Improving fusion call confidence and reliability through an optimized process in low quality RNA from formalin-fixed, paraffin-embedded samples

Abstract

Abstract Identifying fusion genes in solid tumors is crucial for precision diagnosis and treatment of cancer patients. However, poor RNA quality may pose a major challenge to the reliability of fusion detection. In this study, an optimized RNA fusion detection method using targeted next-generation sequencing was developed and validated to detect gene fusions in solid tumors using formalin-fixed, paraffin-embedded (FFPE) samples, where the RNA quality standard DV200 was as low as 20%. Uniquely designed probes that target the fusion junction sequences enhances the detection and realism of classical fusions. Gene fusions in five low-quality RNA samples could only be detected using the designed probe. Archived 104 tumor samples harboring gene fusion were divided into four groups according to RNA quality (DV200) and fusion detection methods. Based on the optimized library construction process, specific probe and bioinformatics analysis process, the RNA fusion panel identified the same gene fusions compared with the DNA level in 14 (100%, group A, DV200 ≥ 40%), 34 (82.9%, group B, DV200 ≥ 40%), 22 (81.5%, group C, 20% ≤DV200 < 40%) and 5 (71.4%, group D, DV200 < 20%) samples, respectively. Taken together, the optimization of the experimental procedure improves the detection of gene fusion in low-quality RNA samples and also contributes to accurate diagnosis and treatment.