RNA hybrid-capture next-generation sequencing has high sensitivity in identifying known and less characterized oncogenic and likely oncogenic NTRK fusions in a real-world standard-of-care setting.

NTRK1, NTRK2, and NTRK3 gene fusions are rare oncogenic driver alterations found in diverse tumor types of adults and children. They are clinically important biomarkers as tumors harboring these genomic alterations have high response rates to targeted therapy. Routine testing for NTRK fusions and treatment with TRK inhibitors has been recommended in multiple tumor types; however, differences between testing technologies used for detecting NTRK fusions can result in variable likelihoods of identification. To assess the prevalence of NTRK fusions in a real-world standard-of-care setting, we analyzed data from 19,591 FFPE samples encompassing 35 solid tumor types submitted for comprehensive genomic profiling (CGP) as part of routine clinical care. CGP testing included DNA hybrid-capture sequencing to detect small variants, copy number alterations, microsatellite instability (MSI), and tumor mutational burden (TMB). RNA hybrid-capture sequencing was concurrently performed to detect fusions and splice variants. Detected NTRK fusions were categorized as oncogenic, likely oncogenic, or variant of unknown significance (VUS) based on the fusion partner, orientation, and breakpoint position. 73 oncogenic or likely oncogenic NTRK fusions were identified in 69 unique tumor specimens across 19 tumor types for a total cohort prevalence of 0.35%. Tumor types with the highest NTRK fusion prevalence included glioblastoma (1.91%), small intestine (1.32%), and head and neck (0.95%) tumors with other solid tumor types ranging from 0.19% (uterine) to 0.63% (breast). We identified diverse intra and inter-chromosomal partner genes for NTRK fusions across all tumor types. Most NTRK fusions were detected in only one tumor specimen, though some recurrent fusions were noted with ETV6, TPM3, LMNA, EML4, TPR, PEAR1, IRF2BP2, and KANK1 fusion partners. Most NTRK fusions were mutually exclusive from other genomic driver alterations, however, almost a third of tumor specimens (29%) contained at least one co-occurring genomic driver, which may affect treatment decisions. The high prevalence of oncogenic and likely oncogenic NTRK fusions detected in our analysis suggests that RNA hybrid-capture-based sequencing for fusion detection is a highly sensitive method for identifying clinically meaningful known and novel NTRK fusions, which may be missed with other detection methods, directly impacting therapeutic options and patient outcomes.