使用三维 (3D) 卷积神经网络 (CNN) 研究全脑 MRI 空间特征对成人型弥漫性神经胶质瘤的预后价值。在一项回顾性多中心研究中，从五个数据集中招募了 1,925 名弥漫性神经胶质瘤患者：SNUH (n=708)、UPenn (n=425)、UCSF (n=500)、TCGA (n=160) 和 Severance (n=132)。 SNUH 和 Severance 数据集作为外部测试集。对比前和对比后 3D T1 加权、T2 加权和 T2-FLAIR 图像被处理为多通道 3D 图像。训练 3D 适应的 SE-ResNeXt 模型来预测总体生存率。使用 Cox 回归评估基于深度学习的预后指数 (DPI)、空间特征衍生的定量评分和已建立的预后标记的预后价值。使用一致性指数（C 指数）和 Brier 评分（BS）进行模型评估。仅 MRI 中位 DPI 生存预测模型的 C 指数分别为 0.709 和 0.677（BS=0.142 和 0.215），生存差异（p < SNUH 和 Severance 数据集分别为（0.001 和 p = 0.002；对数秩检验）。多变量 Cox 分析显示 DPI 是一个重要的预后因素，独立于临床和分子遗传变量：SNUH 和 Severance 数据集的风险比分别=0.032 和 0.036（p < 0.001 和 p=0.004）。多模态预测模型比仅使用临床和分子遗传变量的模型获得了更高的 C 指数：0.783 与 0.774，p = 0.001，SNUH； 0.766 vs. 0.748，p = 0.023，Severance。使用全脑 MRI 的 3D-CNN 模型得出的全局形态特征对于弥漫性胶质瘤具有独立的预后价值。结合临床、分子遗传学和影像数据可产生最佳性能。© 作者 2023。由牛津大学出版社代表神经肿瘤学会出版。版权所有。如需权限，请发送电子邮件至：journals.permissions@oup.com。

To investigate the prognostic value of spatial features from whole-brain MRI using a three-dimensional (3D) convolutional neural network (CNN) for adult-type diffuse gliomas.In a retrospective, multicenter study, 1,925 diffuse glioma patients were enrolled from five datasets: SNUH (n=708), UPenn (n=425), UCSF (n=500), TCGA (n=160), and Severance (n=132). The SNUH and Severance datasets served as external test sets. Pre- and postcontrast 3D T1-weighted, T2-weighted, and T2-FLAIR images were processed as multichannel 3D images. A 3D-adapted SE-ResNeXt model was trained to predict overall survival. The prognostic value of the deep learning-based prognostic index (DPI), a spatial feature-derived quantitative score, and established prognostic markers was evaluated using Cox regression. Model evaluation was performed using the concordance index (C-index) and Brier score (BS).The MRI-only median DPI survival prediction model achieved C-indices of 0.709 and 0.677 (BS=0.142 and 0.215) and survival differences (p < 0.001 and p = 0.002; log-rank test) for the SNUH and Severance datasets, respectively. Multivariate Cox analysis revealed DPI as a significant prognostic factor, independent of clinical and molecular genetic variables: hazard ratio=0.032 and 0.036 (p < 0.001 and p=0.004) for the SNUH and Severance datasets, respectively. Multimodal prediction models achieved higher C-indices than models using only clinical and molecular genetic variables: 0.783 vs. 0.774, p = 0.001, SNUH; 0.766 vs. 0.748, p = 0.023, Severance.The global morphologic feature derived from 3D-CNN models using whole-brain MRI has independent prognostic value for diffuse gliomas. Combining clinical, molecular genetic, and imaging data yields the best performance.© The Author(s) 2023. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.