Genetic landscape of pediatric myelodysplastic syndrome in Japan

Abstract

Abstract Introduction Pediatric myelodysplastic syndrome (MDS) shows distinct features compared to adult MDS, but its genetic characteristics remain incompletely understood, particularly in Asian populations. Furthermore, the morphological distinction between MDS and acute myeloid leukemia (AML) can be challenging in some cases. We conducted genomic analysis of Japanese pediatric MDS cases to reveal their genetic landscape and evaluate the utility of genetic testing for accurate diagnosis. Methods We studied 35 Japanese pediatric cases with morphologically suspected MDS: 28 cases collected through the Japan Children's Cancer Group CHM-14 study and 7 cases collected from 3 hospitals. Morphological evaluation was centrally reviewed by designated experts. All cases underwent targeted panel sequencing of the coding regions of 373 driver genes, the intronic regions of 18 genes involved in driver fusions and 1,316 single nucleotide polymorphism sites for copy number analysis. To study clonal evolution, samples serially collected at intervals ranging from 14 to 292 days were also analyzed for five cases. Results Morphological diagnoses were as follows: acute myeloid leukemia with myelodysplasia-related changes (AML-MRC, N = 3), MDS with excess blasts (MDS-EB, N = 24), refractory cytopenia with multilineage dysplasia (RCMD, N = 3), and refractory cytopenia of childhood (RCC, N = 5). Three cases were therapy-related MDS. The mean number of somatic mutations was lower in RCC than in the other subtypes (1.2 vs. 2.3 per case). The most frequent driver mutations were mutations in RAS/MAPK pathway genes (NRAS/KRAS/PTPN11/NF1/CBL) in 13 cases (37%), and UBTF tandem duplication (UBTF-TD) in 6 cases (17%). Cases with UBTF-TD showed hypercellularity and trilineage dysplasia, particularly in megakaryocyte and erythroid lineages. Splicing factor gene variants, which are common in adult MDS, were observed only in 3 cases (9%). Seven cases harbored AML-associated fusion genes in AML: NUP98::NSD1 in 3 cases and NUP98::DDX10, RUNX1::PRDM16, PICALM::MLLT10, BCR::ABL1 in single cases. Additionally, NPM1 mutations were observed in 2 cases. All these cases exhibited morphological features indistinguishable from those of advanced MDS. One-third of cases with RAS/MAPK pathway gene mutations harbored multiple mutations with low variant allele frequencies, suggesting the existence of multiple subclones. For serially collected samples, subclones with driver mutations caused clonal sweep in four cases, even within 2 weeks without treatment, demonstrating the dynamic clonal evolution in pediatric MDS during the disease progression. Pathogenic germline variants were identified in 6 cases (17%): RUNX1 (N = 3), BLM (N = 1), GATA2 (N = 1), and ADH5/ALDH2 (N = 1). SAMD9/9L mutations were not identified in our cohort. Patients with pathogenic germline variants were younger than those without, with the mean ages of 6.3 and 10.9 years, respectively. In comparison with previous reports from Europe and the United States, our cohort showed fewer germline GATA2 mutations but identified aldehyde degradation deficiency syndrome, which is an East Asian-specific genome instability syndrome. The lower frequency of SAMD9/9L mutations likely results from the small number of RCC cases in our cohort. ConclusionsOur data elucidated the genetic characteristics of pediatric MDS in Japan. Consistent with recent reports from European and American cohorts, UBTF-TD was the most common genetic alteration in pediatric advanced MDS, observed in 6 of 27 cases (22%). Among 27 cases with morphologically suspected advanced MDS, nine cases harbored AML-associated driver alterations, including fusion genes and NPM1 mutation. The distribution of germline variants predisposing to myeloid malignancies may differ between East Asian and Western populations. Our results highlight the limitations of morphological evaluation and underscore the utility of genetic testing in the differential diagnosis between MDS and AML as well as appropriate treatment selection.