Feasibility of Point-of-Care Genomic Profiling in the Diagnosis and Treatment of Cancer of Unknown Primary

INTRODUCTION: Cancer of unknown primary remains a challenging clinical entity. Despite receiving empiric chemotherapy, median overall survival is approximately 6-12 months. Site-specific therapy based on molecular characterization has been shown to improve outcomes; however, feasibility outside of clinical trials, especially in community centers, is lacking. This study explores the application of rapid next-generation sequencing in defining cancer of unknown primary and to identify therapeutic biomarkers. METHODS: A retrospective chart review was performed by identifying pathological samples designated cancer of unknown primary. Next-generation sequencing testing was based on an automated workflow utilizing the Genexus integrated sequencer, validated for clinical use. Genomic profiling was further integrated within a routine immunohistochemistry service, with results reported directly by anatomic pathologists. RESULTS: Between October 2020 and October 2021, 578 solid tumor samples underwent genomic profiling. Among this cohort, 40 were selected based on an initial diagnosis of cancer of unknown primary. The median (range) age at diagnosis was 70 (42-85) and 23 (57%) were female. Genomic data were used to support a site-specific diagnosis in 6 patients (15%). Median turnaround time was 3 business days (IQR: 1-5). Most common alterations identified were KRAS (35%), CDKN2A (15%), TP53 (15%), and ERBB2 (12%). Actionable molecular targeted therapies were identified in 23 (57%) patients, including alterations in BRAF, CDKN2A, ERBB2, FGFR2, IDH1, and KRAS. Immunotherapy-sensitizing mismatch repair deficiency was identified in 1 patient. CONCLUSION: This study supports the adoption of rapid next-generation sequencing among patients with cancer of unknown primary. We also demonstrate the feasibility of integration of genomic profiling with diagnostic histopathology and immunohistochemistry in a community practice setting. Diagnostic algorithms incorporating genomic profiling to better define cancer of unknown primary should be considered for future study.