Immune infiltrate and tumor microenvironment transcriptional programs stratify pediatric osteosarcoma into prognostic groups at diagnosis

Outcomes for adolescents/young adults with osteosarcoma has not improved since decades. Chaotic karyotype of this rare tumor has precluded identification of prognostic biomarkers/stratification. We reasoned that transcriptomic study should overcome this genetic complexity. From the RNA-sequencing of 79 osteosarcomas diagnostic biopsies, we identified stable independent components (Metagenes) recapitulating the tumor microenvironment and clones. Metagene unsupervised classification stratified this cohort into good (G1) and poor (G2) prognostic tumors on overall survival. Multivariate survival analysis ranked this stratification as the most influential variable. Functional characterization associates G1 tumors with innate immunity and G2 tumors with angiogenic, osteoclastic and adipogenic activities as well as PPARï§ pathway upregulation. We next defined a short gene signature to predict G1/G2 tumors from RNA-seq validated within independent cohort of 82 osteosarcomas. We further validated the signature by custom Nanostring panel in 96 additional osteosarcomas. Our study thus proposes new biomarkers to detect hard-to-treat patients as well as related new therapeutic options.EGA study EGAS00001005326