Potential biomarkers of Ewing sarcoma identified through a Europe-wide analysis of prospectively collected samples

Clinical features, aside from disease stage, have proven insufficient to identify patients at extremely high or low risk. There are few established biological markers for EwS. The discovery of new prognostic and/or predictive biomarkers could improve our understanding of tumor heterogeneity, facilitate individual risk stratification, and aid targeted therapies. We initiated a pan-European study using formalin-fixed, prospectively collected, primary tumors prior to chemotherapy from 335 patients in two clinical trial registries, EURO-Ewing 99 and Ewing 2008, to analyze correlation and survival of ten potential prognostic biomarkers. Immunohistochemical (IHC) analyses were conducted for STEAP1, DKK2, and EZH2, while MIR34A and LGALS3BP expression was investigated by quantitative PCR. Genomic alterations, including gain of chromosome 1q, loss of ADAM3A, loss of chromosome 16q, as well as loss of heterozygosity (LOH) and percentage of the genome altered (PGA), were analyzed by FISH and/or SNP-based arrays of genomic DNA. To consider the observed variation in the number of analyzed samples per biomarker (N = 102–300), effect size measures were computed. Correlations between dichotomized biomarker expressions were generally low, except for those reflecting genomic alterations (ϕ >= 0.20 <= 0.38; P < .05). Several relevant correlations were found with clinical features of the underlying EwS patients (ϕ >= 0.20 <= 0.31; P < .05). Most promising prognostic values with clinically relevant effect sizes (OS), adjusted for known prognostic variables, were high MIR34A expression (HR = 0.29; P < .01), LOH (HR = 2.44; P < .05), and PGA (HR = 2.20; P = .05). This pan-European study identified MIR34A, LOH, and PGA as the most promising biomarkers for the prognosis of survival in a well-characterized EWS patient cohort.