Choi Y, Ahn KJ, Nam Y, et al. - One hundred fourteen individuals with pathologically verified glioblastoma were retrospectively chosen from March 2008 to May 2018 in order to ascertain the incremental prognostic value of radiomics on peritumoral T2 hyperintensity in pretreatment glioblastoma. Among the 106 radiomic features, a random forest variable selection generated five appropriate characteristics (two shapes, two gray-level and one first-order characteristic). These radiomic characteristics enhanced survival prediction accuracy when they were added onto clinical and tumor/peritumoral volumetric characteristics. In comparison with the clinical or radiomic model, on the test set, the combined model exhibited a lower mean survival prediction error rate. Therefore, the clinical model with radiomic peculiarities exhibited enhanced survival predictive performance compared with model without radiomic characteristics, hence implying the incremental prognostic value of peritumoral radiomics as MR imaging biomarker in pretreatment glioblastoma.