Skip to main content
NCBI home page
ESMO Open logoLink to ESMO Open
. 2025 Dec 29;11(1):106023. doi: 10.1016/j.esmoop.2025.106023

Clinical utility of pleural effusion supernatant cell-free DNA genotyping in previously treated patients with advanced non-small-cell lung cancer and disease progression: a multicenter retrospective study

S-C Chang 1,2,3, C-Y Huang 4, Y-C Lai 1,3, M-S Hsieh 5,6, C-C Ho 6,7, C-Y Yang 6,7, J-Y Shih 6,7, S-G Wu 6,7,8, P-W Hu 1,3, C-L Hsu 6,7, C-Y Chen 6,9, W-Y Liao 6,7,∗
PMCID: PMC12804021  PMID: 41468685

Abstract

Background

Patients with lung cancer frequently develop pleural effusion as the disease progresses. This study evaluated the utility of cell-free DNA (cfDNA) from pleural effusion supernatant in identifying targetable mutations in non-small-cell lung cancer (NSCLC) patients who developed resistance to prior therapies.

Methods

We conducted a multicenter retrospective study involving 95 patients who experienced disease progression after at least one line of treatment and underwent pleural effusion cfDNA next-generation sequencing testing.

Results

Initial routine molecular testing detected various driver mutations in 70 (73.7%) patients, while 26.3% had no detectable driver mutations. Subsequent cfDNA next-generation sequencing identified additional genomic alterations during disease progression, increasing the detection rate for driver mutations from 73.7% to 85.3%.

For patients whose disease progressed after treatment with first- or second-generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors, 40.7% developed an acquired EGFR T790M mutation, and showed clinical benefit with subsequent administration of osimertinib. Of the 29 patients who progressed following osimertinib treatment, 14 received osimertinib as a first-line therapy, and 15 received osimertinib as a second-line therapy. MET copy number gain (CNG) was identified in 28.6% of patients receiving first-line osimertinib and was associated with a significantly shorter progression-free survival (PFS) of 12.0 months versus 23.0 months for those without MET CNG [hazard ratio (HR) = 18.68, P = 0.00088]. CNG in MET and KRAS was linked to poorer overall survival.

Among the patients for whom second-line osimertinib failed, five acquired the EGFR C797S mutation, and four exhibited MYC CNG. MYC CNG was associated with a median PFS of 8.5 months versus 25.0 months for those without MYC CNG (HR 8.4, P = 0.0046).

Conclusion

CfDNA sampling from pleural effusion is a less invasive and more effective method for genomic profiling in NSCLC with disease progression, offering valuable insights to guide personalized treatment strategies.

Key words: non-small-cell lung cancer, disease progression, pleural effusion, cfDNA

Highlights

  • •

    Pleural effusion cfDNA NGS is feasible and informative for progressive NSCLC.

  • •

    Driver detection increased from 73.7% at baseline to 85.3% with pleural cfDNA NGS.

  • •

    After first-line osimertinib, MET copy number gain was linked to shorter PFS.

  • •

    After second-line osimertinib, MYC copy number gain was linked to shorter PFS.

  • •

    Pleural cfDNA can reveal actionable drivers when tissue or plasma testing is negative.

Introduction

The advent of targeted therapies has made next-generation sequencing (NGS) crucial to non-small-cell lung cancer (NSCLC) treatment because it enables genomic profiling from tissue and blood samples. NGS is widely used in clinical practice to identify mutations, although clinicians often experience challenges such as inadequate tissue samples.1 Nonshedding tumors, in which cell-free DNA (cfDNA) is undetectable in circulation, further limit the utility of blood-based NGS in some cancer cases.2

Evidence indicates that in newly diagnosed NSCLC, sequencing cfDNA in malignant pleural effusions (MPEs) can serve as an alternative to tissue and plasma for identifying driver mutations.3 Pleural effusion supernatant cfDNA was prioritized over plasma ctDNA in this cohort because it directly reflects locoregional tumor shedding within the thoracic cavity and mitigates false-negative plasma results in low-shedding disease. Prior studies demonstrated that pleural supernatant cfDNA yields higher mutation detection rates than plasma and can uncover driver and resistance alterations missed by blood testing, including in cytology-negative effusions.4, 5, 6 In accordance with previous pleural effusion liquid biopsy studies, we refer to our analyte as pleural effusion cfDNA (tumor-derived DNA from pleural effusion supernatant). Tumor-derived fragments are identified bioinformatically by somatic variants and variant allele fractions. This method may enable mutation analysis without requiring invasive tissue biopsies, particularly for patients with high-risk or inaccessible tumors. Although targeted therapy is the established first-line treatment of advanced NSCLC with driver mutations, patient prognosis remains poor due to acquired resistance to targeted agents, which is often associated with the development of MPE.4,7

The present study focused on the utility of using MPE samples in patients with NSCLC who have developed resistance to prior therapies. We aimed to evaluate the feasibility of identifying driver mutations and resistance mechanisms by applying NGS to cfDNA from malignant pleural effusion samples in this subset of patients. The findings provide a practical, less invasive option for genotyping NSCLC patients with drug resistance, supporting personalized therapeutic strategies.

Materials and methods

Patients and data collection

We conducted a multicenter retrospective study involving 95 NSCLC patients with progressive disease who underwent pleural effusion cfDNA NGS. The study included patients with recurrent or advanced NSCLC. Data were collected from patients receiving care at National Taiwan University Hospital (NTUH), National Taiwan University Cancer Center, NTUH Hsinchu Branch, NTUH Yunlin Branch, National Yang Ming Chiao Tung University Hospital (NYCUH), and Taipei Tzu Chi Hospital (TZCH) over a period between December 2021 and June 2024. The Research Ethics Committees of NTUH (No: 202301130RIND), NYCUH (No: 2025A009), and TZCH (No: P00001611) approved the study, which was conducted in compliance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice Guidelines. The need for written informed consent was waived by the institutional review board because of its nature of retrospective analysis. Clinical data were retrospectively recorded for further analysis.

All patients underwent routine and sequential molecular testing using diagnostic tissue or pleural effusion cell block at initial diagnosis. Specifically, tests included epidermal growth factor receptor (EGFR) RT–PCR, anaplastic lymphoma kinase (ALK) immunohistochemical (IHC) analysis, and ROS1 fluorescence in situ hybridization to detect hotspot driver mutations. Patients with negative routine test results—including the EGFR Cobas test, ALK IHC staining, and ROS1 fluorescence in situ hybridization—at initial diagnosis were subsequently considered for NGS testing using tumor tissue or plasma, based on shared decision making between the attending physician and the patient. In Taiwan, NGS testing is not reimbursed by the National Health Insurance (NHI), including during the study period, and is not mandatory for patients with advanced lung cancer.

Pleural effusion collection, cfDNA extraction, and targeted panel sequencing

Pleural fluid was predominantly obtained through single, image-guided thoracentesis; a few patients required short-term pigtail drainage. No cases required video-assisted thoracic surgery (VATS) or open surgery for cfDNA sampling, and no procedure-related complications were reported. Each sample required 40 ml of pleural effusion. All pleural effusion samples were collected and transported in cfDNA blood collection tubes (Streck, USA). The samples were sent to a laboratory certified by the College of American Pathologists. That laboratory analyzed the samples by using a AlphaLiquid100 target capture panel (IMBdx, South Korea). Each pleural effusion sample underwent centrifugation at 2000 g for 10 min. cfDNA was extracted from supernatant using a Maxwell RSC cfDNA Plasma Kit (Promega, Madison, WI) following the manufacturer’s protocol. cfDNA concentrations were determined using a 4200 TapeStation (Agilent Technologies, Santa Clara, CA). The details of this procedure have been described in another study.3 Pleural effusion supernatant was obtained at disease progression, processed immediately upon collection, and submitted to a College of American Pathologists-accredited laboratory for clinical targeted NGS (AlphaLiquid100); testing was ordered as part of routine clinical care rather than as a retrospective batch analysis.

Statistical analysis

Categorical variables are presented as percentage, and continuous variables are presented as medians. Statistical analyses and survival curve plotting were conducted using R software version 4.3.2 (R Foundation for Statistical Computing). Categorical variables were analyzed using chi-square tests, and a one-way analysis of variance or Student’s t-test was used to analyze continuous variables. Survival analysis employed the Kaplan–Meier method to evaluate clinical and survival outcomes, with median values presented with 95% confidence intervals (CIs). Log-rank tests were used to compare differences in clinical and survival outcomes. Statistical significance was defined as a two-sided P value <0.05. Oncoplots were generated using Microsoft Excel. In this study, ‘multi-hit’ was defined as two or more distinct pathogenic or likely pathogenic mutations in the same gene within the NGS panel, excluding EGFR. EGFR variants were displayed separately, and variants of uncertain significance were not counted as hits.

Results

Patient demographics

Between December 2021 and June 2024, 95 patients with metastatic NSCLC who had experienced disease progression after at least one line of systemic treatment were enrolled. The patient characteristics are presented in Table 1. The median patient age was 66 years (range 36-88 years), and 57 patients (60.0%) were women. Adenocarcinoma was the most common histological type (n = 91, 95.8%). Driver mutations detected at initial diagnosis through routine molecular testing were EGFR exon 19 deletion (n = 36, 37.9%), EGFR L858R mutation (n = 22, 23.2%), uncommon or compound EGFR mutations (n = 4, 4.2%), EGFR exon 20 insertion (n = 2, 2.1%), ALK fusion (n = 3, 3.2%), and ROS1 fusion (n = 3, 3.2%). In total, 25 patients (26.3%) had no detectable oncogenic driver mutations in routine tests. Before pleural effusion cfDNA testing, 27/95 (28.4%) patients had received first/second-generation EGFR TKIs, 14/95 (14.7%) had received first-line osimertinib, and 15/95 (15.8%) had received second-line osimertinib (Table 1). All 95 patients had pleural effusion at progression by imaging. Tumor cellularity in pleural effusion samples was analyzed for 54 patients (56.8%), while cell blocks were unavailable for further cellularity analysis in the remaining cases. More than half exhibited <5% tumor cellularity (n = 34, 55.7%) or negative findings (n =7, 7.4%) in pleural fluid. The median concentration of cfDNA in pleural effusion supernatant was 4.29 ng/μl (range 0.018-567.923 ng/μl). DNA concentration did not significantly differ between the samples with high and low pleural effusion cellularity (Supplementary Figure S1, available at https://doi.org/10.1016/j.esmoop.2025.106023).

Table 1.

Characteristics of 95 enrolled non-small-cell lung cancer patients with pleural effusion

Progressive disease (N = 95)
Age, median years (range) 66 (36-88)
Female sex, n (%) 57 (60.0)
Never smoker, n (%) 70 (73.7)
Histological type, n (%)
 Adenocarcinoma 92 (96.8)
 Squamous carcinoma 3 (3.2)
Driver mutations by SoC, n (%)
 EGFR 19 DEL 36 (37.9)
 EGFR L858R 22 (23.2)
 Uncommon or compound EGFR mutations 4 (4.2)
 EGFR 20 ins 2 (2.1)
 ALK fusion 3 (4.3)
 ROS1 fusion 3 (4.3)
 Wild type 25 (26.3)
PE cytology%
 ≥5% 13 (13.6)
 <5% 34 (35.8)
 Negative 7 (7.4)
 Not available 41 (43.2)
Treatment before PE cfDNA test
 After first-/second-generation EGFR TKI 27 (28.4)
 After first-line osimertinib 14 (14.7)
 After second-line osimertinib 15 (15.8)
 After ALK TKI 3 (3.2)
 After ROS1 TKI 3 (3.2)
 Others 33 (34.7)
Open in a new tab

DEL, deletion; ins, insertion; PE, pleural effusion; SoC, standard of care; TKI, tyrosine kinase inhibitor.

NGS testing of pleural effusion supernatant cfDNA following disease progression

In addition to the 70 driver oncogenic mutations identified at initial diagnosis, 5 EGFR mutations and 6 other driver mutations were detected in the pleural effusion samples. The oncoplot for all 95 patients is illustrated in Figure 1. Newly identified EGFR mutations included EGFR L858R + K860I, L858R, exon 19 deletion, A864V, and an exon 20 insertion. Other detected mutations were KIF5B–RET fusion, SND1–BRAF fusion, ERBB2 exon 20 insertion, KRAS G12A, KRAS G12V, and PIK3CA E454K. These mutations were classified as targetable or potentially targetable.5, 6, 7 The detection rate of driver mutations increased from 73.7% (70/95) at diagnosis to 85.3% (81/95) in follow-up analyses.

Figure 1.

Figure 1

Open in a new tab

Oncoplot depicting the driver and co-occurring mutations detected in pleural effusion cfDNA from 95 treatment-resistant patients. #Indicates a driver mutation not detected by standard-of-care tests. Multi-hit status was assigned only for two or more pathogenic/likely pathogenic variants; EGFR mutations and variants of uncertain significance were not included in the multi-hit category. IO/chemo, immunotherapy/chemotherapy; TKI, tyrosine kinase inhibitor; VAF, variant allele frequency.

Disease progression following first- and second-generation EGFR tyrosine kinase inhibitors

In total, 27 patients (28.4%) underwent NGS testing of pleural effusion supernatant after experiencing disease progression following first-line treatment with first- or second-generation EGFR tyrosine kinase inhibitors (TKIs). The oncoplot for these patients is depicted in Figure 2A. Among the 27 patients with EGFR mutations, 11 (40.7%) developed the acquired EGFR T790M mutation, the most common form of resistance to first- and second-generation EGFR TKIs. Figure 2B illustrates the potential mechanisms of acquired resistance, with the most frequent alterations being EGFR T790M, PIK3CA missense mutations, and MET copy number gain (CNG). The major acquired mutations to first- or second-generation EGFR TKIs are summarized as a pie chart in Supplementary Figure S2A, available at https://doi.org/10.1016/j.esmoop.2025.106023.

Figure 2.

Figure 2

Figure 2

Open in a new tab

Genomic alterations detected in pleural effusion cfDNA after disease progression on first- or second-generation EGFR-targeted therapy. (A) Oncoplot of 29 patients resistant to first- or second-generation epidermal growth factor receptor (EGFR)-targeted therapy. (B) Sunburst plot depicting these 29 patients’ driver mutations and potential resistant alterations. CNG, copy number gain.

For the 11 patients with an acquired EGFR T790M mutation in the present study, the median progression-free survival (PFS) following osimertinib treatment had not been reached after a median follow-up of 6 months.

Disease progression following first- or second-line osimertinib

In total, 29 patients who experienced disease progression after first-line (n = 14) or second-line (n = 15) osimertinib treatment underwent pleural effusion supernatant NGS testing (Figure 3A). None of the 14 patients who received first-line osimertinib treatment subsequently developed an acquired EGFR T790M mutation. Genomic alterations potentially representing resistance mechanisms to first-line osimertinib are illustrated in Figure 3B, with MET CNG being the most common (28.6%). The major acquired mutations to first-line osimertinib are summarized as a pie chart in Supplementary Figure S2B, available at https://doi.org/10.1016/j.esmoop.2025.106023.

Figure 3.

Figure 3

Open in a new tab

Genomic alterations detected in pleural effusion cfDNA after disease progression on first-line or second-line osimertinib. (A) Oncoplot for 14 patients resistant to treatment with first-line osimertinib and 15 patients resistant to second-line osimertinib. (B) Sunburst plot of these patients’ driver mutations and potential resistant alterations. CNG, copy number gain.

Patients with MET CNG had a shorter median PFS with first-line osimertinib treatment [12.0 months versus 23.0 months; hazard ratio (HR) 18.68, P = 0.00088] than did those without MET CNG, as illustrated in the Kaplan–Meier curve in Figure 4A and the swimmer plot in Figure 4B. Patients with MET and KRAS CNG also exhibited poorer overall survival following first-line osimertinib treatment, as indicated by the Kaplan–Meier curve in Figure 4C.

Figure 4.

Figure 4

Figure 4

Open in a new tab

Association ofMETandKRAScopy number gain with clinical outcomes after first-line osimertinib. (A) Patients with MET copy number gain (CNG) had significantly shorter progression-free survival to first-line osimertinib. (B) Swimmer plot depicts the treatment sequence of the 14 patients resistant to first-line osimertinib. The red triangle indicates patients with MET CNG. (C) MET and KRAS CNG were associated with poor overall survival after first-line osimertinib treatment. CI, confidence interval; HR, hazard ratio; mPFS, median progression-free survival.

Among the 15 patients who experienced failure of second-line osimertinib treatment, 5 had acquired the EGFR C797S mutation (all in-cis with the T790M mutation), which was the most common resistance mechanism (Figure 3B). Additionally, four patients had MYC CNG, which was associated with a significantly shorter median PFS following second-line osimertinib treatment (8.5 months versus 25.0 months; HR 8.4, P = 0.0046), as illustrated in Supplementary Figure S3A and swimmer plot in Supplementary Figure S3B, available at https://doi.org/10.1016/j.esmoop.2025.106023. The major acquired mutations to first-line osimertinib are summarized as a pie chart in Supplementary Figure S2C, available at https://doi.org/10.1016/j.esmoop.2025.106023.

Patients without targetable mutation

In total, 17 patients had no detectable driver mutations and experienced disease progression following systemic chemotherapy. After the patients experienced disease progression, NGS testing of pleural effusion supernatant identified three driver oncogenic mutations: two KRAS G12X mutations and one PIK3CA E545K mutation. PIK3CA-activating mutations may be treatable in patients without other actionable driver alterations; one case report demonstrated that a patient with the PIK3CA H1047R mutation responded to treatment with alpelisib.8 The TP53 mutation was the most prevalent genetic alteration, occurring in 64.7% of cases (Supplementary Figure S4, available at https://doi.org/10.1016/j.esmoop.2025.106023).

Discussion

This study examined the potential of using cfDNA from pleural effusion supernatant to identify targetable genomic alterations in patients with NSCLC after these patients have been treated. Not all patients underwent NGS before treatment, and 25 patients (26.3%) initially had no detectable driver mutations. Subsequent NGS following disease progression identified an additional 11.6% of driver mutations, some of which were druggable. Indeed, the newly identified alterations were detected in patients who had tested wild type for the same genes at baseline, primarily due to the methodological limitations of earlier single-gene assays. Specifically, variants such as EGFR L858R + K860I, A864V, and exon 20 insertions are not covered by the commercial Cobas EGFR RT–PCR hotspot panel used in routine clinical testing and, therefore, could not have been detected initially. Similarly, fusions and mutations including KIF5B–RET, SND1–BRAF, ERBB2 exon 20 insertion, KRAS G12A/G12V, and PIK3CA E454K were not assessed at diagnosis because these targets were outside the scope of routine clinical assays at that time. Consequently, pleural effusion cfDNA NGS carried out at progression revealed these additional actionable or potentially actionable alterations. Because all patients harboring these newly identified alterations had received only chemotherapy ± immunotherapy and no prior targeted therapy, these variants were classified as driver rather than resistance mutations.

Analysis of cfDNA from pleural effusion supernatant is more sensitive than analysis of cfDNA from plasma, which may be limited by low disease burden or nonshedding tumors.9 In this study, 62.9% of pleural effusions exhibited tumor cellularity <5%, and 13% were cytologically negative. Nonetheless, oncogenic driver mutations were detected in 85.3% of cases. These findings are consistent with those of another study that demonstrated higher sensitivity in mutation detection with cfDNA from pleural effusion supernatant than with genomic DNA from sedimented cells in pleural effusion particularly in cytologically negative samples.5

Limited data exist on targetable genomic alterations and resistance mechanisms in cfDNA from pleural effusions following EGFR TKI or ALK inhibitor treatment. One multicenter study revealed that cfDNA from pleural effusions provides more relevant genomic information than does plasma cfDNA, which is commonly used in clinical trials and practice.6 T790M mutations have been identified in 50%-60% of cases exhibiting acquired resistance to gefitinib or erlotinib and in 47.6% of cases resistant to afatinib.10,11 In our cohort of 27 patients treated with first- or second-generation EGFR TKIs, 40.7% developed the T790M mutation, which was detected using cfDNA from pleural effusion supernatant.

In the AURA 3 trial, patients with advanced NSCLC who acquired the T790M mutation after first-line EGFR TKI treatment achieved longer PFS with osimertinib than with platinum-based chemotherapy (10.1 months versus 4.4 months; HR 0.30, 95% CI 0.23-0.41, P < 0.001).12 In the present study, for the 11 patients with the acquired EGFR T790M mutation, the median PFS following treatment with osimertinib was not reached with a median follow-up period of 6 months, consistent with the findings of the AURA 3 trial. Other common genetic alterations after progression on first- and second-generation EGFR TKIs were PIK3CA-activating missense mutations and MET CNG. These acquired resistance mechanisms are consistent with those observed in patients treated with comparator EGFR TKIs, such as gefitinib or erlotinib, in the FLAURA trial.13,14

Patients treated with osimertinib experienced longer PFS than those treated with other first-generation EGFR TKIs, establishing osimertinib as the preferred standard of care for untreated advanced NSCLC with mutant EGFR.15 However, resistance to osimertinib ultimately develops, and the mechanisms underlying treatment failure vary. Our findings on resistance mechanisms, obtained from cfDNA in pleural effusion supernatant after first- or second-line osimertinib treatment, are consistent with those of the FLAURA13,14 and AURA312 trials. Acquired EGFR mutations such as C797S, amplifications of MET and ERBB2, and small-cell transformation were identified in these studies.13, 14, 15, 16, 17, 18

Following first-line osimertinib treatment, cfDNA from pleural effusion exhibited resistance due to MET CNG, which is associated with a reduction in median PFS of osimertinib (12 months). In the INSIGHT 2 trial, which enrolled patients with EGFR-mutated NSCLC exhibiting MET amplification as a resistance mechanism to first-line osimertinib therapy, the median PFS of first-line osimertinib treatment in the INSIGHT 2 study was 15.4 months (interquartile range 10.3-22.5 months).19 This duration was shorter than the 18.9 months of PFS reported in the FLAURA trial. Additionally, in the resistance analysis subset of the FLAURA trial, no clear correlation was observed between resistance mechanism type and osimertinib or comparator EGFR TKI treatment duration. However, in the osimertinib group, 11 of 15 patients with acquired MET amplification had the shortest treatment durations.14 In our cohort, MET CNG was inferred from targeted-NGS depth ratios in pleural cfDNA; this is methodologically distinct from FISH or liquid NGS-defined MET amplification (e.g. MET/CEP7 ratio ≥2.0, absolute gene copy ≥5, or liquid biopsy next-generation sequencing with a MET plasma gene copy number ≥2.3). Accordingly, we interpret our findings as indicative of CNG-associated signals rather than FISH-confirmed amplification.

Resistance mechanisms to osimertinib differ between first- and second-line therapeutic settings. MET amplification is a common resistance mechanism in first-line settings, whereas the C797S mutation is less common in first-line settings but more common in second-line settings.20 Among the 15 patients for whom second-line osimertinib treatment failed in the present study, 33% (5 patients) developed the EGFR C797S mutation, which occurred in-cis with the T790M mutation. Novel 4G EGFR TKIs are currently being developed to address resistance due to T790M and C797S mutations.21

Our findings also indicate that patients with MYC CNG identified after second-line osimertinib treatment experience shorter PFS with osimertinib. MYC CNGs are associated with resistance to various TKIs, and agents targeting MYC are under clinical investigation.22, 23, 24, 25

Our study has several limitations. Firstly, although the retrospective design relied on multicenter data from a substantial cohort, the findings require further validation in prospective studies with detailed data and coding. Secondly, although 51 of our 95 patients (53.7%) underwent NGS at diagnosis or before treatment progression, our use of different sample types (blood/tissue) and platforms presents challenges in defining resistance mechanisms. Thirdly, National Health Insurance of Taiwan does not cover NGS testing, and the decision to use NGS may be influenced by patient socioeconomic status and physician experience. Finally, the optimal NGS timing, choice of NGS platform, and selection of metastatic sampling sites remain debatable.

In conclusion, this study demonstrates the utility of genomic profiling using cfDNA from pleural effusion supernatant to identify targetable mutations in patients with previously treated metastatic NSCLC. This approach enables the detection of actionable mutations that may not be captured in routine molecular testing. Additionally, the method highlights potential resistance mechanisms, suggesting that posttreatment strategies employing genomic profiling may facilitate a convenient and reliable clinical treatment approach. These findings support the feasibility and clinical utility of pleural effusion cfDNA NGS to complement routine testing at progression; our observations are hypothesis generating and warrant validation in larger, prospective studies with prespecified survival endpoints.

Acknowledgements

We extend our gratitude to Wei-Ru Li, Chia-Lin Tsao, TSH Biopharm Corporation Ltd, and IMBdx Corporation Ltd for their invaluable analytical insights, guidance, and facility support. We thank the staff of the Eighth Core Lab, Department of Medical Research, National Taiwan University Hospital, for their technical support during the study.

Funding

None declared.

Disclosure

CLH reported receiving speaking honoraria from AstraZeneca, Roche, Pfizer, Merck Sharpe & Dohme (MSD) Oncology, Novartis, and Bristol Myers Squibb (BMS). JYS reports grants from Roche; honoraria from AbbVie, ACT Genomics, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, BMS, Chugai Pharmaceutical, Daiichi-Sankyo, Eli Lilly, GSK, Guardant Health, Illumina, Janssen, Lotus, MSD, Merk, Ono Pharmaceutical, Orient EuroPharma, Pfizer, Roche, Synmosa, Takeda, TSH Biopharm and TTY Biopharm, Welgene Biotech, and Zuellig Pharma; and fees for meetings from AstraZeneca, Chugai, Pfizer, and Roche. WYL reported receiving speaking honoraria from AstraZeneca, Roche, Boehringer Ingelheim, Eli Lilly, Pfizer, MSD Oncology, Novartis, TSH Biopharm, BMS, Johnson & Johnson, Bayer, Merck KGaA, and Chugai Pharma Taiwan outside of the submitted work; and received support for attending meetings from AstraZeneca, Chugai Pharma, Boehringer Ingelheim, Pfizer, Merck KGaA, and BMS. All other authors have declared no conflicts of interest.

Supplementary data

Supplementary Figure S1.

Supplementary Figure S1

Open in a new tab

Supplementary Figure S2 A.

Supplementary Figure S2 A

Open in a new tab

Supplementary Figure S2 B.

Supplementary Figure S2 B

Open in a new tab

Supplementary Figure S2 C.

Supplementary Figure S2 C

Open in a new tab

Supplementary Figure S3 A.

Supplementary Figure S3 A

Open in a new tab

Supplementary Figure S3 B.

Supplementary Figure S3 B

Open in a new tab

Supplementary Figure S4.

Supplementary Figure S4

Open in a new tab

References

  • 1.Heitzer E., Haque I.S., Roberts C.E.S., Speicher M.R. Current and future perspectives of liquid biopsies in genomics-driven oncology. Nat Rev Genet. 2019;20:71–88. doi: 10.1038/s41576-018-0071-5. [DOI] [PubMed] [Google Scholar]
  • 2.Rolfo C.D., Madison R.W., Pasquina L.W., et al. Measurement of ctDNA tumor fraction identifies informative negative liquid biopsy results and informs value of tissue confirmation. Clin Cancer Res. 2024;30:2452–2460. doi: 10.1158/1078-0432.CCR-23-3321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Chang S.C., Wei Y.F., Chen C.Y., et al. Profiling cell-free DNA from malignant pleural effusion for oncogenic driver mutations in patients with treatment-naive stage IV adenocarcinoma: a multicenter prospective study. Mol Diagn Ther. 2024;28:803–810. doi: 10.1007/s40291-024-00736-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Heitzer E., van den Broek D., Denis M.G., et al. Recommendations for a practical implementation of circulating tumor DNA mutation testing in metastatic non-small-cell lung cancer. ESMO Open. 2022;7 [Google Scholar]
  • 5.Tong L., Ding N., Tong X., et al. Tumor-derived DNA from pleural effusion supernatant as a promising alternative to tumor tissue in genomic profiling of advanced lung cancer. Theranostics. 2019;9:5532–5541. doi: 10.7150/thno.34070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Jin S., Zhou C., Hou X., et al. A multicenter real-world study of tumor-derived DNA from pleural effusion supernatant in genomic profiling of advanced lung cancer. Transl Lung Cancer Res. 2020;9:1507–1515. doi: 10.21037/tlcr-20-882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Meyer M.L., Fitzgerald B.G., Paz-Ares L., et al. New promises and challenges in the treatment of advanced non-small-cell lung cancer. Lancet. 2024;404:803–822. doi: 10.1016/S0140-6736(24)01029-8. [DOI] [PubMed] [Google Scholar]
  • 8.Daher S., Zer A., Tschernichovsky R., et al. Driver mutation characteristics of phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) in advanced non-small cell lung cancer. Lung Cancer. 2023;178:229–236. doi: 10.1016/j.lungcan.2023.02.023. [DOI] [PubMed] [Google Scholar]
  • 9.Mokánszki A., Bádon E.S., Mónus A., Tóth L., Bittner N., Méhes G. Cell-free DNA from pleural effusion samples: is it right for molecular testing in lung adenocarcinoma? Pathol Oncol Res. 2021;27 [Google Scholar]
  • 10.Kobayashi S., Boggon T.J., Dayaram T., et al. EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. N Engl J Med. 2005;352:786–792. doi: 10.1056/NEJMoa044238. [DOI] [PubMed] [Google Scholar]
  • 11.Wu S.G., Liu Y.N., Tsai M.F., et al. The mechanism of acquired resistance to irreversible EGFR tyrosine kinase inhibitor-afatinib in lung adenocarcinoma patients. Oncotarget. 2016;7:12404–12413. doi: 10.18632/oncotarget.7189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Papadimitrakopoulou V.A., Mok T.S., Han J.Y., et al. Osimertinib versus platinum-pemetrexed for patients with EGFR T790M advanced NSCLC and progression on a prior EGFR-tyrosine kinase inhibitor: AURA3 overall survival analysis. Ann Oncol. 2020;31:1536–1544. doi: 10.1016/j.annonc.2020.08.2100. [DOI] [PubMed] [Google Scholar]
  • 13.Ramalingam S.S., Vansteenkiste J., Planchard D., et al. Overall survival with osimertinib in untreated, EGFR mutated advanced NSCLC. N Engl J Med. 2020;382:41–50. doi: 10.1056/NEJMoa1913662. [DOI] [PubMed] [Google Scholar]
  • 14.Chmielecki J., Gray J.E., Cheng Y., et al. Candidate mechanisms of acquired resistance to first-line osimertinib in EGFR-mutated advanced non-small cell lung cancer. Nat Commun. 2023;14:1070. doi: 10.1038/s41467-023-35961-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Thress K.S., Paweletz C.P., Felip E., et al. Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harboring EGFR T790M. Nat Med. 2015;21:560–562. doi: 10.1038/nm.3854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chen K., Zhou F., Shen W., et al. Novel mutations on EGFR Leu792 potentially correlate to acquired resistance to osimertinib in advanced NSCLC. J Thorac Oncol. 2017;12:e65–e68. doi: 10.1016/j.jtho.2016.12.024. [DOI] [PubMed] [Google Scholar]
  • 17.Kim T.M., Song A., Kim D.W., et al. Mechanisms of acquired resistance to AZD9291: a mutation-selective, irreversible EGFR inhibitor. J Thorac Oncol. 2015;10:1736–1744. doi: 10.1097/JTO.0000000000000688. [DOI] [PubMed] [Google Scholar]
  • 18.Planchard D., Loriot Y., André F., et al. EGFR-independent mechanisms of acquired resistance to AZD9291 in EGFR T790M-positive NSCLC patients. Ann Oncol. 2015;26:2073–2078. doi: 10.1093/annonc/mdv319. [DOI] [PubMed] [Google Scholar]
  • 19.Wu Y.L., Guarneri V., Voon P.J., et al. Tepotinib plus osimertinib in patients with EGFR-mutated non-small-cell lung cancer with MET amplification following progression on first-line osimertinib (INSIGHT 2): a multicentre, open-label, phase 2 trial. Lancet Oncol. 2024;25:989–1002. doi: 10.1016/S1470-2045(24)00270-5. [DOI] [PubMed] [Google Scholar]
  • 20.Fuchs V., Roisman L., Kian W., et al. The impact of osimertinib' line on clonal evolution in EGFRm NSCLC through NGS-based liquid biopsy and overcoming strategies for resistance. Lung Cancer. 2021;153:126–133. doi: 10.1016/j.lungcan.2020.12.039. [DOI] [PubMed] [Google Scholar]
  • 21.Fukuda S., Suda K., Hamada A., et al. Potential utility of a 4th-generation EGFR-TKI and exploration of resistance mechanisms – an in vitro study. Biomedicines. 2024;12:1412. doi: 10.3390/biomedicines12071412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Kenzerki M.E., Ahmadi M., Mousavi P., Ghafouri-Fard S. MYC and non-small cell lung cancer: a comprehensive review. Human Gene. 2023;37 [Google Scholar]
  • 23.Choi W., Jeong K.C., Park S.Y., et al. MYC amplification-conferred primary resistance to capmatinib in a MET-amplified NSCLC patient: a case report. Transl Lung Cancer Res. 2022;11:1967–1972. doi: 10.21037/tlcr-22-176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Wallbillich N.J., Lu H. Role of C-Myc in lung cancer: progress, challenges, and prospects. Chin Med J Pulm Crit Care Med. 2023;1:129–138. doi: 10.1016/j.pccm.2023.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Garralda E., Beaulieu M.E., Moreno V., et al. MYC targeting by OMO-103 in solid tumors: a phase 1 trial. Nat Med. 2024;30:762–771. doi: 10.1038/s41591-024-02805-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES