Abstract
NUP98 rearrangements (NUP98r) are often cryptic and define a high-risk subgroup of pediatric myeloid neoplasms, predominantly acute myeloid leukemia (AML). Despite their clinical significance, the immunophenotypic signatures that enable early recognition of NUP98r remain incompletely characterized.
We retrospectively studied 62 pediatric and young adult patients (age ≤21 years at diagnosis) with NUP98-rearranged myeloid neoplasms diagnosed from 1994 to 2025. We integrated flow cytometry, morphology, cytogenetics, and molecular data and grouped cases by fusion partner—NUP98::NSD1 (n=35), NUP98::KDM5A (n=12), and other NUP98 fusions (NUP98::X; n=15)—and by flow-defined differentiation patterns.
NUP98::NSD1 accounted for 56% of cases and was found in older children, with a median age at diagnosis of 13.2 years. These cases displayed an immature myeloid phenotype, with frequent expression of CD34 (33/35, 94%), CD117 (31/35, 89%), HLA-DR (34/34, 100%), and uniform CD123 positivity in all evaluable cases (22/22, 100%), along with FLT3-ITD and WT1 alterations (22/34 each, 65%) and mostly diploid karyotypes (61%).
NUP98::KDM5A occurred in younger children (median age at diagnosis, 1.9 years) and was associated with erythroid/megakaryocytic differentiation, including CD41/CD61 positivity in 8/10 (80%) and Glycophorin A positivity in 3/9 (33%) evaluable cases. Chromosome 13 abnormalities and RB1 alterations were identified in 60% and 55% of evaluable cases, respectively, and complex karyotypes were present in 63% of cases.
Other NUP98 fusions (NUP98::X; n=15, 24%) showed diverse phenotypes and were enriched for 11p abnormalities (79% of evaluable NUP98::X cases). Most 11p abnormalities were evident on conventional cytogenetics when karyotype data were available, in contrast to NUP98::NSD1 and NUP98::KDM5A cases, which were often cytogenetically cryptic.
In summary, NUP98 fusion partner was associated with recurring, diagnostically useful immunophenotypic, cytogenetic, and molecular patterns. These patterns can facilitate prioritization of RNA-based fusion testing, anticipate partner-specific differentials, and design flow cytometry follow-up strategies.
Keywords: NUP98, pediatric leukemia, immunophenotyping, sequencing
Introduction
Acute myeloid leukemia (AML) accounts for approximately 15–20% of pediatric leukemias and remains a leading cause of cancer-related deaths in children. 1, 2 Despite advances in risk-adapted therapy, relapses are still the main cause of AML-related mortality, and increasing the intensity of traditional chemotherapy is limited by toxicity. Therefore, current pediatric AML protocols rely on combined cytogenetic and molecular risk stratification and flow cytometry–based assessment of measurable residual disease (MRD) to guide treatment intensity and decisions regarding hematopoietic stem cell transplantation (HSCT). 3-6 Large genomic studies have further classified pediatric AML into biologically distinct molecular groups, many of which are characterized by recurrent gene fusions and specific clinical features and outcomes. 7, 8 Among these fusion-defined subgroups, NUP98 rearrangements (NUP98r) are notably associated with a poor prognosis. NUP98, located at 11p15, encodes a nucleoporin within the nuclear pore complex, and NUP98 fusions are observed across various hematologic cancers, including AML, myelodysplastic syndromes, T-lineage acute lymphoblastic leukemia, and mixed-phenotype acute leukemia.7, 9-11 In pediatric AML, NUP98 fusions are present in 5–7% of cases and are consistently linked to high relapse rates and poor survival. 7, 8 Recent comprehensive research further suggests that NUP98 fusion oncoproteins have a unique spectrum of cooperating mutations that collectively contribute to the observed disease phenotypes. 12
The two most common pediatric NUP98 fusion partners are NUP98::NSD1 and NUP98::KDM5A. NUP98::NSD1 is often cryptic by conventional cytogenetics and is frequently accompanied by FLT3-ITD and WT1 alterations; together, these features are associated with aggressive disease biology and poor outcomes. 7, 8, 12-15 In contrast, NUP98::KDM5A occurs mainly in infants and is enriched in acute megakaryoblastic and acute erythroid leukemias, where it is linked to distinctive lineage programs and a worse prognosis 16-19. Current classification frameworks recognize NUP98 rearrangements as genetically important lesions in AML; WHO-HAEM5 recognizes AML with NUP98 rearrangement, whereas ICC places NUP98 among the “AML with other rare recurring translocations” category. 20-22
A practical obstacle is that clinically important NUP98 fusions—particularly NUP98::NSD1 and NUP98::KDM5A—may be missed on routine karyotyping, underscoring the need for timely RNA-based fusion testing, whether through up-front transcriptome-based testing or institution-specific reflex targeted assays at diagnosis. 7, 23, 24 Multiparameter flow cytometry is routinely performed in pediatric acute leukemia for lineage assignment, maturation assessment, and the establishment of leukemia-associated immunophenotypes for MRD monitoring. However, reproducible immunophenotypic signatures that can prompt earlier molecular testing for cryptic NUP98 rearrangements remain incompletely defined. 7, 12, 17 Recent genomic studies have characterized partner-associated molecular correlates, but analyses combining morphology, flow cytometry, cytogenetics, and relapse-associated immunophenotypic evolution remain limited. 7, 12 In this study, we correlate diagnostic flow cytometry with cytogenetic findings and selected recurrent molecular features to (i) define recurrent antigen-expression patterns and maturation profiles among NUP98r leukemias; (ii) determine whether these signatures associate with NUP98 fusion-partner class (NUP98::NSD1, NUP98::KDM5A, and other partners (NUP98::X)); and (iii) provide practical diagnostic clues to prompt timely reflex fusion testing and inform MRD assay design in this high-risk population.
Materials and Methods
Study Cohort
We retrospectively identified 62 pediatric and young adult patients (defined as age ≤21 years at diagnosis; all presented after the neonatal period) with NUP98 rearrangements through collaboration between St. Jude Children’s Research Hospital and The University of Texas MD Anderson Cancer Center from 1994 to 2025, with available material for morphologic review and flow-cytometric immunophenotyping. Overall, 59 cases were de novo, and 3 were therapy-related myeloid neoplasms. Final diagnoses were assigned by integrated hematopathology review using contemporary WHO/ICC terminology. In keeping with WHO-HAEM5, NUP98-rearranged myeloid neoplasms were considered within the genetically defined AML spectrum even when blast percentages were below 20%; ICC 2022 places NUP98 among AML with other rare recurring translocations, generally requiring ≥10% blasts. For case-level continuity, the therapy-related case with <20% blasts was retained in tables under its historical RAEB-T/MDS-AML designation, and one additional case was retained as MDS/MPN after integrated review. Two cases with rare partners (NUP98::RAP1GDS1 and NUP98::BPTF) showed ETP-like/T-lineage-skewed antigen expression. They are included in this study to illustrate the diverse immunophenotypic manifestations associated with NUP98 rearrangements. Specimens were retrospectively reviewed by hematopathologists (MK, JMK, BT, WW). Clinical information and follow-up data were obtained from electronic medical records, including age at diagnosis, sex, treatment history, immunophenotype, cytogenetics, mutation profile, clinical response, and survival status (Tables 1 and 2 and Supplementary Table S1), together with data from a recently published genomic landscape study. 12 This study was approved by the Institutional Review Boards of St. Jude Children’s Research Hospital and The University of Texas MD Anderson Cancer Center.
Table 1.
Clinical and biologic characteristics by NUP98 fusion partner.
| Demographic data | Entire cohort (n=62) | NUP98::NSD1-r (n=35) | NUP98::KDM5A-r (n=12) | NUP98::X-r (n=15) | p Value |
|---|---|---|---|---|---|
| Age at diagnosis, years, median (range) | 10.0 (0.6-20.2) | 13.2 (2.6-20.2) | 1.9 (1.2-10.1) | 4.2 (0.6-16.3) | <0.0001 |
| Sex (female/male) | 23/39 | 9/26 | 7/5 | 7/8 | 0.080 |
| Peripheral blood, median (range) | |||||
| WBC (×109/L) | 15.9 (0.6-460.0) | 102.3 (0.6-460.0) | 10.5 (1.9-81.7) | 14.5 (2.5-416.0) | 0.048 |
| Hgb (g/dL) | 8.7 (4.4-13.0) | 8.7 (4.4-12.3) | 8.2 (5.5-10.1) | 8.2 (6.8-13.0) | 0.86 |
| Platelets (×109/L) | 67 (4-398) | 60 (5-398) | 77 (4-176) | 69 (20-216) | 0.87 |
| Blasts (%) | 32 (0-99) | 47 (0-99) | 10 (1-65) | 39 (4-90) | 0.049 |
| Absolute monocyte count (/mm3) | 300 (0-49464) | 300 (0-49464) | 812 (0-22059) | 185 (0-840) | 0.387 |
| Bone marrow | |||||
| Blasts %, median (range) | 61 (12-99) | 72 (12-95) | 55 (21-70) | 71 (15-99) | 0.519 |
| CNS involvement, n/N (%) | 18/46 (39) | 16/29 (55) | 0/8 (0) | 2/9 (22) | 0.006 |
| Cytogenetics & ploidy | |||||
| Diploid, n/N (%) | 21/50 (42) | 19/31 (61) | 1/8 (13) | 1/11 (9) | 0.008 |
| Balanced translocation, n/N (%) | 18/50 (36) | 10/31 (32) | 2/8 (25) | 6/11 (55) | |
| Complex, n/N (%) | 11/50 (22) | 2/31 (6) | 5/8 (63) | 4/11 (36) | |
| Abnormality 13, n/N (%) | 7/57 (12) | 0/33 (0) | 6/10 (60) | 1/14 (7) | <0.001 |
| Trisomy 8, n/N (%) | 5/56 (9) | 4/33 (12) | 0/10 (0) | 1/13 (8) | 0.814 |
| 11p15, n/N (%) | 11/57 (19) | 0/33 (0) | 0/10 (0) | 11/14 (79) | <0.001 |
| Trisomy 21, n/N (%) | 7/56 (13) | 0/33 (0) | 4/10 (40) | 3/13 (23) | 0.001 |
| Abnormality 5q, n/N (%) | 9*/55 (16) | 4/33 (12) | 2/10 (20) | 3/12 (25) | 0.517 |
| Molecular data, mutations | |||||
| FLT3-ITD, n/N (%) | 23/60 (38) | 22/34 (65) | 0/12 (0) | 1/14 (7) | <0.001 |
| WT1, n/N (%) | 24/59 (41) | 22/34 (65) | 0/12 (0) | 2/13 (15) | <0.001 |
| FLT3-ITD & WT1, n/N (%) | 15/59 (25) | 15/34 (44) | 0/12 (0) | 0/13 (0) | 0.002 |
| RB1 alteration, n/N (%) | 6/56 (11) | 0/33 (0) | 6/11 (55) | 0/12 (0) | <0.001 |
| Therapy (n=50) | |||||
| Chemotherapy, n/N (%) | 49/50 (98) | 31/32 (97) | 8/8 (100) | 10/10 (100) | 0.750 |
| Stem cell transplant, n/N (%) | 34/50 (68) | 24/32 (75) | 5/8 (63) | 5/10 (50) | 0.313 |
| Clinical follow-up (Months) (n=48) | |||||
| Duration, median (range) | 26.9 (0-263.4) | 26.9 (0-263.4) | 11.5 (4.0-31.5) | 48.2 (1.1-145.4) | NA |
| Patient outcome (n=52) | |||||
| Complete remission rate, n/N (%) | 21/52 (40) | 13/32 (41) | 1/8 (13) | 7/12 (58) | 0.123 |
| Death, n/N (%) | 31/52 (60) | 19/32 (59) | 7/8 (88) | 5/12 (42) | 0.123 |
Data are shown as median (range) or n/N (%), with N representing the number of evaluable cases for each variable. Abbreviations: WBC, white blood cell count; Hgb, hemoglobin; CNS, central nervous system; FLT3-ITD, FLT3 internal tandem duplication. Cytogenetic denominators vary by assay. Broad karyotype categories are based on conventional cytogenetic studies at diagnosis and include only cases with sufficient metaphase data for classification. Abnormalities involving 11p and other chromosomal regions are based on conventional cytogenetics supplemented by available genomic copy-number data (WGS, OGM, or chromosomal microarray); denominators reflect cases evaluable by at least one applicable method. *5q abnormalities included 6 present at diagnosis and 3 acquired at relapse/persistent disease. Peripheral blood blast percentages were based on available morphology/differential-count data. RB1 alteration includes sequence-level alteration and/or copy-number loss.
Table 2.
Clinical and biologic characteristics by flow-defined immunophenotype category.
| Demographic data | Entire cohort (n=62) | Myeloid (n=19) | Monocytic (n=9) | Myelomonocytic (n=23) |
Erythroid/megakary ocytic (n=11) |
p Value |
|---|---|---|---|---|---|---|
| Age at diagnosis, years, median (range) | 10.0 (0.6-20.2) | 13.0 (2.3-17.0) | 14.0 (1.8-16.4) | 11.7 (1.8-20.2) | 1.8 (0.6-3.8) | <0.0001 |
| Sex (female/male) | 23/39 | 9/10 | 1/8 | 7/16 | 6/5 | 0.148 |
| Peripheral blood, median (range) | ||||||
| WBC (×109/L) | 15.9 (0.6-460.0) | 12.3 (1.0-460.0) | 102.3 (0.6-222.0) | 18.0 (0.6-412.2) | 8.4 (1.9-15.1) | 0.237 |
| Hgb (g/dL) | 8.7 (4.4-13.0) | 7.8 (4.4-13.0) | 8.7 (7.6-12.3) | 8.7 (4.8-10.1) | 9.0 (7.6-10.1) | 0.577 |
| Platelets (×109/L) | 67 (4-398) | 60 (4-109) | 85 (30-216) | 67 (5-398) | 77 (14-176) | 0.463 |
| Blasts (%) | 32 (0-99) | 43 (2-99) | 60 (0-97) | 34 (2-88) | 8 (1-65) | 0.235 |
| Absolute monocyte count (/mm3) | 300 (0-49464) | 146 (0-16741) | 2133 (0-22059) | 458 (0-49464) | 1008 (157-3820) | 0.276 |
| Bone marrow | ||||||
| Blasts %, median (range) | 61 (12-99) | 57 (12-93) | 56 (15-82) | 75 (20-99) | 55 (21-70) | 0.369 |
| CNS involvement, n/N (%) | 18/46 (39) | 3/15 (20) | 5/8 (63) | 10/17 (59) | 0/6 (0) | 0.012 |
| Cytogenetics & ploidy | ||||||
| Diploid, n/N (%) | 21/50 (42) | 6/15 (40) | 3/8 (38) | 12/21 (57) | 0/6 (0) | 0.119 |
| Balanced translocation, n/N (%) | 18/50 (36) | 6/15 (40) | 4/8 (50) | 6/21 (29) | 2/6 (33) | |
| Complex, n/N (%) | 11/50 (22) | 3/15 (20) | 1/8 (13) | 3/21 (14) | 4/6 (67) | |
| Abnormality 13, n/N (%) | 7/57 (12) | 2/19 (11) | 0/8 (0) | 0/22 (0) | 5/8 (63) | <0.001 |
| Trisomy 8, n/N (%) | 5/56 (9) | 2/18 (11) | 2/8 (25) | 1/22 (5) | 0/8 (0) | 0.302 |
| 11p15, n/N (%) | 11/57 (19) | 4/19 (21) | 2/8 (25) | 5/22 (23) | 0/8 (0) | 0.627 |
| Trisomy 21, n/N (%) | 7/56 (13) | 1/18 (5) | 0/9 (0) | 2/21 (10) | 4/8 (50) | 0.018 |
| Abnormality 5q, n/N (%) | 9*/55 (16) | 4/16 (25) | 2/8 (25) | 2/23 (9) | 1/8 (13) | 0.449 |
| Molecular data, mutations | ||||||
| FLT3-ITD, n/N (%) | 23/60 (38) | 3/17 (18) | 5/9 (56) | 15/23 (65) | 0/11 (0) | 0.002 |
| WT1, n/N (%) | 24/59 (41) | 8/17 (47) | 4/9 (44) | 12/22 (55) | 0/11 (0) | 0.011 |
| FLT3-ITD & WT1, n/N (%) | 15/59 (25) | 2/17 (12) | 4/9 (44) | 9/22 (41) | 0/11 (0) | 0.014 |
| RB1 alteration, n/N (%) | 6/56 (11) | 0/17 (0) | 0/9 (0) | 0/21 (0) | 6/9 (67) | <0.001 |
| Therapy (n=50) | ||||||
| Chemotherapy, n/N (%) | 49/50 (98) | 15/16 (94) | 9/9 (100) | 19/19 (100) | 6/6 (100) | 0.620 |
| Stem cell transplant, n/N (%) | 34/50 (68) | 10/16 (63) | 9/9 (100) | 11/19 (58) | 4/6 (67) | 0.117 |
| Clinical follow-up (Months) (n=48) | ||||||
| Duration, median (range) | 26.9 (0.0-263.4) | 27.7 (0.0-134.7) | 48.4 (7.2-263.4) | 24.1 (13.4-232.6) | 16.4 (4.0-31.5) | 0.033 |
| Patient outcome (n=52) | ||||||
| Complete remission rate, n/N (%) | 21/52 (40) | 9/18 (50) | 6/9 (67) | 6/19 (32) | 0/6 (0) | 0.041 |
| Death, n/N (%) | 31/52 (60) | 10/18 (56) | 3/9 (33) | 12/19 (63) | 6/6 (100) | 0.070 |
Data are shown as median (range) or n/N (%). Denominators vary by assay availability and are indicated where applicable. Abbreviations: WBC, white blood cell count; Hgb, hemoglobin; FLT3-ITD, FLT3 internal tandem duplication. Cytogenetic denominators vary by assay.
Broad karyotype categories are based on conventional cytogenetic studies at diagnosis and include only cases with sufficient metaphase data for classification. Abnormalities involving 11p and other chromosomal regions are based on conventional cytogenetics supplemented by available genomic copy-number data (WGS, OGM, or chromosomal microarray); denominators reflect cases evaluable by at least one applicable method.
5q abnormalities included 6 present at diagnosis and 3 acquired at relapse/persistent disease. Peripheral blood blast percentages were based on available morphology/differential-count data. RB1 alteration includes sequence-level alteration and/or copy-number loss.
Morphologic Assessment
Wright-Giemsa-stained peripheral blood (PB) and bone marrow (BM) aspirate smears and hematoxylin-eosin-stained core biopsy sections were reviewed. A total of 200 leukocytes in PB and 500 cells in BM were examined for differential counts, including blast and erythroid percentages. Peripheral blood blast percentages reported in the tables were based on PB morphology/differential-count data, with correlation to peripheral blood flow cytometry when available. Myeloid blasts were enumerated using the conventional ≥20% threshold as a descriptive measure of blast burden rather than as the sole determinant of final disease classification. Final classification was assigned based on a review of morphologic, cytochemical, immunophenotypic, cytogenetic, and molecular findings according to contemporary WHO-HAEM5 and ICC 2022 criteria. Cases were also subclassified using French-American-British (FAB) criteria as a descriptive shorthand for morphologic differentiation and comparability with prior pediatric AML literature; WHO/ICC categories remained the primary diagnostic classification.
Immunophenotyping Studies
Flow cytometry immunophenotyping was performed on BM aspirates and/or PB using standard multicolor analysis. Antibody panels varied across institutions and time periods, including: CD2, CD3, cytoplasmic CD3, CD4, CD5, CD7, CD8, CD10, CD11b, CD11c, CD13, CD14, CD15, CD19, CD20, CD22, CD25, CD33, CD34, CD36, CD38, CD41, CD42/CD42b, CD45, CD56, CD58, CD61, CD64, CD71, CD79a/cytoplasmic CD79a, CD117, CD123, HLA-DR, CD235a (Glycophorin A), and myeloperoxidase (MPO) (BD Biosciences, San Jose, CA, USA). Data were analyzed using FCS Express (De Novo Software, Pasadena, CA, USA). Cases were categorized into four differentiation groups: myeloid, monocytic/monoblastic, myelomonocytic, and erythroid/megakaryocytic. The myeloid category was defined by blast morphology, MPO positivity, and absence of definitive monocytic differentiation; in MPO-negative cases, the coordinated expression of CD13, CD33, and CD117, together with an overall myeloid blast phenotype, supported myeloid assignment. Monocytic/monoblastic cases were defined by morphology, nonspecific esterase positivity, and expression of CD64, CD14, CD11b, and CD36. Myelomonocytic cases showed concurrent granulocytic/myeloid and monocytic differentiation. Erythroid/megakaryocytic cases were defined by morphology and expression of CD36, CD71, Glycophorin A, CD41, CD42, and/or CD61. Rare cases with ETP-like/T-lineage-skewed antigen expression were retained within the myeloid category when the combined morphologic, cytochemical, and immunophenotypic findings supported myeloid differentiation. Antigen intensity was graded relative to the relevant negative/internal control as follows: dim, fluorescence slightly above and partially overlapping the negative control; moderate, fluorescence clearly greater than the negative control; and bright, fluorescence greater than 1.5 log above the negative control. Partial expression was defined as staining only a subset of the abnormal blast population. For tabulated frequency analyses, complete or partial expression on the abnormal blast population was counted as positive. CD123, CD36, and CD71 were introduced at different times across institutional/historical panels; frequencies of expression are therefore reported only among evaluable cases.
Cytogenetics
Conventional karyotypic analysis was performed on G-banded metaphase cells from unstimulated BM or PB cultures using standard techniques. At least twenty metaphases were analyzed (where possible), and results were interpreted according to ISCN conventions current at the time of testing; historical karyotypes were harmonized on review for consistency in this study. A complex karyotype was defined as ≥3 chromosomal abnormalities (Supplementary Table S2).
Molecular
Fusion status was determined through RNA sequencing (RNA-seq; n=57), optical genome mapping (OGM; DNA-based; n=4), or chromosomal microarray when RNA-seq/OGM was unavailable (n=1) (Supplementary Table S3). DNA-based genomic profiling was conducted in 55 cases and included whole-genome sequencing (WGS; n=43), whole-exome sequencing (WES; n=26, including 25 cases with both WGS and WES), and targeted next-generation sequencing (NGS) panels (n=11). Chromosomal microarray was performed in 1 case. These genomic datasets were generated either through prior research sequencing efforts reported previously 12,23 or as part of routine clinical testing (Supplementary Table S3). Recurrent chromosome-level and focal copy-number abnormalities were assessed using conventional cytogenetics together with available genomic copy-number data from WGS, OGM, or chromosomal microarray. The broader transcriptomic and gene-expression context of NUP98-rearranged leukemia, including fusion-partner– and cooperating-mutation–associated disease phenotypes, was recently reported by Umeda et al. in a larger NUP98-rearranged leukemias cohort. 12 In this pathology-focused study, RNA-seq was used primarily for fusion reporting.
Statistical Analysis
Clinical information and follow-up data were obtained, including treatment history, response, follow-up duration, and survival status. Continuous variables were summarized as median (range) and group-level p-values for continuous variables were calculated in GraphPad Prism using one-way analysis of variance (ANOVA) for comparisons across fusion-partner and immunophenotype groups. Categorical variables were compared using Fisher’s exact test. All tests were two-sided. Given the exploratory nature of this retrospective study and the number of comparisons performed, p-values were not adjusted for multiple comparisons and should be interpreted descriptively. Statistical analyses were performed using GraphPad Prism (GraphPad Software). Swimmer plots were generated in Microsoft Excel (version 2510).
Results
Cohort Composition
The cohort comprised 62 patients, with a median age at diagnosis of 10.0 years (range 0.6-20.2) and a female-to-male ratio of 1:1.7. Fifty-nine (95%) had de novo myeloid neoplasms, and the remaining three patients (5%) had therapy-related myeloid neoplasms. On integrated review, 60 cases met conventional morphologic/immunophenotypic criteria for AML, one therapy-related case with <20% blasts was retained in case-level tables using its historical RAEB-T/MDS-AML designation while discussed as contemporary NUP98-rearranged AML-spectrum disease, and one case was retained as MDS/MPN.
The Spectrum of NUP98 Fusion Partners
The most common NUP98 translocations were NUP98::NSD1 (n=35, 56%) and NUP98::KDM5A (n=12, 19%). For the remaining 15 cases (NUP98::X), we identified nine different partners, including RAP1GDS1 (n=4), DDX10 (n=3), JADE2 (n=2), and single cases of HHEX, TNRC18, ETS1, ZFX, BPTF, and TOP1.
Blast Immunophenotype and Comparison Among Fusion Subgroups
Comprehensive flow-cytometric immunophenotyping classified all 62 baseline cases into four study-defined categories: myeloid (n=19), monocytic (n=9), myelomonocytic (n=23), and erythroid/megakaryocytic (n=11) (Table 2). These categories were assigned using the combined morphologic, cytochemical, and antigen-expression criteria described in the Methods and were the only formal immunophenotypic categories used for cohort-level analyses. Two rare-partner cases with ETP-like/T-lineage-skewed antigen expression were included in the myeloid category after multidisciplinary review.
Across the cohort, blasts frequently expressed myeloid lineage markers, including CD33 (59/62, 95%), CD13 (49/61, 80%), and CD15 (39/56, 70%), with MPO positivity in 36/61 (59%). Progenitor-associated antigens were also frequent, including CD34 (44/62, 71%), CD38 (36/39, 92%), CD117 (53/61, 87%), HLA-DR (53/61, 87%), and CD123 (31/38, 82%). Aberrant lymphoid antigen expression was common, notably CD7 (25/61, 41%) and CD19 (15/61, 25%). CD19 expression was enriched in NUP98::NSD1 and in myeloid/myelomonocytic phenotypes (Table 3) and showed gain/loss in some paired relapse samples (Table 4). CD56 was detected in 8/58 (14%). Megakaryocytic markers (CD41/CD61) were detected in 10/52 (19%) and Glycophorin A in 3/42 (7%) (Table 3 and Figure 1). Among CD19-positive cases with available extended B-lineage marker data, diagnostic samples most often lacked additional B-cell markers: CD10 was positive only in case 56. CD22 and cytoplasmic CD79a co-expression was limited to dim expression in case 32. At relapse, CD10 remained negative in all evaluable CD19-positive samples, whereas CD22 and/or cytoplasmic CD79a were variably detected, usually dim or partial (Table 4).
Table 3.
Flow cytometric marker expression by fusion partner and immunophenotype category.
| Markers | Entire cohort | NUP98::NSD1-r (n=35) | NUP98::KDM5A-r (n=12) | NUP98::X-r (n=15) | Myeloid (n=19) | Monocytic (n=9) | Myelomonocytic (n=23) | Erythroid/megakary ocytic (n=11) |
|---|---|---|---|---|---|---|---|---|
| CD64 (%) | 30/56 (54) | 22/34 (65) | 1/10 (10) | 7/12 (58) | 0/17 (0) | 9/9 (100) | 21/22 (95) | 0/8 (0) |
| CD11b (%) | 45/58 (78) | 31/34 (91) | 5/10 (50) | 9/14 (64) | 12/19 (63) | 8/8 (100) | 20/21 (95) | 5/10 (50) |
| CD14 (%) | 18/56 (32) | 11/33 (33) | 3/11 (27) | 4/12 (33) | 0/17 (0) | 6/9 (67) | 10/21 (48) | 2/9 (22) |
| CD13 (%) | 49/61 (80) | 32/34 (94) | 5/12 (42) | 12/15 (80) | 17/19 (89) | 6/9 (67) | 22/22 (100) | 4/11 (36) |
| CD15 (%) | 39/56 (70) | 30/32 (94) | 1/11 (9) | 8/13 (62) | 9/16 (56) | 9/9 (100) | 21/21 (100) | 0/10 (0) |
| CD33 (%) | 59/62 (95) | 35/35 (100) | 9/12 (75) | 15/15 (100) | 19/19 (100) | 9/9 (100) | 23/23 (100) | 8/11 (73) |
| CD34 (%) | 44/62 (71) | 33/35 (94) | 1/12 (8) | 10/15 (67) | 17/19 (89) | 6/9 (67) | 20/23 (87) | 1/11 (9) |
| CD117 (%) | 53/61 (87) | 31/35 (89) | 10/12 (83) | 12/14 (86) | 17/18 (94) | 4/9 (44) | 22/23 (96) | 10/11 (91) |
| CD38 (%) | 36/39 (92) | 20/20 (100) | 6/9 (67) | 10/10 (100) | 15/15 (100) | 3/3 (100) | 12/12 (100) | 6/9 (67) |
| CD7 (%) | 25/61 (41) | 17/34 (50) | 1/12 (8) | 7/15 (47) | 11/19 (58) | 1/9 (11) | 12/22 (55) | 1/11 (9) |
| CD19 (%) | 15/61 (25) | 13/34 (38) | 0/12 (0) | 2/15 (13) | 7/19 (37) | 2/9 (22) | 6/22 (27) | 0/11 (0) |
| CD56 (%) | 8/58 (14) | 2/32 (6) | 2/11 (18) | 4/15 (27) | 2/19 (11) | 3/9 (33) | 2/20 (10) | 1/10 (10) |
| HLA-DR (%) | 53/61 (87) | 34/34 (100) | 6/12 (50) | 13/15 (87) | 17/18 (94) | 9/9 (100) | 23/23 (100) | 4/11 (36) |
| CD41/CD61 (%) | 10/52 (19) | 1/32 (3) | 8/10 (80) | 1/10 (10) | 1/16 (6) | 0/8 (0) | 0/18 (0) | 9/10 (90) |
| Glycophorin-A (%) | 3/42 (7) | 0/24 (0) | 3/9 (33) | 0/9 (0) | 0/15 (0) | 0/4 (0) | 0/14 (0) | 3/9 (33) |
| CD36 (%) | 26/35 (74) | 16/21 (76) | 7/8 (88) | 3/6 (50) | 4/7 (57) | 5/6 (83) | 10/15 (67) | 7/7 (100) |
| CD71 (%) | 32/41 (78) | 15/22 (68) | 8/9 (89) | 9/10 (90) | 7/10 (70) | 7/8 (88) | 11/15 (73) | 7/8 (88) |
| Myeloperoxidase (%) | 36/61 (59) | 28/35 (80) | 0/12 (0) | 8/14 (57) | 10/18 (56) | 6/9 (67) | 20/23 (87) | 0/11 (0) |
| CD123 (%) | 31/38 (82) | 22/22 (100) | 4/8 (50) | 5/8 (63) | 13/14 (93) | 3/3 (100) | 12/13 (92) | 3/8 (38) |
Data are n/N (%).
Denominators vary due to differences in flow cytometry panels and specimen availability.
Table 4.
Immunophenotype and cytogenetic evolution at relapse.
| Case | Initial IP | First-relapse IP | Other Relapsed IP | Initial Karyotype | Karyotype at relapse | IP category shift | Cytogenetic evolution | Time to first relapse (mo) |
|---|---|---|---|---|---|---|---|---|
| 1 | Myelomonocytic | Myelomonocytic | Myelomonocytic | 46,XX[20] | 46,XX[20] | No | No | 9.1 |
| 2 | Myelomonocytic | Myelomonocytic | Myelomonocytic | 47,XY,+8[14]/46,XY[6] | 47,XY,+8[14]/47,idem,t(3;5)(q21;q11.2)[5]/46,XY[1] | No | Yes | 8.1 |
| 3 | Myeloid (CD19 negative) | Monocytic (CD19 positive; CD10/CD22/CD79a negative) | Monocytic (CD19 positive; CD10/CD22/CD79a negative) | 45-46,XY,del(9)(q22q32)[20] | 46,XY,del(9)(q22q32)[20] | Yes | No | 5.7 |
| 8 | Not available | Myeloid (CD19, CD22, CD79a positive) | 46,XX[20] | 46,XX,del(5)(q13q34),t(8;17)(q13;q25),add(14)(q32),add(22)(q13)[19]/46,XX[1] | NA | Yes | 12.1 | |
| 9 | Not available | Myeloid | Myeloid | 46,XX[20] | 46,XX[20] | NA | No | 3.5 |
| 10 | Not available | Myeloid (CD19 positive, CD7 positive; CD10/CD22 negative, CD79a not tested) | Myeloid (CD19 negative, CD7 negative) | 46,XY,der(5)t(5;12)(q22;q22),add(12)(q22)[20] | 46,XY,add(3)(q27),del(5)(q22),del(12)(q13),add(16)(q24)[5]/45,idem,-Y[4]/46,XX[11] | NA | Yes | 0.9 |
| 11 | Not available | Myelomonocytic (CD19 positive, CD7 positive; CD10 not tested, CD22/CD79a negative) | Myelomonocytic (CD19 negative, CD7 negative) | Not available | 46,XY[20] | NA | NA | 6.4 |
| 12 | Myelomonocytic | Myelomonocytic | 46,XX,del(5)(q23q35)[20] | 46,XX,del(5)(q23q35)[20] | No | No | 10.3 | |
| 15 | Myelomonocytic | Myelomonocytic | 46,XY[20] | 46,XY,t(4;8)(q21;p23),inv(6)(p23q24),t(15;18)(q13;q23)[19]/44,XY,−3,−10[1] | No | Yes | 7.4 | |
| 16 | Myelomonocytic | Myelomonocytic | 46,XY[20] | 46,XY[20] | No | No | 8.4 | |
| 27 | Not available | Myeloid (CD19 positive, CD7 positive; CD10/CD22 negative, CD79a dim) | Myelomonocytic (CD19 positive, CD7 negative; CD10/CD22 negative, CD79a dim) | Not available | 46,XY,t(11;19)(q13;p13)[6]/46,XY,t(3;19)(p10;q10)[4]/46,XY[10] | NA | NA | 20.7 |
| 28 | Monocytic (CD19/CD10/CD22 negative; CD79a not tested) | Myelomonocytic (CD7/CD19 positive; CD10 negative, CD22 partial/moderate, CD79a moderate) | Immature monocytic (CD7 positive; CD19/CD10/CD22/CD79a negative) |
46,XY[20] | 46,XY[20] | Yes | No | 4.8 |
| 29 | Myeloid (CD19/CD10/CD22/CD79a negative) |
Myelomonocytic (CD19 positive; CD10/CD22 negative, CD79a bright) | 46,XY[30] | 46,XY,t(1;17)(p22;p13)[2]/46,XY[39]/46,XX[1] | Yes | Yes | 8.7 | |
| 30 | Myelomonocytic (CD19 negative; CD10/CD22/CD79a not tested) | Myeloid (CD19 positive; CD10/CD22/CD79a negative) |
46,XY[20] | Not available | Yes | NA | 6.2 | |
| 31 | Not available | Myelomonocytic (CD19 positive, CD7 positive; CD10/CD22 negative, CD79a dim) | Myeloid (CD19/CD22 positive, CD7 positive, CD56 positive; CD10 negative, CD22 dim, CD79a negative). Later relapse: CD19 small subset with CD10/CD22/CD79a negative |
46,XY[30] | 46,XY[30] | NA | No | 2.8 |
| 32 | Myelomonocytic (CD19 positive; CD10 negative, CD22/CD79a dim) | Myeloid (CD19 positive; CD10 negative, CD22 dim, CD79a positive) | 46,XY,del(9)(q13q22)[3]/46,XY[17] | 46,XY,add(9)(q12)[20] | Yes | Yes | 4.0 | |
| 34 | Megakaryocytic | Megakaryocytic | 46,XY,−2,add(7)(p15),add(7)(q36),add(12)(p13),add(13)(q14),+mar[12]/46,idem,del(3)(q21)[3]/48,idem,+6,+21[2]/46,XY[3] | 46,XY,t(2;12)(q23;p13),t(7;8)(p15;p23),t(7;13)(q36;q14)[10]/46,XY[10] | No | Yes | 53.6 | |
| 35 | Megakaryocytic | Megakaryocytic | 46,XY,del(13)(q13q22)[20] | 46,XY,del(13)(q13q22)[20] | No | No | 1.1 | |
| 36 | Erythroid | Erythroid | 45,XY,del(9)(p13),der(10;22)(q10;q10), del(13)(q12q14),add(21)(p11.2)[6]/ 46,XY[15] | 45,XY,del(2)(q13q23),inv(9)(p22q13),der(10;22)(q10;q10),del(13)(q12q14),add(21)(p11.2)[20] | No | Yes | 13.2 | |
| 43 | Megakaryocytic | Megakaryocytic | 46,XX,del(13)(q12q14)[13]/46,XX[6] | 46,XX,del(13)(q?12q14)[24]/46,idem,del(2)(q11.2q22)[2]/46,XX[4] | No | Yes | 3.6 | |
| 44 | Myeloid (megakaryocytic markers not tested) | Myeloid with 20% blasts positive for CD42b | 46,XX,del(13)(q12q14)[13]/46,idem,add(2)(q31),−12,+mar[3]/46,XX[4] | 46,XX,del(13)(q12q14)[7]/46,XX[13] | No | Yes | 6.3 | |
| 45 | Myelomonocytic (CD2/CD7/CD19 positive; CD10/CD22/CD79a negative) |
Poorly differentiated (loss of some myeloid markers) with subset CD19/CD22+ and subset CD7; CD10 negative, CD22 dim, CD79a negative | 46,XX,t(4;11)(q21;p15)[5]/46,XX[7] | 46,XX,t(4;11)(q21;p15)[5]/46,XX[7] | Yes | No | 58.3 | |
| 46 | Myeloid (ETP-like features) | Poorly differentiated/Myeloid with T-cell markers (loss of cyCD3) | 46,XX,t(4;11)(q21~24;p15.4),del(6)(q21q23)[19] | 46,XX[25] | Yes | Yes | 25.9 | |
| 53 | Myelomonocytic | Myelomonocytic | 46,XX[20] | 46,XX[20] | No | No | 22.5 | |
| 57 | Myelomonocytic (CD56+) | Myelomonocytic (CD56−) | 49,XY,+10,?inv(11)(p15q?23),+19,+21[2]/46,XY[18] | 49,XY,+10,?inv(11)(p15q?23),+19,+21[2]/46,XY[18] | No | No | 12.1 |
Abbreviations: IP, immunophenotype; ETP, early T-cell precursor; NA, not available. IP category shift indicates a change in the broad immunophenotype category between initial and relapse assessments. Cytogenetic evolution indicates a cytogenetic change at relapse compared with the initial karyotype, including acquisition or loss of abnormalities. Time to first relapse was calculated from the initial diagnostic/initial IP date to the first relapse IP date; when baseline IP was unavailable, the diagnostic date was used.
Figure 1.
Heatmap summarizing case-level clinicopathologic features of pediatric myeloid neoplasms with NUP98 rearrangements. Cases are grouped by fusion partner and annotated for FAB subtype, study-defined broad immunophenotype category (myeloid, monocytic, myelomonocytic, or erythroid/megakaryocytic), selected flow cytometric markers, conventional karyotype category, recurrent copy-number abnormalities, cooperating molecular lesions, and outcome. Rare descriptive modifiers, such as ETP-like or poorly differentiated phenotype, are not separate cohort-level categories. The NPM1 alteration in case 44 is an NPM1::CSF1R rearrangement. See Tables 1-4 and Supplementary Tables S1-S3 for case-level details.
NUP98::NSD1 (n=35) displayed a predominantly immature myeloid immunophenotype with frequent expression of CD34 (94%), CD117 (89%), HLA-DR (100%), and uniform bright CD123 expression (100% of 22 evaluable cases). Granulocytic markers were commonly expressed, including CD13 (94%), CD15 (94%), and MPO (80%). Monocytic-associated antigens were variably expressed, including CD64 (65%), CD11b (91%), and CD14 (33%). Aberrant lymphoid marker expression was frequent (CD7, 50%; CD19, 38%). Erythroid/megakaryocytic markers were largely absent (CD41/CD61, 3%; Glycophorin A, 0%).
NUP98::KDM5A (n=12) cases demonstrated an erythroid/megakaryocytic-biased immunophenotype with low CD34 (8%) and HLA-DR (50%) expression, whereas CD117 was positive in 83% of cases. Granulocytic differentiation was limited (CD13, 42%; CD15, 9%), and MPO was negative in all cases. Megakaryocytic markers CD41/CD61 were detected in 80%, and Glycophorin A was positive in 33% of evaluable cases. CD33 (75%) and CD123 (50% of evaluable cases) were variably expressed, and aberrant lymphoid antigen expression was uncommon (CD7, 8%). CD36 was positive in 7 of 8 (88%) NUP98::KDM5A cases, and CD71 was positive in all M6 cases with evaluable data (3/3, 100%). CD56 expression was infrequent (2/11, 18%) rather than uniformly very bright, and CD38 was retained in most evaluable cases (67%); these features support the distinction from classic RAM/CBFA2T3::GLIS2-associated acute megakaryoblastic leukemia (AMKL).
NUP98::X (n=15) exhibited the greatest immunophenotypic heterogeneity. Most cases demonstrated a myeloid/monocytic immunophenotype with frequent expression of CD13 (80%), CD33 (100%), CD34 (67%), CD117 (86%), and HLA-DR (87%), with variable MPO expression (57%). Monocytic markers were variably expressed (CD64, 58%; CD11b, 64%; CD14, 33%), and aberrant expression of lymphoid antigens was relatively common (CD7, 47%; CD56, 27%). This subgroup included two cases with ETP-like/T-lineage skewed features (NUP98::RAP1GDS1 and NUP98::BPTF), illustrating that rare NUP98 partners can occur in leukemias with marked lineage ambiguity and require integrated classification.
No case in the cohort showed classic hypergranular promyelocytic morphology or overt granulated promyelocyte-like blasts. Three cases - cases 49 (NUP98::DDX10), 51 (NUP98::JADE2), and 53 (NUP98::ZFX) - had immunophenotypic overlap with acute promyelocytic leukemia (e.g. CD34 and HLA-DR partial loss/negative), emphasizing the value of integrating morphology, flow cytometry, cytogenetics, and fusion testing before assigning an acute promyelocytic leukemia mimic.
Baseline Clinical Characteristics and Comparison Among Fusion and Immunophenotypic Subgroups
Baseline demographic, laboratory, cytogenetic, molecular, treatment, and outcome data, stratified by fusion partner, are summarized in Table 1, and corresponding data, stratified by flow-defined immunophenotype, are summarized in Table 2. For variables with incomplete testing or follow-up, percentages are reported as n/N (%), where N represents the number of evaluable cases for that specific variable. Age at diagnosis differed markedly across fusion classes (p<0.0001). NUP98::NSD1 patients were older (median 13.2 years, range 2.6-20.2) compared with NUP98::KDM5A (median 1.9 years, range 1.2-10.1) and NUP98::X (median 4.2 years, range 0.6-16.3). The gender distribution showed male predominance (26/35, 74%) in NUP98::NSD1 and female predominance (7/12, 58%) in NUP98::KDM5A, though the difference was not statistically significant (p=0.080) (Table 1).
When stratified by these flow-defined differentiation categories (as defined in the Methods), median age at diagnosis differed substantially (p<0.0001, Table 2) among the four groups; the erythroid/megakaryocytic group represented the youngest patients (median 1.8 years, range 0.6–3.8), whereas the remaining groups occurred in older children (medians 13.0, 14.0, and 11.7 years, respectively).
Laboratory Findings and Comparison Among Fusion and Immunophenotypic Subgroups
Peripheral blood leukocyte counts (median 15.9×109/L, range 0.6–460.0) varied by fusion partners (p=0.048), with NUP98::NSD1 associated with higher leukocyte counts (median 102.3×109/L) than NUP98::KDM5A (10.5×109/L) or NUP98::X (14.5×109/L) (Table 1). Peripheral blood blast percentages also varied (p=0.049), with lower values in NUP98::KDM5A (median 10%) than in NUP98::NSD1 (47%) or NUP98::X (39%). In paired NUP98::KDM5A samples with concurrent flow cytometry analysis, flow cytometry detected higher blast percentages than morphology in cases 34 and 37 (5.4% vs 2% and 47% vs 15%, respectively), whereas they were concordant in cases 36, 43, and 44 (10% vs 8%, 30.4% vs 35%, and 12% vs 12%, respectively); case 35 was negative by flow and showed only 1% blasts by morphology. Hemoglobin and platelet counts were comparable across fusion classes (p=0.86 and p=0.87, respectively). Absolute monocyte count did not differ significantly by fusion partner (p=0.387). Bone marrow blast percentages (median 61%, range 12–99%) were similar across fusion groups (p=0.519). Central nervous system (CNS) involvement, assessed by cerebrospinal fluid (CSF) cytomorphology (flow cytometry was not routinely performed on CSF), was present in 18/46 (39%) patients and was significantly enriched in NUP98::NSD1 compared with other fusion groups (55% vs 0% vs 22%; p = 0.006; Table 1).
Across immunophenotypic categories, peripheral blood leukocyte, hemoglobin, platelets, blasts, and monocyte counts did not differ significantly (Table 2). CNS involvement differed across lineage categories (p = 0.012; Table 2), with the highest rates of CNS involvement in monocytic (63%) and myelomonocytic (59%) cases, whereas no erythroid/megakaryocytic cases showed involvement.
Morphologic Findings
Representative morphologic and flow cytometric features of the major differentiation categories are illustrated in Figure 2 (myeloid/monocytic/myelomonocytic), Figure 3 (megakaryocytic and erythroid, including lineage-diverse presentations), Figure 4 (uncommon and challenging cases with myeloid and myelomonocytic immunophenotype), and Supplementary Figure S1 (AML case with overlapping features of MDS and MPN, characterized by prominent monocytosis with a mature monocytic immunophenotype).
Figure 2.
Case-based examples illustrating immunophenotypic heterogeneity in NUP98-rearranged pediatric myeloid neoplasms. Representative flow cytometry dot plots (abnormal population highlighted in red) are shown along with corresponding smear morphology, cytochemistry, and/or bone marrow biopsy/immunohistochemistry. (A) Case 50 (myeloid phenotype; NUP98::DDX10). Flow cytometry shows a blast population with an immature myeloid profile. Morphology demonstrates marrow involvement with blasts and some small megakaryocytes (arrow) (×1000), as well as a hypercellular core biopsy (×200). CD42b immunohistochemistry highlights increased numbers of small megakaryocytes, while blasts are negative (×200). (B) Case 51 (monocytic phenotype; NUP98::JADE2). Flow cytometry identifies blasts with high side scatter and monocytic-associated antigen expression (including bright CD33 and CD64), with dim CD19 co-expression. HLA-DR is partially lost, and CD34 and CD117 are absent/very low, at the lymphocyte/internal-control expression level (blue). Although the immunophenotype may raise concern for an APL-like presentation, morphology shows monoblasts rather than hypergranular promyelocytes on aspirate smear (×1000), with MPO-positive cytochemistry (×1000), and marrow biopsy (×200) demonstrates hypercellular marrow with leukemic involvement. (C) Case 57 (myelomonocytic phenotype; NUP98::ETS1, relapse/persistent disease sample). Flow cytometry demonstrates two distinct populations: myeloblasts (red) and monocytic cells (green). Blood and bone marrow smears (×1000) show monoblasts and myeloblasts. MPO cytochemistry is negative in the majority of blasts (×1000), and marrow biopsy confirms disease involvement (×200). Abbreviations: APL, acute promyelocytic leukemia; MPO, myeloperoxidase.
Figure 3.
Case-based examples highlighting phenotypic diversity in NUP98-rearranged pediatric myeloid neoplasms. Representative flow cytometry dot plots (abnormal population highlighted in red) are shown with corresponding smear morphology, cytochemistry, and/or bone marrow biopsy/immunohistochemistry. (A) Case 36 (erythroid/megakaryocytic phenotype; NUP98::KDM5A). Flow cytometry demonstrates an abnormal population with an erythroid/megakaryocytic pattern, including strong expression of erythroid/megakaryocytic-associated antigens (CD36 and CD71). Peripheral blood and bone marrow aspirate show leukemic involvement by blasts with basophilic cytoplasm and occasional cytoplasmic vacuoles and blebs (peripheral blood and bone marrow aspirate, ×1000; bone marrow core biopsy, ×200). MPO cytochemistry is shown for comparison and is negative (×1000). (B) Case 34 (megakaryocytic phenotype; NUP98::KDM5A). Flow cytometry shows a leukemic population with immature features and myeloid-associated antigen expression, including dim CD117 expression and lack of CD33, CD34, and HLA-DR. Megakaryocytic differentiation is highlighted by cytoplasmic CD42b (cyCD42b) expression, with CD71 shown in the same panel, and by dim CD4 expression. Smear morphology shows blasts with occasional cytoplasmic blebs (peripheral blood and bone marrow aspirate, ×1000). The bone marrow biopsy is hypercellular (×200), and CD42b immunohistochemistry supports megakaryocytic lineage (×200). (C) Case 58 (myeloid case with ETP-like features; NUP98::BPTF). Flow cytometry demonstrates an immature blast population with CD34 and CD117 expression and T-lineage-skewed antigen expression, including cytoplasmic CD3 positivity, absent surface CD3, CD4/CD8 double negativity, and CD56 expression. CD5 is negative. On combined morphologic and immunophenotypic review, this case was assigned to the myeloid immunophenotypic category and is shown here to illustrate a rare myeloid case with ETP-like features. Abbreviations: MPO, myeloperoxidase; cy, cytoplasmic; s, surface; ETP-like, early T-precursor-like.
Figure 4.
Immunophenotypic evolution in NUP98-rearranged pediatric myeloid neoplasms. (A) A representative case of NUP98::NSD1 AML showing immunophenotypic shift at relapse. The initial bone marrow (upper panel) displays blasts with a myelomonocytic immunophenotype. Blasts (red) express CD19 and are negative for CD22 and CD56. The relapsed bone marrow (lower panel) shows blasts with a myeloid immunophenotype. A subset of blasts (red) expresses CD19, CD22, cyCD79a, and CD56. (B) Broad immunophenotype category shifts from diagnosis to first relapse (n=19 paired cases; see Table 4). Descriptive modifiers, such as ETP-like features or poorly differentiated phenotype, are displayed as descriptive nodes but not used in cohort-level analyses.
Dysplastic myeloid, megakaryocytic, and erythroid precursors exceeding 10% of the lineage elements were detected in 42 out of 43 cases (98%) with adequate cells for evaluation (myeloid: 34 cases, megakaryocytic: 19 cases, erythroid: 23 cases; Supplementary Table S1). These dysplastic features were observed across fusion partners, including both NUP98::NSD1 and NUP98::KDM5A cases. Erythroid cells showed nuclear-cytoplasmic asynchrony, binucleation, karyorrhexis, irregular nuclear borders, and megaloblastic changes. The most frequent dysplastic feature in the megakaryocyte lineage was small megakaryocytes with single lobes. In NUP98::KDM5A/AMKL cases, small dysplastic megakaryocytes and micromegakaryocytes were often prominent, and distinguishing dysplastic small megakaryocytes from blasts with megakaryocytic differentiation was challenging. The myeloid lineage showed hypergranulation (salmon-colored granules) with some mature hypolobated neutrophils. Monocytosis (19/48, 39%) with an increase in atypical monocytes and left shift in myeloid/monocytic maturation (22/50, 44%) were frequent (Supplementary Table S1; Figure 2 and Supplementary Figure S1).
Cytogenetics and Comparison Among Fusion and Immunophenotypic Subgroups
Broad conventional cytogenetic categorization at diagnosis was available for 50 patients. Recurrent chromosome-level and focal copy-number abnormalities (11p and other genomic regions) were assessed using conventional cytogenetics together with available genomic copy-number data from WGS, OGM, or chromosomal microarray; denominators reflect cases evaluable by at least one applicable method. Among the 50 cases with conventional karyotype categorization, 21/50 (42%) had a diploid karyotype, 18/50 (36%) showed a balanced translocation, and 11/50 (22%) met criteria for a complex karyotype. Distribution differed across fusion-partner subgroups (p=0.008) (Table 1, Figure 1). Diploid karyotypes were most common in NUP98::NSD1 (61%), whereas NUP98::KDM5A showed a high proportion of complex karyotypes (63%), and NUP98::X exhibited a greater proportion of balanced translocations (55%). Using conventional cytogenetics together with available genomic copy-number data, 11p abnormalities were identified in 11/57 (19%) evaluable cases overall and were enriched in the NUP98::X subgroup (79% of evaluable NUP98::X cases; p<0.001), without significant enrichment across immunophenotype categories (Tables 1 and 2). In NUP98::X cases with available diagnostic conventional cytogenetics, these 11p abnormalities were also evident on karyotype and were corroborated by WGS; one additional case (case 50) lacked conventional cytogenetics and showed an 11p abnormality by WGS alone. Chromosome 13 abnormalities were identified in 7/57 (12%) evaluable cases and were heavily enriched in NUP98::KDM5A (6/10, 60%; p<0.001) and the erythroid/megakaryocytic immunophenotype group (5/8, 63%; p<0.001) (Tables 1 and 2).
Somatic (acquired) trisomy 21 appeared in 7/56 (13%) evaluable cases and was confined to NUP98::KDM5A (40%) and NUP98::X (23%), but was absent in NUP98::NSD1 (p=0.001). Trisomy 21 was also more common in erythroid/megakaryocytic cases (50%, p=0.018) (Tables 1 and 2). Trisomy 8 and 5q abnormalities did not vary significantly across fusion-partner or immunophenotype subgroups (Tables 1 and 2). Overall, 5q abnormalities were identified in 9/55 (16%) evaluable cases; 6 were present at diagnosis, whereas 3 were acquired at relapse or persistent disease. Full case-level cytogenetic results are detailed in Supplementary Table S2.
Mutational Profile and Comparison Among Fusion and Immunophenotypic Subgroups
The genomic landscape of NUP98 fusions has recently been characterized in large-scale studies. 7, 12 To associate cooperating mutations with immunophenotypic patterns in our cohort, we focused on FLT3-ITD and WT1, the most recurrent co-alterations in NUP98::NSD1 (Table 1; Supplementary Table S3). In our cohort, FLT3-ITD was detected in 23/60 (38%) and WT1 alterations in 24/59 (41%) evaluable cases (Table 1). Both were strongly enriched in NUP98::NSD1 (FLT3-ITD 22/34, 65%; WT1 22/34, 65%), were absent in NUP98::KDM5A (0/12 for each), and were uncommon in NUP98::X (FLT3-ITD 1/14, 7%; WT1 2/13, 15%) (Table 1). Co-occurrence of FLT3-ITD and WT1 was observed in 15/59 (25%) evaluable cases and occurred exclusively in NUP98::NSD1 (15/34, 44%) (Table 1). Within NUP98::NSD1 cases with evaluable data (Supplementary Table S3), FLT3-ITD positivity was associated with greater monocytic differentiation: FLT3-ITD was present in 5/6 monocytic and 15/19 myelomonocytic cases, compared with 2/9 myeloid cases. WT1 alterations were frequent across NUP98::NSD1 immunophenotypes (monocytic 4/6, myelomonocytic 13/19, myeloid 5/9 evaluable cases) and did not show a distinct flow cytometric signature beyond enrichment within the NUP98::NSD1 subgroup.
Across the overall cohort (Table 2), FLT3-ITD was enriched in monocytic (5/9, 56%) and myelomonocytic (15/23, 65%) categories, less frequent in myeloid (3/17, 18%), and absent in erythroid/megakaryocytic cases (0/11, 0%; p=0.002). WT1 alterations showed a similar distribution (myeloid 8/17, 47%; monocytic 4/9, 44%; myelomonocytic 12/22, 55%; erythroid/megakaryocytic 0/11, 0%; p=0.011). RB1 sequence or copy-number alterations were detected in 6/56 (11%) evaluable cases and were confined to NUP98::KDM5A (6/11, 55%; p<0.001) and to the erythroid/megakaryocytic category (6/9, 67%; p<0.001), overlapping with chromosome 13/13q abnormalities (Tables 1 and 2). The single NPM1 alteration shown in Figure 1 occurred in case 44 as an NPM1::CSF1R rearrangement with a co-occurring GATA1 Pro277fs mutation (Tables 1 and 2). Additional co-mutations were heterogeneous and are provided in Supplementary Table S3.
Immunophenotype and Cytogenetic Evolution at Relapse
Immunophenotype and cytogenetic changes at relapse were assessable in a subset of patients with paired or sequential bone marrow samples (Table 4). Among patients with paired initial and first-relapse immunophenotypes, a shift in the broad immunophenotype category occurred in 7/19 (37%), most often along the myeloid-monocytic spectrum and frequently accompanied by gain or loss of aberrant lymphoid-associated antigens such as CD19 and/or CD7. Cytogenetic evolution was seen in 11/22 (50%) evaluable cases, including acquisition, loss, or other cytogenetic changes consistent with clonal evolution. Some patients developed new cytogenetic abnormalities while maintaining a stable immunophenotypic category, whereas others showed immunophenotypic shift without detectable karyotypic change. These data represent relapse/persistent-disease immunophenotype assessments rather than MRD follow-up measurements.
Treatment and Outcomes
Treatment history was available for 50 patients. Nearly all (49/50, 98%) received chemotherapy. Hematopoietic stem cell transplantation (HSCT) was performed in 34/50 (68%) cases overall, with no major differences in frequency across fusion-partner or immunophenotype subgroups. The median follow-up was 26.9 months (range 0.0-263.4) among cases with available follow-up data (n=48). Outcome data were available for 52 patients. The complete remission (CR) rate was 21/52 (40%). When stratified by fusion partner, CR rates were 41% for NUP98::NSD1, 13% for NUP98::KDM5A, and 58% for NUP98::X (p=0.123). Death occurred in 31/52 (60%) patients, with the highest observed mortality in NUP98::KDM5A (88%) and the erythroid/megakaryocytic immunophenotype (100%). CR differed significantly by immunophenotype (p=0.041), with higher rates in monocytic (67%) and myeloid (50%) cases, lower rates in myelomonocytic cases (32%), and no documented CR in the erythroid/megakaryocytic category (0%). Representative patient-level treatment responses and clinical courses are illustrated in the swimmer plot (Supplementary Figure S2). Notably, these outcome data provide clinical context rather than definitive comparative risk estimates because subgroup sizes were small and treatment spanned 1994-2025, during which induction regimens, transplant approaches, supportive care, and FLT3-inhibitor availability changed substantially.
Discussion
Although NUP98 rearrangements are often grouped as a single adverse genetic category, our clinicopathologic analysis shows that pediatric NUP98-rearranged myeloid neoplasms have fusion partner–associated differentiation programs detected by diagnostic flow cytometry. By integrating morphology, flow cytometry, cytogenetics, and selected molecular correlates, our study complements recent genomic landscape studies and translates them into practical diagnostic patterns for hematopathologists. Because many NUP98 fusions are cryptic by conventional cytogenetics 7, 12, 24, these patterns can promote earlier selection of RNA-based fusion testing and inform flow-based follow-up planning.
Across the cohort, NUP98::NSD1 most often presented as an immature myeloid or myelomonocytic leukemia with frequent CD34, CD117, HLA-DR, and CD123 expression, often accompanied by aberrant CD7 and/or CD19, consistent with other high-risk pediatric AML phenotypes and emphasizing the need to pair flow cytometry with molecular testing. 7, 12, 25 In contrast, NUP98::KDM5A, which occurred predominantly in infants and very young children, showed a pronounced erythroid/megakaryocytic bias, reduced CD34 and HLA-DR, and absent MPO, in keeping with prior reports. 16, 17 These overall patterns are concordant with the partner-associated clinicogenomic distinctions reported by Bertrums et al. and others. 7, 12 Rare-partner cases (NUP98::X) were more heterogeneous but generally retained a myeloid backbone. 7 Recent studies also indicate that NUP98r leukemias can span AML, lower-blast myeloid neoplasms, and occasional lineage-ambiguous or T-lineage/ETP-like presentations, so rare partners such as RAP1GDS1 require especially careful integrated lineage assignment. 12
Cytogenetic and morphologic context further clarified these flow-based patterns. Most cases did not show a karyotype-visible 11p15 abnormality—indicating cryptic NUP98 involvement at the standard cytogenetic level—whereas karyotype-visible 11p15 rearrangements were concentrated in NUP98::X cases. 7, 12, 24 Beyond prognostic stratification, timely recognition of NUP98 rearrangements may also inform eligibility for emerging targeted therapies such as menin inhibitors. 26-28 Together, these morphologic, cytogenetic, and flow-based features support a proposed pattern-based approach for prioritizing molecular evaluation for NUP98 rearrangement (Supplementary Figure S3), which is intended to prioritize cases for targeted RNA fusion testing and, when suspicion persists despite negative or limited initial studies, reflex NUP98 break-apart fluorescence in situ hybridization (FISH) and/or optical genome mapping (OGM), with broad RNA sequencing reserved for unresolved cases. 24 However, unbiased RNA-Seq, if available, is the preferred initial strategy for detecting the full spectrum of these events. The proposed workflow is hypothesis-generating. Extensive validation studies will be necessary to establish the sensitivity, specificity, positive predictive value, or negative predictive value of each parameter.
Morphologic review largely reflected the flow-defined differentiation programs; importantly, dysplastic changes were observed across fusion partners (including both NUP98::NSD1 and NUP98::KDM5A), supporting a combined diagnostic approach rather than relying on a single morphologic feature. In infants and very young children, an erythroid/megakaryocytic immunophenotype should prompt evaluation for NUP98::KDM5A, in addition to other recurrent pediatric AMKL entities, particularly CBFA2T3::GLIS2-associated RAM phenotype and RBM15::MRTFA. 15, 18 In our NUP98::KDM5A cases, retained CD38 in most evaluable cases, lack of uniform bright CD56 pattern, and frequent CD36 positivity were not consistent with RAM immunophenotype commonly associated with CBFA2T3::GLIS2 cases 7, 17
The genomic landscape of NUP98 fusions has been recently characterized in large cohorts. 7, 12 To link cooperating lesions with immunophenotypic patterns, we focused on the highly recurrent FLT3-ITD and WT1 alterations, which were largely confined to NUP98::NSD1 in our cohort. Notably, FLT3-ITD (and WT1) were enriched in cases with monocytic or myelomonocytic differentiation and were absent in erythroid/megakaryocytic leukemias, suggesting that a monocytic/myelomonocytic phenotype should prompt early FLT3/WT1 assessment in parallel with fusion testing. A related entity to consider in cytogenetically normal pediatric myeloid neoplasms with WT1/FLT3-ITD is UBTF tandem duplication, which can overlap with an immature myeloid phenotype and likewise should be included in the initial molecular differential diagnosis. 25, 29
Serial specimens showed that the immunophenotype in NUP98-rearranged disease can evolve. While many cases maintained their overall differentiation category at relapse or persistent disease, a subset shifted along the myeloid–monocytic spectrum and showed antigen drift, and paired cytogenetics often revealed evolution (Table 4). In cases with relatively mature monocytic blast equivalents showing higher CD45 and side scatter and absent CD34/CD117, residual disease may fall outside a traditional blast gate. These findings are useful for MRD panel design as a leukemia-associated immunophenotype established at diagnosis may be insufficient if relapse becomes more monocytic or if aberrant antigens are gained or lost. Flow-based monitoring should, therefore, combine leukemia-associated immunophenotype (LAIP) and different-from-normal approaches 6, 30, emphasize stable backbone markers (e.g., CD33, CD13, CD117, HLA-DR, and when present, CD34 and/or CD123), and deliberately examine both blast and monocytic regions with flexible gating. Available treatment records did not identify menin inhibitor exposure or other differentiation-directed therapy as an explanation for the observed phenotype shifts.
Aberrant CD19 expression was observed in about one-quarter of cases and was more common in NUP98::NSD1, making it a potentially useful diagnostic clue when present on an otherwise myeloid or immature blast population. Because additional B-lineage marker expression was uncommon at diagnosis and variably dim or partial at relapse, these cases do not meet the criteria of B/myeloid mixed-phenotype acute leukemia. Lineage assignment should be based on the overall antigen pattern, MPO and/or myeloid marker expression, and integrated morphologic and molecular findings, analogous to the approach used for acute myeloid leukemia with RUNX1::RUNX1T1. For follow-up flow cytometry, CD19 may improve detection sensitivity when retained, but the observed antigen gain/loss supports using it as a supplementary leukemia-associated immunophenotype feature rather than a primary tracking marker.
This study is limited by its retrospective design, multi-decade time span, small NUP98::KDM5A and NUP98::X subgroup sizes, variability in flow cytometry panels, incomplete longitudinal sampling for some patients, and treatment heterogeneity that precludes definitive outcome analyses. Nevertheless, the clinicobiologic patterns observed here—especially the profiles associated with NUP98::NSD1 and NUP98::KDM5A—provide pathologic context to prior cooperative-group and genomic landscape studies. Together, these data support a pathology-focused approach in which partner-associated flow cytometry patterns prompt timely reflex molecular testing for cryptic NUP98 rearrangements and may guide MRD panel design in this high-risk pediatric NUP98r myeloid-neoplasm population.
Given that most NUP98 rearrangements are cryptic by conventional cytogenetics, partner-associated flow cytometric patterns can provide valuable diagnostic clues that prompt confirmatory molecular studies, including RNA-based fusion testing, in pediatric myeloid neoplasms. NUP98::NSD1 usually exhibits an immature CD34+/CD117+/HLA-DR+/CD123+ phenotype with abnormal CD7/CD19 expression, while monocytic/myelomonocytic differentiation should raise suspicion for co-occurring FLT3-ITD and/or WT1 alterations. In infants, an erythroid/megakaryocytic phenotype should prompt investigation of NUP98::KDM5A alongside CBFA2T3::GLIS2/RAM and RBM15::MRTFA, with chromosome 13/RB1 abnormalities offering supportive evidence. Rare NUP98 fusion partners (NUP98::X) remain diverse, but visible 11p15 rearrangements on karyotypes were enriched in these cases. Since antigenic drift and occasional category shifts can occur at relapse, flow-based monitoring should rely on stable backbone markers and different-from-normal analysis, using CD19 and other aberrancies as supportive tracking features.
Supplementary Material
Supplementary Figure S1. Case 52 (monocytic phenotype; NUP98::JADE2). Representative morphology shows blasts mixed with numerous atypical, monocytic-appearing cells; by morphology alone, many of these cells are difficult to confidently classify as blasts. Flow cytometry (not shown) demonstrates that cells within the conventional “mature monocyte” gate (77% of total cells) have bright CD45 expression and intermediate side scatter, consistent with maturing monocytes. These cells are positive for CD33, HLA-DR, CD11b, CD13, CD14, CD15, CD16, CD36, CD64, dim CD4, and moderate cMPO, with aberrant uniform CD56 expression, indicating that the mature monocytic region largely represents the same leukemic population with a deceptively mature monocytic immunophenotype. This highlights a diagnostic/MRD pitfall in which reliance on morphology or standard maturation-based gating may underestimate leukemic burden. Peripheral blood and bone marrow aspirate, ×1000; hypercellular bone marrow biopsy, ×400; lysozyme immunohistochemistry, ×200; and alpha-naphthyl butyrate (ANB) stain, ×1000.
Supplementary Figure S2. The swimmer plot visualizes selected representative patient-level treatment responses and clinical courses, showing the time from diagnosis (in days) to the last follow-up or death. It annotates key events such as complete remission after induction or salvage therapy, induction failure, HSCT, relapse, and death. The plot illustrates substantial variability in clinical paths, including early induction failure or death in some patients, recurrent relapses in others, and long-lasting clinical courses.
Supplementary Figure S3. Proposed hypothesis-generating algorithm for prioritizing evaluation of NUP98 rearrangement in pediatric myeloid neoplasms, mainly AML. Flow cytometry and cytogenetic findings suggestive of NUP98 rearrangement may prompt targeted RNA fusion testing (e.g., a leukemia fusion panel). If results are negative or samples are limited but suspicion remains, reflex NUP98 break-apart FISH and/or optical genome mapping (OGM) may identify cryptic rearrangements; broad RNA sequencing can be reserved for unresolved cases or persistent clinical suspicion. When an immature myeloid or myelomonocytic immunophenotype raises suspicion for NUP98::NSD1, parallel FLT3-ITD and WT1 testing may support early risk stratification. This descriptive workflow is intended as a practical triage aid when comprehensive up-front RNA-based testing is unavailable; sensitivity, specificity, and predictive values have not been established and require prospective validation. Abbreviations: FISH, fluorescence in situ hybridization; OGM, optical genome mapping; RNA-seq, RNA sequencing; AML, acute myeloid leukemia.
Acknowledgment
We thank current and previous members of the Clinical Immunopathology Laboratory at St. Jude Children’s Research Hospital for their expertise and contributions to this work. J.M. Klco holds a Career Award for Medical Scientists from the Burroughs Wellcome Fund and is a former recipient of the V Foundation Scholar Award (Pediatric).
Funding
This work was supported by the American Lebanese and Syrian Associated Charities of St. Jude Children’s Research Hospital and grants from the NIH (U54 CA243124, Fusion Oncoproteins in Childhood Cancers (FusOnC2) Consortium awarded to J.M. Klco). However, the content does not necessarily reflect the official views of the NIH and is solely the authors’ responsibility.
Data Availability Statement
Sequencing datasets generated through prior research studies are available through the corresponding publications and repositories cited therein. Additional de-identified data supporting the findings of this study are available from the corresponding author on reasonable request, subject to institutional review and patient-privacy restrictions.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Ethics Approval and Patient Consent
This study was approved by the Institutional Review Boards of St. Jude Children’s Research Hospital and The University of Texas MD Anderson Cancer Center. The requirement for informed consent was waived by the Institutional Review Boards for this retrospective study.
References
- 1.Cooper TM, Alonzo TA, Tasian SK, et al. Children’s Oncology Group’s 2023 blueprint for research: Myeloid neoplasms. Pediatr Blood Cancer. 2023;70 Suppl 6(Suppl 6):e30584. 10.1002/pbc.30584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Downing JR, Shannon KM. Acute leukemia: a pediatric perspective. Cancer Cell. 2002;2(6):437–445. 10.1016/s1535-6108(02)00211-8. [DOI] [PubMed] [Google Scholar]
- 3.Rasche M, Zimmermann M, Steidel E, et al. Survival Following Relapse in Children with Acute Myeloid Leukemia: A Report from AML-BFM and COG. Cancers (Basel). 2021;13(10). 10.3390/cancers13102336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Karlsson L, Forestier E, Hasle H, et al. Outcome after intensive reinduction therapy and allogeneic stem cell transplant in paediatric relapsed acute myeloid leukaemia. Br J Haematol. 2017;178(4):592–602. 10.1111/bjh.14720. [DOI] [PubMed] [Google Scholar]
- 5.Tomizawa D, Tsujimoto SI. Risk-Stratified Therapy for Pediatric Acute Myeloid Leukemia. Cancers (Basel). 2023;15(16). 10.3390/cancers15164171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Loken MR, Alonzo TA, Pardo L, et al. Residual disease detected by multidimensional flow cytometry signifies high relapse risk in patients with de novo acute myeloid leukemia: a report from Children’s Oncology Group. Blood. 2012;120(8):1581–1588. 10.1182/blood-2012-02-408336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bertrums EJM, Smith JL, Harmon L, et al. Comprehensive molecular and clinical characterization of NUP98 fusions in pediatric acute myeloid leukemia. Haematologica. 2023;108(8):2044–2058. 10.3324/haematol.2022.281653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Umeda M, Ma J, Westover T, et al. A new genomic framework to categorize pediatric acute myeloid leukemia. Nat Genet. 2024;56(2):281–293. 10.1038/s41588-023-01640-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Michmerhuizen NL, Klco JM, Mullighan CG. Mechanistic insights and potential therapeutic approaches for NUP98-rearranged hematologic malignancies. Blood. 2020;136(20):2275–2289. 10.1182/blood.2020007093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Romana SP, Radford-Weiss I, Ben Abdelali R, et al. NUP98 rearrangements in hematopoietic malignancies: a study of the Groupe Francophone de Cytogenetique Hematologique. Leukemia. 2006;20(4):696–706. 10.1038/sj.leu.2404130. [DOI] [PubMed] [Google Scholar]
- 11.Gough SM, Slape CI, Aplan PD. NUP98 gene fusions and hematopoietic malignancies: common themes and new biologic insights. Blood. 2011;118(24):6247–6257. 10.1182/blood-2011-07-328880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Umeda M, Hiltenbrand R, Michmerhuizen NL, et al. Fusion oncoproteins and cooperating mutations define disease phenotypes in NUP98-rearranged leukemia. Blood. 2025;146(17):2102–2118. 10.1182/blood.2025028993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Niktoreh N, Walter C, Zimmermann M, et al. Mutated WT1, FLT3-ITD, and NUP98-NSD1 Fusion in Various Combinations Define a Poor Prognostic Group in Pediatric Acute Myeloid Leukemia. J Oncol. 2019;2019:1609128. 10.1155/2019/1609128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Tarlock K, Gerbing RB, Ries RE, et al. Prognostic impact of cooccurring mutations in FLT3-ITD pediatric acute myeloid leukemia. Blood Adv. 2024;8(9):2094–2103. 10.1182/bloodadvances.2023011980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.de Rooij JD, Masetti R, van den Heuvel-Eibrink MM, et al. Recurrent abnormalities can be used for risk group stratification in pediatric AMKL: a retrospective intergroup study. Blood. 2016;127(26):3424–3430. 10.1182/blood-2016-01-695551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chisholm KM, Heerema-McKenney AE, Choi JK, et al. Acute erythroid leukemia is enriched in NUP98 fusions: a report from the Children’s Oncology Group. Blood Adv. 2020;4(23):6000–6008. 10.1182/bloodadvances.2020002712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Noort S, Wander P, Alonzo TA, et al. The clinical and biological characteristics of NUP98-KDM5A in pediatric acute myeloid leukemia. Haematologica. 2021;106(2):630–634. 10.3324/haematol.2019.236745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.de Rooij JD, Branstetter C, Ma J, et al. Pediatric non-Down syndrome acute megakaryoblastic leukemia is characterized by distinct genomic subsets with varying outcomes. Nat Genet. 2017;49(3):451–456. 10.1038/ng.3772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Iacobucci I, Wen J, Meggendorfer M, et al. Genomic subtyping and therapeutic targeting of acute erythroleukemia. Nat Genet. 2019;51(4):694–704. 10.1038/s41588-019-0375-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Arber DA, Orazi A, Hasserjian RP, et al. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200–1228. 10.1182/blood.2022015850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dohner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345–1377. 10.1182/blood.2022016867. [DOI] [PubMed] [Google Scholar]
- 22.Khoury JD, Solary E, Abla O, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia. 2022;36(7):1703–1719. 10.1038/s41375-022-01613-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Voss RK, Pastor Loyola VB, Cardenas MF, et al. Clinical experience of using integrated whole genome and transcriptome sequencing as a framework for pediatric and adolescent acute myeloid leukemia diagnosis and risk assessment. Leukemia. 2025;39(12):2946–2956. 10.1038/s41375-025-02774-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yuen LD, Hasserjian RP, Fathi AT, et al. Strategies for identifying NUP98 rearrangements in adult myeloid neoplasms. Haematologica. 2026;111(2):518–534. 10.3324/haematol.2025.288080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Khanlari M, Wang W, Ni Y, et al. Pediatric Myeloid Neoplasms With UBTF Tandem Duplications : Morphologic, Immunophenotypic, and Clinical Characterization. Am J Surg Pathol. 2025;49(4):353–362. 10.1097/PAS.0000000000002350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Carraway HE, Nakitandwe J, Cacovean A, et al. Complete remission of NUP98 fusion-positive acute myeloid leukemia with the covalent menin inhibitor BMF-219, icovamenib. Haematologica. 2025;110(4):1041–1046. 10.3324/haematol.2024.286537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Heikamp EB, Henrich JA, Perner F, et al. The menin-MLL1 interaction is a molecular dependency in NUP98-rearranged AML. Blood. 2022;139(6):894–906. 10.1182/blood.2021012806. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rasouli M, Blair H, Troester S, et al. The MLL-Menin Interaction is a Therapeutic Vulnerability in NUP98-rearranged AML. Hemasphere. 2023;7(8):e935. 10.1097/HS9.0000000000000935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Barajas JM, Umeda M, Contreras L, et al. UBTF tandem duplications in pediatric myelodysplastic syndrome and acute myeloid leukemia: implications for clinical screening and diagnosis. Haematologica. 2024. 10.3324/haematol.2023.284683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Langebrake C, Brinkmann I, Teigler-Schlegel A, et al. Immunophenotypic differences between diagnosis and relapse in childhood AML: Implications for MRD monitoring. Cytometry B Clin Cytom. 2005;63(1):1–9. 10.1002/cyto.b.20037. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure S1. Case 52 (monocytic phenotype; NUP98::JADE2). Representative morphology shows blasts mixed with numerous atypical, monocytic-appearing cells; by morphology alone, many of these cells are difficult to confidently classify as blasts. Flow cytometry (not shown) demonstrates that cells within the conventional “mature monocyte” gate (77% of total cells) have bright CD45 expression and intermediate side scatter, consistent with maturing monocytes. These cells are positive for CD33, HLA-DR, CD11b, CD13, CD14, CD15, CD16, CD36, CD64, dim CD4, and moderate cMPO, with aberrant uniform CD56 expression, indicating that the mature monocytic region largely represents the same leukemic population with a deceptively mature monocytic immunophenotype. This highlights a diagnostic/MRD pitfall in which reliance on morphology or standard maturation-based gating may underestimate leukemic burden. Peripheral blood and bone marrow aspirate, ×1000; hypercellular bone marrow biopsy, ×400; lysozyme immunohistochemistry, ×200; and alpha-naphthyl butyrate (ANB) stain, ×1000.
Supplementary Figure S2. The swimmer plot visualizes selected representative patient-level treatment responses and clinical courses, showing the time from diagnosis (in days) to the last follow-up or death. It annotates key events such as complete remission after induction or salvage therapy, induction failure, HSCT, relapse, and death. The plot illustrates substantial variability in clinical paths, including early induction failure or death in some patients, recurrent relapses in others, and long-lasting clinical courses.
Supplementary Figure S3. Proposed hypothesis-generating algorithm for prioritizing evaluation of NUP98 rearrangement in pediatric myeloid neoplasms, mainly AML. Flow cytometry and cytogenetic findings suggestive of NUP98 rearrangement may prompt targeted RNA fusion testing (e.g., a leukemia fusion panel). If results are negative or samples are limited but suspicion remains, reflex NUP98 break-apart FISH and/or optical genome mapping (OGM) may identify cryptic rearrangements; broad RNA sequencing can be reserved for unresolved cases or persistent clinical suspicion. When an immature myeloid or myelomonocytic immunophenotype raises suspicion for NUP98::NSD1, parallel FLT3-ITD and WT1 testing may support early risk stratification. This descriptive workflow is intended as a practical triage aid when comprehensive up-front RNA-based testing is unavailable; sensitivity, specificity, and predictive values have not been established and require prospective validation. Abbreviations: FISH, fluorescence in situ hybridization; OGM, optical genome mapping; RNA-seq, RNA sequencing; AML, acute myeloid leukemia.
Data Availability Statement
Sequencing datasets generated through prior research studies are available through the corresponding publications and repositories cited therein. Additional de-identified data supporting the findings of this study are available from the corresponding author on reasonable request, subject to institutional review and patient-privacy restrictions.