Bone marrow samples sent for routine immunophenotyping from cytopenic patients with suspected MDS between 01-Nov-2019 and 01-Sep-2022 were included in the study. The definition of cytopenia was at least one of the following: absolute neutrophil count <1.8 × 10⁹/L, platelet count <100 × 10⁹/L, hemoglobin concentration <100 g/L. None of the patients had received vitamin supplementation or growth factors prior to the diagnostic bone marrow aspiration, and none had received therapy for MDS during the follow-up. In all cases, bone marrow smear was evaluated in parallel with FCM, and in most cases, cytogenetics and trephine biopsy were also performed. Flow cytometry and cytomorphology were analyzed independently. Cases with myeloid progenitors ≥5% detected by either bone marrow smear or flow cytometry (FCM) and with inadequate material for cytomorphological analysis were excluded. The patients were followed until the diagnosis of MDS was confirmed. This was based on the following criteria: 1. the presence of cytogenetic aberration except for loss of chromosome Y; 2. ≥5% myeloblast in the trephine biopsy; 3. no improvement in the blood counts during the follow-up, and MDS diagnosis confirmed by a second bone marrow examination or no other etiology was found.

All patients provided informed consent for the diagnostic analysis and for the follow-up. The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional Review Board of Markusovszky University Teaching Hospital.

The monoclonal antibody (mAb) panel consisted of three 8-color tubes (Table 1) designed according to the iMDSFlow guidelines [10, 17]. EDTA was used as an anticoagulant in all samples. Samples were processed within 24 h after the bone marrow aspiration. For labeling, the samples were incubated with mAbs for 15 min at room temperature in the dark according to the manufacturers’ data sheets. Red blood cells were lysed using FACS lysing solution containing 1.5% formaldehyde (BD Biosciences). After washing twice in PBS, the cells were resuspended in 500 μL of PBS and measured within one hour. Samples were measured on FACSCanto II cytometer (BD Biosciences), with 50000 events acquired. Data acquisition and analysis were performed using FACSDiva software (BD Biosciences). A hierarchical gating strategy was employed for data analysis. Initially, doublets were excluded based on forward scatter area (FSC-A) versus forward scatter width (FSC-W). Subsequently, debris were excluded based on forward scatter versus side scatter (FSC vs. SSC). Finally, the main bone marrow cell lineages were gated. Monocytopoesis was identified by CD33^int CD64^int, granulopoesis by CD15⁺, nucleated erythropoetic cells by CD45^dim/negSSC^low/intCD33⁻CD235a⁺CD71⁺. The ratio of myeloid progenitors was enumerated by gating the CD45^dimSSC^low/intCD34⁺CD13⁺ and CD45^dimSSC^low/intCD117 ⁺CD33⁺populations using the total number of nucleated cells as the denominator. In instances where the ratios differed between the two gating strategies, the higher ratio of myeloid progenitors was utilized.

The Ogata score [11] and extended Ogata score [12] were counted and the thresholds of MDS were set to ≥2 for both scores. For Ogata score, four parameters were analyzed (1 point each): the percentage of myeloblasts in all nucleated cells (cut-off ≥2%), the percentage of B-progenitors in all CD34⁺ cells (cut-off ≤5%), the lymphocyte to myeloblast CD45 MFI ratio (≤4 or ≥7.5), and the granulocyte to lymphocyte SSC mode ratio (cut-off ≤6). In the case of extended Ogata score, one extra point was added for the expression of CD7 on myeloid progenitors with a threshold of 20% or of CD56 on monocytes, with a threshold of 30% positive cells.

In addition to these scores, the presence of at least three aberrations in at least two cell compartments was considered indicative of MDS. During the data analysis predefined templates and gates were applied in the FACSDiva software to assess the distinction from the normal pattern. The sensitivity, specificity, and accuracy of Ogata score, extended Ogata score, “3 aberrations in two cell compartments method” and the combination of the Ogata score and “3 aberrations in two cell compartments method” were evaluated.

A total of 179 patients with cytopenia were included in the study. The median age was 72 years (range 28–90). Of these patients, 95 were female. A total of 58 cases were diagnosed as MDS. In the majority of MDS cases the mean corpuscular volume (MCV) was increased or normal in 28 and 27 cases, respectively. In 15 cases, cytogenetic aberrancies in addition to morphological dysplasia were observed. The detected chromosomal abnormalities were complex in 7 cases, del (5q) alone in 4 cases, del (20q) in 2 cases, trisomy 8 (+8) in 1 case, del (9q) in 1 case. In five cases, a ≥5% myeloblast count was found in the trephine biopsy. In 38 cases with morphological dysplasia, there was no improvement in the blood counts during the follow-up period and the diagnosis was confirmed by a second bone marrow examination and/or by excluding other etiologies. Of these 38 patients, ring sideroblasts in 7 cases, increasing ratio of myeloblasts in 3 cases or progression of cytopenia during follow-up in 4 cases, clonal chromosomal (not MDS defining) abnormality in 2 cases, abnormal monocytosis and granulocytosis in 1-1 cases, and dysplasia in more than one lineage in 12 cases were indicative of MDS. The median follow-up period was 2 months (range 0.2–27). MDS was classified according to the 2016 WHO classification: MDS with isolated del (5q) 4 (7%), MDS with single lineage dysplasia (SLD) 5 (9%), MDS with multilineage dysplasia (MLD) 34 (57%), MDS with ring sideroblasts and single lineage dysplasia (RS-SLD) 3 (5%), MDS with ring sideroblasts and with multilineage dysplasia (RS-MLD) 5 (9%), MDS with excess blasts (EB1) 5 (9%), chronic myelomonocytic leukaemia (CMML) 1, myelodysplastic/myeloproliferative neoplasm, unclassifiable (MDS/MPN-U) 1.

The etiology of non-MDS cytopenias is listed in Table 2. The most common causes are vitamin B12 deficiency, toxic effects of drugs, inflammation, and chronic myeloproliferative neoplasia, with primary myelofibrosis being the most prevalent. In many cases, a combination of different causes was diagnosed. In non-MDS cases, a normal MCV was observed most frequently then increased and decreased in 59, 52, and 10 cases, respectively.

The ratio of myeloid progenitor cells (MPCs) was determined by gating the CD45^dimSSC^low/intCD34⁺CD13⁺ or CD45^dimSSC^low/intCD117⁺CD33⁺ populations, which correlated in most of the cases well (Figure 1). In MDS cases, the mean ratio of MPCs with the two different gating strategies was 1.6% (range 0%–4.7%) and 1.6% (range 0.1–4.6%), respectively. A difference in the MPCs exceeding 1% between the two gating strategies was observed in only 2 cases (3.4%) due to loss of either CD34 or CD33.

In non-MDS cases, the ratio of MPCs by the two different gating approaches was 0.8% (range 0.1–2.7%) and 0.7% (range 0.1%–3%), respectively. Myeloid progenitors over 2% were detected in five patients: 2 cases with pancytopenia due to methotrexate, 2 cases with myeloproliferative neoplasia (MPN), and one case with infection.

The sensitivity, specificity, and accuracy of the Ogata score, the extended Ogata score, the “3 aberrations in two cell compartments method” and the combination of the Ogata score with the “3 aberrations in two cell compartments method” were calculated.

The Ogata score was not applicable in only one MDS case due to the loss of CD34 expression on myeloid progenitors. The sensitivity of the Ogata score proved to be 75%. Regarding the 14 false negative cases, 12 of them belonged to the low or very low risk group by R-IPSS, with normal karyotype and myeloid progenitors ≤2% by either method (BM smear, trephine biopsy or FCM). One of the remaining two cases was MDS with fibrosis and complex cytogenetics, while the other was an MDS/MPN with 4% myeloblast in the trephine biopsy. The sensitivity of the extended Ogata score was improved to 81% by identifying an additional 3 MDS cases.

The “3 aberrations in two cell compartments method” yielded a sensitivity of 72%. Sixteen cases were identified as false negatives. Eight of these cases were classified as low or very low risk according to the R-IPSS, while three were placed in the intermediate risk group. In five cases, the R-IPSS could not be calculated due to unsuccessful karyotyping. Two of these cases were classified as MDS-EB1 based on the increased blasts observed in the trephine biopsy.

The sensitivity decreased to 61% when the Ogata score and the “3 aberrations in two cell compartments method” were combined, classifying the cases as MDS if both turned out to be suggestive of MDS (Table 3).

In patients with non-MDS cytopenia all four FCM approaches showed a specificity of at least 62%. The Ogata score and the extended Ogata score exhibited the same specificity. Regarding the Ogata score, score 2 was observed with high frequency in 36 false positive cases. Scores 3 and 4 were more specific and were found in only 10 of non-MDS cases. The decreased B-progenitor cluster size and reduced granulocyte/lymphocyte SSC ratio were detected in 43 (93%) and 33 (72%) of the false positive non-MDS cases, respectively. These two aberrations were observed in cases of any etiology. The decreased B-progenitor cluster size occurred most often in elderly (>70 years), in autoimmune cytopenia and in inflammatory diseases, while reduced granulocyte/lymphocyte SSC ratio was most often found in cytopenia due to toxic effects and inflammation. Abnormal lymphocyte to myeloblast CD45 ratio and increased myeloblast ratio were shown in 15 (32%) and 4 (9%) false positive non-MDS cases, respectively. The most common etiology for the high (3 and 4 points) Ogata scores was pancytopenia due to toxic effects from methotrexate (4 cases). The specificity of the “3 aberrations in two cell compartments method” was 79%. The majority of the false positive cases were attributed to substrate deficiencies, inflammation/infection, and toxic effects. The combination of the Ogata score with the “3 aberrations in two cell compartments method” allowed the exclusion of the majority of non-MDS cases with a specificity of 86% (Table 4). The reasons for false positivity with the combination approach were variable. These included vitamin B12 deficiency (4 cases), inflammation/infection (3 cases), toxic effects (3 cases), combination of different causes (4 cases) (Table 5).

Interestingly, in 12 cases, dysplastic changes in the bone marrow smear were observed without an increase in blasts or cytogenetic aberrancies. These cases demonstrated improvement in blood counts during the follow-up period, which led to the exclusion of MDS. In half of these cases, FCM demonstrated the absence of signs of MDS using any method, while in only three cases, MDS was identified based on the combination of the Ogata score and the “3 aberrations in two cell compartments” method.

The highest accuracy (78%) was achieved by the combination of the Ogata score and the “3 aberrations in two cell compartments method.” The application of the “3 aberrations in two cell compartments method” resulted in almost the same accuracy (77%) (Table 6).

In 9 cases the bone marrow examination was repeated after a median of 11 months (range 7–24). In 3 cases, an increasing ratio of myeloblasts was detected in the second bone marrow sample (2 MDS-EB2 and 1 AML). In 2 cases, both the Ogata and the ELN scores were already abnormal in the first bone marrow sample and the myeloblast ratio did not increase in the second sample. In 3 cases, the Ogata score and in one case the ELN score became abnormal in the second sample. As a result of these changes, both scores became abnormal in two patients.