Real world utilization of comprehensive genomic profiling (CGP) in head and neck squamous cell carcinoma (SCCHN).

Abstract

6021 Background: In Head and Neck Cancer, practice trends of utilizing comprehensive genomic profiling (CGP) vary depending on physician preference and test availability. Here, we examined the impact of CGP on treatment decision-making and patient outcomes in SCCHN at a single institution. Methods: Patients with suspected SCCHN underwent tumor-based CGP using the UCSF500 or TEMPUS xT NGS panel. In addition to demographic and clinicopathologic data, we obtained the reason for CGP when noted, disease stage and line of therapy at the time of CGP, CGP results, and post-CGP treatments. OncoKB was used to annotate actionable mutations in Head and Neck Cancer, in addition to other biomarkers when available (microsatellite instability (MSI), tumor mutation burden (TMB), and EGFR amplification). Results: Between January 2016 and December 2023, 301 unique patient tumor specimens underwent CGP in the setting of suspected or known SCCHN. Sequencing was performed to refine a diagnosis in 57 samples (18.9%). CGP influenced the determination between cutaneous and mucosal primary in 35 samples, clarified the histology in 9, and clarified the origin of a metastatic lesion in 13. For patients with confirmed SCCHN (N = 246), CGP was performed for therapy selection in 193 (78.5%) of patients. Of these, 19 (9.8%) patients were TMB-high (TMB>10), 2 (1.0%) patients were MSI-high, 77 (39.9%) patients were found to have actionable genomic alterations, and clinicians had access to OncoKB-specified agents for 50 (25.9%) of the patients. Anti-PD-1 antibody therapy was administered to 15/20 (75%) of TMB-high and/or MSI-high patients compared to 38 (53.5%) with low TMB or MSI-stable disease. The most common actionable individual gene alterations were in CDKN2A (37.3%), PIK3CA (14.5%), PTEN (7.2%), FBXW7 (6.0%), and HRAS (6.0%) genes. Six of 77 (7.8%) patients with actionable genomic alterations received targeted therapy based on drug availability and the clinical discretion of the treating physician.In these, the overall response rate was 50% (3/6), and the median PFS was 2.6 months. Thus, at this site, the number needed to treat to gain one response in SCCHN patients undergoing sequencing for oncogene-targeted options was 64.5 patients. Conclusions: CGP in HNSCC was used most commonly to increase diagnostic accuracy, often providing diagnostic information that guided therapeutic decision-making. It was used infrequently for therapy selection, but in the few patients selected for targeted therapy based on NGS, objective responses were observed.