Long-term Prognostic Value and Analytical Parameters of the Next-Generation Sequencing–Based Multigene Assay in Hormone Receptor–Positive, HER2-Negative Breast Cancer
Article information
Abstract
Purpose
In hormone receptor (HR)–positive, human epidermal growth factor receptor 2 (HER2)–negative early breast cancer, gene expression testing facilitates treatment decisions. A next-generation sequencing (NGS)–based assay was developed to address test decentralization and underrepresentation of younger/premenopausal patients. We aimed to validate the long-term prognostic value of the NGS-based assay and analyze its quality control (QC) parameters.
Materials and Methods
We analyzed samples from 265 patients with breast cancer with at least 10 years of follow-up. We evaluated the long-term prognostic ability of the NGS-based assay according to the risk groups for distant recurrence, as determined by the Decision Index, and the performance of the QC parameters used for the experimental process.
Results
Among 265 participants, 60.4% were ≤ 50 years old, and 39 (14.7%) experienced distant recurrence within 10 years. In the Decision Index–stratified low- and high-risk groups (n=186, 70.2% and n=79, 29.8%), 10-year distant metastasis–free survival rates were 96.1% (95% confidence interval [CI], 92.1 to 98.1) and 79.3% (95% CI, 68.4 to 86.8), respectively. In patients aged ≤ 50 years, the high-risk group had a hazard ratio of 5.89 (95% CI, 2.84 to 12.20). Analyses including 106 samples that failed the stringent QC criteria showed inferior prognostic value, wherein DV200 and cDNA concentrations were the most crucial parameters.
Conclusion
We validated the prognostic ability of an NGS-based assay to stratify HR-positive/HER2-negative breast cancers and predict the risk of distant recurrence, and confirmed the requirement for stringent QC criteria to ensure its prognostic ability.
Introduction
The 2024 National Comprehensive Cancer Network guidelines for breast cancer recommend gene expression testing to determine systemic adjuvant chemotherapy in patients with hormone receptor (HR)–positive, human epidermal growth factor receptor 2 (HER2)–negative breast cancer [1]. HR-positive/HER2-negative breast cancers are predominantly of the luminal A (50%-60%) and luminal B (< 30%) molecular subtypes [2-4]. While generally associated with favorable prognoses, luminal B cancers have poorer outcomes, particularly with late distant metastases following primary endocrine therapy [2]. Luminal A subtypes exhibit low chemosensitivity [5]; trials, i.e., NSABP-B20 [6] and SWOG8814 [7], indicate multi-agent adjuvant chemotherapy provides no survival benefit. Accurately identifying patients with a low risk of distant metastasis can minimize unnecessary adjuvant systemic therapy and inform appropriate treatment strategies.
The TAILORx and RxPONDER trials for the 21-gene Oncotype DX assay (Genomic Health [now Exact Sciences]) and the MINDACT trial for the 70-gene MammaPrint assay (Agendia) report inconsistent results for predicting chemotherapy response, especially in women aged ≤ 50 years and/or premenopausal women. This has led to guidelines distinguishing gene expression testing and treatment recommendations based on menopausal status [8]. The American Society of Clinical Oncology Guidelines now recommend only the 21-gene assay for premenopausal women or those aged ≤ 50 years [9].
Racial and ethnic disparities significantly influence the etiology of breast cancer, with environmental factors shaping tumor biology [10]. Clinical presentations and molecular subtypes vary across ethnicities [11]. The peak age for breast cancer diagnosis is notably lower in Asian patients (45-50 years) compared to the Western peak age of 70 years [12,13]. Younger patients also exhibit lower mRNA expression of ESR1 and estrogen receptor (ER)–associated genes, along with higher expression of immune-related genes [14]. These patients face greater resistance to endocrine therapies and experience poorer outcomes in luminal B breast cancer, as indicated by elevated Ki-67 expression and higher rates of TP53 mutations [15,16].
These differences in breast cancer characteristics among the Asian population have led to the development of various gene expression assays, including the 95-gene Curebest 95GC Breast assay (Sysmex Corporation) [17], 18-gene Recur-Index assay (Quark Biosciences) [18], and the next-generation sequencing (NGS)–based OncoFREE assay (DCGen) [19]. Among these, the NGS-based assay was the first developed using the NGS platform, enabling local implementation. However, the initial validation cohort had several limitations: (1) inclusion of patients with distant metastasis who had received chemotherapy; (2) a minimal follow-up period of 60 months despite a final median follow-up of 141 months; and (3) statistical analysis not performed at an independent institution.
This study aims to validate the long-term prognostic value of the NGS-based assay and analyze the analytical parameters of samples included in a clinical study that enrolled evaluator-blinded, retrospectively collected samples under the regulations of the Ministry of Food and Drug Safety (MFDS) in South Korea.
Materials and Methods
1. Sample selection criteria
This retrospective, multicenter, evaluator-blinded study complied with Korean MFDS regulations for NGS-based assay approval. The study aimed to evaluate the prognostic ability of the assay for early-stage breast cancer, specifically 10-year distant metastasis–free survival (DMFS) post-surgery. Analyzed samples were sourced from patients who underwent breast cancer surgery, with a median follow-up of 10 years. Formalin-fixed paraffin-embedded (FFPE) tissue specimens were collected from Seoul National University Hospital (SNUH), Samsung Medical Center (SMC), and Asan Medical Center (AMC) (see Supplementary Materials for detailed inclusion and pathologic criteria). For the 10-year survival analysis, patients without distant metastasis were censored at their last available follow-up.
Eligible patients had HR-positive, HER2-negative primary invasive breast cancer with pathologic stage T1b-2/N0. Clinicopathological data, including age, menopausal status, and tumor histopathology, were collected. None of the samples had been utilized in prior studies for the development or validation of the NGS assay [20]. Specimens were reviewed by two pathologists at each institution to confirm the inclusion of invasive tumor portions, after which anonymized specimens were sent for analysis at Seoul National University Bundang Hospital (SNUBH).
2. Sample number calculation
The Korean MFDS designated the negative predictive value (NPV) as the primary endpoint for trial approval. In the initial validation of the NGS-based assay, the NPV was 95.97% for eligible patients [19]. To validate the prognostic ability of the assay, we referenced the MINDACT trial [21], where patients not receiving chemotherapy showed an NPV of 93.7% at 8 years. Considering a 1% annual increase in recurrence rates for patients without axillary lymph node metastasis [22], we set the minimum required 10-year NPV at 91.7%. With a power of 80% and a low-risk population proportion of 67% [19], at least 167 low-risk samples were needed. Therefore, we required a minimum of 250 participants to account for potential dropouts from long-term storage of FFPE tissues.
3. NGS-based assay clinical performance test
Anonymized FFPE tissue sections were prepared by trimming the tissue to include only the tumor areas. RNA was extracted using the RNeasy DSP FFPE Kit (Qiagen). RNA quality was assessed with a Qubit 4 fluorometer (Invitrogen) for quantity, an Agilent 4200 TapeStation (Agilent Technologies) for integrity, and a Nabi-UV/Vis Nano Spectrophotometer (MicroDigital) for purity. Extracted RNA meeting stringent quality control (QC) standards (S1 Table) was deemed satisfactory.
The cDNA library was prepared using the xGen Broad-Range RNA Library Prep Kit (Integrated DNA Technologies). Each library was required to meet at least one of the following criteria: a total cDNA amount of ≥ 100 ng or a mean insert size between 200 and 350 bp. Hybridization with target capture probes was performed using an NGS-based assay Target Capture Solution kit (DCGen). A target-captured library was then selected, amplified, and purified.
NGS was conducted if the total post–polymerase chain reaction amount of the target-captured library was at least 4 nmol/L or the mean library length was between 200-350 bp. Sequencing was performed on a NextSeq 550Dx (Illumina), and the sequencing data were analyzed using NGS-based assay software (DCGen). Data analysis primarily focused on sequencing data that met all stringent QC criteria for NGS data quality.
4. Statistical analyses
Statistical analyses were performed at the Medical Research Collaborating Center of SNUH Biomedical Research Institute. Receiver operating characteristic curves were plotted to evaluate the ability of the risk model to discriminate between patients who developed distant recurrence and those who did not. DMFS and overall survival (OS) rates were estimated by the Kaplan-Meier method, and differences in OS between risk groups were assessed using the log-rank test. Clinical risk was dichotomized as “low” or “high” based on ER status, HER2 status, histologic grade, nodal status, and tumor size (see S2 Table for the exact algorithm). A Cox proportional hazards model including age, tumor size, progesterone receptor (PR) status, clinical risk, and Decision Index (DI; see reference [19] for details) was fitted, with two-sided p < 0.05 considered statistically significant.
Cumulative incidence of metastasis by risk group was analyzed using Gray’s test. Distant metastasis probability was plotted against the DI using Cox regression estimates. To assess the discriminatory performance of the NGS-based assay for predicting metastasis-free survival, we calculated Uno’s C-index and compared its predictive accuracy against conventional clinicopathologic factors—age, tumor size, PR status, and clinical risk score—while accounting for censored observations. All statistical analyses were conducted using SAS 9.4 (SAS Institute Inc.) and SPSS Statistics ver. 25 (IBM Corp.).
5. Study approval
This retrospective study assessed 10-year breast cancer prognosis predictions. The study protocols at SNUH (D-2303-107-1413) and SMC (2023-03-059) were approved with a waiver of written informed consent, given the minimal risk and practical considerations. In contrast, the AMC (2023-1003) required written consent from living participants but did not require it for deceased individuals. Therefore, data collection at AMC was primarily limited to deceased patients due to the difficulty in obtaining consent from living individuals. The clinical performance test protocol was approved by the IRB of SNUBH (E-2305-828-357).
Results
1. Study cohort composition
Initial screening was conducted for 503 patients: 162 from SNUH, 172 from SMC, and 169 from AMC. After excluding 79 patients who did not meet the inclusion criteria, 424 anonymized FFPE specimens were sent to the Analytical Institute at SNUBH. Of the 424 FFPE specimens initially collected, 158 (37%) were excluded for failing predefined analytical criteria (S2 Table)—cDNA library QC (n=22), target-capture library QC (n=4), sequencing QC (n=102), read-alignment failure (n=1), insufficient final concentration (n=15), or post-collection exclusion for not meeting enrollment criteria (n=14)—leaving 266 specimens for testing. One patient was then excluded due to incomplete analytical data, resulting in a final analysis cohort of 265 patients (Fig. 1).
Schematic overview of the study. At three hospitals, the medical records of patients were screened, and those with formalin-fixed, paraffin-embedded samples (n=503) were selected; of these, 424 samples met the inclusion criteria and were anonymized at each hospital before being sent for next-generation sequencing–based assay to Seoul National University Bundang Hospital (SNUBH). Among them, 266 samples passed the stringent quality-control criteria, and one sample was excluded prior to analysis because it did not meet the analytical criteria. AMC, Asan Medical Center; SMC, Samsung Medical Center; SNUH, Seoul National University Hospital.
The median follow-up period for the entire cohort was 128.8 months (all patients had at least 120 months of postoperative follow-up). Among those who developed distant metastasis, the time to event ranged from 11.2 to 208.0 months. Ten-year DMFS was estimated by Kaplan-Meier analysis, with patients who had under 10 years of follow-up censored at their last assessment. No deaths occurred during the follow-up interval. Other clinicopathological information of this cohort is summarized in Table 1. Detailed characteristics of the breast cancer treatment in the enrolled patients are presented in S3 Table.
Demographic and pathological characteristics of formalin-fixed, paraffin-embedded specimens in this clinical performance evaluation
2. Prognostic ability of the DI and the risk of distant recurrence according to risk groups
Thirty-nine and 41 distant metastases were observed during the 10 years and whole period of follow-up, respectively. The median DI was 16.0 (range, –1.0 to 47.2) and the area under the curve (AUC) for distant metastasis according to the DI was 0.757 at 10 years (Fig. 2). The cumulative distant-metastasis rate at 5 and 10 years was 9.8% (95% confidence interval [CI], 6.6 to 13.8) and 14.7% (95% CI, 10.8 to 19.3), respectively. Of the 265 samples analyzed, 186 (70.2%) and 79 (29.8%) were classified into the low- and high-risk groups, respectively. The Kaplan-Meier curves of all samples are shown in Fig. 3. Ten-year DMFS estimates were 96.1% (95% CI, 92.1 to 98.1) and 79.3% (95% CI, 68.4 to 86.8) for the low-risk and high-risk groups, respectively. For 160 patients aged ≤ 50 years, 10-year DMFS estimates were 90.9% (95% CI, 85.9 to 96.2) and 56.4% (95% CI, 42.8 to 74.3) for the low- and high-risk groups, respectively. The probability of distant recurrence at 5 and 10 years increased steadily with an increasing DI (S4 Fig.).
Receiver operating characteristic curve of the Decision Index classified for distant recurrence. AUC, area under the curve; CI, confidence interval.
Association of risk category with distant metastasis–free survival. Kaplan-Meier estimates of survival rates of the low-versus high-risk groups based on the Decision Index (DI). Hazard ratios (HRs) for the high- versus low-risk groups are shown. CI, confidence interval; NGS, next-generation sequencing.
Over the entire follow-up period, four patients died before developing distant metastases and were thus treated as competing events. In the competing-risk analysis, the high-risk group had a significantly increased subdistribution hazard ratio (SHR) for distant metastasis compared with the low-risk group (SHR, 4.33; 95% CI, 2.33 to 8.04; p < 0.001). When stratified by age, patients aged ≤ 50 years had an SHR of 5.89 (95% CI, 2.84 to 12.20; p < 0.001), whereas those aged > 50 years had an SHR of 3.42 (95% CI, 1.04 to 11.20; p=0.034) (S5 Fig.). The OS comparisons across risk groups are presented in the S6 Fig.
3. Evaluation of risk-prediction ability compared to clinicopathologic factors
Uno’s C-indices for variables such as age, pathological characteristics (tumor size and histologic grade), HR status (ER/PR), and clinical risk were evaluated and juxtaposed against the DI (Table 2). The Uno’s C-index of the DI was 0.745 (95% CI, 0.672 to 0.818), suggesting a 74.5% likelihood that individuals with a higher DI score will experience distant metastasis earlier than their counterparts. While the indices for age (p=0.213) and tumor size (p=0.370) did not significantly underperform, the indices for PR status (p < 0.001) and clinical risk (p=0.036) were notably inferior to the DI, underscoring its predictive superiority in these aspects.
Discriminatory ability of the NGS-based assay
4. Validation of predictive power for distant metastasis
Of those categorized as low-risk by the NGS-based assay (n=186), 171 remained metastasis-free over 10 years (Table 3). The 10-year NPV for the low-risk group, as determined by the NGS-based assay, was 91.9% (95% CI, 87.1 to 95.4). When adjusted for sensitivity 61.5%, specificity 75.66%, and a 10-year distant metastasis rate of 9.1% in the cohort used for sample size calculation, the adjusted NPV was 95.2% (95% CI, 92.0 to 98.3), with the lower bound of the 95% CI larger than the minimum required rate at 91.7% established in the statistical analysis plan. The adjusted NPV was 95.5% (95% CI, 91.6 to 99.4) and 94.7% (95% CI, 89.4 to 99.9) for patients aged ≤ 50 years and > 50 years, respectively.
Number and proportion of patients with distant metastasis within 5 and 10 years, stratified by DI (DI ≤ 20, low risk; DI > 20, high risk)
5. Evaluating outcomes and providing QC suggestions based on metrics
Given the utilization of FFPE specimens stored for > 10 years, stringent QC standards were meticulously applied at all stages of the experimental process to mitigate the effects of prolonged storage (S1 Table). This rigorous approach resulted in a high exclusion rate of 37.3%, with 158 of the 424 anonymized specimens being disqualified. Among these, 106 samples were excluded specifically because they failed the NGS data quality criteria. To confirm the utility of these stringent NGS QC criteria, the prognostic ability for distant recurrence according to risk groups was evaluated, including the results of the 106 samples previously excluded under these criteria. The AUC was determined to be 0.687 (S7 Fig.), and the Kaplan-Meier analysis revealed a hazard ratio of 3.38 (95% CI, 2.02 to 5.65; p < 0.001). The AUC and hazard ratio for the samples, including those excluded according to the data quality criteria, were inferior to those for the 265 samples included in the main analysis.
To verify the QC criteria for the wet process, an additional analysis was conducted on 265 samples that were successfully subjected to the experiment. Assuming no significant differences in sample quality across the FFPE collection sites, we identified QC parameters with minimal variability among multiple metrics, including DV200 and cDNA concentration (S8 Table). When comparing the results of the final 265 analyzed samples with those of the 106 samples that passed the wet process QC criteria but failed to meet the dry process standards, DV200 and cDNA concentrations exhibited no substantial differences (data not shown). This analysis, performed on successfully sequenced data, suggests that DV200 and cDNA concentrations are the most critical parameters for the wet process.
Discussion
This study provides a comprehensive evaluation of the long-term prognostic performance and analytical characteristics of the NGS-based multigene assay, specifically developed using HR-positive, HER2-negative breast cancer samples. These findings support the effectiveness of this assay for stratifying patients, including those aged 50 years or younger and premenopausal women, into distinct prognostic categories, enhancing treatment decisions and patient outcomes. Importantly, we also confirmed the significance of the stringent QC standards in ensuring the highest accuracy in predicting distant metastasis.
The NGS-based assay demonstrated 10-year DMFS rates of 96.1% (95% CI, 92.1 to 98.1) for the low-risk and 79.3% (95% CI, 68.4 to 86.8) for the high-risk groups among patients who had not received adjuvant chemotherapy. This level of prognostic discrimination is critical for minimizing overtreatment in patients unlikely to benefit from aggressive systemic therapy.
A notable strength of this study is its evaluator-blinded, retrospective multicenter design, encompassing 265 patients from major hospitals in Korea, conducted under the regulations and surveillance of the MFDS. This comprehensive approach ensured the validity of the analysis and reflected real-world clinical practice, thereby enhancing the generalizability of the findings. By focusing on a predominantly younger Asian patient population (60.4% aged ≤ 50 years), this study addresses a significant gap in breast cancer research, which has historically focused on Western populations.
The unique clinical and molecular characteristics of breast cancer in Asian, Black, and Hispanic patients necessitate the development of race- and ethnicity-specific prognostic tools [10,11]. For patients aged ≤ 50 years who represented a substantial portion of the study cohort and are reflective of patients with breast cancer in Asia, the assay provided an NPV of 95.5% when adjusted for the 10-year distant metastasis rate. It has substantial prognostic value, which is critical given the typically more aggressive disease course and poorer prognosis observed in younger patients with breast cancer.
This study adhered to rigorous QC measures for RNA extraction and sequencing to ensure assay reliability and validity. Tissue storage and fixation time affect the integrity and usability of RNA obtained from FFPE samples [23]. Despite inherent challenges, the successful utilization of anonymized FFPE specimens demonstrated the robustness and applicability of the assay, even after long-term storage. Because NGS-based assays in the clinical setting typically use FFPE surgical specimens developed within a few weeks, they are expected to have superior RNA quality compared with the samples analyzed in this study. The feasibility of the assay for clinical use was previously reported by showing a 100% RNA sequencing success rate for samples with less than 2 months of storage time [19]. Therefore, establishment of stringent yet practical QC parameters for real-world applications is crucial.
Prior to sequencing, DV200 and cDNA concentrations were identified as critical benchmarks, and stringent NGS data QC standards were required to ensure the highest accuracy in prognostic predictions. Although evidence suggests using RNA concentration (> 25.0 ng/μL), library concentration (> 1.7 ng/μL), and DV200 (> 30%-50%) as QC parameters for RNA sequencing from FFPE samples [24], our findings showed that successful sequencing results can be obtained even with lower values. This complexity indicates that relying solely on one or two parameters may not adequately represent overall quality. A tiered QC approach is necessary to ensure comprehensive assessment and quality management of RNA sequencing. While a significant difference in DMFS for the low- versus high-risk groups was observed irrespective of the NGS QC criteria, it is advisable to uphold stringent NGS data QC standards to ensure the highest accuracy in the prediction of distant metastasis.
Stringent standards are especially important to maximize the advantage of the decentralization of NGS-based assays for use in individual laboratories using highly reproducible and robust RNA sequencing and normalization processes [19]. These QC standards can be used as references for the future development of NGS-based assays for other diseases. The value of decentralization strategies using NGS platforms is also highlighted by the fact that NGS versions of the 70-gene assays, Mammaprint and BluePrint (Agendia), were developed to decentralize the test [25,26].
To contextualize our assay’s performance against established tests in Asian cohorts, we compared key prognostic metrics for Oncotype DX [27,28], MammaPrint [29], and the NGS-based assay [19,28] (Table 4). These data illustrate that the NGS-based assay achieves comparable or superior discrimination in Asian breast cancer patients, with a higher C-index and low-risk 10-year DMFS than previously reported for Oncotype DX and MammaPrint in similar populations. This contextual comparison underscores the assay’s relevance and potential to complement or even enhance existing gene-expression tests in Asian clinical practice.
Comparative performance of Oncotype DX, MammaPrint, and the NGS-based assay in Asian breast cancer cohorts, stratified by age
Although this validation confirms the robust clinical utility of the NGS-based assay for risk stratification and therapeutic guidance, we did not assess its cost-effectiveness. Given the higher upfront expense of multigene NGS panels compared with conventional prognostic tools, formal health-economic analyses—such as cost–utility (e.g., cost per Quality-Adjusted Life Year gained) and budget-impact studies—will be essential to evaluate value for payers and healthcare systems. We therefore propose that future research incorporate economic modeling to inform reimbursement policies and support broader implementation of this assay in routine clinical practice.
A limitation of this study is its retrospective sample collection across centers, which introduced a selection bias in the high-risk group, resulting in a higher-than-expected distant metastasis rate compared with the 8%-9.1% reported in prospective trials of HR-positive/HER2-negative, node-negative patients not receiving chemotherapy [30-32]. At one center, the requirement for prospective consent led to the inclusion of many deceased patients (for whom consent was waived), specimens were collected almost exclusively from patients who had already died; of 71 samples, 31 (43.6%) were from deceased patients, for whom consent was waived. This skew toward postmortem samples undermines the generalizability of our cohort and explains why OS in the full cohort was lower than anticipated. Nevertheless, the low-risk group maintained favorable 10-year DMFS and OS rates of 96.1% and 96.2%, respectively.
Another limitation is the heterogeneity of the study population. However, because samples were obtained from multiple large institutions with varying practices and patient demographics, this diversity may in fact strengthen the generalizability of the findings by reflecting real-world variation.
In conclusion, we validated the long-term prognostic value of the NGS-based multigene assay in HR-positive, HER2-negative breast cancer using evaluator-blinded, retrospectively collected samples. We demonstrated the necessity of stringent QC measures to preserve assay performance and highlighted the assay’s strong predictive capacity for distant metastasis. These findings support its clinical utility in guiding treatment decisions and improving patient outcomes.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
This retrospective study assessed the validity of 10-year breast cancer prognosis predictions. Owing to the practical challenges in obtaining consent, the absence of reasons to anticipate consent refusal, and the minimal risk to participants, the IRBs of SNUH (D-2303-107-1413) and SMC (2023-03-059) approved a waiver of the requirement for written informed consent. The IRB of AMC (2023-1003) required written informed consent from patients whose samples were used in the study, unless it was impossible to obtain written consent in cases where the patient had died. The clinical performance test protocol was approved by the IRB of SNUBH (E-2305-828-357).
Author Contributions
Conceived and designed the analysis: Han W, Lee HB, Park KU.
Collected the data: Kim H, Koh J, Lee H, Gong G, Kang E, Ryuh JM, Shin DS, Lee SB, Lee HJ, Kim HK, Shin HC, Han W, Lee HB, Park KU.
Contributed data or analysis tools: Kim H, Koh J, Lee H, Gong G, Oh S, Lee J, Chang HE, Choi Y, Park KU.
Performed the analysis: Kim H, Oh S, Lee J, Chang HE, Choi Y, Kang E.
Wrote the paper: Kim H, Lee HB, Park KU.
Conflicts of Interest
SBL is listed as a co-inventor of patents for the NGS-based assay in this study, owned by and royalties paid from DCGen Co. Ltd. HJL is listed as a co-inventor of patents for the NGS-based assay in this study, owned by and royalties paid from DCGen Co. Ltd. HCS is a member of the board of directors and holds stock and ownership interests in DCGen Co. Ltd., owned by and royalties paid from DCGen Co. Ltd. WH is a member of the board of directors, holds stock and ownership interests in DCGen Co. Ltd., and is listed as a co-inventor on patents for the NGS-based assay evaluated in this study, owned by and royalties paid from DCGen Co. Ltd. HBL is a member of the board of directors, holds stock and ownership interests in DCGen Co. Ltd., and is listed as a co-inventor on patents for the NGS-based assay evaluated in this study, owned by and royalties paid from DCGen Co. Ltd. The other authors have declared that no conflict of interest exists.
Funding
This work was funded by DCGen, Co., Ltd. The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. Additionally, they did not influence the decision to submit the manuscript for publication.
Acknowledgments
We extend our deepest gratitude to the patients who consented to provide FFPE specimens for this study.