MRI Radiomics Reveals Sex Differences in a Preclinical Model of Glioblastoma Immunotherapy Response

Motivation: Individual variations in immunotherapy response make clinical implementation challenging. More personalized treatments could prolong survival. Goal(s): Explore sex-based differences in glioblastoma immunotherapy response in a mouse model using radiomics features extracted from contrast agent concentration maps. Approach: Features were computed from maps of SPIO (relating to labelled macrophages) and Gadolinium (related to blood-brain barrier leakage) in mouse brains over 6 weeks. Important features were input to a logistic regression model for predictions of sex and survival. Results: SPIO concentration features at weeks 1-3 provide accurate prediction of sex (PR AUC>0.8). Features at week 2 accurately predict survival, but only in the male cohort. Impact: Understanding the role of sex-based variations in therapy response may lead to more effective treatment monitoring or personalized treatment plans for glioblastoma and other deadly cancers, ultimately increasing the clinical relevance and uptake of immunotherapies.