Glioma Local Recurrence

Local recurrences in grade 4 adult-type diffuse gliomas often occur within non-enhancing T2 hyperintensity areas after surgery, making it difficult to differentiate between tumors and edema using conventional MRI alone. By incorporating quantitative DCE MRI parameters like Ktrans and Ve, a deep learning model was trained to detect subtle differences in vessel leakiness associated with local recurrence. The model, which improved sensitivity, may enhance risk-adapted radiotherapy planning by better predicting local recurrence in these gliomas.