Real-world analysis of the EGFR mutation detection in plasma ctDNA from patients with non-small cell lung cancer.

e21069 Background: Plasma sample has emerged as a promising surrogate sample for EGFR mutation detection in advanced non-small cell lung cancer (NSCLC). In clinical practice, whether EGFR variants in pretreatment plasma ctDNA of advanced lung cancer can predict prognosis in addition to guiding targeted therapy remains to be further explored. Methods: In this study, we retrospectively enrolled 315 NSCLC patients from Beijing Chest Hospital, 94.9% of whom were in stage IIIB-IV. EGFR gene mutation data from tissue detected by ARMS-PCR (Amoy Diagnostics, Xiamen, China) and paired plasma samples within one month of admission detected by SuperARMS-PCR (Amoy Diagnostics, Xiamen, China) were collected. The patients’ clinicopathological characteristics and treatment information were also collected for further analysis. Differences in PFS of targeted therapy between patients with positive and negative pretreatment plasma ctDNA EGFR mutation were compared. Results: EGFR-positive rate of tissue and plasma ctDNA was 65.1% (205/315) and 43.8% (138/315), respectively. L858R, 19-DEL, T790M, G719X, 20-INS, and L861Q mutations were detected in both samples. The sensitivity, specificity, PPV, NPV and concordance of ctDNA detection was 65.4%,96.4%, 97.1%, 59.9%, and 76.2%, respectively. In the study, patients’ pathological subtype was the factor affected plasma ctDNA EGFR alteration detection accuracy. The concordance in adenocarcinoma patients was obviously higher than that in non-adenocarcinoma (p = 0.02). 92 treatment naïve advanced patients with tissue EGFR mutation were received first and second-generation EGFR tyrosine kinase inhibitors (TKIs) therapies. 59.8% (55/92) patients also had EGFR mutation in plasma ctDNA, 40.2% (37/92) patients had no detectable EGFR mutations in plasma ctDNA. The median PFS in plasma ctDNA EGFR positive patients was 7.0 months versus 14.0 month in plasma ctDNA EGFR negative patients, with significantly difference (p = 0.004). Conclusions: Plasma sample could be a surrogate sample for molecular test when tissue sample was unavailable. The poor prognosis was observed in treatment naïve advanced NSCLC patients with baseline plasma EGFR mutation.