Background

While DNA NGS can detect clinically actionable fusions in tumors, technical and biological limitations may lead to false negatives, preventing patients from receiving approved therapies associated with outcome benefit. RNA NGS is recommended by ESMO guidelines to maximize fusion detection but is not widely used clinically. In a large, real-world dataset, we quantify the benefit of RNA concurrent with DNA NGS for ALK fusion detection in NSCLC.

Methods

We retrospectively analyzed de-identified stage IIIB-C and IV NSCLC samples sequenced with the Tempus xT assay (DNA and whole-exome capture RNA NGS). ALK fusion prevalence was compared to public DNA-seq-based data from Dana Farber Cancer Institute (DFCI, N=4,497) and Memorial Sloan Kettering Cancer Center (MSKCC, N=5,317) with one-sided proportion tests. ALK fusions were limited to known oncogenic partners detectable by FISH or IHC (EML4, KIF5B, KLC1, and PICALM). Therapeutic adoption was analyzed in cases with ≥90 days of first-line medication data.

Results

Among 7,428 NSCLC patients, 2.9% had ALK fusions (n=217), with 83% (n=180) detected by both DNA and RNA, 15% (n=33) from RNA alone, and 2% (n=4) from DNA alone. ALK fusion prevalence in the Tempus Database was higher than the combined MSKCC and DFCI rate (2.4%, n=236/9,814, P=0.036). To investigate RNA vs DNA discordance, we examined exonic EML4-ALK breakpoint locations (n=32). Most cases had canonical breakpoints in RNA (n=26), where 18 had no DNA support in these breakpoints, suggesting complex ALK rearrangements in DNA. The remaining 8 had limited DNA support but were below the clinical threshold for reporting. Of those with medication data (n=7), 6 patients with EML4-ALK fusions detected by RNA alone received targeted therapy post-testing and 5 remained on therapy for ≥100 days.

Conclusions

This large, real-world dataset demonstrates that RNA NGS increases ALK fusion detection by 18% over DNA alone for patients with advanced NSCLC. This increase is likely to translate to a larger number of patients matched to and receiving targeted therapy. Consequently, RNA NGS should be considered for more widespread adoption in the clinic.

VIEW THE PUBLICATION