Magnetic resonance imaging-based radiomic for predicting infiltration levels of CD68+ tumor-associated macrophages in Glioblastomas

Motivation: Predicting Tumor-Associated Macrophages (TAMs) levels using preoperative non-invasive imaging can influence patients with Glioblastoma (Gb) treatment decision-making and evaluate prognosis. Goal(s): This study aimed to combine imaging and radiomics features of preoperative for predicting CD68 + macrophage infiltration. Approach: Retrospective collection 143 patients with Gb. Divided patients into high CD68+TAMs（≥14.8%）and low CD68+TAMs (<14.8%) groups. The radiomics features extraction were based on CE-T1WI and T2WI. Multi-parameter stepwise regression was used to create the models. Results: The combined model, with ADCmin and radiomics features, had the best performance revealing AUCs of 0.865 and 0.825 for the training and testing sets, respectively. Impact: To provide imaging biomarkers for the evaluation of the TAMs infiltration of Gb by using machine learning combined with MR imaging parameters, reveal the relationship between images features and TAMs, and construct an evaluation model to predict macrophage before surgery.