDS-IWG				.011			.6
AML like	1 (1.7)	1 (2.1)	0 (0)		1 (2.6)	1 (1.5)
TP53-complex	4 (6.9)	3 (6.4)	7 (25)		3 (7.7)	4 (6.1)
der(1;7)	0 (0)	0 (0)	0 (0)		0 (0)	0 (0)
7/SETBP1	4 (6.9)	4 (8.5)	4 (14)		5 (13)	3 (4.5)
del(5q)	0 (0)	0 (0)	0 (0)		0 (0)	0 (0)
EZH2-ASXL1	4 (6.9)	1 (2.1)	0 (0)		2 (5.1)	3 (4.5)
IDH-STAG2	4 (6.9)	3 (6.4)	1 (3.6)		2 (5.1)	5 (7.6)
BCOR/L1	1 (1.7)	0 (0)	1 (3.6)		0 (0)	1 (1.5)
bi-TET2	26 (45)	15 (32)	1 (3.6)		18 (46)	23 (35)
U2AF1 157	0 (0)	3 (6.4)	1 (3.6)		0 (0)	3 (4.5)
U2AF1 34	0 (0)	0 (0)	1 (3.6)		0 (0)	0 (0)
SRSF2	2 (3.4)	1 (2.1)	1 (3.6)		1 (2.6)	2 (3.0)
SF3B1	4 (6.9)	7 (15)	6 (21)		2 (5.1)	9 (14)
ZRSR2	0 (0)	1 (2.1)	1 (3.6)		0 (0)	1 (1.5)
CCUS like	4 (6.9)	6 (13)	3 (11)		2 (5.1)	8 (12)
mNOS	4 (6.9)	2 (4.3)	1 (3.6)		3 (7.7)	3 (4.5)
IPSS-M				.033			.008
Very low	10 (17)	9 (18)	1 (3.6)		3 (7.5)	16 (24)
Low	19 (33)	17 (35)	9 (32)		12 (30)	24 (36)
Moderate low	8 (14)	7 (14)	2 (7.1)		3 (7.5)	12 (18)
Moderate high	5 (8.6)	4 (8.2)	1 (3.6)		5 (13)	4 (6.0)
High	12 (21)	5 (10)	6 (21)		12 (30)	5 (7.5)
Very high	4 (6.9)	7 (14)	9 (32)		5 (13)	6 (9.0)
CPSS-mol				.002			.017
Low	17 (29)	20 (41)	0 (0)		10 (25)	27 (40)
Intermediate 1	16 (28)	9 (18)	8 (29)		6 (15)	19 (28)
Intermediate 2	16 (28)	14 (29)	15 (54)		14 (35)	16 (24)
High	9 (16)	6 (12)	5 (18)		10 (25)	5 (7.5)

Bold values are P values < .05.

CCUS, clonal cytopenia of undetermined significance; CMML, chronic myelomonocytic leukemia; CPSS-mol, clinical/molecular CMML-specific Prognostic Scoring System; ICC, International Consensus Classification; IPSS-M: molecular International Prognostic Scoring System; MDS-IWG, myelodysplastic syndrome International Working Group molecular taxonomic system; WHO5, fifth edition of the World Health Organization classification.

^aValues are represented as n (%).

^bPearson χ² test.

^cP values in this column compare transient oligomonocytosis cases versus persistent oligomonocytosis and overt monocytosis cases grouped together.

^dCMML-like genetics refers to multihit TET2 alterations, TET2 and SRSF2 comutation, or RAS pathway gene mutations (KRAS, NRAS, CBL, and PTPN11). Some cases will have a combination of these findings, and the patient numbers in the subsequent “multihit TET2,” “TET2 and SRSF2,” and RAS pathway rows therefore do not sum to this total.

Applying WHO5, approximately 54% of patients with MD-CMML had a CMML-like genetic signature. In contrast, applying the more stringent ICC morphologic criteria enriched CMML-like genetic signatures to 76% in the combined CCMUS and MD-CMML cohorts, but the frequency of CMML-like genetics in the many unclassified cases remained high, at 41% (Supplementary Table S1). RAS pathway mutations were associated with marrow hypercellularity (P = .004) and were therefore enriched in ICC-defined MD-CMML. Cases meeting WHO5 criteria but not ICC criteria were overall enriched for lower-risk IPSS-M and CPSS-mol categories, although including WHO5-only MD-CMML cases also doubled the number of “very high”–risk MD-CMML cases by IPSS-M (Table 2). The molecular and genetic features of the cohort are displayed in Figure 2, and case-level hematologic, genetic, and classification information is available in the Supplementary Material.

Figure 2.

Genetic features and classification of the cohort organized by monocytosis tier. Overt monocytosis indicates absolute monocyte count ≥ 1 × 10⁹/L and relative monocytes ≥10%; oligomonocytosis is defined as absolute monocyte count ≥ 0.5 × 10⁹/L and relative monocytes ≥10%. AML, acute myeloid leukemia; CCMUS, clonal cytopenia and monocytosis of undetermined significance; CMML, chronic myelomonocytic leukemia; CPSS-mol, clinical/molecular CMML-specific Prognostic Scoring System; EB, excess blasts; IB, increased blasts; ICC, International Consensus Classification; IPSS-M, molecular International Prognostic Scoring System; LB, low blasts; MDS, myelodysplastic syndrome (neoplasm); MDS-IWG, myelodysplastic syndrome International Working Group molecular taxonomic system; MPN, myeloproliferative neoplasm; MLD, multilineage dysplasia; NOS, not otherwise specified; RS, ring sideroblasts; SLD, single-lineage dysplasia; WHO4r, fourth revised edition of the World Health Organization classification of hematologic malignancies; WHO5, fifth edition of the World Health Organization classification.

Case Examples

Representative examples of cases illustrating diagnostic disagreements among the 3 classification systems are illustrated in Figure 3. The examples include a case classifiable as MD-CMML by both WHO criteria but not by ICC (Fig. 3A, B), a case with 3 different classifications across systems (Fig. 3C, D), and cases classified as MD-CMML by ICC and WHO5 criteria because of persistent oligomonocytosis, but without CMML-like genetics, with increased blasts, and with short survival (Fig. 3E–H). Please see the figure caption for additional details.

Figure 3.

Example photomicrographs of cases and classification. (A, B) Patient with monocytosis of 1 × 10⁹/L with normocellular marrow, dyserythropoiesis, and biallelic TET2 mutation, diagnosed as myelodysplastic-type CMML by WHO4r and WHO5, unclassifiable by ICC. (C, D) Patient with persistent oligomonocytosis and normocellular marrow with many ring sideroblasts and SF3B1 mutation, diagnosed as myelodysplastic syndrome with ring sideroblasts and single-lineage dysplasia by WHO4r, CMML-1 by WHO5, and unclassifiable by ICC. (E, F) Patient with persistent oligomonocytosis, hypercellular marrow with multilineage dysplasia, monosomy 7, and increased blasts with approximately 3.5-month survival, diagnosed as myelodysplastic syndrome with excess blasts-2 by the WHO4r criteria and CMML-2 by ICC and WHO5 criteria. (G, H) Patient with persistent oligomonocytosis, multilineage dysplasia, increased blasts, and biallelic TP53 mutation and complex karyotype with approximately 4.5-month survival, diagnosed as myelodysplastic syndrome with excess blasts-2 by WHO4r and as CMML-2 by ICC and WHO5. The cases illustrated in (E-H) had some circulating conventional myeloblasts in the blood but also had increased mature-appearing monocytes. (A, C, E, and G, ×100, hematoxylin and eosin), (B, F, and H, ×1000, Wright-Giemsa), and (D, ×1000, Perls Prussian blue.) CMML, chronic myelomonocytic leukemia; ICC, International Consensus Classification; WHO4r, fourth revised edition of the World Health Organization classification of hematologic malignancies; WHO5, fifth edition of the World Health Organization classification.

Survival

We saw no significant difference in overall survival in patients with oligomonocytosis vs those with monocyte counts ≥ 1 × 10⁹/L (Fig. 4A). In contrast, significant differences in overall survival were seen among patients with persistent monocytosis on the basis of their MDS-IWG genetic category (P = .001), CPSS-mol risk group (P < .0001), and IPSS-M risk group (P < .0001; Fig. 4B). CMML-2 cases had poor survival when using either ICC or WHO5 criteria (Supplementary Fig. S2). In univariable analysis, cases meeting criteria for MD-CMML by both ICC and WHO5 criteria had shorter survival than those meeting WHO5 criteria alone (P = .00083), but this finding did not retain statistical significance in multivariable models that included either the IPSS-M or CPSS-mol risk stratification.

Figure 4.

Survival characteristics. (A) Overall survival among all patients with persistent monocytosis shows no difference between those meeting traditional (n = 58) or oligomonocytic thresholds (n = 49). (B) Molecular International Prognostic Scoring System stratification shows clear differences in survival in patients with persistent monocytosis (n = 107). (C) Myelodysplastic syndrome genetic subtypes did not show differing survival based on the presence or absence of monocytosis at a ≥ 0.5 × 10⁹/L threshold (only genetic categories with >10 cases in each monocytosis tier are shown). (D) Analysis of the external molecular International Prognostic Scoring System data set shows no differences in survival across genetic subtypes based on monocytosis tiers (only genetic categories with >20 cases in each tier are shown). For (A) and (D), oligomonocytosis is defined as absolute monocyte count ≥ 0.5 × 10⁹/L and relative monocytes ≥10%; overt monocytosis is defined as absolute monocyte count ≥1 × 10⁹/L and relative monocytes ≥10%. CCUS, clonal cytopenia of undetermined significance; mNOS, molecularly not otherwise specified.

We further sought to determine whether the presence or absence of monocytosis (defined at a threshold ≥ 0.5 × 10⁹/L with at least 10% monocytes) affected survival. We assessed molecular-subtype–specific survival in our internal data set for the MDS-IWG–defined biTET2 and SF3B1 groups because at least 10 patients in those groups had monocyte counts on either side of the oligomonocytosis threshold. No significant survival differences were seen within the genetic groups on the basis of monocytosis (Fig. 4C). We confirmed this finding in the very large external IPSS-M data set, studying molecular subtypes with at least 20 patients in each monocytosis tier (no monocytosis, oligomonocytosis, and overt monocytosis ≥ 1 × 10⁹/L). Although significant differences in survival are seen across genetic subtypes, the monocytosis tier did not significantly affect survival (Fig. 4D).

Discussion

The recognition that myeloid neoplasms with oligomonocytosis have a propensity to progress to CMML and exhibit characteristic genetic features of CMML has led to significant changes in CMML diagnostic criteria to accommodate this biologic insight. Although both ICC and WHO5 have lowered the diagnostic CMML monocytosis threshold to 0.5 × 10⁹/L (assuming that clonality can be documented below a monocyte threshold 1 × 10⁹/L), their other diagnostic criteria differ. WHO5 largely did not modify diagnostic criteria for CMML versus those present in WHO4r, other than requiring morphologic dysplasia for patients with oligomonocytosis. The ICC, however, has added requisite morphologic criteria for a CMML diagnosis that was not present in WHO4r, most importantly requiring age-adjusted marrow hypercellularity because of myeloid proliferation.

Age-adjusted marrow hypercellularity has historically been ill-defined. Many practicing hematopathologists are taught and have promulgated a “100–patient age” rule of thumb, but this has never been well supported empirically, and older expert consensus-based recommendations for normal bone marrow cellularity were largely derived from a single autopsy study that included only 10 patients >70 years.^21–23 In recent years, publications have attempted to more robustly define normal marrow cellularity ranges over the age spectrum and have found far more physiologic variation that would be suggested by the simplistic rubric above.^18,24 Although the initial ICC publication did not define thresholds for marrow hypercellularity in CMML,⁶ the recently published ICC book has proposed specific age-adjusted thresholds for marrow hypercellularity based largely on data from Wong et al.^17,18 For our cohort nearly entirely composed of patients aged >60 years, the new ICC criteria would require marrow cellularity >60% for CMML to be considered.

We find that these collective changes introduced in the WHO5 and ICC CMML criteria lead to varied and significant challenges in clinical practice in the classification of myeloid neoplasms with <20% blasts and WBC count < 13 × 10⁹/L. First, approximately 20% of patients in both our cohort and the IPSS-M cohort with MDS as defined by WHO4r had oligomonocytosis per new thresholds at the time of bone marrow examination. However, in our internal cohort, 36% of these potential new MD-CMML patients had only transient, mild monocytosis. The patients with transient monocytosis had genetic features as a group more similar to MDS than CMML compared with patients with persistent oligomonocytosis, supporting that they should not be classified as MD-CMML. Although monocytosis persistence has been a historical component of CMML diagnosis (although it has been removed as an explicit criterion by ICC), the lowering of monocyte threshold to 0.5 × 10⁹/L will cause MD-CMML to be considered in a large proportion of possible MDS patients and will complicate diagnostic reporting.

Next, the expansive MD-CMML definition adopted by WHO5 compels many cases that do not have biologic features of CMML to be lumped into the MD-CMML category. If revision of diagnostic thresholds was meant to make space for oligomonocytic CMML as a continuum of CMML, often representing an earlier stage of disease, this has come at the expense of including high-grade neoplasms without CMML-like biology in the category. MD-CMML cases with blasts >5% nearly doubled in our cohort when WHO5 criteria were applied vs WHO4r, and only 1 of 10 cases in this category exhibited a CMML-like genotype.

We find that the changes to CMML criteria made by the ICC engender a different set of diagnostic difficulties. The ICC has instituted morphologic criteria for CMML that did not exist in the WHO4r. This has had the commendable effect of significantly enriching the MD-CMML category for cases with a CMML-like genotype versus the WHO5. However, the introduction of morphologic criteria for CMML without corresponding adjustments of criteria elsewhere in the ICC system appears to have created a diagnostic gap in the classification because cases with persistent monocytosis > 0.5 × 10⁹/L are explicitly excluded from all MDS subtypes. We find that 13% of our entire cohort is unclassifiable by ICC criteria if taken literally; it may be that most of these cases should still be considered to represent MDS that happen to exhibit monocytosis (because they would have been diagnosed under WHO4r), but 55% of WHO4r MD-CMML patients no longer meet morphologic ICC criteria for CMML, and the appropriate classification of these cases is less clear.

We applied the recently proposed MDS-IWG genetic classification to our cases.²⁰ As expected, there is enrichment of the bi-TET2 genetic subtype in both WHO5 and ICC-defined MD-CMML. In our cohort, >10 patients in both the MDS-IWG SF3B1 and bi-TET2 categories had monocyte counts above or below the oligomonocytic threshold, and we did not see differences in survival within these genetic groups when stratified by monocytosis. The expansive MD-CMML definition endorsed by the WHO5 would more than double the number of CMML cases in the SF3B1 genetic class. Because previous studies have suggested that SF3B1-mutated CMML is rare and exhibits a low-risk, myelodysplastic phenotype and features that overlap with those of SF3B1-mutated MDS,²⁵ it is not evident that labeling more cases in the genetic SF3B1 class as MD-CMML is clinically helpful or desirable.

We extended our genetic subtype analysis to the much larger IPSS-M data set, where we again showed no difference in survival within MDS-IWG genetic subgroups at different levels of monocytosis. In particular, we show that cases in the TP53-complex genetic group have exceptionally poor outcomes irrespective of the monocytosis level. Although this is likely expected, it seems that the emphasis of the classification should be placed on the TP53 mutation and karyotype in these cases rather than on monocytosis, especially in the oligomonocytic setting.

Our findings largely align with those reported by Baumgartner et al,²⁶ who reclassified 2130 cases of oligomonocytosis using WHO5 and ICC 2022 classifications. In their study, 1.5 times more oligomonocytic CMMLs were identified by WHO5 versus ICC criteria compared with 3.5 times more oligomonocytic CMML cases in our study by WHO5 than by ICC criteria. Similarly, Baumgartner et al²⁶ found that approximately 20% of patients with WHO4r-defined CMML were no longer CMML per ICC morphologic criteria, whereas in our cohort, this percentage is higher at 55%. The reason for increased discordance among classification systems in our patient cohort is not entirely clear, but it may be because of the relatively stringent (but ICC endorsed) trephine core biopsy-based hypercellularity thresholds used in our study rather than the aspirate smear-based cellularity estimation used in the study by Baumgartner et al.²⁶ As in our cohort, Baumgartner et al²⁶ found that ICC-defined CMML are more enriched for CMML-like genetic signatures than are WHO5-defined CMML and that oligomonocytic CMML cases by WHO5 criteria exhibit mutational features intermediate between MDS and CMML.

There are several limitations of our study. Our study was retrospective, and cellularity and dysplasia assessment were performed at the time of diagnosis by multiple board-certified hematopathologists. This reflects real-world practice, and using previously generated data also avoided possible bias in cellularity assessment because cellularity was not a diagnostic criterion when the cases were originally evaluated. The patient cohort is relatively small and was heterogeneously treated, which complicates assessment of survival; however, many of our findings are corroborated in the large IPSS-M data set and in the study by Baumgartner et al.^19,26 All patients included in our study had a pathogenic mutation as identified by myeloid-directed sequencing studies, and so rare monocytic myeloid neoplasms without mutations but with cytogenetic abnormalities would therefore not have been included in our cohort. Furthermore, we restricted our analysis exclusively to patients with WBC count < 13 × 10⁹/L. Although all oligomonocytic CMML patients would fall into this category by definition, we have not studied MP-CMML, and some data such as the frequency of bone marrow hypercellularity in MD-CMML patients in our cohort likely do not extend to MP-CMML. At our institution, we do not perform monocyte partitioning by flow cytometry or routine CD14 immunohistochemistry on core biopsies; these are not necessary for classification by WHO5 or ICC but had such studies been done, it is possible that they would have militated against classification as CMML in some cases (although the cases would still have met the formal diagnostic criteria).

In conclusion, our manuscript highlights the current practical diagnostic difficulties faced by hematopathologists and hematologist/oncologists when assigning WHO5 and ICC diagnoses in patients with myeloid neoplasms and monocytosis. Practitioners will encounter many cases that previously would have been considered to be MDS but that might now meet MD-CMML criteria because of oligomonocytosis. We confirm that monocytosis persistence appears to be important, but this may not always be information that is available to the pathologist, depending on the practice setting. The morphologic criteria introduced by the ICC have increased the “specificity” of the diagnosis, enriching for CMML-like genetics, RAS pathway mutations, increased monocyte counts, and decreased hemoglobin concentration versus the more expansive WHO5 criteria. This collectively suggests that ICC morphologic criteria may be successful at identifying patients at a more clinically advanced stage of disease than using WHO5 criteria alone. However, by ICC criteria, many patients who likely do have a disease on the CMML genetic spectrum are left without a clear classification if diagnostic criteria are applied strictly, and genetically informed risk models ultimately predict patient outcomes better than do the morphologic features.

As we now have comprehensive proposed molecular classifications of MDS and MDS/MPN, it seems that we should reconsider traditional and arbitrary morphologic and laboratory disease criteria based on dichotomized continuous variables (marrow cellularity, monocytosis thresholds, etc). For myeloid neoplasms with MDS-like WBC data, it seems that terminology such as “with monocytosis” could be introduced as a diagnostic qualifier analogous to that now used for myeloid neoplasms arising in the therapy-related or germline-predisposition setting in the ICC.⁶ This would allow for continued study of these cases without shoehorning neoplasms with divergent biology together into the same group. These cases could potentially be monitored for acquisition or expansion of RAS-mutated subclones, which are important biologic markers of monocytic expansion and progression in bona fide CMML^20,27–29 and were enriched in our cohort in monocytic patients with higher bone marrow cellularity.

Supplementary Material

Supplementary Material

The online version contains supplementary material available at https://doi.org/10.1016/j.modpat.2025.100823.

Data Availability

The data sets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.