Introduction

Bone marrow examination (BME) and the assessment of immature platelet fraction (IPF) are two distinct but complementary diagnostic tools in hematology. BME, which comprises both marrow aspiration and bone biopsy, is an invasive procedure requiring local anesthesia and specific expertise. In contrast, IPF can be assessed rapidly and non-invasively using a peripheral blood sample. The normal range of IPF, the percentage of the total number of platelets that are immature, is typically 1%–7%1.

Whereas BME provides a comprehensive assessment of hematopoiesis, including the cellularity, megakaryocyte morphology, and presence of malignant cells, IPF offers insight into thrombopoietic activity because it reflects recent platelet production in the bone marrow. Recent studies have highlighted the potential of combining these methods to improve the diagnosis and prognosis of various hematologic disorders2,3,4,5,6,7. However, the relationships between BME findings and IPF with respect to a broad spectrum of hematologic conditions remain underexplored. Most studies to date have focused on specific conditions or have involved small patient cohorts, limiting the generalizability of their findings. In addition, the correlations between IPF and specific BME findings, such as megakaryocyte morphology and dysplasia, have not been comprehensively evaluated in a large, diverse patient sample.

In the present study, we aimed to comprehensively analyze the relationships between BME findings and IPF in a large, diverse cohort of patients with various hematologic disorders. By evaluating these relationships, we sought to assess the diagnostic utility of the use of IPF in conjunction with BME for patients with hematologic conditions, and to assess whether IPF could serve as a non-invasive surrogate for specific BME parameters. The open-ended, exploratory approach permitted an unbiased evaluation of these relationships, and the results provide novel insight that could be used to refine diagnostic algorithms and improve the care of patients with hematologic disorders.

Results

Patient characteristics

The study cohort comprised 1,552 patients with a median age of 73 (range: 20–95) years, of whom 59.4% were male. The most common indications for BME were diagnoses of myelodysplastic syndrome (MDS) (n = 364, 23.4%), bone marrow infiltration with diffuse large B-cell lymphoma (n = 300, 19.3%), and multiple myeloma (n = 160, 10.3%). Table 1 presents the demographic, clinical, and laboratory characteristics of the entire sample and each subgroup (those with MDS, acute myeloid leukemia (AML), acute lymphocytic leukemia (ALL), or immune-mediated thrombocytopenia (ITP)), enabling readers to fully appreciate the breadth and depth of the data analyzed in the study.

Table 1 Characteristics of the participants.
Full size table

Correlations between IPF and BME findings

We revealed a significant correlation between IPF and bone marrow megakaryocyte number in patients with MDS (p < 0.001; Fig. 1) but not in the entire cohort (Supplementary Fig. 1). However, the low R² value of 0.251 indicates a weak relationship between these variables. This weak correlation was particularly evident in patients with megakaryocyte counts of ≤ 15, for whom there was considerable scatter in the data, likely reflecting the heterogeneity of the underlying diseases. This variability suggested that the number of megakaryocytes identified during BME does not always correspond directly with the IPF or the degree of thrombocytopenia. To investigate this relationship further and its clinical implications, we focused on patients with MDS, who formed the largest subgroup of the study sample. The results showed a stronger correlation for the IPF number (R² = 0.454, p < 0.001) compared with the IPF percentage, which supports the association between the IPF and megakaryocyte proliferation in MDS (Fig. 2) and the entire cohort (Supplementary Fig. 2).

Fig. 1
figure 1

Correlation between the IPF value and bone marrow megakaryocyte count in patients with MDS. Scatter plot, on which the x-axis indicates the bone marrow megakaryocyte count (Meg#) and the y-axis indicates the IPF percentage (IPF%). Each point represents a patient, and a regression line with R and p-values is shown. IPF immature platelet fraction; MDS myelodysplastic syndromes.

Full size image
Fig. 2
figure 2

Correlation between the IPF value and bone marrow megakaryocyte count in patients with MDS. Scatter plot, on which the x-axis indicates the bone marrow megakaryocyte count (Meg#) and the y-axis indicates the IPF value (IPF#). Each point represents a patient, and a regression line with R and p-values is shown. IPF immature platelet fraction, MDS myelodysplastic syndromes.

Full size image

Correlations among the IPF, megakaryocyte count, and platelet count

The analysis of correlations among the IPF, megakaryocyte count, and platelet count revealed distinct patterns in the study cohort, as shown in the Supplementary Figures.

Megakaryocyte count vs. platelet count

A weak but statistically significant positive correlation (R2 = 0.269, p < 0.001) was identified, indicating that higher megakaryocyte numbers are generally associated with greater platelet production (Supplementary Fig. 3). Examination of the data scatter (e.g., for ≤ 15 megakaryocytes) indicated greater variability in the platelet counts of patients with lower vs. higher megakaryocyte counts.

IPF vs. platelet count

An inverse relationship was identified (R2 = − 0.209, p < 0.001), which implies that a high IPF may be associated with a low mature platelet count in specific contexts, possibly because of rapid platelet turnover or bone marrow dysfunction (Supplementary Fig. 4).

Megakaryocyte count vs. IPF

A strong positive correlation was identified (R2 = 0.572, p < 0.001), which indicates that higher megakaryocyte numbers are generally associated with greater platelet production (Supplementary Fig. 5). However, there were some outliers, which suggests a dysregulation of thrombopoiesis in specific subtypes (e.g., MDS with fibrosis). The moderate R2 values (i.e., 0.269 and 0.572, respectively) indicate that the IPF and megakaryocyte count can explain only a proportion of platelet dynamics and that other factors, such as genetic mutations and microenvironment, must play important roles. Disease stratification (e.g., MDS vs. myeloproliferative neoplasm) is critical, because the correlations varied substantially with disease subtype. These findings underscore the complexity of platelet regulation in hematologic disorders and imply that IPF can serve as a marker of thrombopoietic activity, rather than being a standalone predictor.

IPFs of patients with specific hematologic disorders (Supplementary tables 1 and supplementary Fig. 6)

ITP

All IPF measurements in patients with ITP were performed at the time of the initial diagnosis, before any therapeutic interventions. Patients who had previously undergone treatment, for example with corticosteroids, intravenous immunoglobulin, or thrombopoietin receptor agonists, were excluded, to ensure that the IPF values reflected the untreated disease state. Patients diagnosed with ITP had significantly higher IPFs than those with other diseases (median: 14.8% vs. 3.0%, respectively; p < 0.001). The high IPF of patients with ITP reflects greater thrombopoietic activity in response to peripheral platelet destruction.

MDS

Interestingly, the patients with MDS had a wide range of IPFs, but these were generally high (median: 6.8%, range: 0.40%–43.5%). There was a significant correlation between the IPF and bone marrow megakaryocyte number (p < 0.001) in these patients. However, although the low R² value of 0.251 indicates a weak positive relationship between these variables, the 95% confidence interval for the correlation was 0.152–0.345, which is relatively narrow, implying that the estimate was reliable. This paradoxical finding suggests the presence of ineffective thrombopoiesis in MDS, such that there is greater production of immature platelets that fail to mature properly. Further analysis of the patients with MDS showed that the IPF was highly specific for megakaryocyte abnormalities (p < 0.001, Fig. 3).

Fig. 3
figure 3

IPFs of patients with MDS. A box-and-whisker plot, on which the x-axis indicates the presence of abnormal megakaryocytes (Meg) and the y-axis indicates the IPF percentage (IPF%) and the IPF value (IPF#). The median, interquartile range, and outliers for each category are shown. IPF immature platelet fraction, MDS myelodysplastic syndromes.

Full size image

IPF of patients with acute leukemia

In the context of acute leukemia, IPF measurements for patients with ALL or AML were made at the time of the initial diagnosis, before any therapeutic interventions. This design eliminated the confounding effects of treatments that might have altered thrombopoietic activity. The IPF distinguished patients with ALL from those with AML, and patients with ALL typically had lower IPFs (median: 2.8%, range: 0.66%–4.6%) than those with AML (median: 3.4%, range: 1.0%–9.6%; Supplementary Fig. 7). Although the IPF showed potential use as an index to monitor thrombopoiesis, its diagnostic utility for acute leukemia subtypes is limited by the overlapping values, and there was no statistically significant difference in the IPF values for patients with ALL or AML (Supplementary Fig. 8).

Discussion

The findings of this comprehensive study of 1,552 patients who underwent BME and IPF analysis provide valuable insight into the diagnostic utility of combined use of these tools for various hematologic disorders. The IPF was associated with both the BME findings and the final diagnoses. In patients with MDS, there was a significant correlation between the IPF value and bone marrow megakaryocyte number (R2 = 0.251, p < 0.001), which supports the hypothesis that the IPF is a reliable index of thrombopoietic activity. This relationship indicates that the IPF may represent a non-invasive marker of bone marrow megakaryocyte function and its abnormalities and could be particularly useful for patients for whom BME is contraindicated or delayed.

Several previous studies have explored the relationship between IPF and BME findings. A study by Briggs et al.. showed the reproducibility and stability of the IPF value measured using the Sysmex XE-2100 analyzer, suggesting that it could be used as an index of bone marrow megakaryocyte activity4. Another study showed that the IPF correlates with the number of bone marrow megakaryocytes (n = 19, R2 = 0.57)7, supporting the idea that the IPF can serve as a non-invasive marker of bone marrow megakaryocyte function, and, thus, platelet production. The IPF provides real-time information about the rate of platelet production in the bone marrow, with a high IPF implying a high level of platelet turnover or active thrombopoiesis and a low IPF implying a lower level of platelet production. In addition, measuring the IPF assists with the differentiation of the various causes of thrombocytopenia. For instance, in ITP, in which peripheral platelet destruction is the primary pathology, the IPF is typically high, whereas in conditions characterized by impaired platelet production, such as aplastic anemia, it is usually low.

The IPF can also be used to predict imminent platelet recovery in various clinical scenarios, including following chemotherapy or hematopoietic stem cell transplantation, and in the presence of other conditions that involve thrombocytopenia. This use of the predictive ability of the IPF value could reduce the prevalence of unnecessary platelet transfusions. Furthermore, in patients with certain hematologic disorders, a change in the IPF could be used to evaluate the response to therapy before changes in the absolute platelet count become apparent. Finally, determining the IPF provides a noninvasive means of assessing bone marrow megakaryocyte activity, potentially obviating the need for invasive BME in some cases. By providing additional information about platelet dynamics, the IPF can improve understanding of various hematologic conditions and assist with their diagnosis, prognosis, and treatment. The integration of IPF determination into routine clinical practice would represent a significant advance in the field of hematology by providing a more nuanced approach to the management of patients with platelet disorders.

Several large population studies have shown that whereas the platelet count tends to decrease with age, the IPF itself does not differ significantly between adults of various ages8 or between men and women9. Verdoia et al. found that although platelet counts decline with age, the mean IPF remains stable, and no significant difference in IPF was identified between men and women8. Negro et al. showed that sex is not an independent predictor of high IPF, and that the reference range for IPF does not require adjustment for sex9. Our data are consistent with these findings: neither age (using a cutoff of 60 years) nor sex showed a significant association with the IPF in the present cohort (data not shown). In neonates, the appropriate IPF reference range is influenced by gestational age, but no adjustment for age is required for adults10. Thus, in the present and previous studies, age and sex have been shown not to influence the IPF, and, therefore, age-specific reference ranges are not required in clinical practice.

The marked difference between the IPFs of patients with ITP and those of other patients (median: 14.8% vs. 3.0%, respectively; p < 0.001) highlights the potential for the IPF to be used to differentiate these conditions. This finding could markedly affect clinical decision-making and reduce the need for invasive BME in patients with clear-cut ITP. However, it is important to note that whereas the IPF is valuable information, it should be interpreted in conjunction with other clinical and laboratory parameters. In a study by Araki et al.., the use of a cutoff of 7.7% for the IPF was associated with a sensitivity of 86.8% and specificity of 92.6% for the differentiation of ITP from aplastic anemia5, and Jung et al. found that a cutoff of 7.3% could distinguish ITP from aplastic anemia with a sensitivity of 54.0% and specificity of 92.2%11. In addition, Strauss et al.. found that the IPFs of children with platelet production defects were low, whereas they were very high in patients with ITP12.

The IPFs of patients with MDS varied substantially (median: 6.8%, range: 0.4%–43.5%), but IPF demonstrated high specificity for megakaryocyte abnormalities (p < 0.001). These paradoxical findings may reflect the ineffective thrombopoiesis that characterizes MDS, and the data imply that IPF could serve as an additional diagnostic marker for this heterogeneous group of disorders. Future studies should explore whether IPF values within specific ranges are associated with specific MDS subtypes or prognosis. In one study, some patients with MDS had very high IPFs despite the absence of severe thrombocytopenia13. Those with chromosomal abnormalities, including monosomy 7 and complex abnormalities involving chromosomes 7 or 5q, often have platelet counts > 50 × 109/L with IPFs of ≥ 10%13.

The IPF demonstrated high specificity for the differentiation of ALL from AML (p < 0.001). However, the overlap of the ranges of values, as illustrated by the lack of a statistically significant difference in the IPF between patients with ALL or AML, limit its diagnostic utility for acute leukemia subtypes. This finding suggests that IPF could be a valuable addition to the diagnostic workup of acute leukemia, permitting the earlier initiation of appropriate treatment. However, further research is needed to validate this finding in large cohorts and to better understand the biological mechanisms underlying the differences in IPF between patients with ALL and AML. Although previous studies have not demonstrated the utility of IPF for the differentiation of ALL and AML, they have provided some insight into the relevance of IPF for acute leukemia. In a study of children with thrombocytopenia, the IPFs of patients with vs. without ALL were much higher12. Another study showed that the mean IPF of 12 patients with acute leukemia was 16.2%, which is much higher than normal14. Although our cohort showed modest differences in IPF values between patients with AML and ALL, this observation is primarily exploratory and lacks direct clinical significance. Definitive differentiation between AML and ALL relies on cytomorphology and immunophenotyping, and IPF should not be considered a diagnostic tool for distinguishing these disorders. Any observed differences may reflect variations in platelet count distributions and underlying disease biology, rather than being useful for routine clinical practice.

IPF could be used with BME findings to refine the diagnostic algorithms used in hematology. For instance, in patients with thrombocytopenia, an initial IPF measurement could be used to guide decision-making regarding whether to proceed with BME. In the context of suspected acute leukemia, the IPF could be included in the initial panel of tests to aid with rapid disease classification. However, the prognostic value of the IPF for various hematologic disorders should be explored in longitudinal studies. For example, serial IPF measurements could be used to monitor the response to treatment of patients with ITP or to predict relapse in those with acute leukemia.

Several seminal studies have advanced understanding of the IPF in various hematological settings, including ITP, MDS, and bone marrow transplantation.​.

For example, Kashiwagi et al.15. elucidated the clinical correlations of the IPF in ITP, while Takami et al.16. and Abe et al.17. expanded knowledge regarding transplantation and MDS. Briggs et al.18. established robust reference ranges and a methodology for IPF measurement in healthy and diseased cohorts. These studies provide further mechanistic and clinical perspectives on the role of the IPF in thrombopoiesis.

Although the present study provides valuable insight, it is important to acknowledge its limitations. The single-center design and strict inclusion criteria for untreated patients resulted in a modestly-sized cohort of patients with ITP (n = 20). Although this is consistent with the cohort size in previous studies (e.g., Araki et al.5., larger multicenter studies are needed to validate the findings in diverse populations. The single-center design of the present study may also limit the generalizability of the findings, and the follow-up period may have been insufficient to capture the long-term outcomes. In addition, the effects of concurrent treatments on the IPF were not comprehensively analyzed, and this should be a subject for future research. It is important to note that although the IPF represents a promising biomarker, it should be used in conjunction with other clinical and laboratory parameters. Previous studies have suggested several potential clinical applications for the IPF. For example, it could be used as a diagnostic tool in place of unnecessary bone marrow aspirations in some patients with thrombocytopenia19. IPF determination may also serve as an early predictor of bone marrow recovery after chemotherapy and stem cell transplantation20 to help these patients avoid unnecessary platelet transfusions20.

Importantly, the clinical value of bone marrow megakaryocyte counts should be interpreted with caution because this parameter is semiquantitative and subject to limited reproducibility between observers and studies. Variations in sample preparation, staining, and counting methods may introduce additional bias. This limitation highlights the importance of interpreting megakaryocyte findings in conjunction with other clinical and laboratory data, and of drawing definitive conclusions with caution. Although previous studies suggest that simultaneous evaluation of IPF and circulating thrombopoietin levels can enhance the mechanistic understanding of thrombopoietic activity17, thrombopoietin measurements were not comprehensively performed in all patients in our study.​ This omission limits further biological interpretation within our cohort. Future investigations integrating both parameters may provide additional valuable insights into the regulation of platelet production in hematological disorders.

In conclusion, the present comprehensive analysis of 1,552 patients undergoing BME and IPF analyses yielded several important findings. First, we identified a statistically significant correlation between the IPF value and bone marrow megakaryocyte number (R2 = 0.251, p < 0.001) in patients with MDS; however, the relationship was weak. Second, patients with ITP had a much higher median IPF compared with other patients (14.8% vs. 3.0%, respectively; p < 0.001). Patients with MDS displayed wide variability in their IPF values, but the IPF was highly specific for megakaryocyte abnormalities (p < 0.001). These findings underscore the potential for the use of a combination of IPF with BME findings to improve diagnostic accuracy in hematology. The present study provides a strong foundation for refining existing diagnostic approaches, particularly for conditions such as MDS and acute leukemia. These results highlight the need for a nuanced approach to the interpretation of BME findings in the context of thrombocytopenia, particularly when the underlying disease, such as MDS, involves complex pathophysiology. Further research is warranted to elucidate the mechanisms underlying the identified variation and to validate the use of the IPF, in the form of prospective studies of patients with MDS.

Methods

The study protocol was approved by the institutional review board of St. Mary’s Hospital, Japan (approval number: 24–0103), and the study was performed in accordance with the institutional guidelines and the principles of the Declaration of Helsinki. The need for informed consent was waived because of the retrospective, observational study design and the anonymization of the data in the institutional electronic database. Nevertheless, we obtained informed consent using an opt-out approach via the hospital website (there were no requests for exclusion).

Patient selection

We performed a retrospective study of data collected for 1,552 consecutive patients who underwent both BME and IPF measurement at St. Mary’s Hospital, Japan, between January 1, 2020, and December 31, 2024. To ensure consistency, only the first set of BME and IPF data obtained at the time of the initial diagnosis, before any therapeutic interventions, were included. Patients who subsequently received treatment were not excluded from the cohort but their treatment-related data were not analyzed because of incomplete documentation across the study period. We included patients with various hematologic abnormalities, ensuring that data for a diverse cohort would be available for a comprehensive analysis. At St. Mary’s Hospital, IPF measurement is a routine part of the hematologic workup and is performed in addition to BME for all patients who are referred for BME. This protocol was implemented at the start of the study period.

The decision to include the IPF as a standard test was based on emerging evidence of its diagnostic and prognostic value for various hematologic disorders. The institutional guidelines recommend IPF measurement in the following situations: (1) as part of the initial workup for patients undergoing BME, (2) for the monitoring of the response to treatment in patients with certain conditions, such as ITP, and (3) in cases of unexplained thrombocytopenia, even when BME is not immediately indicated. This comprehensive approach ensured that we obtained IPF data for all patients who under BME; thereby, minimizing selection bias for the study cohort. The 1,552 patients included in the analysis were consecutive patients who met the inclusion criteria, and there were no exclusions on the basis of IPF data availability. For the analysis that compared patients with ALL and those with AML, only newly diagnosed treatment-naïve patients were included. Patients with relapsed or refractory disease and those who had previously undergone treatment were excluded to ensure that the analyzed IPF values reflected the untreated disease state.

Data collection

All patients aged ≥ 18 years who were referred for BME were eligible for inclusion. The demographic data, medical history, and complete blood count were recorded for each patient. The follow-up period was from January 1, 2020, to December 31, 2024. All BME-related and IPF measurements were performed at the time of the initial diagnosis, before any therapeutic interventions. Patients who had previously undergone chemotherapy, received an immunosuppressive agent, or underwent another treatment that could have affected hematopoiesis, were excluded. This approach ensured that the reported IPF values and megakaryocyte counts reflected the untreated disease state and eliminated the confounding effects of previous therapy.

Inclusion and exclusion criteria

Eligible patients were those with newly diagnosed hematologic disorders who underwent BME and concurrent measurement of the IPF before the initiation of any treatment. Patients were excluded if they had a history of hematopoietic stem cell transplantation, therapy-related (secondary) leukemia, or prior exposure to chemotherapy, immunosuppressive agents, or other treatments that could affect hematopoiesis. Patients with incomplete medical records or who declined participation were also excluded. This protocol ensured that the analysis was conducted using data from untreated patients, allowing accurate assessment of IPF and bone marrow characteristics without confounding treatment effects.

IPF measurement

Before the BME procedure, the IPF for each patient was measured using an automated hematology analyzer (Sysmex XN-1000 hematology analyzer; Sysmex, Tokyo, Japan). All blood samples for the IPF measurement were collected in EDTA tubes and processed within 4 h of collection. Prior to analysis, specimens were stored and maintained at room temperature (18–25 °C). If delayed analysis was anticipated, samples were promptly refrigerated at 2–8 °C to preserve stability, in accordance with published guidelines21,22,23.​ Specimens were mixed gently at the time of collection and immediately before analysis to prevent platelet clumping, and quality controls were performed at least once daily using commercial control materials, in accordance with international guidelines for automated hematology analyzers23. IPF values increase with prolonged sample storage or excessive temperature variation; therefore, strict adherence to these protocols ensured the validity and reliability of all reported measurements. The IPF value is reported as the percentage of the total platelet count. IPF is a proprietary parameter trademarked by Sysmex Corporation. In this study, IPF refers specifically to a Sysmex-specific parameter reflecting the proportion of reticulated platelets (RP), calculated by flow cytometric detection of RNA-containing platelets using Sysmex hematology analyzers. RPs are the youngest circulating platelets containing residual RNA, and their proportion reflects the rate of thrombopoiesis. RPs can be quantified by flow cytometric analysis on the basis of their high RNA content, resulting in strong fluorescence after RNA-specific dye staining.

Biologically, elevated RP fractions indicate increased platelet production, as seen in conditions with accelerated bone marrow response to peripheral thrombocytopenia, while lower RP fractions may suggest impaired platelet production. Although the IPF and RPs are related parameters reflecting the population of newly released, RNA-rich platelets, these parameters are not identical. The IPF is a Sysmex-specific hematology analyzer parameter estimating the proportion of immature platelets using flow cytometric detection of RNA-containing platelets, while RPs are most commonly measured by fluorescence-based flow cytometry for residual platelet RNA.​ These measurements are generally correlated in peripheral blood; however, IPF values can be overestimated in the presence of megakaryocytes, such as in congenital macrothrombocytopenias (e.g., MYH9-related disorders) because platelet size affects fluorescence signal intensity. Thus, clinicians should interpret IPF values with caution in these conditions, and this limitation should be recognized when using IPF as a surrogate for RPs.

The reference range for the IPF depends on the hematology analyzer and the institution, and there is currently no international standard.​ Therefore, each laboratory should establish its own reference ranges for the IPF. At our institution, the reference range for healthy individuals (n = 35), as measured by the Sysmex XN-1000 analyzer, is a median IPF value of 3.5% (range: 1.0%–6.8%).​ This range was determined using samples from healthy adult volunteers and provides a basis for interpreting the results of the IPF values in this study. The IPF results reported in the present study were obtained using the Sysmex XN series hematology analyzer, which is an older-generation model. It is worth noting that IPF values may vary depending on the analyzer model because of differences in measurement technology and algorithms. Therefore, caution is warranted when comparing IPF results across different analyzer generations and institutions.

BME

Standard protocols were used to perform bone marrow aspirations and biopsies. Bone marrow aspirates were stained with May–Grünwald Giemsa, and biopsies were fixed in formalin, decalcified, and stained with hematoxylin and eosin. Special stains were used as needed.

Bone marrow analysis

Two independent hematopathologists, who were blinded to the IPF results, evaluated the bone marrow samples. The following parameters were assessed: (1) cellularity, categorized as hypocellular, normocellular, or hypercellular relative to age-adjusted norms; (2) megakaryocyte number, quantified per low-power field and classified as low, normal, or high; (3) megakaryocyte morphology (dysplastic features were recorded, including hypolobation, hyperlobation, and micromegakaryocytes); (4) erythroid and myeloid lineages (evaluated for maturation, dysplasia, and the presence of blasts); and (5) the presence of malignant cells or other pathologic findings. In instances of discrepancies in the findings, a third hematopathologist reviewed the samples, and a consensus was sought through discussion. This approach minimized inter-observer variability and ensured robust, reproducible morphological evaluation.

Diagnostic classification

Final diagnoses were made on the basis of clinical data, peripheral blood findings, BME results, and the results of cytogenetic analysis and molecular studies, in accordance with the World Health Organization classification of hematopoietic and lymphoid tissues.

Statistical analysis

All statistical analyses were performed using EZR (Saitama Medical Center, Saitama, Japan; http://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmedEN.html)24, a graphical user interface for R (The R Foundation for Statistical Computing; www.r-project.org) that extends the functionality of R Commander by adding statistical functions. Analyses were performed using R version 4.3.1 and R Commander version 2.9-1.9. Pearson’s correlation coefficient was used to assess the relationships between the IPF value and bone marrow megakaryocyte number in the entire cohort and in patients with MDS. The IPFs for the diagnostic groups were compared using the Mann–Whitney U test. Receiver operating characteristic curve analysis was initially used to identify potential predictors and determine optimal cutoff values for continuous variables. p < 0.05 was considered statistically significant.