A retrospective analysis examining liquid biopsies with tissue biopsies in patients with advanced non–small cell lung cancer.

Abstract

e20683 Background: Non-small cell lung cancer (NSCLC) is genetically diverse, often harboring actionable mutations (AM) that can be targeted. It is critical that providers accurately and quickly identify AM through next-generation sequencing (NGS). Tissue-based biopsies (Tbx) are invasive and have long turnaround times. Blood-based biopsy (Bbx), a less invasive and faster technique, analyzes circulating tumor DNA from blood samples. Our first aim is to report TBx and Bbx concordance rates stratified based on different AM. Our second aim is to report the lead time provided by Bbx compared to Tbx in starting treatment. Methods: This single-institution, retrospective, IRB-approved study analyzed patients treated between January 1, 2016 and December 31, 2025. Only patients whose charts were complete and contained both a Tbx and a Bbx obtained within 30 days of one another at the time of initial diagnosis were included. We included only cases with fully scanned NGS reports available in the electronic medical record—not summaries, interpretations, or partial documentation. Results: 66 patients were analyzed. Median age was 69 years, with 41% male and 59% female. Ethnicities include Asian (26%), African American (9%), Hispanic (3%), and Caucasian (51%). 51% reported current or prior tobacco use. 68% had a tissue biopsy followed by liquid biopsy. 67% had an AM finding on Bbx with predominant reason for ‘negative result’ being ‘ctDNA not detected”; while 55% had an AM finding on Tbx, with most common reason for ‘negative result’ being insufficient tissue. 24% of patients had an AM solely present on Bbx while 9% of patients had an AM solely present on Tbx. Concordance rates between Tbx and Bbx were stratified based on mutations: EGFR exon 19 (95%)(84.9% - 98.7%), EGFR exon 21 (95%)(84.9% - 98.7%), ALK (95%) (84.9% - 98.7%), ROS (98%) (89.7% - 99.9%), MET (92%)(80.6% - 96.8%), BRAF (90%)(78.5% - 95.8%). Time to report for treatment for AM based on Bbx was a median of 8 days while Tbx was 30 days. Amongst the patients with AM finding on Bbx, 39% were started on targeted treatments based on BBx results prior to availability of results from Tbx with a median lead time of 13.5 days. Conclusions: Our study adds valuable real-world data to the existing literature. While concordance between the two modalities is extremely high for all types of AM, our data also shows that mutations can be solely picked up on Bbx but not on Tbx and vice versa. This finding supports the complementary use of both Tbx and Bbx at initial diagnosis as necessary for thoracic oncologists. Discordance in mutation detection by Bbx and Tbx can be due to tumor heterogeneity, sensitivity of the assays, as well as RNA quality and quantity. Since our cohort spanned nearly a decade, discordance could also have been influenced by the evolution of NGS assays and tech. Our data also showcases the lead-time advantage with Bbx, enabling early initiation of treatment.