This study showed that BWF liquid biopsy samples more reliably reflected the tumor mEGFR status than did plasma samples when using ddPCR. Additionally, BWF could be used to analyze both L858R and E19del mutations and produced more sensitive results than did plasma at different stages of lung cancer.
Although tissue biopsy is the gold standard for molecular genotyping in lung cancer, liquid biopsy may play an important role as a complementary method for targeted gene detection and the prediction of the clinical course or outcome [15, 16], as well as for the detection of lung cancer at an early stage [17]. Liquid biopsy is a relatively non-invasive, safe, and simple procedure. Therefore, it can be performed in patients with poor medical conditions or when the location or size of lung cancer makes biopsy difficult.
Many studies have shown the feasibility of liquid biopsy for lung cancer, but some limitations remain [13, 18]. Blood samples are mainly used for liquid biopsy, including determination of ctDNA, circulating tumor cells, platelets, exosomes, and microRNAs, which provide specific genetic information on the tumor. Although the proportion of ctDNA can vary depending on the tumor burden, stage, vascularization, and biological features, the value is generally only 0.1–1.0%. Furthermore, ctDNA has a relatively short half-life, from approximately 16 to 150 min [19]. Thus, if sample processing is delayed, the experimental results would not reflect the precise ctDNA level. To overcome this limitation, we froze BWF and plasma within 3 and 8 h of collecting samples, respectively, and used Streck tubes, which maintain the stability of cfDNA for up to 14 days and CTC for up to 7 days. Recently, Krug el al. [20] demonstrated that compared with that of ctDNA, combined use of exosomal RNA (exoRNA) and cfDNA allowed the detection of activating EGFR mutations and T790M mutation with improved sensitivity (96% and 90%, respectively). In particular, this approach resulted in the largest improvement in sensitivity (26% to 74%) in patients with intrathoracic metastatic diseases. Therefore, measuring exoRNA and ctDNA in BWF specimens may greatly improve the detection of EGFR mutations, even at an early stage of lung cancer; however, further studies are needed to evaluate this possibility.
BAL plays a supporting role in the diagnosis of lung cancer [21]. Since the 1980s, various studies have shown the usefulness of BAL in the diagnosis of lung malignancies [22]. BAL showed a diagnostic yield of 33–90% in diffuse malignant pulmonary infiltrates, although the value differed depending on the cancer type. In particular, in NSCLC, such as squamous cell carcinoma and adenocarcinoma, the diagnostic yield was 50% and 77%, respectively [22]. Park et al. [23] suggested that BALF might be effective for determining the EGFR mutation status. Although their study involved a small number of subjects (n = 20), a high concordance rate (91.7%) was observed between BALF and tissue for detecting EGFR mutations using PNA-mediated PCR clamping and PANAMutyper with fluorescence melting curve analysis. However, there were only three patients with early-stage lung cancer, and they did not show any difference in the detection rate between BALF and plasma. Our study included 38 patients with an early-stage lung cancer, and the sensitivity of mEGFR detection was 65% using BWF and only 15.0% using plasma. The use of BWF showed great improvements in the sensitivity and diagnostic yield compared with that of the plasma in the case of patients with an early stage of lung cancer. Therefore, we suggest that BWF rather than plasma be used to detect mutational variations, regardless of the lung cancer stage.
There are various methods of mEGFR detection in lung cancer using liquid biopsy samples. ddPCR, which is based on the generation of ~20,000 droplets, is one of the powerful advanced techniques for detecting rare gene mutations. Sacher el al. [24] reported a sensitivity of 82% and 74% in the detection of E19del and L858R mutations, respectively, via ddPCR using liquid biopsy samples from patients with advanced lung cancer, and Thress el al. [25] reported a sensitivity of 90% and a specificity 100% in the case of the L858R mutation. In our study, ddPCR also showed high sensitivity and specificity using both plasma and BWF for the detection of mEGFRs in advanced-stage cancer.
Our study showed that BWF could be substituted for tissue biopsy samples to confirm EGFR mutations, and the use of BWF may shorten the time from the diagnosis to treatment by avoiding delays for biopsy and confirmation of biopsy results. BWF specimen collection is easier and safer in comparison with lung biopsy because only simple bronchoscopy needs to be performed in the former, for which hospitalization is not required.
Additionally, we investigated the relationship between tumor size and the mEGFR detection rate. Although a significant agreement between liquid biopsy and tissue biopsy samples was expected regarding the mEGFR status for larger-sized tumors, no significant association was noted between the tumor size and mEGFR detection rate for liquid biopsy and tissue biopsy samples (data not shown). These findings suggest that the diagnostic yield of ddPCR using BWF liquid biopsy samples mainly depends on the disease stage rather than on tumor size. However, further studies with larger sample sizes are needed.