Importance: Genomic analyses solely focused on detecting mutations do not benefit most pediatric cancer patients. Alternate genomic approaches are needed to identify additional treatment biomarkers and therapeutic targets. Objective: To evaluate the performance of our Comparative Analysis of RNA Expression (CARE) pipeline in nominating druggable targets in pediatric patients with difficult-to-treat or rare cancers. Design, Setting, and Participants: Our cohort study, the Comparative Analysis of RNA Expression to Improve Pediatric and Young Adult Cancer Treatment (CARE IMPACT), was conducted collaboratively by the UC Santa Cruz Treehouse Childhood Cancer Initiative and Stanford University School of Medicine. From June 4, 2018, to September 24, 2020, UCSC Treehouse obtained and processed RNA sequencing (RNA-Seq) data for 35 tumor samples from 33 children and young adults with a relapsed, refractory or rare cancer. Each tumor RNA-Seq dataset underwent CARE analysis to reveal activated cancer driver pathways and nominate treatment options. We compare the CARE pipeline to other gene expression outlier detection approaches and discuss challenges and opportunities for the clinical implementation of RNA-Seq-based gene expression outlier detection for pediatric cancer patients. Exposures: Patients underwent tumor RNA-Seq analysis and standard-of-care tumor DNA profiling. UCSC Treehouse compared each patient's tumor RNA-Seq profile with over 11,000 uniformly analyzed tumor profiles from public data repositories. These comparisons reveal candidate cancer genes and pathways that represent potential therapeutic targets. Main Outcome(s) and Measure(s): Proportion of patients for whom CARE provided useful clinical and biological information for patient care. Impact of comparator cohort choice on outlier findings. Results: Of our 33 patients, 31 (94%) had CARE IMPACT findings of potential clinical significance. These findings were implemented in 5 patients, 3 of which had defined clinical benefit. We demonstrated that composition of comparator cohorts determines which outliers are detected and that large and diverse cohorts containing data from tumors similar to the patients produce the most clinically relevant outlier results. Conclusions and Relevance: Comparative RNA-Seq analysis may identify additional cancer driver pathways and druggable targets in patients with rare or difficult-to-treat pediatric cancers relative to standard-of-care DNA profiling. This study highlights the clinical utility of CARE for pediatric tumors and underscores the need for further evaluation of this approach to improve patient outcomes.