3055 Background: Acute lymphoblastic leukemia (ALL) ranks among the most prevalent pediatric cancers. Molecular testing, including cytogenetics, has been widely used for prognosticating pediatric ALL. However, cytogenetic testing in pediatric ALL is incapable of detecting submicroscopic genetic changes and suffers from variable sensitivity, which may lead to heterogeneous outcomes within cytogenetic subgroups. In addition to conventional prognostic biomarkers, long non-coding RNAs (lncRNAs) are expressed in both a disease- and tissue-specific manner, emerging as promising biomarkers. However, well-established lncRNA prognostic biomarkers for pediatric ALL have yet to be identified. Therefore, in this study, we discovered three singular lncRNA prognostic biomarkers and a prognostic panel composed of multi-lncRNAs for pediatric ALL using our newly developed BAMBI method. Methods: This study utilized RNA-seq data from 561 distinct individuals, comprising 647 blood or bone marrow biospecimens obtained from pediatric patients with ALL from the dbGaP TARGET dataset. Among these individuals with RNA-seq and survival data, 25.8% deceased during follow-up. The average age at enrollment was 8.1 years. Of this cohort, 36.0% were female, and 61.5% were Caucasian white. Initially, our lncRNA identification tool, Flnc (Li Z., et al., Noncoding RNA. 2022), was employed to identify and quantify both novel and known lncRNAs expressed in pediatric ALL. Subsequently, our machine-learning based method, BAMBI, was utilized to discover lncRNAs as new prognostic biomarkers in pediatric ALL. Results: Three lncRNAs (ZBTB47 and NKTR antisense RNA, AC097359.2’s new isoform, and a novel lncRNA near STXBP5) were identified as individual putative biomarkers for prognosis. Each of these three lncRNA biomarkers exhibited predictive power for survival, with accuracy ranging from 68.2% to 72.8% and AUROC scores ranging from 67.7% to 75%. Furthermore, we established three lncRNA biomarker panels composed of these three lncRNAs along with a few additional lncRNAs, which showed improved prediction power with accuracy ranging from 78.3% to 79.6%, and AUROC scores ranging from 82.1% to 85.2%. Conclusions: We discovered both singular lncRNA and panels of multi-lncRNAs as prognostic biomarkers of pediatric ALL patients. These putative lncRNA biomarkers could be further validated through independent cohorts or prospective clinical study designs.