Phase II trial of imatinib mesylate in patients with PDGFRA/B‐negative hypereosinophilic syndrome

Summary

The role of imatinib in PDGFRA/B‐negative hypereosinophilic syndromes (HES) is controversial because of the heterogeneity of HES and the scarcity of prospective studies. We conducted a phase II clinical trial to evaluate the efficacy of imatinib in PDGFRA/B‐negative HES. Thirty‐two patients were treated with imatinib (100–400 mg daily), and the molecular basis of their response was identified using whole‐exome sequencing (WES) and whole‐transcriptome sequencing (WTS). The haematological response rate was 46.9%, with a complete haematological response (CHR) rate of 18.8%. The median time to response was 1.5 months. Among the six patients who achieved CHR, five maintained it until the 24th cycle of imatinib and one lost response after 20 months. The median progression‐free survival was 4.3 months. WES and WTS were conducted for 11 patients. The number of non‐silent mutations did not differ between responders and non‐responders. Nine differentially expressed genes, including SNORD15A, were downregulated in responders. STAT5B::RARA, PAK2::PIGX, and FIP1L1::CHIC2 fusions were identified in patients with sustained responses, and RNF130::BRAF and WNK1::KDM5A fusions were identified in non‐responders. Imatinib, along with an appropriate biomarker, could be a promising option for PDGFRA/B‐negative HES.

INTRODUCTION

Hypereosinophilic syndromes (HES) encompass a diverse group of disorders characterized by elevated eosinophil levels and subsequent organ damage. ¹ , ² Reactive HES, caused by secondary causes, is common; after excluding this, it is important to identify primary HES with clonality. ³ With advancements in molecular diagnostic techniques, the World Health Organization and International Consensus Classification recently revised the classification systems for eosinophilic diseases. Myeloid/lymphoid neoplasms with eosinophilia and tyrosine kinase gene fusion arise from rearrangements in genes encoding tyrosine kinases such as PDGFRA, PDGFRB, and FGFR1. ⁴ To date, dozens of different fusion genes have been reported. ⁵ Identifying fusion genes is crucial in HES, as tyrosine kinase inhibitors (TKIs) are active in these cases, with imatinib proving particularly effective in patients harbouring PDGFRA or PDGFRB fusion genes. ⁶ , ⁷ , ⁸ , ⁹ In cases of idiopathic HES, where clonality and other primary causes have been excluded, corticosteroids remain the mainstay of therapy and are effective; however, resistance and side effects become problematic with long‐term use. ¹⁰ , ¹¹ Recently, mepolizumab, a biological agent targeting interleukin‐5 (IL‐5), demonstrated efficacy in idiopathic HES. ¹² , ¹³ , ¹⁴

Studies on the efficacy of imatinib in PDGFRA/B‐negative HES have reported response rates ranging from 9% to 60%. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ The heterogeneous patient populations in retrospective studies may have contributed to these inconsistent results, highlighting the need for prospective studies to confirm the role of imatinib in these populations. Furthermore, patients who respond well to imatinib imply the presence of undiscovered imatinib‐sensitive genetic alterations beyond the PDGFRA/B rearrangements. Therefore, we conducted a phase II study to evaluate the efficacy and safety of imatinib in patients with PDGFRA/B‐negative HES. Additionally, we investigated the molecular characteristics of patients using whole‐exome sequencing (WES) and whole‐transcriptome sequencing (WTS).

MATERIALS AND METHODS

Study design and participants

This investigator‐initiated, open‐label, multicentre, single‐arm, prospective phase II study was conducted at three hospitals in South Korea between January 2014 and December 2020. Patients aged 20 years or older with HES were eligible if they tested negative for BCR::ABL1, FGFR1, and PDGFRB rearrangement, as well as for CHIC2 gene (located on chromosome 4q12) deletion by fluorescence in situ hybridization or RT‐PCR on bone marrow (BM). HES was defined and diagnosed based on the end‐organ manifestations that could be directly related to absolute eosinophil count (AEC) of more than 1.5 × 10⁹/L on two or more occasions and/or tissue eosinophilia without a secondary cause. ²⁰ All reactive causes of eosinophilia were carefully excluded prior to study enrollment. Patients who received hydroxyurea or interferon treatment within 4 weeks prior to enrollment were also excluded. Corticosteroid co‐administration was allowed if the daily dose was 15 mg/day or lower of prednisolone.

Study protocol

Imatinib mesylate was administered as a single dose, starting with 100 mg daily for the first week. Thereafter, the daily dose was increased to 400 mg at the investigator's discretion if a haematological response was not achieved. If the patient exhibited steroid dependency (requiring a concurrent prednisolone equivalent dose of >15 mg/day), they were considered non‐responsive to imatinib, and the patient was withdrawn from the study. If deemed beneficial for the patient, imatinib could be continued for a maximum of 24 cycles, with each cycle lasting 28 days. Dose modification of imatinib was performed based on the worst grade of toxicity, according to the approved protocol (Supporting Information, Methods).

Endpoints and assessment

The primary endpoint was the overall haematological response (OHR) rate. OHR was defined as the sum of the complete haematological response (CHR) and partial haematological response (PHR) rates. CHR was defined as the normalization of white blood cell counts with AEC of less than 0.7 × 10⁹/L and the disappearance of signs of organ involvement or symptoms for ≥4 weeks. PHR was defined as a decrease in the AEC ≥50% and/or symptomatic improvement. The secondary endpoints included progression‐free survival (PFS), toxicities, and genetic characteristics related to the response. PFS was defined as the time from imatinib administration to the confirmation of no response, disease progression, or death from any cause, and the data were censored at the most recent disease assessment. Adverse events (AEs) were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03.

WES and WTS were conducted to investigate the genetic characteristics associated with the haematological response to imatinib. These tests were exclusively performed on the 11 patients who consented to the test, and blood sampling for the examination occurred prior to the initiation of imatinib. The procedures for WES and WTS are detailed in Supporting Information, Methods section.

Statistical analysis

The sample size was calculated using Simon's optimal two‐stage design. We expected a 20% OHR rate for imatinib. With a power of 80% and an alpha of 5%, 10 patients were required for the first stage. If one or more patients showed a response in the first stage, the study progressed to the second stage, which required 29 patients. Four or more responses are required for this drug to be evaluated in further studies. Considering a 10% dropout rate, 32 patients were required.

Categorical variables were compared between the two groups using Fisher's exact test or the Chi‐square test, and p‐values were two‐sided. Differences in the means or medians of the continuous variables between the two groups were assessed using the Wilcoxon rank‐sum test. Survival curves were estimated using the Kaplan–Meier method. The statistical software “R” version 4.3.1 (www.r‐project.org) was used for all statistical analyses. p‐values of <0.05 were considered statistically significant.

RESULTS

Patient and disease characteristics

A total of 32 patients were enrolled; their baseline characteristics are shown in Table 1. The median age was 50 years (range, 20–74), and 20 patients (62.5%) were men. Serum vitamin B12 and IgE levels were measured in 14 and 18 patients respectively, and were elevated in 5 and 12 patients, respectively. The median AEC and BM eosinophil % were 2.89 × 10⁹/L (range, 1.57–13.9) and 18.4% (range, 5.2–54.9), respectively. BM findings were consistent with HES in all patients, with no evidence of mast cell infiltration, significant fibrosis, or associated haematological malignancy.

TABLE 1.

Baseline characteristics of study patients, according to haematological response.

	All patients N = 32	Responder N = 15	Non‐responder N = 17	p‐Value
Median age, years (range)	50 (20–74)	52 (20–74)	49 (23–67)	0.655
Age >60 years, n (%)	8 (25.0)	4 (26.7)	4 (23.5)	1.000
Sex, n (%)				0.170
Male	20 (62.5)	7 (46.7)	13 (76.5)
Female	12 (37.5)	8 (53.3)	4 (23.5)
Body mass index, median (range)	23.7 (17.1–36.6)	23.9 (18.9–36.6)	23.4 (17.1–28.8)	0.852
History of eosinophilia, months, median (range)	10.7 (2.3–183.9)	7.8 (2.4–72.1)	20.5 (2.3–183.9)	0.213
<6 months, n (%)	7 (21.9)	3 (20.0)	4 (23.5)	1.000
6 months to <1 year, n (%)	11 (34.4)	7 (46.7)	4 (23.5)	0.316
≥1 year, n (%)	14 (43.8)	5 (33.3)	9 (52.9)	0.448
Previous treatment, n (%)
Cyclophosphamide	2 (6.2)	1 (6.7)	1 (5.9)	1.000
Hydroxyurea	7 (21.9)	4 (26.7)	3 (17.6)	0.851
Corticosteroid	31 (96.9)	14 (93.3)	17 (100)	0.949
Clinical manifestation, n (%)
Heart	2 (6.2)	1 (6.7)	1 (5.9)	1.000
Lung	12 (37.5)	5 (33.3)	7 (41.2)	0.927
Liver	8 (25.0)	4 (26.7)	4 (23.5)	1.000
Skin	12 (37.5)	6 (40.0)	6 (35.3)	1.000
Lymph node enlargement	3 (9.4)	2 (13.3)	1 (5.9)	0.909
Intestine	7 (21.9)	2 (13.3)	5 (29.4)	0.503
Nerve	3 (9.4)	2 (13.3)	1 (5.9)	0.909
Peripheral blood cell count, median (range)
Haemoglobin, g/dL	13.6 (10.1–16.3)	13.0 (10.1–15.4)	14.5 (10.6–16.3)	0.209
Platelet, ×109/L	269 (107–493)	234 (161–472)	298 (107–493)	0.051
White blood cell, ×109/L	12.0 (5.6–21.4)	10.4 (5.6–21.4)	12.9 (9.0–20.6)	0.179
Absolute neutrophil count	4.56 (1.63–11.33)	4.24 (1.63–7.35)	4.98 (2.05–11.33)	0.137
Absolute lymphocyte count	2.28 (0.41–4.71)	2.00 (1.03–4.71)	2.33 (0.41–3.92)	0.950
Absolute eosinophil count	2.89 (1.57–13.90)	3.00 (1.57–13.90)	2.59 (1.60–10.76)	1.000
Conventional cytogenetics, n (%)
Normal	25 (78.1)	11 (73.3)	14 (82.4)	0.804
Y deletion	2 (6.2)	1 (6.7)	1 (5.9)
Not assessed	5 (15.6)	3 (20.0)	2 (11.8)
BM M:E ratio, median (range)	3.6 (1.3–17.9)	4.0 (1.9–14.0)	3.4 (1.3–17.9)	0.827
BM blast %, median (range)	0.6 (0–2.5)	0.7 (0–2.0)	0.4 (0–2.5)	0.183
BM eosinophil %, median (range)	18.4 (5.2–54.9)	15.6 (5.2–54.9)	26.8 (10.3–52.7)	0.038
BM cellularity, n (%)				0.497
Hypercellular	1 (3.1)	0 (0)	1 (5.9)
Normocellular	24 (75.0)	13 (86.7)	11 (64.7)
Hypocellular	4 (12.5)	1 (6.7)	3 (17.6)
Not assessed	3 (9.4)	1 (6.7)	2 (11.8)

Abbreviations: BM, bone marrow; HES, hypereosinophilic syndrome; M:E ratio, myeloid:erythroid ratio.

Efficacy

A response evaluation was conducted for all 32 patients. Based on the best response during the study, the CHR rate was 18.8% (6/32). The PHR rate was 28.1% (9/32), resulting in an OHR rate of 46.9% (95% CI, 29.1–65.3), with the median time to response was 1.5 months (range, 0.5–4.9) (Figure 1A). Among the six patients who achieved CHR, the median time to CHR was 1.75 months (range, 1–11), with considerable variability among the patients. No differences in the baseline characteristics between responders and non‐responders were observed, except for the BM eosinophil %, which was lower in the responders (median, 15.6% vs. 26.8%, p = 0.038). Among the six patients who achieved CHR, all but one maintained CHR until the 24th cycle of imatinib; one patient lost response after 20 months of imatinib treatment. Of the five patients who sustained CHR up to the 24th cycle, three continued to maintain their CHR without additional treatment following imatinib discontinuation. Among the nine patients who achieved PHR, two lost their response at 8 and 18 months after imatinib treatment. Of the five patients who maintained PHR through the 24th cycle, three continued to sustain their response. The median PFS was 4.3 months (95% CI, 3.0–not reached), and the 6‐month PFS rate was 45.7% (95% CI, 31.2–67.1) (Figure 1B). No patient died during the study.

FIGURE 1.

Clinical outcome of imatinib mesylate. (A) Cumulative incidence of haematological response. (B) Progression‐free survival. CHR, complete haematological response.

The dose of imatinib for responders and non‐responders was increased similarly in a stepwise manner (Figure 2; Table S1). However, the treatment duration was shorter for non‐responders, as imatinib was discontinued earlier in cases of no response. Of the 15 responders, 10 increased to the full dose (400 mg/day), and four experienced dose reduction. Twelve patients (37.5%) received concomitant corticosteroids at least once; six were non‐responders, and six were responders (including one who achieved CHR). The patient who achieved CHR initially started with prednisolone 15 mg/day, then reduced to 5 mg/day at imatinib cycle 2, and subsequently maintained at 2.5 mg/day while sustaining CHR. Ten patients (31.2%) completed the treatment for up to 24 cycles. The reasons for end‐of‐trial in the remaining 22 patients were as follows: 16 (72.7%) due to no response, 3 (13.6%) due to loss of response, 2 (9.1%) due to withdrawal of consent, and 1 (4.5%) due to an AE.

FIGURE 2.

Kinetics of the absolute eosinophil count (AEC) and dose–response relationship during 24 cycles of imatinib mesylate and oral corticosteroid (OCS) based on the response. Data points are the mean and standard deviation. (A) AEC (all patients). (B) Imatinib dose (all patients). (C) Oral prednisolone dose (12 patient who received OCS).

Toxicity

AEs occurred in 22 (68.8%) patients, with grades 1 or 2 accounting for 59.4% and grades 3 or 4 accounting for 9.4% (Table 2). The most frequently reported AEs were skin rash, myalgia or muscle cramps, and oedema, all of which were grade 1 or 2. haematological toxicity was rare, with only two cases of neutropenia: one grade 2 and one grade 3. Imatinib was abandoned in one patient because of a grade 3 complete atrioventricular block.

TABLE 2.

Summary of adverse event.

Adverse events	Grade 1–2	Grade 3–4	Total
Haematological toxicity, n (%)
Neutropenia	0	2 (6.2)	2 (6.2)
Non‐haematological toxicity, n (%)
Oedema	7 (21.9)	0	7 (21.9)
Skin rash	7 (21.9)	0	7 (21.9)
Myalgia/muscle cramps	7 (21.9)	0	7 (21.9)
Abdominal pain	6 (18.8)	0	6 (18.8)
Nausea/vomiting	4 (12.5)	0	4 (12.5)
Fatigue	4 (12.5)	0	4 (12.5)
Infections	3 (9.4)	0	3 (9.4)
Headache	2 (6.2)	0	2 (6.2)
Alopecia	2 (6.2)	0	2 (6.2)
Anxiety	2 (6.2)	0	2 (6.2)
Diarrhoea/constipation	1 (3.1)	0	1 (3.1)
Chest pain	1 (3.1)	0	1 (3.1)
Pleural effusion	0	1 (3.1)	1 (3.1)
Peripheral neuropathy	1 (3.1)	0	1 (3.1)
Blurred vision	1 (3.1)	0	1 (3.1)
Complete atrioventricular block	0	1 (3.1)	1 (3.1)

WES and WTS analysis

WES and WTS were performed on 11 patients (five responders and six non‐responders). WES datasets were analysed to explore the mutational profiles of the responders and non‐responders. The median coverage of depth was a median of 159.9X (137.6–239.7X) for HES and 148.1X (107.2–175.6X) for matched normal genomes (Table S2). A total of 121 somatic mutations (113 single‐nucleotide variants and eight indels) were identified in the HES genome (Table S3). There were no significant differences in the number of non‐silent mutations between responders (median of eight mutations, range 3–11) and non‐responders (median of six mutations, range 3–10) (Figure 3A). Mutations in PCDHB13 and SIGLEC16 were recurrently detected in non‐responders (two of six, 33%), whereas mutations in DSPP and HRNR were recurrently detected in responders (two of five, 40%).

FIGURE 3.

Genetic characteristics related to the response. (A) Whole‐exome sequencing analysis between responders and non‐responders. (B) Heatmap of differentially expressed genes between responders and non‐responders.

Transcriptome analysis was conducted using 11 WTS datasets. The heatmap of the differentially expressed genes (DEGs) showed a discrepancy bewteen the responders and non‐responders (Figure 3B). A total of 12 genes was deregulated in the responder group compared with the non‐responder group (adjusted p‐value <0.05, log₂ fold change >1); three genes were relatively upregulated, and nine genes were relatively downregulated in the responders (Table 3). DEGs upregulated in responders were enriched for the hallmark heme metabolism‐related pathway in GESA (adjusted p‐value <0.05, normalized enrichment score >1.3). In addition, we identified seven fusion transcripts reported in the fusion database, ²¹ but none of the PDGFRA fusion transcripts were observed. STAT5B::RARA fusion was identified in one patient with PHR, and PAK2::PIGX and FIP1L1::CHIC2 fusions were identified in one patient with CHR. Both patients maintained their responses throughout the study. RNF130::BRAF and WNK1::KDM5A fusion were identified in the non‐responder group.

TABLE 3.

Results of RNA sequencing analysis.

Dysregulated genes in the responders as compared with the non‐responders
Gene	Log2 fold change	p‐Value	FDR
Upregulated (N = 3)
ABC7‐43041300I9.1	22.384	<0.001	<0.001
RP11‐734I18.1	4.924	<0.001	0.008
RFPL4AL1	6.400	<0.001	0.020
Downregulated (N = 9)
RP11‐34P13.15	−25.496	<0.001	<0.001
SNORD15A	−23.347	<0.001	<0.001
AC132186.1	−21.718	<0.001	<0.001
PMS2P2	−1.697	<0.001	0.006
HOXA2	−2.033	<0.001	0.007
RNF17	−6.538	<0.001	0.013
GTF2IRD2P1	−4.600	<0.001	0.013
SNORA53	−5.915	<0.001	0.020
DKK3	−3.165	<0.001	0.028

List of fusion transcripts identified by whole‐transcriptome sequencing
Gene1	Gene2	chr1	Breakpoint1	str1	chr2	Breakpoint2	str2	Fusion type
Responder
RARA	STAT5B	17	38 474 700	+	17	40 369 482	−	UTR/CDS (truncated)
PIGX	PAK2	3	196 731 093	+	3	196 782 627	+	Intra‐chromosomal
CHIC2	FIP1L1	4	54 930 349	−	4	54 292 039	+	In‐frame
Non‐responder
PIGX	PAK2	3	196 731 093	+	3	196 782 627	+	Intra‐chromosomal
WNK1	KDM5A	12	863 490	+	12	495 140	−	In‐frame
BPTF	PITPNC1	17	67 875 019	+	17	67 578 185	+	Intra‐chromosomal
RNF130	BRAF	5	179 440 061	−	7	140 508 795	−	In‐frame

Abbreviations: CDS, coding sequence; chr, chromosome; FDR, false discovery rate; str, strand; UTR, untranslated region.

DISCUSSION

In this prospective phase II trial, we present data on the efficacy of imatinib mesylate in 32 patients with PDGFRA/B‐negative HES, along with genetic information obtained through WES and WTS. The haematological response rate to imatinib was 46.9%, with a CHR rate of 18.8%, and the response was stable. Acceptable and manageable toxicity profiles were observed.

In PDGFRA/B‐positive HES, the CHR to imatinib is well‐established at 99%, demonstrating its effectiveness. ¹⁹ , ²² , ²³ Following the report by Cools et al. that imatinib may also be effective in patients with PDGFRA/B‐negative HES, ²⁴ subsequent studies have evaluated the efficacy of imatinib in this population; however, the response rates varied. Two previous prospective phase II trials of imatinib included 15 and 8 patients with PDGFRA/B‐negative HES, with haematological response rates of 40% and 50%, respectively. ¹⁷ , ¹⁸ In contrast, in a prospective study involving 18 patients with PDGFRA‐negative HES, ¹⁶ and in another prospective study involving 36 patients with PDGFRA/B‐negative HES, ¹⁹ the haematological response rates were 16% and 25%, respectively. This discrepancy among studies is likely attributable to the relatively small sample size and heterogeneous patient selection. Therefore, this study aimed to evaluate the efficacy of imatinib in a sufficient number of patients. The median time to haematological response was 1.5 months, indicating a prompt response, which is in line with previous studies. ¹⁷ , ¹⁸ , ¹⁹ Furthermore, 12 of the 15 responders maintained their response, indicating that sustained response, in addition to rapid response, can be expected in a significant proportion of responders.

The rapid and durable response observed in some patients with PDGFRA/B‐negative HES suggests the presence of unknown imatinib‐sensitive genetic alterations. Although we excluded CHIC2 deletion and PDGFRB rearrangements, the possibility of other fusion partners with PDGFRA remains, as several have been reported. Curtis et al. also reported the discovery of two novel fusion genes (STRN::PDGFRA fusion for t(2;4) and ETV6::PDGFRA for t(4;12)). ²⁵ Our data suggest several fusion genes that might be associated with the response to imatinib. The STAT5B::RARA fusion gene, identified in one of our imatinib responders, has been previously reported in patients with acute promyelocytic leukaemia. ²⁶ It has been reported that the STAT5B::RARA fusion can lead to the activation of the STAT5B transcription factor. ²⁷ Imatinib has been reported to have an inhibitory effect on the phosphorylation of STAT5, suggesting that the STAT5 pathway might be a potential target for imatinib. ²⁸ The RNF130::BRAF fusion induces constitutive activation of the kinase domain of BRAF. ²⁹ As a result, the downstream MAPK signalling pathway is aberrantly activated, potentially diminishing the efficacy of imatinib. ³⁰ , ³¹ , ³² In addition, several studies have reported that when WNK1 is fused with other genes, it increases the expression of the fused gene. ³³ , ³⁴ KDM5A is associated with the resistance to TKIs. ³⁵ , ³⁶ Therefore, the WNK1::KDM5A fusion observed in a non‐responder in our study might have contributed to the non‐responsiveness to imatinib. In our study, the downregulation of nine genes, including SNORD15A, and the upregulation of three genes were associated with a favourable response to imatinib. Fujita et al. reported the potential involvement of small nucleolar RNAs, including SNORD17 and SNORD15A, in telomere biology. ³⁷ Considering the dose‐dependent inhibition of telomerase activity by imatinib, our study aligns with previous reports, suggesting that the inhibition of cell proliferation by suppressing telomerase activity, not mediated by tyrosine kinases, could be another mode of action of imatinib. ³⁸ , ³⁹ Further research is required regarding the perspective of fusion analysis through RNA sequencing.

Recently, given the pivotal role of IL‐5 in eosinophil activation, biological agents targeting IL‐5 or its receptor, such as mepolizumab and benralizumab, have become the focus of increased investigation. ⁴⁰ , ⁴¹ Although the most appropriate dose and predictive biomarkers for response have not yet been clearly identified, mepolizumab has significantly reduced the onset and risk of flares in PDGFRA‐negative HES while maintaining manageable AEs. ¹² , ⁴² , ⁴³ Benralizumab has also demonstrated efficacy in reducing AEC in PDGFRA‐negative HES, and a phase III trial is currently underway. ⁴⁴ Therefore, in PDGFRA/B‐negative HES, imatinib should be considered with appropriate patient selection. Unlike monoclonal antibodies such as mepolizumab, which target cell surface antigens, imatinib, as a TKI, targets the underlying disease mechanism. Thus, imatinib can be effective when disease progression is associated with the pathways it modulates. However, there is a scarcity of research on genetic alterations associated with imatinib response in PDGFRA/B‐negative HES, and detailed genomic analyses beyond PDGFR fusion are indeed limited in clinical settings. Therefore, predicting the imatinib response based on clinical characteristics can be beneficial. In our study, no differences in baseline characteristics, including AEC, were observed between responders and non‐responders, which is consistent with the findings of previous studies. ¹⁸ , ¹⁹ However, responders had a significantly lower % of BM eosinophils. Whether the difference in BM eosinophil levels is a result of genetic alterations related to the imatinib response remains unclear. Khoury et al. reported that features of primary myeloid HES might serve as surrogate markers for imatinib response in PDGFRA‐negative HES. ⁴⁵ Further research on predictive biomarkers is warranted. On the other hand, considering that most responders achieved rapid response, a short imatinib trial may easily identify patients who would benefit from imatinib. Intermesoli et al. also reported that a short trial of imatinib (100 mg/day for 2 weeks) can rapidly confirm imatinib‐sensitive HES. ¹⁵

This study has several limitations. First, the sample sizes used in WES and WTS were small. This could have resulted in low statistical power. Furthermore, due to limited sample availability, independent assays for the identified fusion genes could not be conducted; therefore, information regarding their functional relevance is missing. Second, pathological confirmation through tissue biopsy was not performed in the assessment of haematological response.

Despite these limitations, our study is the first prospective trial of imatinib in patients with PDGFRA/B‐negative HES. Our study addresses the long‐standing problem that retrospective studies with a small number of patients may not adequately evaluate the efficacy of imatinib. Furthermore, we identified several DEGs and fusion transcripts that may be associated with the response to imatinib. Given the rarity of genomic analyses related to the imatinib response in PDGFRA/B‐negative HES, the present study holds value as it suggests further research on potential genomic biomarkers.

In conclusion, imatinib demonstrated favourable outcomes in some patients with PDGFRA/B‐negative HES. The responses were rapid and sustained among the responders, with good compliance. Several DEGs and fusion transcripts have been suggested to be potentially associated with the imatinib response. Further comprehensive investigations with larger sample sizes are required to validate the predictive genetic biomarkers for imatinib in PDGFRA/B‐negative HES.

AUTHOR CONTRIBUTIONS

Dong Hyun Kim: Data curation, formal analysis, investigation, writing—original draft, writing—review and editing. Seokhyeon Kim: Formal analysis, investigation, writing—original draft. Seonyang Park, Ja Min Byun, Junshik Hong, Dong‐Yeop Shin, Inho Kim, Soo Mee Bang, Jeong‐Ok Lee, Ji Yun Lee, Sang‐A Kim, Ki Hwan Kim: Investigation, writing—review and editing. Yeun‐Jun Chung: Methodology, writing—review and editing. Seung‐Hyun Jung: Methodology, supervision, writing—original draft. Youngil Koh, Sung‐Soo Yoon: Conceptualization, investigation, supervision, writing—review and editing. All the authors checked and approved the final version of the manuscript.

FUNDING INFORMATION

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (Ministry of Science, ICT, and Future Planning) (grant number NRF‐2021R1A4A2001553). Imatinib mesylate was provided free of charge by Boryung Pharma (Korea). This study is an investigator‐sponsored trial, and the study was conducted independently of the funder. The funder had no role in study design, data collection, data analysis, or data interpretation.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS STATEMENT

This trial was conducted in accordance with the Declaration of Helsinki, Good Clinical Practice Guidelines, and National Policies on Bioethics and Human Biological Specimens. The study protocol was approved by the Institutional Review Board of each participating institution.

PATIENT CONSENT STATEMENT

Written informed consent was obtained from all patients before their participation.

CLINICAL TRIAL REGISTRATION (INCLUDING TRIAL NUMBER)

This trial was approved and registered with the Korean Ministry of Food and Drug Safety (No. 20130196318).

Supporting information

Data S1.

Kim DH, Kim S, Park S, Byun JM, Hong J, Shin D‐Y, et al. Phase II trial of imatinib mesylate in patients with PDGFRA/B‐negative hypereosinophilic syndrome. Br J Haematol. 2024;205(6):2305–2314. 10.1111/bjh.19828

DATA AVAILABILITY STATEMENT

The datasets generated during the current study are available from the corresponding author upon reasonable request.