Key considerations for comprehensive validation of an RNA fusion NGS panel

Abstract Objectives Validation of RNA-based NGS assays for the detection of therapeutically targetable gene fusions is challenging. Here, we report systematic validation and quality control monitoring of our targeted fusion panel for the detection of 17 clinically relevant fusion transcripts across several tumor types. We implemented this RNA Fusion Panel as a reflex test for tumors lacking DNA driver mutations. Design Forty-four formalin-fixed, paraffin-embedded (FFPE) or fresh-frozen lung, brain, soft tissue and skin tumors were used to determine the accuracy of the assay. Additional fusion-positive specimens and a calibrated reference standard were used to establish the precision, reproducibility and sensitivity of the assay. All aspects of the validation, including quality control metrics, were performed according to New York State guidelines. Results For the RNA fusion panel, accuracy, reproducibility and precision studies were above 99%. Reproducibility and sensitivity studies with the reference standard were helpful in identifying inconsistencies. The limit of detection for most RNA fusion transcripts was 50 copies. Application of the RNA fusion assay as a reflex test to 450 tumor samples lacking DNA driver mutations resulted in a 10% increase in diagnostic yield with minimal additional processing time. Conclusions The validated RNA fusion panel provides clinical utility in therapy selection for patients with solid tumors. By using a sequential testing approach, the RNA fusion assay complements the DNA hotspot assay in identifying clinically relevant variants across many tumor types with minimal additional increase in processing time.