Classification and prognostic stratification based on genomic features in myelodysplastic neoplasms, myeloproliferative neoplasms, and their overlapping conditions

doi:10.21203/rs.3.rs-4352959/v1

Download PDF

Article

Classification and prognostic stratification based on genomic features in myelodysplastic neoplasms, myeloproliferative neoplasms, and their overlapping conditions

https://doi.org/10.21203/rs.3.rs-4352959/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

In this study, we analyzed clinical and genomic data from 1,585 patients diagnosed with myeloid neoplasms (MNs), including myeloproliferative neoplasms (MPN, n = 715), myelodysplastic neoplasms (MDS, n = 698), MDS/MPN (n = 94), and aplastic anemia (AA, n = 94). We identified ten distinct genomic groups that redefine MN classification using unsupervised genomic clustering through the Dirichlet Process (DP), correlating specific genetic mutations with survival outcomes and disease subtypes. Notably, groups DP1 and DP5, characterized by JAK2 and CALR mutations, respectively, showed a very favorable prognosis among patients with MPN. Groups DP2, DP7, and DP9 demonstrated a very adverse prognosis across MN subtypes. Specifically, DP2 encompasses MDS patients with TP53 mutations and complex karyotypes, DP9 is distinguished by acute myeloid leukemia-related mutations, including NPM1, and DP7 includes patients with SETBP1 mutations, indicating heterogeneous MN phenotypes. DP10 and DP8, linked to SF3B1, DDX41 mutations or chromosome 1q derivatives present a favorable risk profile. Our research emphasizes the critical role of genomic insights in enhancing the classification, prognostic accuracy, and therapeutic stratification of MNs. The survival improvement observed with transplantation in the very adverse risk groups underscores the potential of genomic classifications to inform personalized treatment strategies, signifying a significant step toward the integration of genomics into MN clinical management.

Recent genomic studies have introduced an alternative paradigm for the classification of myeloid neoplasms (MNs), focusing on shared and distinct genomic patterns (1–3). Landmark studies in myeloproliferative neoplasms (MPN) (4), myelodysplastic neoplasms (MDS) (5, 6), and their overlapping diseases (7) demonstrate the potential of genomic classification systems for disease subclassification and personalized prognostication. This has led to a fundamental shift towards genetics-based approaches in characterizing these diseases, with recent classification systems increasingly incorporating genomic attributes to align more closely with their underlying biology (2, 8). Nevertheless, current frameworks still prioritize clinical features over biological characteristics in distinguishing major classes of MNs (3, 8). The morphological and phenotypic criteria have subjective elements, leading to inter-observer variability in diagnosis (5). This results in biological heterogeneity within disease groups and hinders insights into pathogenic mechanisms that traverse boundaries. Additionally, the diagnostic thresholds for clinical and morphological features are not sufficiently discriminative, as they often reflect the outcomes of driver genetic variations (9). Moreover, molecular markers are intermingled in both MDS and MPN, such that a comprehensive genomic characterization is still lacking. Therefore, the current system often faces challenges in patients with ill-defined phenotypes, such as admixed features of myelofibrosis, dyspoiesis, and increased blasts at the advanced phase.

In this study, we present the first comprehensive genomic analysis encompassing a broad range of MNs in a Korean population, profiling frequently mutated genes and cytogenetic abnormalities in 1,585 patients. Through unsupervised genomic grouping, we identified novel subgroups reflecting fundamental disease relationships and reinforced pre-existing disease categories, enabling enhanced individualized prediction of clinical outcomes. This approach facilitates the proposal of an integrated classification system based on genetics rather than subjective clinical criteria. This study aims to contribute to the development of a more refined classification system based on underlying genomic features.

Study cohort

Clinical and genomic data were collected from the Clinical Data Warehouse of Catholic Medical Center in Seoul, Republic of Korea (http://cohort.cmcnu.or.kr). De-identified data from patients, who performed diagnostic NGS study on somatic mutations for MNs from 2017 to 2021 was extracted (sTable1). According to the 2022 WHO criteria, diagnosis of MPN, MDS (including clonal cytopenias of undetermined significance (CCUS)), MDS/MPN, and their morphologically defined subtypes were determined. We also included patients initially thought to have MNs, but instead presented with histopathologically confirmed AA. Patients with non-targeted disease, or insufficient clinical data for complete blood count, bone marrow findings, and karyotypes were excluded. Finally, a total of 1,585 patients were included in the modeling (sTable2). In addition, a validation cohort of 150 patients was sequentially collected during 2022 (sTable3). The institutional review board of the Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea, approved the studies (KC22RISI0217).

Genomic data

DNA samples from peripheral blood or bone marrow at diagnosis were used. Mutation data was obtained using targeted NGS panels as previously described (10, 11). We used two versions of the NGS panel depending on the accessed time points. As 87 genes were common between both NGS panels, we investigated the results further. The pathogenicity of the detected variants was determined using AMP (12) and ClinGen criteria (13). Cytogenetic data was obtained from G-banding and selective FISH tests. Detailed methods are described in the Supplementary Materials.

Statistics

Fisher’s exact test was employed to identify co-occurrence and mutually exclusive relationships among genetic abnormalities. Bayesian network analysis was performed to identify genomic associations and causalities. The Dirichlet process was applied to define genomic subgroups. We examined the timing of mutation acquisition using Bradley-Terry models of pairwise comparisons of clonal fractions. Survival analyses were conducted using the Kaplan-Meier method, which estimates the probability of survival over time. To assess differences in survival rates between various subgroups, we employed the log-rank test. Overall survival was defined as time from diagnosis to the last follow-up or death from any cause. Univariate and the subsequent multivariate Cox proportional hazard (PH) model was applied to compare clinical and hematological characteristics among different genomic subgroups. The detailed methods are described in the Supplementary Materials.

Cohort characteristics

The training cohort consisted of 715 patients with MPN, 698 patients with MDS (including 19 CCUS), 78 patients with MDS/MPN, and 94 patients with AA. MPNs were comprised of polycythemia vera (PV, n = 127), essential thrombocythemia (ET, n = 278), primary myelofibrosis (PMF, n = 233), secondary MF (n = 60), chronic neutrophilic leukemia (CNL, n = 4), juvenile myelomonocytic leukemia (JMML, n = 5), chronic eosinophilic leukemia (CEL, n = 1), and MPN not otherwise specified (NOS, n = 7). Patients with MDS included MDS with low blasts (LB, n = 294), LB and ring sideroblasts (LB-RS, n = 48), hypoplastic MDS (n = 55), increased blasts-1 (IB-1, n = 163), IB-2 (n = 108), and IB-fibrosis (n = 11). MDS/MPN included patients with chronic myelomonocytic leukemia (CMML, n = 62), MDS/MPN with neutrophilia (MDS/MPN-N, n = 7), MDS/MPN-NOS (n = 8), and MDS/MPN-RS and thrombocytosis (MDS/MPN-RS-T, n = 1). The validation cohort consisted of 77 patients with MPN, 55 patients with MDS, 7 patients with MDS/MPN, and 11 patients with AA (Fig. 1A). In the survival analysis of the training cohort, patients with MPN showed the most prolonged survival with an estimated survival time of 14.8 years (95% CI: 14.0–15.6), followed by AA (12.4 years, 10.0–14.8), MDS (8.6 years, 7.5–9.6), and MDS/MPN (5.7 years, 4.4–6.9) (Fig. 1B). Detailed demographics and hematological and clinical features are summarized in sTable 3–4 and sFigure 1.

Profile of gene mutations and chromosomal abnormalities.

Among the patients in the training cohort, 1,271 patients (80.2%) presented with one or more mutations with a median mutation number of 2 (range: 1–8). The average mutation number per patient was highest in MDS/MPN (3.2 ± 1.9), followed by MPN (1.9 ± 1.7), MDS (1.7 ± 1.6), and AA (0.2 ± 0.5). Mutations were identified in 82 of 87 genes, and 36 genes were mutated in more than 1% of patients. Among them, JAK2 (42.2%), ASXL1 (18.4%), TET2 (13.1%), U2AF1 (10.0%), CALR (9.7%), TP53 (8.7%), DNMT3A (6.0%), SF3B1 (5.9%), and RUNX1 (5.0%) were frequently mutated. Germline confirmed that DDX41 mutations were identified in 62 patients (3.9%). Normal karyotype was reported in 1,066 patients (67.3%), and 17 types of chromosomal events were identified in more than 1% of them, including + 8 (8.3%), -20/20q- (6.2%), -7/7q- (5.8%), + 1/1q+ (4.8%), and − 5q/non-isolated 5q- (3.6%). By disease types, any chromosomal abnormality was observed in 376 patients with MDS (53.9%), 20 patients with MDS/MPN (25.6%), 108 patients with MPN (15.1%), and 15 patients with AA (16.0%). The average number of chromosomal abnormalities per patient was highest in MDS (2.1 ± 4.0), followed by MDS/MPN (0.6 ± 1.3), MPN (0.4 ± 1.3), and AA (0.2 ± 0.5).

Next, we focused on 53 genomic abnormalities that occurred in more than 1% of patients. Consequently, genomic data from 1,359 patients were used for further genomic grouping (Fig. 1C and sFigure2).

Genomic associations

Pairwise association analysis (sFigure3) and Bayesian network analysis (sFigure4) revealed patterns of co-occurrences and exclusive pairings among gene mutations and chromosomal abnormalities. JAK2 mutations were strongly and exclusively associated with CALR and DDX41 mutations, as well as chromosomal abnormalities, such as + 8, -5/non-isolated 5q-, and − 17/17p-/t17. CALR mutations were exclusively associated with U2AF1. TP53 mutations showed significant associations predominantly with chromosomal abnormalities: -5/non-isolated 5q-, -7/7q-, and − 18. NPM1 mutations commonly co-occurred with BCOR and RUNX1. SETBP1 mutations frequently co-occurred with CSF3R and were often found with GATA2, NRAS, NF1, and PTPN11 mutations. SRSF2 mutations showed moderate associations with CUX1, IDH2, and STAG2.

Characteristics of the genomic groups

Using the Dirichlet process (DP), we identified 10 distinct genomic groups, each characterized by unique features (Fig. 2 and sFigure 5–7). DP1 was the largest genomic group, containing 456 patients characterized by activating mutations in JAK2. DP5 included 148 patients with CALR mutations. DP2 (n = 118) was characterized by TP53 mutations and/or complex karyotypes frequently involving − 5/5q- (not isolated − 5q), -7/7q-, -17/17p-/t17, -9/9q-, and − 13/13q-. DP9 included 21 patients with AML-like mutation patterns, specifically NPM1 and co-occurring BCOR, IDH2, DNMT3A, and WT1 mutations. DP7 (n = 40) showed a higher frequency of SETBP1 mutations with co-occurring CSF3R and/or RAS pathway mutations. DP4 (n = 93) displayed a higher frequency of mutations in TET2 and SRSF2, with co-occurring STAG2, NRAS, KRAS, ZRSR2, and IDH2. DP3 (n = 177) was associated with U2AF1, RUNX1, BCOR, and PHF6 mutations, and chromosomal changes, such as + 8. DP6 (n = 35) was linked to mutations in CBL, ETV6, and KMT2D. DP10 was the second largest group (n = 224), showing marked genomic heterogeneity, associated with mutations in SF3B1 or DDX41. DDX41 (n = 46) was totally exclusive to SF3B1 (n = 58), and most DDX41 cases (n = 43) accompanied germline DDX41 mutations. Rare cases of isolated − 5q, -Y, or MPL mutations were also included. DP8 (n = 47) was notably associated with chromosomal abnormalities, including isolated + 1/1q + and various derivative chromosomes involving chromosome 1. The most prevalent common derivative chromosome identified was derivative (1;7), accounting for 33% of cases. Additionally, derivative chromosome 1 was observed to involve other chromosomes, including 6, 12, 13, 19, 10, 14, 15, and 22. Lastly, we assigned patients without specific genomic changes to DP0 (n = 226).

Bradley-Terry modeling was used to estimate the mutation acquisition order of each genomic group (sFigure 8). Mutations driving MPN, specifically JAK2, CALR, and MPL, were acquired early in genomic groups DP1, DP5, and DP10. Conversely, the JAK2 mutation occurred relatively later in other groups, DP3, DP4, and DP6. In DP2, TP53 occurred earlier than epigenetic regulators (TET2, ASXL1, and DNMT3A) and splicing factors (U2AF1 and SF3B1). Conversely, epigenetic regulators (BCOR and EZH2) and splicing factors (ZRSR2 and U2AF1) were acquired early in the disease development process in DP3 and DP6 groups. In DP4, mutations in signaling pathways, such as CBL, KRAS, NRAS, JAK2, and CSF3R evolved at the late stage.

Clinical characteristics of the defined genomic groups

In DP1, which was enriched with JAK2 mutation, most PV patients characterized by erythrocytosis and/or panmyelosis were included. Conversely, DP5, enriched with CALR mutations, predominantly included ET patients presenting with isolated thrombocytosis. In DP2, enriched with TP53 mutations and complex karyotypes, most patients were diagnosed with MDS, exhibiting significant thrombocytopenia and increased blasts. However, the DP9 group predominantly included patients diagnosed with MDS, who were reclassified as AML according to updated WHO classification criteria (3) due to the presence of NPM1 mutations. In DP7, enriched with SETBP1 and ASXL1 mutations, patients typically present with hypercellular BM and may exhibit neutrophilia and/or monocytosis. Diagnoses observed in this group include MDS, CMML, PMF, and other MPNs, such as CNL and JMML. DP4, enriched with TET2 and SRSF2 mutations prevalent in MDS and CMML, is characterized by relatively older age and monocytosis. DP3 and DP6 predominantly included patients diagnosed with MDS, irrespective of blast percentage. Patients in DP6 typically presented with hypercellular BM and some cases of PMF were also observed. In DP10, patients enriched with SF3B1 mutations typically exhibited older age and ring sideroblasts. Many cases were diagnosed with MDS-LB-RS, and all cases of MDS/MPN-RS-T were assigned to DP10. Additionally, other cases of MDS and PMF were included in DP10, particularly those enriched with DDX41 mutations. Most patients in group DP8, which was enriched with chromosome 1q abnormalities, were diagnosed with MDS and exhibited increased mean corpuscular volume. Lastly, group DP0, which showed no genetic abnormalities, mostly consisted of patients diagnosed with AA and hypoplastic MDS. These patients were characterized by a younger age and hypocellular BM.

Survival analysis according to newly defined groups and stem cell transplantations

The 5-year survival rate (5-ysr) varied markedly across genomic groups. DP1 and DP5 showed the highest 5-ysr at 98.0% and 95.4%, respectively. Conversely, DP2 and DP9 had the lowest 5-ysr at 30%. DP7 had a 44.9% 5-ysr, closely followed by DP4 at 46.1%. Moderate 5-ysrs were observed in DP3 and DP6 at 55.8% and 68.6%, respectively, while DP10 and DP8 had higher 5-ysrs of 73.4% and 78.8%, respectively. In DP10, DDX41 and SF3B1 showed indifferent survival outcomes (P = 0.3832). Finally, DP0 showed a high 5-ysr of 80.6%. Thereafter, we stratified the 10 DP groups into six distinct risk categories, based on survival probabilities (Fig. 3) and hazard ratios (sFigure 9). The Very favorable MN, comprising DP1 and DP5, showed an estimated survival of 15 years (95% CI: 10.9–12.3). In contrast, DP2, DP7, and DP9, categorized as Very adverse MN demonstrated a notably lower median survival of 1.7 years (95% CI: 1.3–9.3). DP4, classified as Adverse MN had a median survival of 4.1 years (95% CI: 3.0–9.0). The Intermediate MN, encompassing DP3 and DP6, revealed a median survival of 7.7 years (95% CI: 4.9–10.7). Favorable MN, consisting of DP10 and DP8, exhibited a median survival of 14.6 years (95% CI: 9.8–17.7). Additionally, DP0, lacking specific genomic abnormalities, exhibited an estimated survival of 12.4 years (95% CI: 11.2–13.7).

We also identified that approximately half of the Very adverse MN cases received hematopoietic stem cell transplantation (HSCT; 42.4% in DP2, 52.4% in DP9, and 47.5% in DP7), while the Very favorable MN had a limited number of transplanted patients (5% in DP1 and 13.5% in DP5). When analyzing patients after separating them according to whether they received HSCT, our defined risk categories continued to show distinctive prognostic impact, specifically in patients who did not undergo transplantation (sFigure 9). Meanwhile, in the Very adverse MN group, the impact of transplantation on survival was pronounced, resulting in improved survival outcomes (P < 0.001, HR = 3.1 in DP2; P = 0.03, HR = 4.1 in DP9; P = 0.01, HR = 3.9 in DP7).

We further assessed the risk categories alongside clinicohematological features using the Cox Proportional Hazards (PH) model. The results revealed that the Very adverse MN had the most significant impact on prognosis, exhibiting the highest hazard ratio of 3.65 (95% CI: 2.16–6.16). Additionally, we found that age, sex, blast percentage, cellularity, and the number of mutations all influenced overall survival, although to a lesser degree than the risk categories (Table 1).

Table 1

Univariate and multivariate analysis of clinical and genetic factors for overall survival.
Variables		Univariate
Variables		P	HR	95% CI	P	HR	95% CI
Genomic risk groups	Very favorable MN [DP1 and DP5 respect to DP0]	< 0.01	0.20	0.10–0.38	0.24	0.58	0.24–1.42
	Very adverse MN [DP2, DP7 and DP9 respect to DP0]	< 0.01	6.85	4.46–1053	< 0.01	3.65	2.16–6.16
	Adverse MN [DP4 respect to DP0]	< 0.01	3.36	1.98–5.72	0.85	1.07	0.55–2.09
	Intermediate MN [DP3 and DP6 respect to DP0]	< 0.01	2.41	1.54–3.78	0.18	1.45	0.84–2.50
	Favorable MN [DP8 and DP10 respect to DP0]	0.09	1.51	0.94–2.43	0.27	0.73	0.42–1.28
Clinical subtypes	MDS [respect to AA]	< 0.01	3.78	1.78–8.02	0.22	1.79	0.71–4.55
	MDSMPN [respect to AA]	< 0.01	3.70	1.52–9.01	0.81	1.15	0.37–3.58
	MPN [respect to AA]	0.04	0.43	0.15–0.98	0.21	0.46	0.13–1.53
Age		< 0.01	1.04	1.03–1.05	< 0.01	1.03	1.02–1.04
Sex [respect to Female]		< 0.01	1.77	1.38–2.26	< 0.01	1.55	1.17–2.06
Hemoglobin		< 0.01	0.76	0.73–0.80	< 0.01	0.81	0.76–0.86
Platelet		< 0.01	1.00	0.99-1.00	0.16	1.00	1.00–1.00
WBC Count		0.80	1.00	0.98–1.01
Neutrophil (%)		< 0.01	0.97	0.97-1.00	0.01	0.99	0.98-1.00
Monocyte (%)		< 0.01	1.02	1.01–1.03	0.01	0.98	0.97-1.00
PB blast (%)		< 0.01	1.09	1.05–1.12	0.55	1.02	0.96–1.07
BM blast (%)		< 0.01	1.09	1.07–1.10	< 0.01	1.04	1.02–1.06
BM Cellularity (%)		< 0.01	1.01	1.00-1.01	0.01	1.01	1.00-1.01
BM Fibrosis grade		< 0.01	0.77	0.67–0.89	0.55	1.07	0.86–1.32
Dyserythropoiesis		< 0.01	2.30	1.79–2.96	0.31	0.83	0.58–1.19
Dysgranulopoiesis		< 0.01	2.52	1.95–3.26	0.67	0.92	0.62–1.36
Megakaryocyte dysplasia		< 0.01	2.08	1.62–2.66	0.31	0.84	0.60–1.18
Ring sideroblasts		0.05	1.01	1.00-1.02
Transplantation		0.02	1.34	1.05–1.70	< 0.01	0.63	0.47–0.85
Chromosomal abnormality number		0.22	0.92	0.81–1.05
Mutation number		< 0.01	1.30	1.20–1.39	< 0.01	1.14	1.04–1.24
HR hazard ratio, CI confidence interval, MN myeloid neoplasm, MDS myelodysplastic neoplasm, MPN myeloproliferative neoplasm, WBC white blood cell, PB peripheral blood, BM bone marrow
Variables assigned to be significant prognostic factors in univariate analysis were included in the multivariate analysis.

Independent validation of the genomic groups

We categorized patients from a prospectively collected independent validation cohort (Fig. 3) and assigned 143 (95.3%) patients to each category as follows: DP1 (n = 58), DP0 (n = 17), DP10 (n = 14), DP3 (n = 12), DP2 (n = 11), DP5 (n = 11), DP4 (n = 5), DP6 (n = 5), DP8 (n = 5), DP7 (n = 3), and DP9 (n = 2). The Very adverse MN showed a significantly higher death rate compared to the Very favorable MN (HR 15.63) and Favorable MN (HR 14.64) (sFigure 10). The Cox PH model was unavailable in the validation cohort due to the limited number of patients.

This study represents a groundbreaking exploration of the genomic landscape of MNs within a Korean cohort, analyzing an extensive dataset from 1,585 patients. Moving beyond the scope of previous research, which targeted specific MN categories (2, 4, 5, 7, 9), our study encompassed a broad spectrum of the disease, unveiling a comprehensive understanding of its genomic diversity. This approach not only broadens the existing knowledge base but also sets a new benchmark for future genomic analyses of MNs.

We elucidated the genomic association in MNs and identified 11 genomic subgroups, ranging from DP1 to DP10, each exhibiting unique genomic characteristics, alongside a distinct DP0 lacking these characteristics. Certain subgroups, notably DP1 and DP5, are distinguished by their unique mutational profiles ─ JAK2 in DP1 and CALR in DP5 ─ and associated clinical and laboratory features of MPNs, leading to a very favorable prognosis.

Conversely, other subgroups, such as DP2, are not confined to a specific disease entity, encompassing a range of conditions, including MDS irrespective of blast percentage, MPN, and MDS/MPN, and even AA. These are characterized by the presence of TP53 mutations and/or complex karyotypes, associated with a very adverse prognosis. We further clarified the biological and prognostic importance of NPM1 mutations, now classified as a distinct AML entity, applicable regardless of blast percentage in both MDS and AML (3, 8). Although limited to 21 cases, individuals classified within DP9 stood out from other subgroups due to their very adverse outcomes, indicating the need for a distinct therapeutic strategy for MDS with NPM1 (14, 15). Additionally, the discovery of subgroup DP9, spanning a variety of MNs beyond MDS to include MDS/MPN, validates the recent 5th WHO classification and underscores the need for its expanded use in disease classification. Subgroup DP7, distinguished by SETBP1 mutations, has been associated with leukemic evolution in progenitors harboring ASXL1, NRAS, and CSF3R mutations via specific pathways (16–18), indicating a very adverse prognosis. This subgroup appears across various MNs, including MDS, CMML, CNL, and JMML. Given its association with high-risk, HSCT is recommended, irrespective of phenotype presentation. Beyond delineating three subgroups based on genomic profiles and their association with a very adverse prognosis, our analysis suggests that patients within these groups may benefit from HSCT.

DP4, characterized by TET2 and/or SRSF2 mutations, included a significant number of patients (n = 93) across MDS, MPN, and MDS/MPN, exhibiting an adverse prognosis. While the prognosis for DP4 was better than that for subgroups DP2, DP9, and DP7, it was poorer compared to other subgroups. We also identified two subgroups: DP3, characterized by mutations in U2AF1, + 8, RUNX1, and BCOR; and DP6, marked by mutations in CBL, ETV6, and KMT2D. The majority were classified within MDS, with a smaller proportion in MPN and even fewer in MDS/MPN, all of which exhibited an intermediate prognosis.

DP10 was characterized by various mutations associated with favorable prognoses, including SF3B1, DDX41, isolated 5q-, MPL or –Y mutations. Our findings provide additional insight that patients with DDX41 mutations share a similar prognostic status with those harboring SF3B1 mutations or isolated 5q-. This suggests a potential need to re-evaluate current risk stratification systems to incorporate DDX41 status (19, 20). In addition, + 1/1q + and der(1;7) represent common chromosomal abnormalities in MNs (21, 22), with der(1;7) being the most prevalent, accounting for 1.5–6% of patients with MDS and AML. While der(1;7) leads to both 7q- and 1q+, a prior study suggested categorizing der(1;7) as a distinct entity due to its comparatively better outcomes than those observed in cases with only − 7/7q- (21). Our study builds upon these findings, offering additional evidence to justify classifying der(1;7) as a subgroup with a favorable prognosis.

The last group, DP0, identified by the absence of specific genomic features, predominantly includes cases of AA and hypoplastic MDS. Distinguishing between hypoplastic MDS and AA presents a challenge due to the scarcely observable cells in the BM. However, our results underscore the distinct biological underpinnings of hypoplastic MDS, affirming its similarity to AA. This insight helps bridge the conceptual gap between malignancies and immune-mediated disorders (6, 23).

This study has several limitations. AML cases were excluded during the study design due to the prevalence of gene fusions and well-defined mutations in many cases. However, including myelodysplasia-related AML may have offered valuable insights into the connection between genomic risk groups and increased blast counts. Furthermore, the validation cohort, derived from the same institution, was relatively small. Consequently, there is a need for multicenter, large-scale studies to further validate and expand upon our findings.

In conclusion, our study aimed to elucidate the intricate genomic landscape underlying cytogenetic and molecular genetic abnormalities across a spectrum of overlapping MNs. We successfully identified subgroups that transcend specific disease diagnoses, instead grouping them by overarching disease categories. These subgroups not only highlight genetic events pivotal in disease development and prognosis but also inform the selection of patients who could benefit from HSCT. Our study addresses the prognostic needs of relatively uncommon disease categories, which are frequently overlooked in existing prognostication systems. Our results provide valuable insights into these lesser-studied MN categories, laying the foundation for the development of personalized therapeutic strategies.

Acknowledgements: This research was supported by a grant from the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (RS-2023-00244625).

Author contributions: J.M.L. and G.L. conceptualized and directed the project, collected and analyzed data, interpreted findings, and drafted the manuscript. A.A. and J.J. provided assistance in interpreting genomic data. Y.J.K., S.P., and D.K. contributed to the interpretation of clinical data for myelodysplastic neoplasms. S.E.L. and T.Y.K. focused on the interpretation of clinical data for myeloproliferative neoplasms, while S.S.P. specialized in analyzing data for aplastic anemia. B.C., N.G.C., J.W.L., J.W.Y., and S.J. specifically assisted in interpreting clinical data for pediatric patients. Y.K. critically reviewed and revised the manuscript, providing essential feedback. M.K. initiated and designed the study, overseeing all aspects of its execution. All authors engaged in discussing the results and contributed to the refinement of the final manuscript.

Data availability statement: The variant data supporting this study's findings are included within the supplementary table. The raw sequencing data for the validation cohort will be openly available in NCBI SRA repository. For the data not available in the repository are available on the reasonable request from the first author J.M.L. and corresponding author, M.K.

Tran TB, Siddon AJ. Molecular findings in myeloid neoplasms. Int J Lab Hematol. 2023.
Duncavage EJ, Bagg A, Hasserjian RP, DiNardo CD, Godley LA, Iacobucci I, et al. Genomic profiling for clinical decision making in myeloid neoplasms and acute leukemia. Blood. 2022; 140(21): 2228-2247.
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): 1703-1719.
Grinfeld J, Nangalia J, Baxter EJ, Wedge DC, Angelopoulos N, Cantrill R, et al. Classification and personalized prognosis in myeloproliferative neoplasms. N Engl J Med. 2018; 379(15): 1416-1430.
Kewan T, Durmaz A, Bahaj W, Gurnari C, Awada H, Ogbue O, et al. Molecular patterns identify distinct subclasses of myeloid neoplasia. Nat Commun. 2023; 14(1): 3136.
Bersanelli M, Travaglino E, Meggendorfer M, Matteuzzi T, Sala C, Mosca E, et al. Classification and personalized prognostic assessment on the basis of clinical and genomic features in myelodysplastic syndromes. J Clin Oncol. 2021; 39(11): 1223.
Palomo L, Meggendorfer M, Hutter S, Twardziok S, Ademà V, Fuhrmann I, et al. Molecular landscape and clonal architecture of adult myelodysplastic/myeloproliferative neoplasms. Blood. 2020; 136(16): 1851-1862.
Arber DA, Orazi A, Hasserjian RP, Borowitz MJ, Calvo KR, Kvasnicka H-M, et al. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022; 140(11): 1200-1228.
Patwardhan PP, Aarabi M, Aggarwal N. Genomics of myelodysplastic/myeloproliferative neoplasm. Semin Diagn Pathol. 2023; 40(3):195-201.
Lee J-M, Kim Y-J, Park S-S, Han E, Kim M, Kim Y. Simultaneous monitoring of mutation and chimerism using next-generation sequencing in myelodysplastic syndrome. J Clin Med. 2019; 8(12): 2077.
Lee J-M, Lee H, Eom K-S, Lee S-E, Kim M, Kim Y. Impact of integrated genetic information on diagnosis and prognostication for myeloproliferative neoplasms in the next-generation sequencing era. J Clin Med. 2021; 10(5): 1033.
Li MM, Cottrell CE, Pullambhatla M, Roy S, Temple-Smolkin RL, Turner SA, et al. Assessments of Somatic Variant Classification Using the Association for Molecular Pathology/American Society of Clinical Oncology/College of American Pathologists Guidelines: A Report from the Association for Molecular Pathology. J Mol Diagn. 2023; 25(2): 69-86.
Horak P, Griffith M, Danos AM, Pitel BA, Madhavan S, Liu X, et al. Standards for the classification of pathogenicity of somatic variants in cancer (oncogenicity): joint recommendations of Clinical Genome Resource (ClinGen), Cancer Genomics Consortium (CGC), and Variant Interpretation for Cancer Consortium (VICC). Genet Med. 2022; 24(5): 986-998.
Montalban-Bravo G, Kanagal-Shamanna R, Sasaki K, Patel K, Ganan-Gomez I, Jabbour E, et al. NPM1 mutations define a specific subgroup of MDS and MDS/MPN patients with favorable outcomes with intensive chemotherapy. Blood Adv. 2019; 3(6): 922-933.
Patel SS, Ho C, Ptashkin RN, Sadigh S, Bagg A, Geyer JT, et al. Clinicopathologic and genetic characterization of nonacute NPM1-mutated myeloid neoplasms. Blood Adv. 2019; 3(9): 1540-1545.
Makishima H, Yoshida K, Nguyen N, Przychodzen B, Sanada M, Okuno Y, et al. Somatic SETBP1 mutations in myeloid malignancies. Nat Genet. 2013; 45(8): 942-946.
Carratt SA, Braun TP, Schonrock Z, Smith BM, Coleman DJ, Kong G, et al. Oncogenic SETBP1 Mutations Combine with Activating Mutations in CSF3R to Produce a Highly Proliferative, Lethal Leukemia through Aberrant Myc Signaling. Blood. 2020; 136: 51-52.
Tanaka A, Nishimura K, Saika W, Kon A, Koike Y, Tatsumi H, et al. SETBP1 is dispensable for normal and malignant hematopoiesis. Leukemia. 2023; 37(9): 1802-1811.
Bataller A, Loghavi S, Gerstein Y, Bazinet A, Sasaki K, Chien KS, et al. Characteristics and clinical outcomes of patients with myeloid malignancies and DDX41 variants. Am J Hematol. 2023; 98(11): 1780-1790.
Makishima H, Bowman TV, Godley LA. DDX41-associated susceptibility to myeloid neoplasms. Blood. 2023; 141(13): 1544-1552.
Sanada M, Uike N, Ohyashiki K, Ozawa K, Lili W, Hangaishi A, et al. Unbalanced translocation der (1; 7)(q10; p10) defines a unique clinicopathological subgroup of myeloid neoplasms. Leukemia. 2007; 21(5): 992-997.
Okuda R, Ochi Y, Chonabayashi K, Hiramoto N, Sanada M, Handa H, et al. Molecular Landscape of Myeloid Neoplasms with Der (1; 7)(q10; p10). Blood. 2022; 140(Supplement 1): 6902-6904.
Calabretto G, Attardi E, Teramo A, Trimarco V, Carraro S, Mossuto S, et al. Hypocellular myelodysplastic syndromes (h-MDS): from clinical description to immunological characterization in the Italian multi-center experience. Leukemia. 2022 Jul; 36(7): 1947-1950.

There is NO conflict of interest to disclose.

LeukemiaMDSMPNsupplementary.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Classification and prognostic stratification based on genomic features in myelodysplastic neoplasms, myeloproliferative neoplasms, and their overlapping conditions

Status:

Version 1

Abstract

Figures

INTRODUCTION

METHODS

Study cohort

Genomic data

Statistics

RESULTS

Cohort characteristics

Genomic associations

Characteristics of the genomic groups

Clinical characteristics of the defined genomic groups

Survival analysis according to newly defined groups and stem cell transplantations

Independent validation of the genomic groups

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1