Understanding the Unmet Needs in Diagnostics and Therapeutics of Chronic Lymphocytic Leukemia in Korea
Article information
Abstract
Purpose
Since the U.S. Food and Drug Administration approval of ibrutinib in 2012, the treatment landscape for chronic lymphocytic leukemia (CLL) and small lymphocytic lymphoma (SLL) has undergone a paradigm shift. Nevertheless, the disease remains incurable, posing ongoing clinical challenges. In Korea, these challenges are further compounded by unique characteristics of patients and national healthcare system.
Materials and Methods
The multicenter retrospective analysis included 519 patients diagnosed with CLL/SLL between 2006 and 2024 across three major Korean institutions.
Results
The median age at diagnosis was 62 years. Among 267 patients who received first-line therapy, 68.9% (184/267) were treated with immunochemotherapy, 19.4% (52/267) with cytotoxic chemotherapy, and 11.2% (30/267) with a Bruton tyrosine kinase (BTK) inhibitor. Since the Korean approval of ibrutinib in 2016, BTK inhibitor use has steadily increased. Subgroup analyses demonstrated that patients receiving BTK inhibitor–based therapy had more favorable outcomes compared to those treated with immunochemotherapy.
Conclusion
This study provides the most comprehensive real-world reflection of current diagnostic and therapeutic practices in Korea. Through comparative analysis with international data, it offers valuable insights into the limitations of current CLL/SLL care and may inform future strategies to optimize management in the Korean context.
Introduction
Chronic lymphocytic leukemia (CLL)/small lymphocytic leukemia (SLL) is the most common leukemia in Western countries [1], but has been associated with different geographic distributions. The reported age-standardized incidence rate of CLL/SLL is 4.75/100,000 persons per year in the United States [2], 4.92/100,000 in Europe [3], and 5.4 to 6.3/100,000 in Latin America [4], while East Asian countries have been correlated with significantly lower incidence ranging from 0.1 to 0.52 [5,6]. Interestingly, from annual percentage change (APC)’s point of view the greatest changes are observed in East Asian population: for example, Korea has seen CLL/SLL APC of 4.17% during the years 1999 to 2010 while APC remained steady at 0.68% in the United States during the same time period [7]. The reasons behind such discrepancies have never been completely disclosed, and scientific questions remain as to whether we are dealing with the same CLL/SLL or CLL/SLL in East Asians are indeed associated with different characteristics leading to different treatment responses.
On the other hand, the CLL/SLL treatment paradigm has shifted significantly in the past decade. Prior to the advent of novel agents, younger patients with CLL/SLL requiring therapy were typically treated with immunochemotherapy regimens, such as fludarabine, cyclophosphamide, and rituximab (FCR) regimen [8] or bendamustine and rituximab (BR) regimen [9]. The introduction of two main pillars in molecularly targeted therapy, Bruton tyrosine kinase (BTK) inhibitors [10] and B-cell lymphoma 2 (BCL-2) antagonists [11], has substantially improved the treatment and prognosis of CLL/SLL.
Since the U.S. Food and Drug Administration (FDA) approval of the first BTK inhibitor ibrutinib in 2012, CLL/SLL treatment has undergone multiple transformative shifts within just a decade. Historically, the introduction of more effective therapies has correlated with gradual improvements in overall survival (OS), suggesting that incorporating these novel treatments into clinical practice will likely contribute to a substantial rise in disease prevalence. In Korea, this impact is expected to be even more pronounced, given the significantly younger median age at initial CLL/SLL diagnosis (63 years) [12], and the higher APC [7]. Additionally, despite the landmark advancements in CLL/SLL treatment, the disease remains incurable, posing a long-term challenge. Considering Korea’s unique healthcare context—where the majority of the cancer treatment is covered by the national insurance—it is clear that CLL/SLL will present an increasing burden on national health policies and financial resources.
Therefore, we have carried out this study to generate recent, high-quality epidemiologic data that will serve as a foundation for future resource allocation and CLL/SLL care planning, while also establishing essential baseline data to support more innovative therapeutic approaches.
Materials and Methods
1. Patients
This was a longitudinal cohort study conducted between January 2006 and January 2024 at Seoul St. Mary’s Hospital, Seoul National University Hospital, and Korea University Guro Hospital. CLL/SLL was diagnosed according to International Workshop on Chronic Lymphocytic Leukemia (iwCLL) guidelines [13], and patients aged 18 years or older were retrospectively identified and included. Their medical records were reviewed and analyzed for demographics, baseline disease characteristics, details of treatment, treatment outcomes, and survival. When available, results from cytogenetic and molecular genetics data were collected. The initial cohort consisted of 561 patients. However, 42 patients were excluded from the final analysis due to issues such as nationality, transfer to other institutions, loss to follow-up, and incomplete test results or medical records. Consequently, a total of 519 patients were included in the final analysis (Fig. 1).
Patients for analysis of data. CLL, chronic lymphocytic leukemia; SLL, small lymphocytic lymphoma.
This study was conducted in accordance with the Declaration of Helsinki and ethical standards of the Institutional Review Boards of each hospital. All authors had access to the study data and reviewed and approved the study.
2. Treatment and definitions
Treatment regimens were classified as follows: immunochemotherapy (including FCR and obinutuzumab-chlorambucil), chemotherapy (which consists of only traditional cytotoxic chemotherapy drugs, such as chlorambucil or fludarabine), BTK inhibitors, and BCL-2 inhibitor. The response evaluation was done according to iwCLL guidelines.
3. Targeted next-generation sequencing
In patients with available samples, targeted next-generation sequencing (NGS) was performed, using the Theragen Cancer CGP panel on the Illumina NovaSeq 6000 platform. A total of 39 genes associated with CLL/SLL pathogenesis were analyzed. Variant filtering included a minimum variant allele frequency threshold of 1%, and only non-synonymous variants were retained unless previously reported as pathogenic in the Catalogue Of Somatic Mutations In Cancer (COSMIC). Low-quality variant calls were excluded. Functional annotation and pathogenicity prediction were conducted using Combined Annotation Dependent Depletion (CADD), Protein Variation Effect Analyzer (PROVEAN), and Splice Artificial Intelligence (SpliceAI) to identify variants with potential clinical significance.
4. Statistical analysis
Progression-free survival (PFS) was defined as the time from the start of a therapy to either disease progression or death from any cause. OS was defined as the time from the diagnosis to death from any cause. Survival rates were estimated using the Kaplan-Meier method, and the differences between groups were evaluated using the log-rank test.
Survival outcomes were stratified by first-line treatment regimen, treatment start date (before vs. after 2016; For analytical purposes, 2016 was used as the cutoff year, reflecting the point at which ibrutinib became an approved treatment option in Korea), CLL International Prognostic Index (CLLIPI), and TP53 mutation status. All statistical analyses were performed using EZR Version 1.52 software (Saitama Medical Center, Jichi Medical University), a graphical user interface for R ver. 4.0.0 (R Foundation for Statistical Computing) [14].
Results
1. Patients’ characteristics
Baseline clinical and laboratory features of all 519 patients are summarized in Table 1. The median age at diagnosis was 62 years (range, 28 to 95 years), and 61.1% were male. Overall, 57.0% of patients were diagnosed before the age of 65 years.
Baseline characteristics
Rai staging at diagnosis revealed that stages I (32.4%) and II (27.2%) were the most frequent, whereas advanced disease (stages III and IV) was observed in 11.4% and 10.2% of patients, respectively. According to the Binet classification, stage A was identified in 44.9% of cases, followed by stage B in 38.9% and stage C in 15.0%. CLL-IPI scores were available for 40.9% of the cohort, among whom 16.0% and 16.8% were classified as low- and intermediate-risk, respectively, while high-risk and very high-risk categories comprised 7.3% and 0.8% of patients.
Laboratory abnormalities at presentation included anemia (hemoglobin < 11 g/dL) in 19.8% of patients and thrombocytopenia (platelet count < 100×109/L) in 10.6%. Elevated β2-microglobulin levels (> 3.5 mg/dL) were reported in 8.1%, and absolute lymphocyte counts exceeding 15×109/L were observed in 48.2% of cases.
Among the patients with available molecular genetic data, nine out of 290 (3.1%) had del(17p), 18 out of 253 (7.1%) had tumor protein p53 (TP53) mutations, and 10 out of 146 (6.8%) showed mutated immunoglobulin heavy-chain variable region (IGHV) status. Cytogenetic data were available for 298 patients, and complex karyotype was identified in 34 (11.4%) of them.
2. Treatment patterns
During the median follow-up of 45.3 (range, 0.1 to 236.4 months) months, 267 patients (51.4%) ultimately required treatment (Table 2). Across the cohort, 252 patients (48.6%) did not receive any CLL/SLL-directed therapy during the observation period, whereas 190 (36.6%) patients received one line of therapy, 56 (10.8%) received two lines, and 21 (4.0%) received three or more lines (Table 2).
Time to treatment, treatment indications, and lines of therapy
For those who did require treatment, the median time from diagnosis to first treatment was 4 months (range, 0 to 115 months). The most common indications for treatment initiation were progressive cytopenia (anemia 49.4%, thrombocytopenia 32.5%) and bulky disease (27.5%).
Types of treatment administered are presented in Fig. 2. Immunochemotherapy was the most common first-line approach (68.9%), with FCR regimen accounting for 48.7% of cases (Fig. 2A, S1 Table). Chemotherapy alone was administered in 19.5%, primarily chlorambucil-based regimens (12.0%). BTK inhibitors were used in 11.2% of patients, and BCL-2 inhibitor–based therapy was rarely applied (0.3%). While immunochemotherapy has remained the predominant treatment modality, the use of BTK inhibitors in the first-line setting has increased over time, following the approval of ibrutinib for CLL/SLL treatment in Korea (S2 Fig.).
Treatment regimens administered by line of therapy. (A) First-line treatment. (B) Second-line treatment. (C) Third-line treatment. BCL2i, BCL-2 inhibitor; BTKi, Bruton tyrosine kinase inhibitor; CTx, chemotherapy; IC, immunochemotherapy.
This trend continues in later lines of therapy, with BTK inhibitors being the most frequently prescribed regimen in the second-line setting (53.2%), followed by immunochemotherapy (26.0%) and chemotherapy alone (14.3%) (Fig. 2B, S3 Table). BCL-2 inhibitor–based regimens were administered in 6.5% of cases. In the third-line setting (n=21) (Fig. 2C, S4 Table), immunochemotherapy remained common (38.1%), alongside BTK inhibitors (23.8%) and BCL-2 inhibitor–based regimens (19.0%).
Table 3 presents the number of cases, treatment response, duration of response, and progression status for each treatment modality, irrespective of the line of therapy. Disease progression was documented in 31.6% of patients treated with immunochemotherapy, 70.1% of those receiving chemotherapy, 14.5% of patients treated with BTK inhibitors, and 10.0% of those treated with BCL-2 inhibitors. Richter’s transformation occurred in 9.4% of patients treated with immunochemotherapy and 6.0% of those treated with chemotherapy. No cases of Richter’s transformation were observed among patients who received BTK inhibitor or BCL-2 inhibitor–based therapies.
Response and outcomes by treatment
3. Impact of clinical variables on survival
Fig. 3 and 4 show the OS and PFS curves, respectively. For the whole group, median OS was not reached (Fig. 3A) and the median PFS was 82.9 months (95% confidence interval [CI], 67.6 to 107.9) (Fig. 4A).
Overall survival. (A) Total population. (B) Subgroups stratified by age at initial diagnosis. (C) Subgroups stratified by treatment initiation date. (D) Subgroups stratified by first-line treatment regimens. (E) Subgroups stratified by Chronic Lymphocytic Leukemia International Prognostic Index (CLL-IPI). (F) Subgroups stratified by TP53 mutation status.
Progression-free survival. (A) Total population. (B) Subgroups stratified by age at initial diagnosis. (C) Subgroups stratified by treatment initiation date. (D) Subgroups stratified by first-line treatment regimens. (E) Subgroups stratified by Chronic Lymphocytic Leukemia International Prognostic Index (CLL-IPI). (F) Subgroups stratified by TP53 mutation status. BTK, Bruton tyrosine kinase.
Advanced age at initial diagnosis was definitively associated with inferior outcome; Patients who were diagnosed at the age of 65 years or older showed increased risk of death (hazard ratio [HR], 2.86; 95% CI, 1.60 to 5.10; p < 0.001) and progression (HR, 2.16; 95% CI, 1.40 to 3.33; p < 0.001) (Figs. 3B and 4B).
When stratified by year of treatment initiation, patients who began therapy prior to 2016 experienced inferior survival compared with those treated from 2016 onward. Although the median OS was not reached in either group, compared to pre-2016 group, patients in post-2016 group had a significantly decreased risk of death (HR, 0.49; 95% CI, 0.25 to 0.96; p=0.03) (Fig. 3C). A similar pattern was observed for PFS, with median PFS of 78.9 months in the pre-2016 group and 82.9 months in the post-2016 group (HR, 0.92; 95% CI, 0.58 to 1.49; p=0.72) (Fig. 4C).
Regarding treatment modality, individuals receiving BTK inhibitor–based first-line therapy demonstrated more favorable outcomes compared to those treated with immunochemotherapy. While the median OS was not reached in either cohort, the estimated 36-month OS rate was higher in the BTK inhibitor group (96.0%) than in the immunochemotherapy group (91.6%) (HR, 0.32; 95% CI, 0.04 to 2.34; p=0.26) (Fig. 3D). For PFS, the median duration was not reached in the BTK inhibitor group, whereas it was 91.3 months (95% CI, 77.3 to 127) among patients receiving immunochemotherapy. The corresponding 36-month PFS rates were 91.8% and 79.7%, respectively (HR, 0.60; 95% CI, 0.14 to 2.49; p=0.48) (Fig. 4D).
4. Impact of molecular features on survival
Survival outcomes differed according to prognostic risk categories as defined by the CLL-IPI. CLL-IPI was available in 212 patients, with high and very risk scores observed in 19.8% (42/212). Increasing risk scores were associated with progressively poorer OS and PFS. Compared to the low-risk group, the high- and very high-risk group had a significantly increased risk of death (HR, 5.16; 95% CI, 1.71 to 15.56; p=0.004) (Fig. 3E). The intermediate-risk group did not show a statistically significant difference in survival compared to the low-risk group (HR, 1.26; 95% CI, 0.36 to 4.40; p=0.72) (Fig. 3E). A similar trend of declining PFS was observed with higher CLL-IPI scores; The high- and very high-risk group had a significantly increased risk of progression compared to the low-risk group (HR, 7.49; 95% CI, 2.48 to 22.64; p < 0.001) (Fig. 4E).
Patients harboring TP53 mutations exhibited particularly adverse outcomes. Patients with TP53 mutation had significantly worse survival compared to those without mutation (HR, 5.56; 95% CI, 1.62 to 19.11; p=0.006) (Fig. 3F). PFS difference was not statistically significant (HR, 2.48; 95% CI, 0.87 to 7.03; p=0.08) (Fig. 4F), but median PFS was 41.6 months (95% CI, 3.8 to not estimable) in patients with TP53 mutations and 78.9 months (95% CI, 60.2 to 105.3) in those without (Fig. 4F).
Targeted NGS was conducted in the subset of patients (n=85) with adequate samples. Among the patients who underwent NGS, 70 (82.4%) were sampled prior to the initiation of therapy. Fig. 5 illustrates the mutational landscape of these patients, integrating results from targeted NGS and fluorescence in situ hybridization (FISH) analysis.
Mutational landscape of 85 chronic lymphocytic leukemia patients based on targeted next-generation sequencing and fluorescence in situ hybridization (FISH) analysis.
The median number of mutations per patient was 6 (range, 1 to 13). Splice-site variants were the most common mutation type, followed by missense, frameshift, nonsense, and inframe insertions or deletions. Among the 39 genes analyzed, the most frequently altered genes included MGA (48%), DNMT3A (45%), ATM (39%), CHD2 (38%), and SETD2 (38%). Other frequently mutated genes included SPEN (33%), KMT2D (31%), NOTCH1 (28%), SF3B1 (28%), CARD11 (18%), and TP53 (2%).
To investigate the clinical relevance of somatic mutations, we evaluated differences in PFS according to the mutation status of the most frequently altered genes (MGA and DNMT3A), as well as well-characterized or biologically relevant genes associated with CLL/SLL pathogenesis, including ATM, NOTCH1, SF3B1, and CARD11. Kaplan-Meier analysis of PFS stratified by mutation status of six selected genes (S5 Fig.) revealed no statistically significant differences between mutated and wild-type groups for any individual gene. Although not statistically significant, mutations in ATM (HR, 1.78; 95% CI, 0.76 to 4.17; p=0.18), SF3B1 (HR, 1.66; 95% CI, 0.72 to 3.80; p=0.22), and NOTCH1 (HR, 1.65; 95% CI, 0.72 to 3.72; p=0.22) were each associated with a trend toward inferior PFS, consistent with their known prognostic roles in CLL/SLL; the lack of statistical significance may be attributable to limited sample size or low event rates.
5. Impact of age on treatment regimen and response
Fig. 6A presents a Sankey diagram illustrating the distribution of first-line treatment regimens and response categories according to age at first-line treatment, dichotomized at 65 years. An important consideration in interpreting this diagram is that, among patients treated with BTK inhibitors (n=30), response assessments including bone marrow evaluation were incomplete in 20 cases due to ongoing treatment. These cases were conservatively categorized as having achieved at least a partial response (PR), although it is likely that a subset would ultimately meet criteria for complete response (CR) upon formal assessment.
First-line treatment regimens and outcomes by age groups. (A) Total. (B) Older than 65 years. (C) Younger than 65 years. BTK, Bruton tyrosine kinase; CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.
Fig. 6B and C stratify the diagram by age group. In patients aged ≥ 65 years, the proportion receiving immunochemotherapy as first-line treatment was markedly lower than in those aged < 65 years. (63.8% vs. 73.7%). This treatment gap was compensated by the preferential use of BTK inhibitors in older patients (15.3% vs. 7.2%). The rate of achieving CR or PR did not appear substantially inferior in the ≥ 65-year group compared with their younger counterparts (85.8% vs. 90.4%).
Discussion
This study represents the largest cohort of Korean patients with CLL/SLL reported to date, confirming previously described clinical features and providing more reliable data on various aspects including cytogenetic profiles. Table 4 summarizes the baseline features and first-line treatment of CLL/SLL in comparison with data from other countries. Compared to real-world cohorts from the United States and Europe, the Korean population demonstrated a similar or quite higher prevalence of adverse prognostic markers. For example, the proportion of patients with unmutated IGHV—generally associated with an unfavorable prognosis—was notably higher in the Korean cohort, in line with previous domestic study [12]. Interestingly, neighboring countries such as China [15] and Japan [20] reported considerably lower frequencies of unmutated IGHV even compared to Western populations [16,21].
While our study included CLL-IPI–based risk stratification, comparisons with previously reported cohorts were limited due to the relative paucity of published data incorporating CLL-IPI as a staging tool. In a retrospective, multicenter study of independent cohort of newly diagnosed CLL/SLL conducted across five Italian institutions [22], more than half of the patients (471/858, 54.9%) were classified as low risk, a substantially greater proportion than in our cohort (83/212, 39.2%). In contrast, the frequency of high and very high-risk categories was comparable between the two populations (173/858, 20.2% vs. 42/212, 19.8%) although the prevalence of high-risk mutation was higher in our cohort. This discrepancy is likely attributable to the inclusion of age in the scoring system, as the median age at diagnosis in our cohort was 62 years—considerably younger than the 65-70 years typically reported in Western populations.
Although our cohort exhibited a biological risk profile at least as unfavorable as those in Western populations, access to BTK inhibitors was disproportionately limited. The proportion of patients who received a BTK inhibitor as first-line therapy was 11.2% in our study, whereas it reached 31.2% and 45.0% in Europe and the United States, respectively—substantially higher than in Korea. These observations suggest the presence of a significant unmet clinical need in the current management of CLL/SLL in Korea.
When patients were stratified by age at first-line treatment (< 65 vs. ≥ 65 years), those aged ≥ 65 years received immunochemotherapy markedly less frequently, with this gap offset by the increased use of BTK inhibitor–based regimens. While comorbidities, polypharmacy, and increased risk of treatment-related complications are common in elderly patients (though not captured in this dataset), it can be reasonably inferred that the total number of treatment cycles and cumulative drug doses administered were not higher—and likely lower—than those in younger patients. Despite these differences, the rate of overall response in the older cohort was largely comparable to that observed in younger individuals.
BTK inhibitors have demonstrated substantial clinical benefit in CLL/SLL as above, supporting their role as key frontline options where available. However, given the chronic and largely incurable nature of CLL/SLL, many patients had received continuous treatment, and the limitations of long-term BTK inhibitor monotherapy are increasingly recognized, including treatment-related toxicities, cumulative quality-of-life burden, the development of resistance, and the broader societal and economic costs of prolonged therapy.
Fixed-duration strategies have emerged to address these challenges. In treatment-naïve CLL/SLL, phase III trials of time-limited ibrutinib-venetoclax regimens have shown higher rates of undetectable minimal residual disease (MRD) and longer PFS than chemoimmunotherapy, with acceptable safety profile [23,24]. More recent phase III trials extending this approach to next-generation BTK inhibitor backbones—acalabrutinib [25] or zanubrutinib [26]—in combination with BCL-2 inhibitors, with or without anti-CD20 antibodies, have likewise demonstrated deep remission with durable disease control across clinically relevant biologic risk strata, including IGHV and TP53. Collectively, these results have been reflected in international treatment guidelines, which now recognize fixed-duration BTK inhibitor plus BCL-2 inhibitor therapy as a key frontline option for CLL/SLL.
The potential value of fixed-duration therapy may be especially relevant in East Asian settings, including Korea, given the younger median age at diagnosis observed in our cohort compared with many Western series (Table 4), potentially increasing lifetime exposure and cumulative burden of indefinite therapy. However, considering potential differences in baseline demographics, disease biology, and healthcare context, the efficacy and tolerability of such regimens should be prospectively confirmed in this population. In this regard, our findings underscore the importance of generating prospective, regionally representative evidence to ensure that fixed-duration targeted approaches are appropriately validated for patients in Asia.
With the expanding use of time-limited targeted regimens, molecular and response-based biomarkers are increasingly incorporated into risk assessment and response evaluation. Molecular profiling—including NGS, FISH, and conventional cytogenetics—supports identification of high-risk disease biology and may inform treatment selection and intensity [27]. In parallel, achievement of undetectable MRD after therapy is a validated prognostic marker for durable remission and is increasingly integrated into response-adapted treatment frameworks [14,24,26].
In this context, we performed targeted NGS in a subset of our cohort as an exploratory analysis to characterize the mutational spectrum and to assess concordance with established cytogenetic risk markers and clinically relevant outcomes. Previous studies investigating CLL/SLL mutational landscapes across different ethnic groups have revealed distinctive patterns, with Asians generally showing higher frequencies of ATM and TP53 mutations, but lower rates of SF3B1 and NOTCH1 mutations compared to White populations. In our cohort, the mutation frequency of several genes, including ATM and KMT2D, was consistent with prior domestic studies [28,29], supporting its biological representativeness. Interestingly, the prevalence of TP53 mutations was lower than previously reported in the Korean population. This discrepancy may reflect differences in patient selection or disease stage distribution and underscores the biological heterogeneity inherent in real-world populations.
Survival analysis of frequently mutated or well-characterized genes did not demonstrate statistically significant survival differences associated with individual mutations, likely due to the limited sample size. Nevertheless, survival trends were directionally consistent with prior studies, supporting the established prognostic relevance of these mutations. The relatively small cohort size may have limited the statistical power to detect certain differences; a challenge commonly encountered in studies involving non-White populations. This underscores the need for broader implementation of NGS testing not only in Korea but also across other non-White populations.
However, in real-world clinical practice as evidenced by this study, even basic genetic assessments such as FISH or targeted sequencing are underutilized. Despite the large cohort size, nearly half of patients did not undergo FISH analysis, and IGHV sequencing was performed in 146 out of 519 patients. This not only represents a limitation of the study but also reflects a systemic gap in Korean CLL/SLL care. To address this, comprehensive strategies—including clinicians and patients’ education, reimbursement reform, and standardized testing protocols—are needed to expand access to guideline-recommended molecular diagnostics.
In conclusion, this study provides the most comprehensive real-world reflection of current diagnostic and therapeutic practices in Korea. Through comparative analysis with international data, it offers valuable insights into the limitations of current CLL/SLL care and may provide a foundation for future strategies aimed at optimizing care delivery, guiding clinical decision-making, and informing healthcare policies. Continued multicenter collaboration and systematic data collection will be essential to refine paradigms and address the unique challenges of CLL/SLL management in Asia.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
This study was conducted in accordance with the Declaration of Helsinki and the ethical standards of the Institutional Review Boards of each hospital (Seoul National University Hospital: h-2410-008-1574, Seoul St. Mary’s Hospital: KC25RIDI0445, Korea University Guro Hospital: 2025GR0306). The requirement for written informed consent was waived by the IRB owing to the retrospective nature of the study.
Author Contributions
Conceived and designed the analysis: Yu ES, Kwon S, Byeon S, Byun JM, Park SS, Eom KS, Choi CW.
Collected the data: Yu ES, Kwon S, Byeon S, Byun JM, Park SS.
Contributed data or analysis tools: Yu ES, Kwon S, Byeon S, Byun JM, Park SS.
Performed the analysis: Yu ES, Kwon S, Byeon S, Byun JM, Park SS, Eom KS, Choi CW.
Wrote the paper: Yu ES, Kwon S, Byeon S, Byun JM, Park SS, Eom KS, Choi CW.
Conflicts of Interest
Conflict of interest relevant to this article was not reported.