NIMG-20. MULTI-HABITAT RADIOMICS UNRAVELS DISTINCT PHENOTYPIC SUBTYPES OF GLIOBLASTOMA WITH CLINICAL AND GENOMIC SIGNIFICANCE

Abstract

Abstract BACKGROUNDS We aimed to evaluate the potential of radiomics as an imaging biomarker for GBM patients and explore the molecular rationale behind radiomics by radio-genomics approach. METHODS A total of 144 primary GBM patients were included in this study as a training cohort. Using multi-parametric MR images, radiomics features were extracted from multi-habitats of the tumor. We applied Cox-LASSO algorithm to build a survival prediction model and validated this model using an independent validation cohort (56 patients from Vienna). With the selected radiomics features, GBM patients were consensus clustered to reveal inherent phenotypic subtypes. The subtypes were further explored in terms of genomic signatures. RESULTS GBM patients were successfully stratified by the radiomics risk score, a weighted sum of radiomics features, corroborating the potential of radiomics as a prognostic biomarker. Using consensus clustering, we identified three distinct subtypes which significantly differed in the prognosis (‘heterogenous enhancing’, ‘rim-enhancing necrotic’, and ‘cystic’ subtype). Multi-variate cox regression analysis confirmed that radiomics subtype as an independent prognostic factor. Transcriptomic traits enriched in individual subtypes were in accordance with imaging phenotypes summarized by radiomics. For example, rim-enhancing necrotic subtype was well described by radiomics profiling (T2 autocorrelation & flat shape) and highlighted by the inflammatory genomic signatures, which well correlated to its phenotypic peculiarity (necrosis). CONCLUSIONS The present study confirmed the feasibility of radiomics as an imaging biomarker for GBM patients with comprehensible biologic annotation. Imaging subtypes derived from radiomics successfully recapitulate the genomic underpinnings of GBM tumors and in turn reinforce their potential as a prognostic biomarker.