Oncogenic fusion detection is an essential part of clinical diagnosis and management of non-small cell lung carcinoma. Numerous methods are available for detection of oncogenic fusions in the clinical laboratory, although RNA sequencing has rapidly gained prominence. Accordingly, however, multiple different RNA sequencing assays exist, with diverse methodologies and varying performance characteristics. Here, we report our single-institutional clinical experience with a testing algorithm for non-small cell lung carcinoma that uses amplicon-based DNA/RNA sequencing followed by reflex hybridization-capture based RNA sequencing if the initial testing is negative for oncogenic drivers. A total of 1,211 non-small cell lung carcinoma specimens were received for molecular testing, and 120 (∼10%) were reflexed for hybridization-capture based RNA sequencing. Of the 120 cases tested, oncogenic fusions were identified in 9 and included clinically actionable fusions involving ALK, BRAF, NRG1, NTRK3, ROS1, and RET. None of these fusions were detected by the amplicon-based assay. Review of the 20,900 non-small cell lung cancer cases in the AACR Project Genie v15.1 publicly available database (registration required) revealed that of the 1081 cases harboring fusions, 893 (82.6%) could theoretically be detected by the amplicon-based assay. Overall, this study shows that the addition of reflex hybridization-capture based RNA sequencing could improve detection of rare and novel oncogenic fusions, maximizing patient eligibility for appropriate targeted therapies or clinical trials.