The NUP98 gene, located on chromosome 11p15, encodes a protein of the nucleoporin family of carriers that transport substrates into the nucleus and are responsible for regulating transcription and mitotic progression [1]. It is involved in various chromosomal translocations, and proteins produced by NUP98 fusions aberrantly regulate gene expression, leading to uncontrolled cell proliferation and impaired differentiation, ultimately resulting in oncogenesis [2].

Since the description of the translocation involving chromosome 11p15 in the blast phase of chronic myeloid leukemia in 1982, NUP98 rearrangements (NUP98r) have been reported as fusions with more than 30 partner genes in hematological malignancies, including myeloid and lymphoid neoplasms [3^-6]. NUP98r are largely divided into two categories depending on whether they are fused with homeobox (HOX) genes (e.g., HOXA9, HOXA11, HOXA13, HOXC11, HOXC13, and PRRX1) or non-HOX genes (e.g., NSD1, KDM5A, DDX10, and PSIP1) [7, 8]. These partners share similar leukemic phenotypes, often resulting in poor outcomes, particularly in pediatric patients [9]. The recent expansion of RNA-based whole-transcriptome sequencing has led to increased detection of NUP98r, even in cases with cytogenetically cryptic fusions (e.g., NUP98::NSD1 and NUP98::KDM5A), underscoring their clinical significance.

Considering their distinctive clinicopathologic characteristics, NUP98r were recently included as recurrent genetic abnormalities defining AML in the 2022 WHO classification as well as the International Consensus Classification (ICC) [10, 11]. However, until recently, the 11p15 translocation identified through chromosomal analysis was often overlooked, and thorough evaluation for NUP98r has not been frequently conducted. Moreover, data regarding the laboratory and clinical characteristics associated with rare NUP98r partner genes are insufficient. Therefore, we investigated the clinical and genetic characteristics of patients with myeloid neoplasms who showed NUP98r among those identified as having 11p15 translocation through chromosomal analysis.

Patients

We retrospectively reviewed the results of bone marrow (BM) chromosomal analyses conducted between 2011 and 2023 and identified 15 patients showing 11p15 translocation. Subsequently, NUP98r were evaluated in these patients using FISH and/or reverse transcription (RT-)PCR. All patients were diagnosed or revised according to the 2016 WHO classification of myeloid neoplasms [12], and patient clinical and laboratory information was obtained from electronic medical records. For survival analysis, patients with AML aged ≥18 yrs who were diagnosed between July 2018 and December 2022 were used as control groups, including patients with AML with a normal karyotype (NK) (N=26), AML with myelodysplasia-related (MR) (N=58), and AML with mutated TP53 (N=33), with the latter two groups reclassified according to the ICC [11, 13]. This study was approved by the Institutional Review Board of Samsung Medical Center, Seoul, Korea (approval No.: 2023-08-152), and informed consent was waived.

Cytogenetic studies

Conventional chromosomal analysis was performed on heparinized BM aspirates using a standard G-banding technique following short-term culturing without mitogen, and at least 20 metaphases were analyzed for karyotyping. FISH was performed on BM aspirates using a NUP98 break-apart FISH probe (MetaSystems, Altlussheim, Germany) according to the manufacturer’s instructions. Briefly, after spreading cells onto slides, the cells and probe were denatured at 73±1°C for 5 mins and incubated at 37°C overnight. The cells were incubated in 50% formamide/ 2XSSC solution and counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI). FISH signals were analyzed using DAPI, FITC, and rhodamine triple-band pass filter sets, and at least 200 interphase cells were analyzed.

Molecular genetic studies

RNA was extracted from BM aspirates using a QIAamp RNA Blood Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using a HiSenScript RH(–) RT PreMix Kit (Intronbio, Seongnam, Korea) according to the manufacturer’s instructions. For each chromosomal abnormality observed in individual patients, candidate exons of partner genes were selected based on previous reports [8, 14^-19], and specific primers were designed. Selected NUP98 partner genes included DDX10 (11q22), HOXA9/HOXA11/HOXA13 (7p15), PSIP1 (9p22), PRRX1 (1q24), and HOXC11/HOXC13/RARG (12q13). In cases where the partner gene was undisclosed or where insufficient sample was available, molecular analysis was not performed. After confirmation of PCR amplification, the breakpoint of each fusion transcript was verified through Sanger sequencing, and data were analyzed using Sequencher (v.5.4; Gene Codes, Ann Arbor, MI, USA).

Genomic DNA was extracted from BM aspirates using a DNA extraction kit (Promega, Madison, WI, USA). Fragment analysis for FLT3-internal tandem duplication (ITD) was performed on an ABI Prism 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) using in-house primers, and the data were analyzed using GeneMapper Software 4.0 (Applied Biosystems). Targeted next-generation sequencing (NGS) of 49 myeloid neoplasm-related genes was conducted using IDT xGen pre-designed/custom probes (Integrated DNA Technologies, Coralville, USA) on a NextSeq 550Dx instrument (Illumina, San Diego, CA, USA). Reads were aligned to the GRCh37/hg19 human reference genome using Burrows–Wheeler Aligner-MEM, and variant calling, annotation, and analysis were carried out using the in-house pipeline.

Statistical analysis

Statistical analyses were performed using the Statistical Software Package for Social Sciences (IBM SPSS Statistics v.27; IBM, Armonk, NY, USA). Relapse-free survival was determined from the time of morphologic complete remission of AML to disease relapse, death from any cause, or last follow-up. Overall survival (OS) was determined from the time of initial diagnosis to death from any cause or last follow-up. Survival analysis was performed using Kaplan–Meier plots, and differences in survival were compared using the log-rank test. Values are expressed as the median with interquartile range (IQR). Statistical significance set at P<0.05.

Characteristics of patients with NUP98r

Among the 15 patients with 11p15 translocation, 11 had NUP98r, and their clinicopathological characteristics are summarized in Table 1. The median patient age was 44 yrs (range, 4–77 yrs), and 64% of patients were female. The median white blood cell count, Hb level, and platelet count were 6.68×10⁹/L (IQR, 2.93–88.79 ×10⁹/L), 85 g/L (IQR, 71–98 g/L), and 40×10⁹/L (IQR, 30–83 ×10⁹/L), respectively. All patients exhibited anemia, and nine (82%) showed thrombocytopenia.

Table 1 Clinicopathological characteristics of patients with non-cryptic NUP98 rearrangement

Case No.	Age (yrs)	Sex	WBC(×109/L)	Hb(g/L)	PLT(×109/L)	Diagnosis	Etiology	FAB classification	Immunophenotype	Dysplasia (lineage)	CR	Allogeneic PBSCT	RFS	Outcome(OS)
1	4	M	5.75	68	34	AML	de novo	M5	CD11c+, CD13+, CD33+, CD34+dim, CD64+dim, CD117+, cMPO+, HLA-DR+	+ (E/M)	+	+	+(12 mo)	D(23 mo)
2	45	M	95.13	84	40	AML	de novo	M5	CD13+, CD33+, CD64+, CD117+dim, cMPO+, HLA-DR+	+ (M)	+	+	–(41 mo)	A(42 mo)
3	50	F	2.47	100	56	MDS	therapy-related (OC)	NA	NA	–	NA	–	–	D(5 mo)
4	37	F	214.37	74	38	CMML-2; AML (28 mo)*	de novo	M5	CD11c+dim, CD13+, CD33+, CD34+	+ (E/G/M)	+	+	+(12 d)	D(32 mo)
5	77	F	15.68	61	4	AML	therapy-related (OC)	M5	CD13+, CD33+, CD64+, CD66c+, CD117+, cMPO+	+ (E)	NA	–	–	D(4 mo)
6	58	F	2.54	98	96	AML	de novo	M1	CD13+, CD33+, CD34+, CD117+, cMPO+, HLA-DR+	–	+	+	–(106 mo)	A(107 mo)
7	55	F	3.31	98	154	MDS-EB2; AML (4 mo)*	de novo	M5	CD4+dim, CD11c+dim, CD13+, CD33+, CD64+, CD117+, cMPO+, HLA-DR+, nTdT+dim	+ (E/G)	+	–	–(6 mo)	D(12 mo)
8	14	M	82.44	121	26	AML	de novo	M1	CD7+, CD13+, CD33+, CD34+, cMPO+	–	NA	–	–	D(3 d)
9	60	F	1.29	67	5	AML	therapy-related (OC)	M1	CD7+, CD13+, CD33+, CD34+dim, CD64+, CD117+, cMPO+, HLA-DR+	–	NA	–	–	D(30 d)
10	29	F	233.81	98	70	AML	de novo	M4	CD2+dim, CD11c+, CD13+, CD19+dim, CD33+, CD34+, CD64+dim, cCD79a+dim, CD117+, cMPO+, HLA-DR+, nTdT+dim	+ (M)	+	+	+(4 mo)	D(8 mo)
11	58	M	6.68	85	573	AML	de novo	M4	CD13+, CD14–/+, CD33+, CD34+/+dim, CD64–/+, CD117+/–, cMPO+, HLA-DR+	+ (G/M)	+	+	+(5 mo)	D(13 mo)

*Time to leukemic transformation.

Abbreviations: WBC, white blood cell; PLT, platelet; FAB, French-American-British; CR, complete remission; PBSCT, peripheral blood stem cell transplantation; RFS, relapse-free survival; OS, overall survival; M, male; F, female; aE, erythroid; M, megakaryocytic; G, granulocytic; D, dead; A, alive; mo, months; d, days; NA, not available; MDS, myeloid dysplastic neoplasm; CMML, chronic myelomonocytic leukemia; MDS-EB2, MDS with excess blasts; OC, ovarian cancer; BC, breast cancer.

At initial diagnosis, eight (73%) patients had AML, one (9%) had chronic myelomonocytic leukemia (CMML)-2, and two (18%) had myelodysplastic syndrome (MDS); one with MDS-excess blasts (EB)-2, and the other diagnosed as having MDS based on cytogenetics despite the absence of dysplasia. The frequency of non-cryptic NUP98r was 0.81% in AML, 0.40% in MDS, and 1.67% in CMML. Three (27%) patients developed the diseases as a result of prior chemotherapy for ovarian or breast cancer. Among non-AML patients, those with CMML-2 and MDS-EB-2 progressed to AML after 28 months and 4 months, respectively. Ultimately, 10 (91%) patients were diagnosed as having AML. In the French-American-British classification of AML, M5 (acute monoblastic/monocytic leukemia) was the most common subtype (N=5), followed by M1 (AML without maturation) (N=3) and M4 (acute myelomonocytic leukemia) (N=2), and half of the patients with AML exhibited negative or dim CD34 expression. Dysplasia in at least one lineage was observed in seven (63.6%) patients.

Genetic characteristics of NUP98r

NUP98::DDX10 was the most common (N=3), followed by NUP98::HOXA9 (N=2), NUP98::PSIP1 (N=2), NUP98::PRRX1 (N=1), and NUP98::HOXC11 (N=1). Representative FISH, RT-PCR, and sequencing results of patients with NUP98r are presented in Fig. 1, and detailed cytogenetic and molecular genetic characteristics are provided in Table 2.

Figure 1. Representative FISH and molecular analysis results for various NUP98r. (A) FISH analysis using a NUP98 break-apart probe showed negative (two fusions) and positive (one red, one green, and one fusion) signals in case No. 11. (B) Agarose gel electrophoresis of fusion transcripts between NUP98 and each partner gene (lane 1, internal control using GAPDH [amplicon size, 131 bp], lane 2, fusion transcripts using primer pairs of NUP98 and partner genes). (C) Confirmation of the breakpoint of each fusion transcript via Sanger sequencing.
Abbreviation: NUP98r, NUP98 rearrangement.

Table 2 Cytogenetic and molecular genetic characteristics of patients with non-cryptic NUP98 rearrangement

Case No.	Dx	Karyotype	NUP98 FISH (%)	Candidate partner gene	Identified partner gene	Breakpoint	FLT3-ITD	Targeted NGS
1	AML	46,XY,inv(11)(p15q22)[19]/46,XY[1]	NA	DDX10	DDX10	NUP98 exon 14::DDX10 exon 7	ND	NA
2	AML	46,XY,inv(11)(p15q22)[12]/46,idem,del(5)(q22q33)[5]/46,XY[3]	1F1R1G (59.5%)	DDX10	DDX10	NUP98 exon 14::DDX10 exon 7	ND	KRAS c.35G>T, p.G12V (33.8%); WT1 c.1356_1358delinsTTC, p.Q452_C453delinsHS (30.0%)
3	MDS	46,XX,inv(11)(p15q22)[14]/46,XX[6]	1F1R1G (58.0%)	DDX10	DDX10	NUP98 exon 14::DDX10 exon 7	NA	NA
4	CMML-2; AML	46,XX,t(7;11)(p15;p15)[19]/46,XX[1]	1F1R1G (91.5%)	HOXA9, HOXA11, HOXA13	HOXA9	NUP98 exon 12::HOXA9 exon 1	ND	NA
5	AML	46,XX,t(7;11)(p15;p15)[20]	1F1R1G (92.5%)	HOXA9, HOXA11, HOXA13	HOXA9	NUP98 exon 11::HOXA9 exon 1	ND	ND
6	AML	49,XX,+4,t(7;11)(p15;p15),+8,+15[20]	1F1R1G (83.5%)	HOXA9, HOXA11, HOXA13	NA*	NA	ND	NA
7	MDS-EB2; AML	46,XX,t(9;11)(p22;p15)[20]	1F1R1G (78.5%)	PSIP1	PSIP1	NUP98 exon 9::PSIP1 exon 7	NA	NA
8	AML	46,XY,t(9;11)(p22;p15)[19]/46,XY[1]	1F1R1G (94.0%)	PSIP1	PSIP1	NUP98 exon 9::PSIP1 exon 5	ND	PHF6 c.902_903insGGGATAA, p.Y301* (90.0%)
9	AML	46,XX,t(1;11)(q24;p15)[19]/46,XX[1]	1F1R1G (82.0%)	PRRX1	PRRX1	NUP98 exon 11::PRRX1 exon 2	ND	NA
10	AML	46,XX,t(11;12)(p15;q13)[5]/46,XX[15]	1F1R1G (80.5%)	HOXC13, HOXC11, RARG	HOXC11	Inconclusive†	D	NA
11	AML	47,XY,t(11;16)(p15;q23),+21[17]/46,XY[3]	1F1R1G (43.5%)	Unknown	NA‡	NA	ND	ASXL1 c.1934dup, G646Wfs*12 (31.9%); DNMT3A c.917G>A, p.W306* (41.8%); DNMT3A c.2111T>G, p.V704G (39.6%); IDH2 c.419G>A, p.R140Q (38.4%); STAG2 c.1416+2T>C, p.? (81.2%)

*RT-PCR was not available because of the lack of a sample.

^†Multiple bands were observed on RT-PCR electrophoresis, and sequencing yielded inconclusive results.

^‡Targeted RT-PCR could not be performed for candidate partner genes because the partner gene of the NUP98 rearrangement at the 16q23 locus was unknown.

Abbreviations: Dx, diagnosis; FLT3-ITD, FLT3-internal tandem duplication; NGS, next-generation sequencing; MDS, myeloid dysplastic neoplasm; CMML, chronic myelomonocytic leukemia; MDS-EB2, MDS with excess blasts; NA, not available; ND, not detected, D, detected.

All three patients with inv(11)(p15q22) harbored a fusion between NUP98 exon 14 and DDX10 exon 7 (Fig. 1). Among them, two were diagnosed as having AML, and one as having MDS. t(7;11)(p15;p15) was detected in three patients, with two diagnosed as having AML and one as having CMML-2. In the 7p15 region, known partner genes of NUP98 included HOXA9, HOXA11, and HOXA13, and a fusion between NUP98 exon 12 or 11 and HOXA9 exon 1 was identified in two patients. However, in one patient, the partner gene could not be identified because of insufficient samples. Two patients with t(9;11)(p22;p15) were diagnosed as having MDS-EB2 and AML, respectively. They had a fusion between NUP98 exon 9 and PSIP1 exon 5 or 7. Lastly, in one patient each with AML with t(1;11)(q24;p15) and t(11;12)(p15;q13), NUP98::PRRX1 and NUP98::HOXC11 were detected, respectively. In a patient with t(11;16)(p15;q23), NUP98r were confirmed via FISH. However, there was no previously known partner gene of NUP98 in the 16q23 region, suggesting a potential novel partner gene. Unfortunately, further evaluation could not be conducted because of insufficient sample.

Besides NUP98r, three patients exhibited additional cytogenetic abnormalities, such as 5q deletion, +4, +8, +15, or +21; and FLT3-ITD was detected in only one out of nine patients with AML. Targeted NGS was performed for four patients with AML. A WT1 variant was observed in one patient with NUP98::DDX41, accompanied by a KRAS variant. Additionally, one patient with NUP98::PSIP1 had a PHF6 variant, and the other patient with NUP98r involving 16q23 harbored five variants in ASXL1, DNMT3A, IDH2, and STAG2.

Prognostic impact of NUP98r

During the median follow-up of 12 months (range, 0–107 months), four (36%) patients had relapsed, and nine (82%) had expired. The median OS of patients with NUP98r was 12.0 months (95% confidence interval [CI], 3.4–29.6 months), and the 5-yr OS was 18.2% (95% CI, 5.2%–63.7%) (Fig. 2). When considering only patients with de novo NUP98r, the median OS was 13.0 months (95% CI, 0–28.2 months) and the 5-yr OS was 25.0% (95% CI, 7.5%–83.0%) (Supplemental Data Fig. S1). Notably, one pediatric patient with AML with NUP98::PSIP1 expired only 3 days after the diagnosis because of rapidly evolving disseminated intravascular coagulation (DIC) (i.e., brain infarction and hemorrhage).

Figure 2. Kaplan–Meier plot showing the survival probability of patients with NUP98r and other AML subgroups.
Abbreviations: AML-NK, AML with normal karyotype; AML-MR, AML with myelodysplasia-related; AML-mutated TP53, AML with mutated TP53.

Finally, we compared the survival of patients with NUP98r with that in other AML subgroups (Fig. 2). Patients with AML with mutated TP53 had the poorest survival (P=0.025), and the survival rate of patients with NUP98r was similar to that of patients with AML with MR (P=0.997). Although not statistically significant, the survival rate of patients with NUP98r appeared to be inferior to that of patients with AML with NK (P=0.188).

We identified various rare non-cryptic NUP98r in patients with myeloid neoplasms and demonstrated their poor prognostic impact. NUP98r are most frequently reported in pediatric AML, occurring in 4%–7% of patients, with a slightly higher incidence in male patients [9, 20, 21]. Although research on NUP98r in adult AML is limited, the reported frequency of NUP98r is 2.5% [1]. We found a significantly lower frequency of NUP98r, which appeared more frequently in adults, suggesting that it is a characteristic of non-cryptic NUP98r.

The disease phenotypes associated with NUP98r encompass various hematologic malignancies, including AML, MDS, chronic myeloid leukemia (CML), CMML, mixed-phenotype acute leukemia, and T-cell acute lymphoblastic leukemia [8]. Regarding the phenotypes of AML, in a recent large-scale pediatric AML study, NUP98::KDM5A was found to be associated with erythroid (M6) and megakaryocytic (M7) phenotypes, but no consistent immunophenotypic correlations were observed [9]. In contrast, Hollink, et al. [7] found an association between NUP98::NSD1 and monocytic differentiation (M4 and M5) in adult patients with AML but not in pediatric patients. In our study, patients predominantly exhibited monocytic differentiation. Moreover, approximately half of our patients had accompanying leukocytosis. Notably, previous studies have reported leukocytosis in 78%–85% of NUP98r patients [1, 7], suggesting it is a characteristic feature of NUP98r. NUP98r were also frequently reported in therapy-related myeloid neoplasms, which were observed in 27% of patients in our study.

The most common partner genes involved in NUP98r are NSD1 and KDM5A, collectively accounting for 6.2% of pediatric patients with AML [9]. As these genes are located at the distal ends of chromosome 5q and chromosome 12p, respectively, they are challenging to detect using conventional cytogenetics [8, 22]. Conversely, NUP98r other than NUP98::NSD1 and NUP98::KDM5A are rare and are mostly identified through conventional cytogenetics, with NUP98::HOXA9 being frequently reported [8, 9]. Therefore, we focused on characterizing these rare and non-cryptic NUP98r, excluding common NUP98::NSD1, NUP98::KDM5A, and other cryptic NUP98r. NUP98::DDX10 was the most common, followed by NUP98::HOXA9 and NUP98::PSIP1.

RNA-sequencing has revealed various NUP98 fusion exon junctions. Bertrums, et al. [9] suggested an association between exon junctions and fusion partners, reporting that breakpoints of NUP98 fusions were located at NUP98 exons 12 and 13 in 85% of patients, with additional breakpoints spanning from exon 11 to exon 29. Notably, they observed an enrichment of breakpoints at NUP98 exon 12 for NUP98::HOX gene fusions. In our study, NUP98::HOX gene fusions (i.e., NUP98::HOXA9 and NUP98::PRRX1) exhibited breakpoints at NUP98 exons 11 and 12, whereas NUP98::non-HOX gene fusions (i.e., NUP98::DDX10 and NUP98::PSIP1) showed breakpoints at NUP98 exons 9 and 14. Interestingly, in pediatric patients with AML with rare NUP98r excluding NUP98::NSD1, breakpoints at NUP98 exon 13 and improved outcome trends have been reported [9]. However, in our study, none of the patients exhibited breakpoints at exon 13 of NUP98.

FLT3-ITD has been reported to be associated with NUP98:: NSD1 in AML, with a frequency of 48%–91%, predicting poor prognosis when observed with NUP98::NSD1 and/or WT1 variants [7^-9, 20]. WT1 variants have also been frequently reported in AML with various NUP98r [8]. In our study, FLT3-ITD and WT1 variants were observed in one out of nine and one out of four patients with AML, respectively. As expected, FLT3-ITD appeared to be uncommon in NUP98r other than NUP98::NSD1.

In the 2022 WHO classification and ICC for myeloid neoplasms, NUP98r has been newly added as a defining genetic abnormality of AML [10, 11]. Specifically, the 2022 WHO classification has conferred significant clinical significance by diagnosing AML regardless of the blast percentage when NUP98r is present [10]. In this regard, all patients in our study could be classified as AML with NUP98r from the initial diagnosis. However, according to the ICC, AML with NUP98r is diagnosed when blasts comprise ≥ 10% [11], and in our study, 10 out of 11 patients met this criterion. Especially in our study, the clinical utility of the criterion of ≥ 10% blasts was supported by the leukemic transformation of both patients with MDS-EB2 and CMML-2. However, there was uncertainty regarding the diagnosis of one patient with low-grade MDS harboring NUP98::DDX10 as having AML according to the 2022 WHO classification, warranting caution in diagnosis. This patient, undergoing treatment for metastatic ovarian cancer, showed no evidence of dysplasia or blast increase and died because of ovarian cancer progression [23].

NUP98r is associated with a dismal prognosis in AML [1, 7^-9, 14, 20, 21]. While most studies were based on high-frequency NUP98r, such as NUP98::NSD1 and NUP98::KDM5A, our study added prognostic significance to rare NUP98r. In two studies of pediatric AML, the 5-yr OS rates were 35% and 48% for patients with NUP98r, respectively, versus 64% and 72% in those without NUP98r [9, 20]. In adult AML with NUP98r, although studies are limited, the median OS has been reported as 13 months [1, 14], which is lower than the median survival of 33 months and the 5-yr OS of 63% in de novo patients with AML aged 40–59 yrs [24]. Similarly, our study demonstrated a poor prognosis in patients with rare NUP98r, with a median OS of 12 months and a 5-yr OS rate of 18.2%. These results are in line with previous findings in patients with AML with MR, who are known to have a poor prognosis [25]. In one patient, the concurrent presence of DIC was associated with an extremely poor prognosis, highlighting the need for caution and further research into its relevance.

Our study has several limitations. We did not comprehensively investigate the characteristics of cryptic NUP98r, and only a few patients were included because of the rarity of non-cryptic NUP98r. Additionally, partner genes in two patients with NUP98r were not identified, and NGS was conducted in only one-third of the patients.

In conclusion, our study identified the clinical and genetic characteristics of patients with myeloid neoplasms harboring rare and non-cryptic NUP98r. Currently, patients with NUP98r are believed to be underdiagnosed. As cancer therapy advances toward personalized treatment tailored to individual genetic profiles, the importance of comprehensive assessment for identifying NUP98r in patients with myeloid neoplasms is emphasized, particularly given its association with poor prognosis.

We are grateful to the Daishin Songchon Foundation for the contribution of a Medical Research Fund to Samsung Medical Center (SMX1230881).

Park MS and Kim BR analyzed the data and drafted the manuscript; Jang JH and Jung CW provided clinical information; Kim HJ provided expertise regarding the laboratory data and supervised the study; Kim HY conceptualized the study and edited the manuscript. All authors have read and approved the final manuscript.

None declared

This research was supported by a grant provided by the National Research Foundation of Korea, funded by the Korean Government (NRF-2022R1F1A1073963).

Supplementary materials can be found via https://doi.org/10.3343/alm.2024.0190