The use of RNA sequencing to assist the diagnosis of pediatric cancers via classification of disease subtypes is now a routine clinical test in Quebec. We have previously demonstrated that clinical transcriptomics can be adapted to a real-time low-cost clinical decision support tool using nanopore RNA sequencing, enabling precise disease subtype identification in under 5 minutes for ~AU$350. Following validation of this approach on newly recruited and bio-banked patient samples, we identified multiple known and novel unannotated long non-coding RNA (lncRNA) genes that are significantly associated with cancer subtypes, revealing significant gaps in reference transcriptome annotations and their potential role as key oncogenic regulators. Using CRISPR interference (CRISPRi) and Cas13-mediated knockdowns, followed by deep single molecule sequencing using the latest nanopore technology, we confirmed the oncogenic functions of 5 of these lncRNAs with clear sub-cellular localisation using single molecule FISH, demonstrating their critical functions in gene expression regulation and signaling pathways. We also performed RNA modification analysis for m6A, pseudouridine, inosine and m5 using the latest Oxford Nanopore direct RNA base calling software, which identified thousands of isoforms with differential post-transcriptional modifications that do not display significant changes in abundance following knock-down. This multi-omic approach integrates differential RNA isoform expression and RNA modifications, offering a robust resource for novel biomarker discovery and personalized therapeutic targeting in leukemia. Our findings highlight the potential of real-time clinical epitranscriptomics to advance cancer diagnostics and treatment strategies.