Intratumoral heterogeneity poses a significant challenge to the diagnosis and treatment of recurrent glioblastoma. This study addresses the need for non-invasive approaches to map heterogeneous landscape of histopathological alterations throughout the entire lesion for each patient. We developed BioNet, a biologically-informed neural network, to predict regional distributions of two primary tissue-specific gene modules: proliferating tumor (Pro) and reactive/inflammatory cells (Inf). BioNet significantly outperforms existing methods (p < 2e-26). In cross-validation, BioNet achieved AUCs of 0.80 (Pro) and 0.81 (Inf), with accuracies of 80% and 75%, respectively. In blind tests, BioNet achieved AUCs of 0.80 (Pro) and 0.76 (Inf), with accuracies of 81% and 74%. Competing methods had AUCs lower or around 0.6 and accuracies lower or around 70%. BioNet’s voxel-level prediction maps reveal intratumoral heterogeneity, potentially improving biopsy targeting and treatment evaluation. This non-invasive approach facilitates regular monitoring and timely therapeutic adjustments, highlighting the role of ML in precision medicine.

瘤内异质性对复发性胶质母细胞瘤的诊断和治疗构成了重大挑战。本研究解决了非侵入性方法的需求，以绘制每位患者整个病灶中组织病理学改变的异质景观。我们开发了 BioNet，一种生物学信息神经网络，用于预测两个主要组织特异性基因模块的区域分布：增殖肿瘤 （Pro） 和反应性/炎症细胞 （Inf）。BioNet 的性能明显优于现有方法 （p < 2e-26）。在交叉验证中，BioNet 实现了 0.80 （Pro） 和 0.81 （Inf） 的 AUC，准确率分别为 80% 和 75%。在盲测中，BioNet 的 AUC 分别为 0.80 （Pro） 和 0.76 （Inf），准确率分别为 81% 和 74%。竞争方法的 AUC 较低或约为 0.6，准确度较低或约为 70%。BioNet 的体素级预测图揭示了肿瘤内异质性，有可能改进活检靶向和治疗评估。这种非侵入性方法有助于定期监测和及时调整治疗，突出了 ML 在精准医疗中的作用。