Abstract
Background
Childhood myelodysplastic neoplasm (cMDS) is a rare condition with distinct biological and clinical characteristics. To characterize the epidemiological features of cMDS, including incidence, demographics, and underlying etiologies. We also aimed to identify prognostic factors and assess long-term outcomes associated with allogeneic hematopoietic stem cell transplantation (allo-HSCT) in patients with cMDS.
Methods
This nationwide Korean study analyzed 484 patients < 19 years of age diagnosed with cMDS between 2003 and 2021, using data from the Korean National Health Insurance Service database.
Results
Annual incidence ranged from 1.7 to 4.3 per million children, with 7.2% had a prior history of malignancy. A total of 180 patients underwent allo-HSCT for cMDS. The 5-year overall survival (OS) rates were 76.6% in the transplant group and 80.5% in the non-transplant group. Among non-transplanted patients, key prognostic factors included later period of diagnosis (hazard ratio [HR] 0.372, P = 0.012), history of prior malignancy (HR 3.329, P = 0.002), and chemotherapy for cMDS (HR 3.701, P = 0.002). Only CBT was significantly associated with increased mortality in univariate analysis of the transplant group (HR, 3.911; P = 0.008). The 10-year cumulative incidence of leukemia was 20.5%, with a significantly higher rate in the cMDS with increased blasts (cMDS-IB) group (53.3% vs. 19.4%, P = 0.002).
Conclusions
While many patients achieved stable disease control without allo-HSCT, monitoring remains essential due to the risk of leukemic transformation, particularly in therapy-related and cMDS-IB subtypes.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12885-026-16502-4.
Keywords: Childhood, MDS, Prevalence, Allo-HSCT, Long term outcomes, Leukemic transformation
Introduction
Myelodysplastic neoplasm (MDS) is a clonal hematopoietic disorder characterized by peripheral cytopenia, ineffective hematopoiesis, and an increased risk of progression to acute myeloid leukemia (AML). Although MDS predominantly affects the elderly, childhood MDS (cMDS) is exceedingly rare, with an annual incidence of only 1 to 4 cases per million children [1]. cMDS presents with distinct biological and clinical features, frequently involving hypocellular bone marrow (~ 80%) [2, 3] and may arise de novo or in association with predisposing factors such as prior exposure to cytotoxic agents, inherited bone marrow failure syndromes, or genetic predisposition syndromes [4]. The classification of cMDS has evolved, from pediatric adaptations of the original World Health Organization (WHO) classification to the 2022 WHO and international consensus classifications, which incorporate refinements based on histopathological features and blast percentages [5–7].
Allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains the standard of care for high-risk cMDS, particularly in cases involving increased blasts, therapy-related cMDS, or low-blast disease with high-risk cytogenetic abnormalities [8]. However, due to the rarity of cMDS, data on its clinical characteristics, long-term treatment outcomes, and prognostic factors following allo-HSCT remain limited. Current understanding is primarily based on analyses from relatively small or single-institution cohorts [9–15].
Furthermore, population-based studies investigating the true incidence, demo graphic patterns, and underlying etiologies of cMDSare particularly limited in the literature [4, 16–20]. To date, no comprehensive population-based study of cMDS has been conducted in Korea. Here, we performed a population-based analysis using Korea’s National Health Insurance Service (NHIS) database. The objective was to characterize the epidemiological features of cMDS, including incidence, demographics, and underlying etiologies. Additionally, we aimed to identify prognostic factors and assess long-term outcomes associated with allo-HSCT in patients with cMDS.
Materials and methods
Data source and study population
Korea operates a universal single-payer national health system, and the NHIS maintains comprehensive national records of all covered inpatient and outpatient visits, procedures, and prescriptions. The NHIS database includes personal demographic information, health insurance claim codes (for procedures and prescriptions), diagnostic codes based on the Korean Standard Classification of Diseases, 7th Revision, which is aligned with the International Classification of Diseases, 10th Revision (ICD-10), as well as death records from the Korean National Statistical Office and general health checkup data [21].
This retrospective cohort study utilized claims data from the NHIS database spanning from 2002 to 2022. The study population consisted of individuals under the age of 19 who were diagnosed with cMDS, identified by ICD-10 code D46, between 2003 and 2021 (n = 644). To include newly diagnosed cMDS patients, we excluded individuals diagnosed with cMDS in 2002 (n = 30), and those who had a diagnosis of leukemia (ICD-10 codes C91–C95) prior to baseline (n = 141). The final study cohort consisted of 484 patients (Supplementary Fig. 1).
Measurements
Childhood MDS with increased blasts (cMDS-IB) was defined as having an ICD-10 code of D46.2 at the time of diagnosis. All other diagnoses (D46.0–D46.9) were classified as cMDS with low blasts (cMDS-LB). Transplantation procedures were categorized as allogeneic bone marrow (BM), peripheral blood (PB), or cord blood (CB) transplants based on relevant procedure codes (Supplementary Table 1). Patients who underwent transplantation after leukemic transformation were excluded from the transplantation outcome analysis. History of malignancy was defined as having any ICD-10 code within the C category prior to the initial diagnosis of cMDS. In transplant patients, chemotherapy for cMDS was defined as administration of any of the following agents more than one month before transplantation: cytarabine, mitoxantrone, thiotepa, daunorubicin, busulfan, etoposide, thioguanine, fludarabine, idarubicin, azacitidine, decitabine, or lenalidomide. In non-transplant patients, chemotherapy was defined as treatment with any of above agents from the cMDS diagnosis until the development of leukemia. Conditioning regimens were identified using claim codes (Supplementary Table 2) within 14 days before transplantation and were classified as either total body irradiation (TBI)-based or busulfan-based. The incidence of leukemia was defined by the appearance of ICD-10 codes C91-C95 after the initial diagnosis of cMDS.
Demographic data, including age, sex, area of residence (metropolitan vs. provincial), and household income (based on insurance premium percentiles: 0–25, 26–50, 51–75, 76–100, and unknown), were obtained from the insurance eligibility database. Metropolitan areas were defined as cities with populations exceeding one million.
Statistical analysis
The annual incidence of cMDS in the general population aged ≤ 18 years was estimated using data from the Korean Statistical Information Service [22]. Participant characteristics were summarized as mean ± standard deviation or median (interquartile range, IQR) for continuous variables depending on the normality of the data distribution and as number (percentage) for categorical variables. Group comparisons were conducted using independent two-sample t-tests or Mann-Whitney U tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Non-transplant patients were followed until death within five years of cMDS diagnosis or until the end of the study period (December 31, 2022), whichever occurred first. Transplant patients were followed from the date of transplantation. Survival curves were generated using the Kaplan–Meier method and compared using log-rank tests. Furthermore, patients who underwent transplantation after leukemic transformation were excluded from the analyses of transplant characteristics and survival outcomes. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated using Cox proportional hazards regression models to identify factors associated with 5-year mortality in both groups. Multivariable analyses included covariates that were statistically significant in univariable analyses. The cumulative incidence of leukemic transformation was estimated for all patients, including those excluded from the survival analyses, using the cumulative incidence function with death before leukemic transformation treated as a competing event. Group comparisons were performed using Gray’s test. The proportional hazards assumption was evaluated using log(-log) survival plots and Schoenfeld residuals. A P-value of < 0.05 was considered statistically significant. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
Results
Study population and incidence of cMDS
A total of 484 patients were newly diagnosed with cMDS between 2003 and 2021. The most common subtype was cMDS-LB, accounting for 96.9% of cases (469/484). The annual incidence of cMDS during this period ranged from 1.7 to 4.3 per million children aged 0–18 years, with a mean annual incidence of 2.5 per million (95% CI: 2.3–2.7 per million). No consistent trend in incidence was observed over time (Fig. 1, Supplementary Table 3).
Fig. 1.
Number and annual incidence of cMDS between 2003 and 2021. Blue bars indicate the annual number of newly diagnosed cMDS patients (left y-axis). The orange line represents the annual incidence rate of cMDS per million children (right y-axis). The incidence rate was calculated based on the total population of children aged 0–18 years in Korea for each corresponding year
The overall male-to-female ratio was 1.2, and the median age at diagnosis was 10 years (Table 1). When stratified by age group, the most common age at diagnosis was 15–18 years (26.9%), followed closely by the 5–9 years group (26.2%). Patients with cMDS-IB had a lower median age at diagnosis than those with cMDS-LB, although the number of cMDS-IB cases was small (n = 15). This difference showed a trend toward statistical significance (median age: 6 years vs. 10 years, P = 0.066). Geographically, the distribution of cMDS cases across prefectures did not significantly differ by level of urbanization. However, patients with cMDS-IB were more likely to reside in metropolitan areas compared to those with cMDS-LB (60.0% vs. 46.1%, P = 0.011).
Table 1.
Characteristics of cMDS by cMDS-IB and cMDS-LB
| Characteristics | Overall | cMDS-LB | cMDS-IB | P-value |
|---|---|---|---|---|
| No. participants | 484 | 469 | 15 | |
| Age at diagnosis, yrs | ||||
| Median (Q1 - Q3) | 10 (5–15) | 10 (5–15) | 6 (1–13) | 0.066 |
| 0–4 | 108 (22.3) | 101 (21.5) | 7 (46.7) | 0.113 |
| 5–9 | 127 (26.2) | 123 (26.2) | 4 (26.7) | |
| 10–14 | 119 (24.6) | 118 (25.2) | 1 (6.7) | |
| 15–18 | 130 (26.9) | 127 (27.1) | 3 (20.0) | |
| Sex | 0.898 | |||
| Male | 266 (55.0) | 258 (55.0) | 8 (53.3) | |
| Female | 218 (45.0) | 211 (45.0) | 7 (46.7) | |
| Year of diagnosis | 0.534 | |||
| 2003–2012 | 264 (54.5) | 257 (54.8) | 7 (46.7) | |
| 2013–2021 | 220 (45.5) | 212 (45.2) | 8 (53.3) | |
| Household income | 0.521 | |||
| 0–25 | 92 (19.0) | 88 (18.8) | 4 (26.7) | |
| 26–50 | 69 (14.3) | 65 (13.9) | 4 (26.7) | |
| 51–75 | 128 (26.4) | 125 (26.5) | 3 (20.0) | |
| 76–100 | 182 (37.6) | 178 (37.9) | 4 (26.7) | |
| Unknown | 13 (2.7) | 13 (2.8) | 0 (0) | |
| Residence | 0.011 | |||
| Provincial | 258 (53.3) | 253 (53.9) | 5 (33.3) | |
| Metropolitan | 225 (46.5) | 216 (46.1) | 9 (60.0) | |
| Unknown | 1 (0.2) | 0 (0) | 1 (6.7) | |
| History of prior malignancy | 35 (7.2) | 33 (7.0) | 2 (13.3) | 0.296 |
| Chemotherapy | 53 (10.9) | 46 (9.8) | 7 (46.7) | < 0.001 |
| Cytarabine | 39 (73.6) | 33 (71.7) | 6 (85.7) | 0.660 |
| Mitoxantrone | 4 (7.6) | 4 (8.7) | 0 (0) | > 0.999 |
| Daunorubicin | 9 (17.0) | 6 (13.0) | 3 (42.9) | 0.086 |
| Etoposide | 19 (35.9) | 18 (39.1) | 1 (14.3) | 0.400 |
| Thioguanine | 4 (7.6) | 4 (8.7) | 0 (0) | > 0.999 |
| Fludarabine | 3 (5.7) | 3 (6.5) | 0 (0) | > 0.999 |
| Idarubicin | 11 (20.8) | 10 (21.7) | 1 (14.3) | > 0.999 |
| Azacitidine | 4 (7.6) | 4 (8.7) | 0 (0) | > 0.999 |
| Decitabine | 3 (5.7) | 3 (6.5) | 0 (0) | > 0.999 |
| Lenalidomide | 0 (0) | 0 (0) | 0 (0) | - |
| Stem cell sourcea | 0.731 | |||
| Bone marrow | 25 (13.9) | 25 (14.4) | 0 (0) | |
| Peripheral blood | 145 (80.5) | 138 (79.8) | 7 (100) | |
| Cord blood | 10 (5.6) | 10 (5.8) | 0 (0) |
Values are presented as the number of patients (%), mean ± standard deviation, and median (interquartile range)
a Analysis of 180 patients who underwent allo-HSCT for MDS
Among all patients with cMDS, 7.2% had a prior history of malignancy, and the mean interval between the diagnosis of the primary malignancy and the subsequent development of cMDS was 3.0 years. Of these patients, 26 had prior solid tumors and 9 had prior hematologic malignancies. The most common preceding malignancies were central nervous system (CNS) tumors (n = 7), followed by bone tumors (n = 5) and soft tissue tumors (n = 4) (Supplement Table 4). The mean age at cMDS diagnosis in these patients was 11.1 years (95% CI: 6.1–16.1 years), which was older than in patients with de novo cMDS.
Characteristics of treatment and transplantations
Chemotherapy was administered to 10.9% of the total patient population (Table 1). The proportion of patients receiving chemotherapy was significantly higher in the cMDS-IB group compared to the cMDS-LB group (46.7% vs. 9.8%, P < 0.001). Among the chemotherapeutic agents used, cytarabine was the most commonly administered (73.6%), followed by etoposide (35.9%), idarubicin (20.8%), and daunorubicin (17.0%).
Of the 484 patients, 215 underwent allo-HSCT. Among them, 35 were excluded from further analysis because transplantation occurred after a diagnosis of leukemia. Therefore, the final cohort for cMDS outcome analysis consisted of 180 patients in the transplant group and 269 in the non-transplant group (Table 2). When classified by hematopoietic cell source, PB stem cells were the most commonly used (80.6%), followed by BM (13.9%) and CB (5.6%). The use of PB as a stem cell source increased in the latter half of the study period (2013–2021) compared with the first half (2003–2012) (33.8% vs. 66.2%). In contrast, the use of BM (72.0% vs. 28.0%) and CB (70.0% vs. 30.0%) decreased over the same period. In transplants using BM or PB, busulfan-based conditioning regimens were more commonly employed than TBI-based regimens (BM: 36.0% vs. 4.0%; PB: 46.9% vs. 17.9%). Conversely, in cord blood transplants (CBT), TBI-based regimens were more frequently used than busulfan-based regimens (60.0% vs. 30.0%).
Table 2.
Characteristics of MDS with transplantation by stem cell source
| Characteristics | Non-transplant group | Transplant group | P-valuea | |||
|---|---|---|---|---|---|---|
| Overall | Bone marrow | Peripheral blood | Cord blood | |||
| No. participants | 269 | 180 | 25 | 145 | 10 | |
| Age at diagnosis, yrs | 0.749 | |||||
| 0–9 | 133 (49.4) | 87 (48.3) | 12 (48.0) | 69 (47.6) | 6 (60.0) | |
| 10–18 | 136 (50.6) | 93 (51.7) | 13 (52.0) | 76 (52.4) | 4 (40.0) | |
| Age at allo-HSCT, yrs | 0.900 | |||||
| 0–9 | - | 63 (35.0) | 8 (32.0) | 51 (35.2) | 4 (40.0) | |
| ≥ 10 | - | 117 (65.0) | 17 (68.0) | 94 (64.8) | 6 (60.0) | |
| Sex | 0.220 | |||||
| Male | 140 (52.0) | 101 (56.1) | 18 (72.0) | 78 (53.8) | 5 (50.0) | |
| Female | 129 (48.0) | 79 (43.9) | 7 (28.0) | 67 (46.2) | 5 (50.0) | |
| Year of diagnosis | < 0.001 | |||||
| 2003–2012 | 172 (63.9) | 74 (41.1) | 18 (72.0) | 49 (33.8) | 7 (70.0) | |
| 2013–2021 | 97 (36.1) | 106 (58.9) | 7 (28.0) | 96 (66.2) | 3 (30.0) | |
| History of prior malignancy | 19 (7.1) | 11 (6.1) | 0 (0) | 10 (6.9) | 1 (10.0) | 0.278 |
| Subtype | 0.731 | |||||
| cMDS-IB | 5 (1.9) | 7 (3.9) | 0 (0) | 7 (4.8) | 0 (0) | |
| cMDS-LB | 264 (98.1) | 173 (96.1) | 25 (100) | 138 (95.2) | 10 (100) | |
| Household income | 0.572 | |||||
| 0–25 | 52 (19.3) | 325 (17.8) | 1 (4.0) | 28 (19.3) | 3 (30.0) | |
| 26–50 | 41 (15.2) | 23 (12.8) | 4 (16.0) | 18 (12.4) | 1 (10.0) | |
| 51–75 | 71 (26.4) | 52. (28.9) | 8 (32.0) | 41 (28.3) | 3 (30.0) | |
| 76–100 | 99 (36.8) | 67 (37.2) | 12 (48.0) | 52 (35.9) | 3 (30.0) | |
| Unknown | 6 (2.2) | 6 (3.3) | 0 (0) | 6 (4.1) | 0 (0) | |
| Residence | 0.530 | |||||
| Provincial | 150 (55.8) | 86 (47.8) | 15 (60.0) | 67 (46.2) | 4 (40.0) | |
| Metropolitan | 119 (44.2) | 93 (51.7) | 10 (40.0) | 77 (53.1) | 6 (60.0) | |
| Unknown | 0 (0) | 1 (0.6) | 0 (0) | 1 (0.7) | 0 (0) | |
| Chemotherapy for cMDSb | 12 (4.5) | 30 (16.7) | 3 (12.0) | 23 (15.9) | 4 (40.0) | 0.123 |
| Type of conditioning | 0.005 | |||||
| TBI-based | - | 33 (18.3) | 1 (4.0) | 26 (17.9) | 6 (60.0) | |
| Busulfan-based | - | 80 (44.4) | 9 (36.0) | 68 (46.9) | 3 (30.0) | |
| TBI and Busulfan | 5 (2.8) | 0 (0) | 5 (30.5) | 0 (0) | ||
| Others | - | 62 (34.4) | 15 (60.0) | 46 (31.7) | 1 (10.0) | |
Values are presented as the number of patients and percentages
a Comparative analysis according to stem cell source (BM vs. PB vs. CB)
b In patients who underwent hematopoietic stem cell transplantation, only chemotherapy administered for cMDS prior to transplantation was included in the analysis
Survival outcomes and associated factors
Among the 180 patients who underwent allo-HSCT for the treatment of cMDS, the 5-year overall survival (OS) probability was 76.6% (Fig. 2A). The 5-year OS probability among the 269 patients who did not undergo transplantation was 80.5%. Results from both univariable and multivariable analyses of factors associated with mortality, as well as survival curves stratified by these factors, are presented in Table 3, and Fig. 2B-D.
Fig. 2.
(A) Overall survival (OS) after allogeneic hematopoietic stem cell transplantation in the transplant group and after diagnosis of myelodysplastic syndrome (MDS) in the non-transplant group. (B–D) OS according to the following clinical factors in the non-transplant group: (B) period of diagnosis, (C) presence of therapy-related MDS, and (D) presence of increased blasts. Follow-up was calculated from the date of transplantation in the transplant group and from the date of cMDS diagnosis in the non-transplant group
Table 3.
Factors affecting 5-year mortality using the Cox proportional hazard model
| Variables | Non-transplantation (n=269) | Transplantation (n=180) | ||||||
|---|---|---|---|---|---|---|---|---|
| No. death | Univariable | P-value | Multivariable | P-value | No. death | Univariable | P-value | |
| HR (95% CI) | HR (95% CI) | HR (95% CI) | ||||||
| Age at diagnosis | ||||||||
| 0–9 | 28 | Reference | 14 | Reference | ||||
| 10–19 | 23 |
0.761 (0.438–1.321) |
0.331 | 25 |
1.733 (0.901–3.335) |
0.100 | ||
| Age at allo-HSCT | ||||||||
| 0–9 | 12 | Reference | ||||||
| ≥ 10 | 27 |
1.223 (0.619–2.414) |
0.689 | |||||
| Sex, Female | 23 |
0.887 (0.511–1.541) |
0.671 | 18 |
1.128 (0.601–2.118) |
0.707 | ||
| Year of diagnosis | ||||||||
| 2003–2012 | 43 | Reference | Reference | 21 | Reference | |||
| 2013–2021 | 8 |
0.327 (0.154–0.696) |
0.004 |
0.372 (0.173–0.802) |
0.012 | 18 |
0.637 (0.378–1.073) |
0.090 |
| Stem cell source | 0.008 | |||||||
| Bone marrow | - | 5 | Reference | |||||
| Peripheral blood | - | 29 |
1.141 (0.486–2.678) |
|||||
| Cord blood | - | 5 |
3.911 (1.313–11.650) |
|||||
| History of prior malignancy | 9 |
3.754 (1.824–7.726) |
< 0.001 |
3.329 (1.584–6.996) |
0.002 | 4 |
1.749 (0.622–4.924) |
0.290 |
| Subtype, cMDS-IB | 3 |
3.777 (1.176–12.134) |
0.026 |
1.478 (0.399–5.479) |
0.559 | 2 |
1.233 (0.297–5.119) |
0.773 |
| Household income | 0.624 | 0.869 | ||||||
| 0–25 | 10 | Reference | 6 | Reference | ||||
| 25–50 | 7 |
0.860 (0.327–2.260) |
4 |
0.845 (0.238–2.995) |
||||
| 50–75 | 10 |
0.693 (0.292–1.689) |
14 |
1.465 (0.563–3.812) |
||||
| 75–100 | 22 |
1.160 (0.549–2.449) |
15 |
1.173 (0.455–3.025) |
||||
| Unknown | 2 |
1.830 (0.401–8.350) |
0 | - | ||||
| Residence | 0.266 | |||||||
| Provincial | 27 | Reference | 20 | Reference | ||||
| Urban | 24 |
1.155 (0.666–2.002) |
0.608 | 18 |
0.805 (0.426–1.523) |
|||
| Unknown | 0 | - | 1 |
40.76 (0.546–30.439) |
||||
| Chemotherapy for cMDS | 9 |
5.957 (2.879–12.326) |
< 0.001 |
3.701 (1.597–8.576) |
0.002 | 10 |
1.734 (0.845–3.559) |
0.134 |
| Type of conditioning | 0.237 | |||||||
| Others | - | 8 | Reference | |||||
| TBI-based | - | 9 |
2.474 (0.952–6.425) |
|||||
| Busulfan-based | - | 22 |
2.146 (0.955–4.823) |
|||||
| TBI & Busulfan | - | 0 | - | |||||
Multivariable analysis includes variables with P-values < 0.05 in univariable analysis
In the non-transplant group, a later period of diagnosis (2013–2021) was associated with significantly reduced mortality compared to the earlier period (2003–2012), while history of malignancy, the cMDS-IB subtype, and receipt of chemotherapy for MDS were associated with increased mortality. In multivariate analysis, a later period of diagnosis remained significantly associated with reduced mortality (HR, 0.372; 95% CI, 0.173–0.802; P = 0.012), while history of malignancy (HR, 3.329; 95% CI, 1.584–6.996; P = 0.002) and chemotherapy for MDS (HR, 3.701; 95% CI, 1.597–8.576; P = 0.002) were independently associated with increased mortality. In the transplant group, only CBT was significantly associated with increased mortality in the univariate analysis (HR, 3.911; 95% CI, 1.313–11.650; P = 0.008).
Development of leukemia after cMDS diagnosis
A total of 97 patients developed leukemia following a diagnosis of cMDS, including 89 patients from the cMDS-LB group and eight from the cMDS-IB group. This cohort includes the 35 patients who were excluded from the survival analyses due to undergoing transplantation after leukemic transformation. Accounting for competing risks, the 10-year cumulative incidence of leukemia was 20.5% (Fig. 3A). Most cases of leukemic transformation occurred within three years of the initial cMDS diagnosis, with a 3-year incidence of 18.9%. The incidence of leukemia was significantly higher in the cMDS-IB group compared to the cMDS-LB group. At 10 years, the cumulative incidence was 53.3% in the cMDS-IB group versus 19.4% in the cMDS-LB group (P = 0.002) (Fig. 3B).
Fig. 3.
(A) Cumulative incidence of leukemia in all study patients. (B) Cumulative incidence of leukemia stratified by myelodysplastic syndrome subtype
Discussion
We conducted a nationwide study of 484 patients with cMDS, finding an annual incidence of 2.5 per million children in Korea. Most were diagnosed with cMDS-LB, with 7.2% having a prior history of malignancy. Notably, many with cMDS-LB achieved stable disease without allo-HSCT, with 5-year OS rates of 76.6% for the transplant group and 80.5% for the non-transplant group. The 10-year incidence of leukemia was 20.5%.
Similar to other landmark population-based analyses [16–18], our nationwide cohort provides unique insights through comprehensive and unbiased case identification over an extended timeframe. Leveraging the NHIS database enabled us to generate robust risk estimates and clinical outcome data for cMDS, extending beyond childhood and into adulthood. This prolonged follow-up is a key methodological strength, allowing us to capture the full natural history of cMDS and document long-term outcomes with greater accuracy.
In our study, the cMDS-IB subtype accounted for only 3.1% of cases, which is lower than the 10–25% range reported in previous studies [23, 24]. This discrepancy can be explained by several factors. First, following WHO guidelines, patients with Down syndrome showing MDS-like features were likely classified under “myeloid leukemia associated with Down syndrome” rather than cMDS-IB. Previous research has demonstrated that excluding Down syndrome patients reduce advanced MDS diagnoses by approximately 50% [25]. Additionally, in the Korean ICD system, refractory anemia with excess blasts in transformation (RAEB-t) is included as a subcategory of AML (ICD code C92.0), thus RAEB-t cases would have been excluded from our cMDS-IB cohort. Furthermore, the significantly higher rate of cMDS-IB diagnoses in metropolitan areas suggests potential underdiagnosis in non-metropolitan regions, likely due to differences in diagnostic infrastructure or specialist availability. The low proportion of cMDS-IB cases likely contributed to the lower overall incidence of leukemic transformation observed in our cohort (20.0%) compared to previous reports (31% in pediatric and young adult MDS patients, and approximately 40% in children) [13].
While the cumulative incidence of therapy-related MDS/AML ranges from 5% to 11% in children treated with standard solid-tumor protocols [8, 26, 27], our study was unable to directly assess therapy-related etiology due to the limitations of claims-based data. Instead, we identified prior history of malignancy as a significant adverse prognostic factor, particularly in the non-transplant group, with a median interval of 3.0 years from prior malignancy diagnosis to cMDS onset. These findings underscore the importance of evaluating prior malignancy history in the diagnostic and prognostic evaluation of cMDS.
Allo-HSCT, the only curative treatment for cMDS, is typically indicated for patients with high-risk features, such as transfusion dependence, severe neutropenia, excess blasts, monosomy 7, del(7q), complex karyotype, or therapy-related disease [8, 9]. Transplantation may be considered at any disease stage, regardless of cMDS subtype, pretransplant treatment, or remission status, and offers the potential to prevent progression to AML [13, 28]. Historically, HSCT has achieved cure rates of 60–80% in refractory cytopenia of childhood (RCC) and approximately 50–60% in advanced MDS [9–15]. Patients with cMDS-LB generally show more favorable outcomes compared to those with excess blasts [11]. The European Working Group on Childhood MDS (EWOG-MDS) reported 3-year OS and event-free survival rates of 84% and 74%, respectively, among RCC patients who underwent reduced-intensity conditioning (RIC) with unrelated donor transplants [29]. Similarly, the Center for International Blood and Marrow Transplant Research (CIBMTR) reported 8-year disease-free survival (DFS) rates of 65% and 40% for matched and mismatched unrelated donor bone marrow transplants, respectively, in RCC [30]. In our study, the 5-year post-transplant OS rate for patients with cMDS-LB was approximately 77%, consistent with these previously reported outcomes.
Treating patients with cMDS-IB remains particularly challenging. The EWOG-MDS study reported a 5-year OS rate of 63% in children with advanced MDS who underwent HSCT from HLA-matched sibling or unrelated donors [31]. In the CIBMTR study, 8-year DFS rates following unrelated donor transplants were 48% for patients with refractory anemia with excess blasts (RAEB) and 28% for those with RAEB-t [30]. While transplant-related mortality is the leading cause of treatment failure after myeloablative conditioning in RCC [10], relapse is the predominant cause of failure in advanced MDS [31, 32]. Notably, in our study, patients with cMDS-IB in the non-transplant group had poorer survival compared to those with cMDS-LB. In contrast, no significant survival difference was observed between subtypes in the transplant group. However, this finding should be interpreted with caution due to the small number of patients with excess blasts.
The optimal conditioning regimen prior to HSCT in MDS remains a subject of debate. In our cohort, approximately one-third of transplant recipients did not receive TBI or busulfan, suggesting the use of RIC. However, due to the lack of detailed information on drug or radiation dosages, the actual proportion of patients receiving RIC may be higher. Busulfan was the most frequently used conditioning agent, while TBI was more commonly employed in CBT, consistent with previous findings that TBI is administered in approximately half of cases [33]. We found no significant differences in survival based on conditioning regimen, although the absence of dosage data limits definitive conclusions regarding conditioning intensity. Previous studies have suggested that myeloablative conditioning regimens—including agents such as busulfan, cyclophosphamide, and melphalan—are preferred in patients with increased blasts [8], while RIC regimens incorporating thiotepa, treosulfan, fludarabine, or aplastic anemia–oriented approaches may be more appropriate for patients with RCC [29, 34, 35].
Although CB is a valuable alternative for patients lacking a matched sibling or suitable unrelated adult donor, outcomes following CBT are generally inferior to those achieved with BM or PB stem cell transplants [36]. In our study, CB was associated with increased mortality in univariate analysis. However, considering the characteristics of this study, this association may have been influenced by factors beyond graft source, including disease status, conditioning regimen, or transplant center experience. CBT outcomes have improved over time with advances in technique. A large registry-based study reported an increase in 3-year DFS from 27% before 2001 to 50% in subsequent years [33]. More recently, a study in children with high-risk myeloid malignancies showed improved graft-versus-host disease-free, relapse-free survival following T cell–replete CBT, although only 10 patients in the cohort had MDS [37]. Further data in pediatric MDS populations are needed to validate these findings.
In our study, 10.9% of patients received chemotherapy for MDS; however, it was not associated with improved survival. Rather, it was associated with a higher mortality rate in the non-transplant group. However, due to limited availability of detailed clinical information on treatment indications (e.g., baseline disease risk) and treatment-related mortality (e.g., cause of death), the contribution of chemotherapy to mortality cannot be definitively determined in this study. likely due to both the presence of adverse prognostic factors in patients who received chemotherapy and increased treatment-related toxicity. Although some reports suggest that cytarabine-based induction therapy may result in favorable responses [38], its role in advanced MDS remains controversial [8]. Among non-transplant patients, AML-type chemotherapy is associated with poor outcomes (OS < 30%) due to high induction failure and relapse rates [31], and is therefore not routinely recommended. The hypomethylating agent azacitidine has been shown to induce bone marrow responses and improve event-free survival [28, 39], although its effect on post-HSCT OS remains uncertain [13].
This study has several inherent limitations related to the use of NHIS data. Important clinical and transplant-related variables, including cytogenetic findings, genomic profiles (such as germline predisposition syndromes and genetic mutations), donor type, and conditioning intensity, were not available, limiting adjustment for potential confounders and raising the possibility of residual confounding. Future large-scale studies incorporating genetic variation, a core factor in the diagnosis, treatment, and prognosis of cMDS, are warranted to complement population-based findings. Additionally, the absence of data on relapse and transplant-related mortality limits our ability to fully interpret the causes of treatment failure. Furthermore, the findings from the multivariable Cox regression analyses should be interpreted cautiously, particularly for the transplant subgroup, given the limited number of outcome events and the resulting potential for model instability. Because variable selection was based on univariable significance in order to limit the number of covariates and reduce overparameterization, the possibility of overfitting cannot be excluded. Nevertheless, the study’s strengths include the use of a nationwide, population-based database encompassing nearly the entire Korean population. This enabled a robust analysis of clinical characteristics and outcomes in a rare disease population that is otherwise difficult to study at scale.
In conclusion, this nationwide Korean study provides comprehensive epidemiological data, treatment outcomes, and prognostic insights for cMDS in a large cohort. Our findings highlight that while allo-HSCT remains a cornerstone of curative treatment, a substantial proportion of patients, particularly those with cMDS-LB, who represented the majority of our cohort, achieved stable disease control without undergoing transplantation. However, careful risk stratification is critical, as therapy-related etiology and increased blast percentage were associated with significantly poorer survival in non-transplanted patients and a higher risk of leukemic transformation, respectively.
Supplementary Information
Acknowledgements
None.
Abbreviations
- allo-HSCT
Allogeneic hematopoietic stem cell transplantation
- AML
Acute myeloid leukemia
- BM
Bone marrow
- CB
Cord blood
- CBT
Cord blood transplants
- CIBMTR
Center for International Blood and Marrow Transplant Research
- CIs
Confidence intervals
- cMDS
Childhood MDS
- cMDS-IB
Childhood MDS with increased blasts
- cMDS-LB
Childhood MDS with low blasts
- DFS
Disease-free survival
- EWOG-MDS
European Working Group on Childhood MDS
- HR
Hazard ratios
- ICD-10
International Classification of Diseases, 10th Revision
- MDS
Myelodysplastic neoplasm
- NHIS
National Health Insurance Service
- OS
Overall survival
- PB
Peripheral blood
- RAEB
Refractory anemia with excess blasts
- RCC
Refractory cytopenia of childhood
- TBI
Total body irradiation
- WHO
World Health Organization
Authors' contributions
ESY and JY was responsible for study conceptualization and design, data interpretation, and writing of manuscript, design of figures/tables. MK was responsible for data collection, data analysis, data interpretation, visualization, and writing of manuscript, design of figures/tables. All authors participated in the writing and review of this article, and approved the final submitted version.
Funding
This study was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (RS-2025-23523376) and a grant funded by Korea University Medicine & Seoul Clinical Laboratories (SCL) Grant. This study was also supported by Korea university Guro hospital (Korea research-driven hospital) and grant funded by Korea University Medicine (K2407191).
Data availability
All the collected medical information was stored in National Health Insurance Service (NHIS) database. The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This retrospective nationwide cohort study was conducted following the principles of the Declaration of Helsinki. This study was approved by the Institutional Review Board of Korea University Guro Hospital, which waived the requirement for informed consent due to the use of publicly available de-identified data (2023GR0313). The study was also conducted with approval to use the NHIS database (NHIS-2024-1-524).
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Eun Sang Yi, Email: eunsangyi@korea.ac.kr.
Jung Yoon, Email: unoaotro@korea.ac.kr.
References
- 1.Niemeyer CM, Baumann I. Classification of childhood aplastic anemia and myelodysplastic syndrome. Hematol Am Soc Hematol Educ Program. 2011;2011:84–9. [DOI] [PubMed] [Google Scholar]
- 2.Baumann I, Führer M, Behrendt S, Campr V, Csomor J, Furlan I, et al. Morphological differentiation of severe aplastic anaemia from hypocellular refractory cytopenia of childhood: reproducibility of histopathological diagnostic criteria. Histopathology. 2012;61(1):10–7. [DOI] [PubMed] [Google Scholar]
- 3.Yang W, Zhang P, Hama A, Ito M, Kojima S, Zhu X. Diagnosis of acquired bone marrow failure syndrome during childhood using the 2008 World Health Organization classification system. Int J Hematol. 2012;96(1):34–8. [DOI] [PubMed] [Google Scholar]
- 4.Hasle H. Myelodysplastic and myeloproliferative disorders of childhood. Hematol Am Soc Hematol Educ Program. 2016;2016(1):598–604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Khoury JD, Solary E, Abla O, Akkari Y, Alaggio R, Apperley JF, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia. 2022;36(7):1703–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Arber DA, Orazi A, Hasserjian RP, Borowitz MJ, Calvo KR, Kvasnicka HM, et al. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rudelius M, Weinberg OK, Niemeyer CM, Shimamura A, Calvo KR. The International Consensus Classification (ICC) of hematologic neoplasms with germline predisposition, pediatric myelodysplastic syndrome, and juvenile myelomonocytic leukemia. Virchows Arch. 2023;482(1):113–30. [DOI] [PubMed] [Google Scholar]
- 8.Locatelli F, Strahm B. How I treat myelodysplastic syndromes of childhood. Blood. 2018;131(13):1406–14. [DOI] [PubMed] [Google Scholar]
- 9.Patel SS. Pediatric Myelodysplastic Syndromes. Clin Lab Med. 2021;41(3):517–28. [DOI] [PubMed] [Google Scholar]
- 10.Starý J, Locatelli F, Niemeyer CM. Stem cell transplantation for aplastic anemia and myelodysplastic syndrome. Bone Marrow Transpl. 2005;35(Suppl 1):S13–6. [DOI] [PubMed] [Google Scholar]
- 11.Chisholm KM, Bohling SD. Childhood Myelodysplastic Syndrome. Clin Lab Med. 2023;43(4):639–55. [DOI] [PubMed] [Google Scholar]
- 12.Wlodarski MW, Hirabayashi S, Pastor V, Starý J, Hasle H, Masetti R, et al. Prevalence, clinical characteristics, and prognosis of GATA2-related myelodysplastic syndromes in children and adolescents. Blood. 2016;127(11):1387–97. quiz 518. [DOI] [PubMed] [Google Scholar]
- 13.McCall D, Abuasab T, Rodriguez-Sevilla JJ, Mohamed SF, Patnaik A, Devireddy K, et al. Characteristics and outcomes of children, adolescent, and young adult patients with myelodysplastic neoplasms: A single-center retrospective analysis. Leuk Res. 2024;144:107563. [DOI] [PubMed] [Google Scholar]
- 14.Suo P, Wang S, Xue Y, Cheng Y, Kong J, Yan C, et al. Unmanipulated haploidentical hematopoietic stem cell transplantation for children with myelodysplastic syndrome. Pediatr Transpl. 2020;24(7):e13864. [DOI] [PubMed] [Google Scholar]
- 15.Yoo JW, Im HJ, Kim H, Koh KN, Kang SH, Min SY, et al. Improved outcomes of allogeneic hematopoietic stem cell transplantation including haploidentical transplantation for childhood myelodysplastic syndrome. Bone Marrow Transpl. 2020;55(8):1595–603. [DOI] [PubMed] [Google Scholar]
- 16.Hasle H, Kerndrup G, Jacobsen BB. Childhood myelodysplastic syndrome in Denmark: incidence and predisposing conditions. Leukemia. 1995;9(9):1569–72. [PubMed] [Google Scholar]
- 17.Hasle H, Wadsworth LD, Massing BG, McBride M, Schultz KR. A population-based study of childhood myelodysplastic syndrome in British Columbia, Canada. Br J Haematol. 1999;106(4):1027–32. [DOI] [PubMed] [Google Scholar]
- 18.Passmore SJ, Chessells JM, Kempski H, Hann IM, Brownbill PA, Stiller CA. Paediatric myelodysplastic syndromes and juvenile myelomonocytic leukaemia in the UK: a population-based study of incidence and survival. Br J Haematol. 2003;121(5):758–67. [DOI] [PubMed] [Google Scholar]
- 19.Aguilera DG, Vaklavas C, Tsimberidou AM, Wen S, Medeiros LJ, Corey SJ. Pediatric therapy-related myelodysplastic syndrome/acute myeloid leukemia: the MD Anderson Cancer Center experience. J Pediatr Hematol Oncol. 2009;31(11):803–11. [DOI] [PubMed] [Google Scholar]
- 20.Cho HW, Choi YB, Yi ES, Lee JW, Sung KW, Koo HH, et al. Therapy-related myeloid neoplasms in children and adolescents. Blood Res. 2016;51(4):242–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Song SO, Jung CH, Song YD, Park CY, Kwon HS, Cha BS, et al. Background and data configuration process of a nationwide population-based study using the korean national health insurance system. Diabetes Metab J. 2014;38(5):395–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Korea Statistical Information Servive (Statistics Korea). Projected Population by Age(Korea): Korea Statistical Information Servive. https://kosis.kr/index/index.do.
- 23.Hasle H, Niemeyer CM. Advances in the prognostication and management of advanced MDS in children. Br J Haematol. 2011;154(2):185–95. [DOI] [PubMed] [Google Scholar]
- 24.Sahoo SS, Pastor VB, Goodings C, Voss RK, Kozyra EJ, Szvetnik A, et al. Clinical evolution, genetic landscape and trajectories of clonal hematopoiesis in SAMD9/SAMD9L syndromes. Nat Med. 2021;27(10):1806–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Starý J, Baumann I, Creutzig U, Harbott J, Michalova K, Niemeyer C. Getting the numbers straight in pediatric MDS: distribution of subtypes after exclusion of down syndrome. Pediatr Blood Cancer. 2008;50(2):435–6. [DOI] [PubMed] [Google Scholar]
- 26.Le Deley MC, Leblanc T, Shamsaldin A, Raquin MA, Lacour B, Sommelet D, et al. Risk of secondary leukemia after a solid tumor in childhood according to the dose of epipodophyllotoxins and anthracyclines: a case-control study by the Société Française d’Oncologie Pédiatrique. J Clin Oncol. 2003;21(6):1074–81. [DOI] [PubMed] [Google Scholar]
- 27.Bhatia S, Krailo MD, Chen Z, Burden L, Askin FB, Dickman PS, et al. Therapy-related myelodysplasia and acute myeloid leukemia after Ewing sarcoma and primitive neuroectodermal tumor of bone: A report from the Children’s Oncology Group. Blood. 2007;109(1):46–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Winters AC, Maloney KW, Treece AL, Gore L, Franklin AK. Single-center pediatric experience with venetoclax and azacitidine as treatment for myelodysplastic syndrome and acute myeloid leukemia. Pediatr Blood Cancer. 2020;67(10):e28398. [DOI] [PubMed] [Google Scholar]
- 29.Strahm B, Locatelli F, Bader P, Ehlert K, Kremens B, Zintl F, et al. Reduced intensity conditioning in unrelated donor transplantation for refractory cytopenia in childhood. Bone Marrow Transpl. 2007;40(4):329–33. [DOI] [PubMed] [Google Scholar]
- 30.Woodard P, Carpenter PA, Davies SM, Gross TG, He W, Zhang MJ, et al. Unrelated donor bone marrow transplantation for myelodysplastic syndrome in children. Biol Blood Marrow Transpl. 2011;17(5):723–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Strahm B, Nöllke P, Zecca M, Korthof ET, Bierings M, Furlan I, et al. Hematopoietic stem cell transplantation for advanced myelodysplastic syndrome in children: results of the EWOG-MDS 98 study. Leukemia. 2011;25(3):455–62. [DOI] [PubMed] [Google Scholar]
- 32.Kikuchi A, Hasegawa D, Ohtsuka Y, Hamamoto K, Kojima S, Okamura J, et al. Outcome of children with refractory anaemia with excess of blast (RAEB) and RAEB in transformation (RAEB-T) in the Japanese MDS99 study. Br J Haematol. 2012;158(5):657–61. [DOI] [PubMed] [Google Scholar]
- 33.Madureira AB, Eapen M, Locatelli F, Teira P, Zhang MJ, Davies SM, et al. Analysis of risk factors influencing outcome in children with myelodysplastic syndrome after unrelated cord blood transplantation. Leukemia. 2011;25(3):449–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bortnick R, Wlodarski M, de Haas V, De Moerloose B, Dworzak M, Hasle H, et al. Hematopoietic stem cell transplantation in children and adolescents with GATA2-related myelodysplastic syndrome. Bone Marrow Transpl. 2021;56(11):2732–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Inagaki J, Fukano R, Kurauchi K, Noguchi M, Tanioka S, Okamura J. Hematopoietic stem cell transplantation in children with refractory cytopenia of childhood: single-center experience using high-dose cytarabine containing myeloablative and aplastic anemia oriented reduced-intensity conditioning regimens. Biol Blood Marrow Transpl. 2015;21(3):565–9. [DOI] [PubMed] [Google Scholar]
- 36.Wlodarski MW, Sahoo SS, Niemeyer CM. Monosomy 7 in Pediatric Myelodysplastic Syndromes. Hematol Oncol Clin North Am. 2018;32(4):729–43. [DOI] [PubMed] [Google Scholar]
- 37.Horgan C, Mullanfiroze K, Rauthan A, Patrick K, Butt NA, Mirci-Danicar O, et al. T-cell replete cord transplants give superior outcomes in high-risk and relapsed/refractory pediatric myeloid malignancy. Blood Adv. 2023;7(10):2155–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Kobos R, Steinherz PG, Kernan NA, Prockop SE, Scaradavou A, Small TN, et al. Allogeneic hematopoietic stem cell transplantation for pediatric patients with treatment-related myelodysplastic syndrome or acute myelogenous leukemia. Biol Blood Marrow Transpl. 2012;18(3):473–80. [DOI] [PubMed] [Google Scholar]
- 39.Waespe N, Van Den Akker M, Klaassen RJ, Lieberman L, Irwin MS, Ali SS, et al. Response to treatment with azacitidine in children with advanced myelodysplastic syndrome prior to hematopoietic stem cell transplantation. Haematologica. 2016;101(12):1508–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All the collected medical information was stored in National Health Insurance Service (NHIS) database. The data that support the findings of this study are available from the corresponding author upon reasonable request.