Features of Failed Genomic Testing in Advanced NSCLC Specimens

Abstract Introduction/Objective Lung cancer continues to be the leading cause of cancer death in the United States. For individuals with advanced lung cancer, genomic testing has the potential to direct patients and their care teams toward a targeted therapy or a specific clinical trial. In some instances, however, submitted specimens fail to meet test standards or do not survive the analytic process. The implications of a failed genomic test for patients can be grave. In this study, we aim to identify common features among non-small cell lung cancer (NSCLC) samples submitted from our medical center that failed FoundationOne testing (Foundation Medicine Inc. Cambridge, MA). Methods From all 366 NSCLC samples listed in the FoundationOne database, we identified 31 (8.5%) unique and accessible accession numbers submitted from our institution that failed processing between 2013-2018. These 31 samples were compared for common features, including NSCLC subtype, tumor location, sampling method, tumor size, submitting team, and reported error type. Results From all of our samples, 45% (14/31) were adenocarcinoma, 29% (9/31) were squamous cell carcinoma (SCC), 22% (7/31) were NSCLC otherwise unspecified, and 3% (1/31) was adenoid cystic carcinoma. The majority of samples (20/31) were sampled from primary sites including core biopsies (n=9), FNA (n=5), lobectomy specimens (n=4), and bronchial washings (n=2). The remaining metastatic samples came from sites including lymph node, bone, brain, and adrenal gland. Per FoundationOne: the majority of samples (27, 87%) failed following sequencing, while the remaining four (4, 13%) samples failed in the analytic phase. Conclusion At this time, it remains unclear why many of these samples failed sequencing. Our next steps include comparing the degree of necrosis from samples and comparing our failed sample pool to a successful sample pool of equal size to remove potential confounding factors.