Comparison between liquid biopsy and tissue NGS in two cohorts of advanced NSCLC

e21043 Background: Molecular testing is mandatory in advanced non-small cell lung cancer (NSCLC) due to prognostic and therapeutic implications. Correlation between blood and tissue next-generation sequencing (NGS) has recently been demonstrated. Methods: A retrospective observational study of two cohorts of NSCLC patients who underwent liquid biopsy at the time of advanced disease diagnosis was conducted. In cohort 1 (N = 129), liquid biopsy was performed using a 74 genes panel (Guardant360 CDx). In cohort 2 (N = 149), NGS was performed in peripheral blood using a 63 genes panel (FoundationOne Liquid CDx). For tissue NGS analysis, Oncomine Precision Assay and Oncomine Focus Assay, with 50 and 52 genes, respectively, were performed. Data were analyzed in both cohorts independently, and compared by statistical descriptive methods. Results: In cohort 1, 74/129 (57.4%) were men and 55/129 (42.6%) were non-smokers or had a ≤ 20-pack-year tobacco history. The most frequent histology was adenocarcinoma (99/129; 76.7%). In 14/129 (10.9%) patients, a second determination of NGS in peripheral blood was necessary and 1/129 (0.8%) had a total of three determinations. Tissue NGS was not performed in 32/129 (24.8%) because of insufficient sample. In cohort 2, 99/149 (66.4%) were men and 48/149 (32.2%) were non-smokers or had a ≤ 20-pack-year tobacco history. The most frequent histology was adenocarcinoma (99/149; 66.4%). Tissue NGS was not performed in 40/149 (26.8%) due to insufficient sample (26/40; 65%) or technical difficulties to perform a biopsy (14/40; 35%). In cohort 1, pathogenic mutations were detected in 119/129 (92.2%) and in 62/97 (63.9%) patients (p < 0.001), with median time to obtain results of 11 (IQR 9-14) and 29 (IQR 22-39) days for blood and tissue NGS, respectively (p < 0.001). Actionable mutations included alterations of levels I-II according to ESCAT and were found in 40/129 (31%) and 33/97 (34%) patients using blood and tissue NGS, respectively (p = 0.63). In cohort 2, circulating tumor DNA was not detected in 5/149 (3.5%) patients. Pathogenic mutations were found in 123/149 (82.6%) and in 71/109 (65.1%) patients (p = 0.001), with median time to obtain results of 15 (IQR 14-15) and 27 (IQR 16-27) days for blood and tissue NGS, respectively (p < 0.001). Actionable mutations were detected in 41/149 (27.5%) and 41/109 (37.6%) using blood and tissue NGS, respectively (p = 0.08). Targeted therapy was given to 31/129 (24%) and 39/149 (26.2%) patients in cohort 1 and 2, respectively. Conclusions: To our knowledge, this is the first retrospective observational study where liquid biopsy and tissue NGS are compared in two different cohorts of patients with advanced NSCLC. Time to obtain results was shorter for blood NGS than tissue sequencing, and liquid biopsy provided molecular information in cases in which tissue NGS is not available. Further studies to identify patients in which liquid biopsy could replace tissue analysis are needed.