Abstract
Purpose – Comprehensive genomic profiling (CGP) is strongly advocated by organizations like ASCO and National Comprehensive Cancer Network (NCCN) to tailor precision-guided therapies for patients with cancer. However, these guidelines often lack specificity regarding panel size or essential features required within CGP panels.

Materials and Methods – A pilot pan-cancer study of patients (N = 85) treated in a value-based community health care setting who received next-generation sequencing (NGS)–based CGP from Tempus AI, Inc and an expanded analysis of de-identified patient records from the Tempus database (N = 465) for patients treated within a community health care network. Rates of potentially actionable findings defined as variants matching US Food and Drug Administration –approved therapies (somatic variants/resistance markers, Tumor Mutational Burden high [TMB-H] or Microsatellite Instability high [MSI-H]), or identification of pathogenic/likely pathogenic germline findings. Enrollment in clinical trials was included in the expanded cohort.

Results – In our pilot cohort, potentially actionable alterations were detected in 49% of patients (n = 42). Of them, 12% (n = 5) of patients had a finding that was identified only by advanced features—including tumor-normal match, RNA sequencing, and/or liquid biopsy reflex testing—and may have been missed by the existing in-network NGS options. In the expanded cohort, potentially actionable alterations were detected in 31% (n = 146) of patients. Of them, 12% (n = 17) had findings identified solely by advanced features, including three patients enrolled in clinical trials.

Conclusion – This study indicates that the CGP expanded treatment options in both the pilot and expanded cohort, with improvements directly attributable to specific features like tumor-normal matching, liquid biopsy reflex ordering, RNA sequencing, and Just-in-Time clinical trial matching.