Novel Bronchoscopy Method for Molecular Profiling of Lung Cancer: Targeted Washing Technique
Article information
Abstract
Purpose
There have been efforts to find alternative samples other than standard samples of tissue or plasma for mutational analyses for lung cancer patients. However, no other sample or technique has replaced the mutational analyses using standard samples. In this prospective study, we assessed a novel bronchoscopy method, named as targeted washing technique, for detecting the epidermal growth factor receptor (EGFR) mutation.
Materials and Methods
A 3.0-mm ultrathin bronchoscope was precisely navigated to the target lung lesion with the assistance of virtual bronchoscopic navigation and fluoroscopy. Once the bronchoscope is placed in front of target lung lesion, 0.9% normal saline was instilled for targeted washing. EGFR testing using targeted washing fluid (TWF) was compared to standard methods using plasma or tumor tissue.
Results
In 41 TWF samples, the T790M mutation was detected in tissue, plasma, and TWF samples at rates of 22.0%, 9.8%, and 29.3%, respectively. The overall EGFR T790M detection rate using tissue, plasma, or TWF samples was 36.6%, with TWF samples increasing the T790M mutation detection rate by up to 10%. The accuracy of T790M mutation detection using TWF sample was 82.9% compared with standard samples. Four patients were found to have the EGFR T790M mutation solely through EGFR testing using TWF, which repeated rebiopsies using either plasma or tissue finally confirmed to have the T790M mutation.
Conclusion
We demonstrated the clinical potential of targeted washing technique for molecular testing, which can be a good option to overcome spatial heterogeneity, low sensitivity of plasma sample or technical limitations in collecting tumor tissues.
Introduction
The dawn of precision medicine has ushered in a new era, marked by a thorough understanding of oncogene addiction and the consequent emergence of targeted therapeutic agents [1,2]. Notably, there has been a quantum leap in the treatment landscape of non–small-cell lung cancer (NSCLC), particularly concerning genetic aberrations such as epidermal growth factor receptor (EGFR) mutations, anaplastic lymphoma kinase rearrangements, ROS1 rearrangements, and beyond [3]. The introduction of novel targeted agents has also given rise to concurrent challenges in companion diagnostics and the selection of appropriate samples [4].
The introduction of liquid biopsy techniques using plasma samples has streamlined the process of obtaining mutational information, providing a more accessible alternative compared to analyses conducted using tissue biopsy [5]. Specifically, liquid biopsy has been employed for detecting EGFR mutations in NSCLC patients, which was widely utilized for identifying T790M mutations after the first-line treatment with first or second-generation EGFR–tyrosine kinase inhibitors (EGFR-TKIs) [6]. However, there have been concerns about relatively high false-negative results in liquid biopsy using plasma samples [7]. In contrast, tissue biopsy has raised concerns regarding the invasiveness of procedures and limitations in overcoming spatial heterogeneity [8,9].
To overcome the limitations or drawbacks of mutational analyses using standard samples of plasma or malignant tumor tissue, there have been efforts to find alternative samples for mutational analyses for NSCLC, such as sputum or pleural fluid [10-12]. However, no other sample or technique has replaced the mutational analyses using plasma or tissue. In this study, we assessed a novel bronchoscopy method of targeted washing technique for detecting the EGFR mutation.
Materials and Methods
1. Study subjects
From April 2022 to September 2023, a prospective observational pilot study was conducted to assess the feasibility of a targeted washing technique for detecting EGFR mutation at Pusan National University Hospital, Busan, a university-affiliated, tertiary referral hospital in Busan, South Korea. The inclusion criteria were as follows: (1) patients aged 20 years or older, (2) individuals with advanced NSCLC with EGFR mutations, (3) confirmation of progression based on Response Evaluation Criteria in Solid Tumor (RECIST) criteria after treatment with EGFR-TKIs, and (4) absence of contraindications for bronchoscopy.
2. Targeted washing technique
A representative case of targeted washing technique is shown in Fig. 1. Before bronchoscopy, local anesthesia was applied to the oropharynx using 4% lidocaine with an atomizer. Subsequently, all study participants received intravenous midazolam and fentanyl to achieve conscious sedation. Initially, a conventional bronchoscopy was conducted using a 4-mm flexible bronchoscope (BF-P260F, Olympus) to administer lidocaine for local anesthesia to the tracheobronchial tree. If no endobronchial lesion was identified during conventional bronchoscopy, an ultrathin bronchoscope (BF-MP190F, Olympus) with a diameter of 3.0-mm was navigated precisely to the target lung lesion as close as possible based on guidance from thin-section chest computed tomography (CT) images and with the assistance of virtual bronchoscopic navigation (LungPoint, Broncus Medical). To accurately locate the target lung lesion, a radial probe endobronchial ultrasound (EBUS) (UM-S20-17S, Olympus) was advanced through a 1.7-mm-diameter working channel of the ultrathin bronchoscope under X-ray fluoroscopy guidance. Once the target lung lesion was identified, the radial probe EBUS was withdrawn, and 3 mL of 0.9% normal saline was instilled for targeted washing. After a 3-second interval, the washing fluid was retrieved by gently pulling the bronchoscopy proximally. Targeted washing procedures were repeated until more than 5 mL of washing fluid had been collected. The scheme of targeted washing technique is presented in Fig. 2.
Representative case of targeted washing in patients who progressed from the treatment with afatinib. (A) Lung cancer with a size of 31×17 mm (arrow) was found in the right middle lobe with multiple bone metastases. (B) After 30 months of afatinib treatment, disease progression was identified with the development of numerous hematogenous metastases in both lungs and an increase in the size of bone metastases. At the time of progression, right middle lobe mass was found to be 24×11 mm (arrow). (C) Three-dimensional images of virtual bronchoscopy navigation were created to precisely locate the target lung lesion. (D) Targeted washing was performed at the medial segment of right middle lobe for epidermal growth factor receptor (EGFR) testing using a 3.0-mm-sized ultrathin bronchoscope under the assistance of fluoroscopy. The T790M mutation was detected in EGFR testing using targeted washing fluid, whereas neither the tissue nor plasma sample could detect the T790M mutation.
Schema of targeted washing technique. (A) Initially, an ultrathin bronchoscope with a diameter of 3.0 mm (indicated by the black arrow) was precisely navigated to the target lung lesion as closely as possible, based on thin-section chest computed tomography images and with the assistance of virtual bronchoscopic navigation and fluoroscopy. (B) Subsequently, a radial probe endobronchial ultrasound (EBUS) (red line) was advanced through the working channel of the ultrathin bronchoscope to accurately locate the target lung lesion. (C) Finally, the radial probe EBUS was withdrawn, and 3 mL of 0.9% normal saline (blue lines) was instilled. After a 3-second interval, the washing fluid was retrieved for collecting cell-free DNA derived from the lung tumor. (D) Fluoroscopy image during targeted washing; an ultrathin bronchoscope is placed in front of the target lung lesion (black arrows).
Following the collection of targeted washing fluid (TWF), bronchoscopic biopsies of the primary lung tumor were performed using either forceps biopsy or cryobiopsy based on the physicians’ discretion, aiming for histologic diagnosis and tissue next generation sequencing (NGS). In cases where the target lung lesion could not be identified, EBUS-guided transbronchial needle aspiration was employed to sample mediastinal lymph nodes suspected of metastasis based on axial CT scans.
3. DNA isolation from TWF and droplet digital polymerase chain reaction
Cell-free DNA (cfDNA) was extracted from TWF using QIAamp DSP Circulating NA Kit (Qiagen) according to the manufacturer’s instructions. Using cfDNA from TWF, droplet digital polymerase chain reaction (ddPCR) was conducted using a ddPCR system (QX200 Droplet Digital PCR System, Bio-Rad) and an EGFR mutation analysis kit (Droplex EGFR Mutation Test v2, Gencurix) following the manufacturer’s recommended protocol. Detailed information of EGFR testing using TWF samples was provided in Supplementary Materials.
4. Conventional EGFR mutation analysis
Using tissue specimen, EGFR mutation tests were performed using an EGFR Mutation Detection Kit (PNA clamp, Panagene). Cobas EGFR Mutation Test v2 (cobas, Roche Molecular System) was used on the plasma sample.
5. Statistical analysis
Categorical variables were presented as number (%) and continuous variables were described as mean (range). Detection rate of molecular test was calculated as the number of cases with the presence of oncogenic alteration divided by the total number of cases. The detection of T790M mutation through standard EGFR testing is defined as any identification of EGFR T790M mutation utilizing tissue or plasma specimens. Statistical analyses were conducted using R software for Windows ver. 4.2.3 (R Foundation for Statistical Computing).
Results
1. Study patients
During the study period, a total of 41 TWF samples were collected from 34 patients. Among them, seven patients underwent repeated re-biopsy after the progression of subsequent therapy. The clinical characteristics of the 34 patients are summarized in Table 1. The median age was 72 years (range, 50 to 86 years), and 79.4% were female. The mean longest diameter of primary lung cancer was 41 mm (range, 9 to 100 mm). Second-generation EGFR-TKIs were used in 67.7% of the study subjects, while 29.4% received first-generation EGFR-TKIs. One patient was treated with lazertinib, a third-generation EGFR-TKI, as part of a clinical trial.
Patient characteristics
2. Bronchoscopy results
For conscious sedation, midazolam (mean, 3.7 mg; range, 2 to 7 mg) and fentanyl (mean, 53 μg; range, 25 to 125 μg) were administered in 40 cases of bronchoscopy, while one received 90 mg of propofol. The mean volumes of instilled and retrieved normal saline were 15.4 mL (range, 10 to 20 mL) and 8.2 mL (range, 5 to 15 mL), respectively. After the collection of TWF, forceps biopsy was performed for the primary lung tumor in 21 cases (51.2%), cryobiopsy in five cases (12.2%), and a combination of forceps biopsy and cryobiopsy in 10 cases (24.4%) out of a total of 41 bronchoscopies. Using radial probe EBUS, primary lung tumor could not be detected in five patients (12.2%); three received EBUS-guided transbronchial needle aspiration for mediastinal lymph node and only targeted washings were done without tissue sampling in two patients. As results, malignant tumor cells were successfully collected in 35 patients (85.4%), whereas tumor tissue could not be collected in six patients (14.6%).
Mean duration of bronchoscopy was 22.5 minutes (range, 6.3 to 44.6 minutes). There was no immediate complication after targeted washing of primary tumor. Balloon hemostasis was performed in two patients who received transbronchial cryobiopsy for peripheral lung lesion. None of patients suffered iatrogenic pneumothorax, procedure-related infection, or life-threatening complication.
3. EGFR testing results
Table 2 presents the results of EGFR testing across 41 cases. Using tissue samples, EGFR mutation was identified in 80.5% (33 out of 41) of cases, while the T790M mutation was detected in 22.0% (9 out of 41). In plasma samples, EGFR mutation positivity was observed in 56.1% (23 out of 41) of cases, with T790M mutations detected in 9.8% (4 out of 41). Standard EGFR testing results, combining tissue and plasma EGFR testing, indicated that 26.8% showed positive T790M mutation (11 out of 41). Out of the total 41 TWF samples, EGFR mutations were detected in 70.7% (29 out of 41), while the T790M mutation was identified in 29.3% (12 out of 41) of cases. The overall EGFR T790M detection rate using tissue, plasma, or TWF samples was 36.6% (15 out of 41), with EGFR testing using TWF samples increasing the T790M mutation detection rate by up to 10%. Detailed yield of each bronchoscopic procedure is presented in S1 Table.
EGFR testing results
Table 3 shows an agreement analysis of the detection of EGFR T790M mutation between the TWF and standard specimens. The sensitivity and specificity in the T790M mutation detection using the TWF sample were 72.7% (95% CI, 43.4 to 90.3) and 86.7% (95% CI, 70.3 to 94.7), respectively. The accuracy of T790M mutation detection of the TWF sample was 82.9% (95% CI, 68.7 to 91.5) compared with standard samples.
Agreement analysis of the detection of EGFR T790M mutation between the TWF and standard specimens
4. Early detection of T790M mutation in TWF sample
Four patients were found to have the EGFR T790M mutation solely through EGFR testing using TWF samples. Subsequently, all of them underwent cytotoxic chemotherapy, and repeated rebiopsies using either plasma or tissue were performed after progression from chemotherapy. As a result, all four patients were confirmed to have the T790M mutation in tissue or plasma samples. Consequently, their treatment was changed to third-generation EGFR-TKIs (osimertinib or lazertinib).
Discussion
As personalized medicine in the oncology field develops, re-biopsy is essential in the treatment process [13-15]. However, tissue re-biopsy is often difficult due to various situations, such as fibrosis of cancer due to previous treatment or deterioration of patients’ performance [7,16]. Uozu et al. [8] reported that 45% of patients were found to encounter difficulty in collecting tumor tissue when the tumor had progressed after first-line EGFR-TKI treatment based on RECIST criteria. Similarly, 14.6% of cases (6 out of 41) were unable to collect appropriate tissue samples for EGFR testing in the present study. Moreover, molecular testing using plasma samples is simple and least invasive; however, its low sensitivity has been regarded as a significant limitation [15,17]. Therefore, there has been a great need for a more sensitive and less invasive method for molecular testing, including EGFR testing, comparable to tissue or plasma samples. Our findings indicated that TWF samples for EGFR testing are less invasive and yield higher detection rates compared to standard specimens for the detection of the T790M mutation (T790M detection rate: tissue, 22.0%; plasma, 9.8%; and TWF, 29.3%).
We previously reported that the diagnostic yield of extracellular vesicles (EVs) from bronchial washing fluid samples [18]. In 55 bronchial washing fluid-derived samples, the overall detection sensitivity of EGFR mutations was found to be 90% with 100% specificity. Based on these findings, we conclude that EVs from bronchial washing fluid represent a promising source for liquid biopsy. However, the lack of standardized protocols and the variability between different isolation techniques make it difficult to apply EVs in clinical practices yet [19-22]. Furthermore, the process involves the cumbersome steps of isolating EVs from samples and subsequently extracting DNA. Therefore, we isolated cfDNA directly from TWF and employed ddPCR for the detection of EGFR mutations in this study, which was found to be an effective method for detecting the EGFR mutation.
There have been previous reports in conventional bronchial washing for detecting EGFR mutation. Lee et al. [23] showed that ROC for sensitivity was 0.895 for bronchial washing fluid detecting sensitive EGFR mutations (19del or L858R). Zhang et al. [24] also demonstrated the sensitivity of EGFR mutation detected in bronchial washing fluid was 92.5%. Additionally, Roncarati et al. [25] performed NGS analysis using bronchial washing fluid and showed a concordance rate of over 90% when compared to tissue [25]. However, little is known about the utility of bronchial washing fluid samples for re-biopsy, such as detecting the EGFR T790M mutation. Target agents or immunochemotherapy has been demonstrated to effectively shrink tumor volume compared to the conventional chemotherapy [26-28]; therefore, even after progression from first-line treatment, the overall tumor volume is often much less compared to the initial presentation. This could potentially affect the sensitivity and specificity of molecular tests using conventional re-biopsy specimens, such as plasma or tumor tissue. Our results demonstrate that targeted washing techniques are a promising method for collecting cfDNA for molecular testing, even in situations of lower tumor burden during re-biopsy.
Additionally, we noticed the early detection of T790M from TWF in the present study. Strictly speaking, it corresponds to a false positive when assessed based on results from tissue or plasma samples. However, given the subsequent confirmation of the T790M mutation in tissue or plasma specimens, it is deemed highly probable that early detection occurred in the TWF, in which case the aforementioned results align with the mutation’s presence. Many studies reported that tumors can secrete circulating tumor DNA (ctDNA) into the bloodstream before they are visible on imaging, and symptoms of disease are detected [29,30]. Our results suggest that mutational testing using TWF can be a reasonable option for overcoming spatial heterogeneity, low sensitivity of plasma samples, or technical limitations in collecting tissue samples.
Traditionally, deep washing has been utilized for various lung diseases, such as tuberculosis; however, the bronchoscope used for deep washing has typically been a conventional bronchoscope with a diameter of 5 to 6 mm. We developed a targeted technique for acquiring ctDNA from areas surrounding lung tumors, in which the distal tip of the bronchoscope must be positioned directly in front of the tumor, as shown in Fig. 2. To achieve precise localization of the lung lesion, we employed an ultrathin bronchoscope with a 3 mm diameter in combination with radial probe EBUS and fluoroscopy. Regarding the volume of saline instillation, we repeatedly instilled 3 mL of 0.9% normal saline, as the bronchial washing was performed in peripheral lesions where the bronchoscope was positioned just before the lung tumor, leaving very limited space within the bronchus. In our previous experience, instilling a large volume of saline led to endobronchial bleeding, likely due to an abrupt increase in pressure within the peritumoral space and bronchial tree. To mitigate this risk, we opted for repeated instillations of 3 mL of 0.9% normal saline.
This study had several limitations. First, this was a single-center study with available samples from only 41 cases. EGFR T790M was detected exclusively in TWF but not in tissue samples in six cases (14.6%). Due to the small number of study subjects, determining the specific candidates who could benefit from TWF was not feasible. Further studies are required to identify patients who may derive the most benefit from this procedure. Second, the methods of EGFR detection varied depending on the tissue, plasma, and TWF sample. Third, our data showed a lower rate of T790M mutation than other previous studies. It may be explained that the L858R mutation accounts higher portion of our data. Finally, the utility of TWF samples was solely assessed in EGFR testing; hence, further studies are warranted to validate TWF samples for various oncogenic drivers, including next-generation sequencing.
In conclusion, our distinctive technology, named as targeted washing technique, for molecular testing in NSCLC patients showed less invasive, high yield of EGFR T790M mutation detection compared with standard methods. We anticipated that our technique can be readily applied in various kinds of molecular testing in NSCLC, not just EGFR mutation.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
The study and protocol were approved by the Institutional Review Board of Pusan National University Hospital (No. 2204-006-113) and registered as the Clinical Trials.Gov (NCT05517083). The study was conducted in accordance with the principles of the Declaration of Helsinki and written informed consent was obtained from all study participants.
Author Contributions
Conceived and designed the analysis: Kim MH, Eom JS.
Collected the data: Kim MH, Seong H, Jang H, Kim S, Yoo W, Kim SH, Mok J, Lee K, Kim KU, Lee MK.
Contributed data or analysis tools: Kim MH, Seong H, Kim SH.
Performed the analysis: Kim MH, Seong H, Kim SH, Eom JS.
Wrote the paper: Kim MH, Eom JS.
Conflicts of Interest
Conflict of interest relevant to this article was not reported. Yuhan Cooperation provided research funding, but had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Funding
The present study was supported by Yuhan Cooperation.
Acknowledgments
We would like to thank Editage (www.editage.co.kr) for English language editing.