Abstract Introduction: Next-generation sequencing (NGS) has become a critical diagnostic assay to identify pathologic SNVs, CNVs, and gene rearrangements. However, it has not been routinely used to assess expression levels of target genes that could indicate patient populations responsive to therapeutics. This is largely due to the absence of reliable bioinformatics tools for the assessment of expression levels within RNA assays. We have developed a novel within-sample distribution-based method that assesses the relative extremity of expression for individual genes. Our predominate focus has been on assessing the expression of NTRK1, NTRK2, NTRK3, ROS1, and ALK, however this method can be readily applied to other genes and NGS platforms. Methods: Within the kinase domain, the expression of a targeted region is represented by the number of unique deduplicated reads for NGS studies and normalized probe expression values (based on spike-in controls) for NanoString studies. A Poisson distribution is used to represent primer expression with the parameters estimated via maximum likelihood. The interquartile range (IQR) of the entirety of the sample's read counts is calculated and only those that do not exceed the third quartile bound by more than 150% of the IQR are considered during parameter estimation. A probability is then assigned to each of the primers based on their read counts. A geometric mean of the individual primer probabilities represents the probability value for the entire gene. Expression values are reported as -log10 (p-value) and cutoffs of 6 and 1 were used to call significant expression for NGS and NanoString platforms, respectively. Results: The NGS-based approach correctly identified significant expression in all gene rearrangement positive cell lines (n = 11). In cell lines not harboring gene rearrangements, the NGS and NanoString platforms showed 100% concordance in calling significant expression (n = 12). In a gene rearrangement negative FFPE cohort, concordance between NGS and NanoString platforms was 97%-99% for target genes (n = 102). ROS1 and ALK were most commonly found to be significantly expressed in FFPE samples (5% and 2%). Conclusions: We have developed a statistical-based approach to detecting expression levels for RNA-based NGS assays. This can be applied to cohort studies to not only identify clinical samples that may benefit from targeted kinase inhibitor therapeutics but also correlate with predicted outcome of disease. Citation Format: Robert Shoemaker, Aaron Berlin, Amy Diliberto, Heather Ely, Marissa Chen, Danielle Murphy, Jason Christiansen, Vince Reuter, Abel Licon. A novel, statistical-based method to determine RNA expression by next-generation sequencing in clinical FFPE samples. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 5274.