Accurate RET Fusion Detection in Solid Tumors Using RNA Sequencing Coverage Imbalance Analysis

Accurate detection of oncogenic gene fusions is becoming increasingly important given the availability of highly effective targeted therapies. However, their identification in clinical practice remains challenging due to the rarity of individual events, diversity of partner genes, and variability of breakpoint locations. Conventional approaches such as immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) lack multiplexing capacity and demonstrate variable sensitivity and specificity, while direct identification of fusion transcripts in whole-transcriptome sequencing (RNA-seq) profiles provides broader applicability but limited sensitivity, as fusion junctions are frequently supported by a minimal number of reads or even no reads at all. In this study, a novel approach was employed to accurately detect clinically actionable <i>RET</i> (REarranged during Transfection) fusions. This approach entailed the measurement of the imbalance in RNA-seq read coverage of potential fusion oncogenes at their 3' and 5' exons. A total of 1327 experimental solid tumor RNA-seq profiles were screened, including 154 non-small cell lung cancer and 221 thyroid cancer samples. The <i>RET</i> status was validated in 78 selected cases by targeted NGS and Sanger sequencing. An analysis of the coverage imbalance was conducted, which enabled the accurate discrimination between true and false positive <i>RET</i> fusions. This approach outperformed other methods and yielded 100% sensitivity and specificity with optimized thresholds. The findings were validated using an independent cohort of 79 thyroid cancer cases, confirming the reliability of the results. Among the 18 <i>RET</i> fusion-positive samples, one was identified as an extremely rare case (<i>RUFY3::RET</i>), and two were determined to be novel fusions (<i>FN1::RET</i>, <i>PPP1R21::RET</i>). The findings of this study demonstrate that exon coverage imbalance analysis serves as a robust complement to computational RNA-seq analysis pipelines for the detection of clinically relevant <i>RET</i> fusions.