LAUSR

Abstract Background Circulating tumor DNA (ctDNA) from liquid biopsy is a source of tumor genetic material for EGFR testing in NSCLC when resistance to I-II generations TKI-therapies occurs or upfront, in absence of tissue biopsy. A multi-regional survey for patients treated with EGFR TKI from January to December 2018 was conducted regarding use of liquid biopsy for NSCLC in clinical practice. Methods Aim of the study was to describe EGFR testing workflow by liquid biopsy in clinical setting. Seven major lung-cancer centers in North Eastern Italy participated to the survey. Results Overall 316 patients (360 samples) have been screened for ctDNA EGFR, with a medium number of 1.1 samples/patient. All institutions reported EGFR as the main gene tested for clinical purposes in plasma samples of NSCLC, other genes (i.e. KRAS/BRAF) were occasionally tested upon oncologist request. Seven out of seven (100%) centers used commercially available real time CE-IVD tests. NGS or droplet digital PCR were used by one center each, as confirmation methods. In all institutions blood drawing was performed in the same hospital of EGFR testing (in 6/7 - 86%- centers in the Medical Oncology Unit), and plasma separated within 2 hours. EGFR testing was performed in the Surgical Pathology Unit and report edited within 24 hours in 3/7 (43%) centers and within 3-5 working days in 4/7 (57%) centers. Among all, 108 (34%) patients were tested at the time of diagnosis with an EGFR mutation rate of 15%. At progression to TKI treatment, 208 (66%) patients were tested. T790M positive rate was 55/122 (45%). Inconclusive cases (negative for both T790M and known actionable mutation) were 86 (41%). All centers declared that histo/cytological re-biopsy was suggested in these cases. Conclusions Real-world experiences about EGFR testing in liquid biopsies revealed homogeneous habits among interviewed centers, according to national guidelines and literatures data. The consistent number of liquid biopsies at the time of diagnosis mainly refers to some reference centers that collect specimens from several institutions. However it might impose a revision of bronchoscopy workflows in order to obtain better samples suitable both for cito-histological diagnosis and molecular profiling. Legal entity responsible for the study The authors. Funding Has not received any funding. Disclosure All authors have declared no conflicts of interest.