Multi-omics evaluation of relapsed pediatric cancers: What information do these sequential analyses yield?

10047 Background: While cure rates for children with cancer have significantly improved, relapses remain a challenge, requiring deeper understanding to address them. Nowadays, genomic analyses are widely used at diagnosis and in relapse settings, becoming a standard-of-care in pediatric. The aim of this study is to describe the genomic evolution of relapsed pediatric tumors in search of clonal selection and pathway identification. We also want to assess the clinical value of these new data obtained in relapsed tumors. Methods: This is a retrospective analysis from canadian pediatric oncology precision medicine projects. We selected patients aged < 30 years with sequencing data available at diagnosis and relapse. Clinical and genomic data were collected, and each patient was paired with a non-relapsed patient. The incidences of genomic alterations were compared in the two populations and for each patient. For patients who relapsed, patient-adjusted longitudinal mixed models assessed differentially expressed genes at relapse vs diagnosis. Gene set enrichment analyses were performed, using GLMMSeq results, to study metabolic pathways that undergo significant dysregulation over time (p < 0.05). Electronic surveys were sent to the treating physicians of relapsed patients. Results: A total of 45 patients with 1 or more relapses were compared with 44 patients without relapse. Longitudinal analysis was performed on 35 relapsed patients. Our population has a median age of 10 y.o., a majority had leukemia (47 %) or sarcoma (31%). Among relapsed patients, the mutational burden at diagnosis was 0.82 mut/MB and 1.21 mut/MB at first relapse, compared with 0.47 mut/MB in non-relapsed patients (p = 0.02). 33% of relapsed patients had or acquired a TP53 alteration, compared with 16% without relapse (p = 0.084). MAPK pathway alterations were more prevalent among relapsed patients (p = 0.004). Longitudinal analyses showed enrichment in MAPK pathway at relapse. Other pathways were also significantly enriched at relapse (Wnt, TP53, TNFa, TGFb), while immune pathways (immunoglobulin/lymphocyte complex and activation) were downregulated. Although, 45% of clinicians considered that genomic analysis at relapse was useful, only 18% actually integrated the genomic data into clinical decisions due to better options available than targeted therapies. Indeed, when targeted therapies were used as proposed, it was mostly for a 2 nd or 3 rd relapse and for sarcoma. Conclusions: This study describes the evolution of the genomic landscape, showing an enrichment in mutation and pathways, and increased mutational burden in relapsed pediatric tumors. Longitudinal differential analyses brought more information about genomic evolution than the usual genomic reports sent to clinicians. Overall, the analyses performed are useful for clinicians, and a small subset of patients benefited from this information to guide therapies.