Deep learning iTB map maps glioma infiltration beyond contrast enhancement

[medRxiv] Researchers developed and validated the infiltrative Tumor Burden (iTB) map, a deep learning biomarker that uses multiparametric MRI (mpMRI) to identify infiltrative tumor within non-enhancing FLAIR-hyperintense regions, according to a medRxiv preprint. The retrospective study analyzed preoperative mpMRI from 80 patients with WHO grade II-IV gliomas and 24 patients with brain metastases, using spatially localized stereotactic biopsies from non-enhancing glioma regions as histopathologic ground truth and peritumoral brain metastasis regions as non-tumor controls, and trained a 3D convolutional neural network. Model development included 121 biopsy-localized samples from 56 glioma patients and 104 non-tumor samples from 26 brain metastasis patients. According to the authors, augmentation optimization raised independent test-set accuracy from 74.5% to 85.5% and AUC from 0.859 to 0.942 (p = 0.008). Ensemble-derived iTB maps showed low infiltration probability in brain metastases and extensive elevated infiltration probability in glioblastoma. The authors said the biomarker may improve surgical planning, radiotherapy targeting and assessment of glioma burden.