From the Departments of Radiology (L.G., D.G., C.C., T.M.S., U.N., N.F., A.M.D., J.D.R.), Pathology (V.G.), Radiation Medicine and Applied Sciences (C.R.M., T.M.S., J.H.G.), Neurologic Surgery (T.B.), Bioengineering (T.M.S.), and Neurosciences (J.D.S., D.P., A.M.D.), University of California San Diego, 9500 Gillman Dr, La Jolla, CA 92093; Cortechs.ai, San Diego, Calif (G.M., N.W.); Department of Translational Neurosciences, Pacific Neuroscience Institute and Saint John's Cancer Institute at Providence Saint John's Health Center, Santa Monica, Calif (S.K.); and Department of Biophysics, Medical College of Wisconsin, Milwaukee, Wis (K.M.S.).
Radiol Artif Intell. 2024 Sep;6(5):e230489. doi: 10.1148/ryai.230489.
Purpose To develop and validate a deep learning (DL) method to detect and segment enhancing and nonenhancing cellular tumor on pre- and posttreatment MRI scans in patients with glioblastoma and to predict overall survival (OS) and progression-free survival (PFS). Materials and Methods This retrospective study included 1397 MRI scans in 1297 patients with glioblastoma, including an internal set of 243 MRI scans (January 2010 to June 2022) for model training and cross-validation and four external test cohorts. Cellular tumor maps were segmented by two radiologists on the basis of imaging, clinical history, and pathologic findings. Multimodal MRI data with perfusion and multishell diffusion imaging were inputted into a nnU-Net DL model to segment cellular tumor. Segmentation performance (Dice score) and performance in distinguishing recurrent tumor from posttreatment changes (area under the receiver operating characteristic curve [AUC]) were quantified. Model performance in predicting OS and PFS was assessed using Cox multivariable analysis. Results A cohort of 178 patients (mean age, 56 years ± 13 [SD]; 116 male, 62 female) with 243 MRI timepoints, as well as four external datasets with 55, 70, 610, and 419 MRI timepoints, respectively, were evaluated. The median Dice score was 0.79 (IQR, 0.53-0.89), and the AUC for detecting residual or recurrent tumor was 0.84 (95% CI: 0.79, 0.89). In the internal test set, estimated cellular tumor volume was significantly associated with OS (hazard ratio [HR] = 1.04 per milliliter; < .001) and PFS (HR = 1.04 per milliliter; < .001) after adjustment for age, sex, and gross total resection (GTR) status. In the external test sets, estimated cellular tumor volume was significantly associated with OS (HR = 1.01 per milliliter; < .001) after adjustment for age, sex, and GTR status. Conclusion A DL model incorporating advanced imaging could accurately segment enhancing and nonenhancing cellular tumor, distinguish recurrent or residual tumor from posttreatment changes, and predict OS and PFS in patients with glioblastoma. Segmentation, Glioblastoma, Multishell Diffusion MRI © RSNA, 2024.
开发和验证一种深度学习（DL）方法，以检测和分割胶质母细胞瘤患者治疗前后 MRI 扫描中的增强和非增强细胞肿瘤，并预测总生存期（OS）和无进展生存期（PFS）。
本回顾性研究纳入了 1297 例胶质母细胞瘤患者的 1397 次 MRI 扫描，包括内部 243 次 MRI 扫描（2010 年 1 月至 2022 年 6 月）用于模型训练和交叉验证以及四个外部测试队列。两位放射科医生根据影像学、临床病史和病理发现对细胞肿瘤进行分割。将灌注和多壳扩散成像的多模态 MRI 数据输入到 nnU-Net DL 模型中以分割细胞肿瘤。量化了分割性能（Dice 评分）和区分复发性肿瘤与治疗后变化的性能（受试者工作特征曲线下面积[AUC]）。使用 Cox 多变量分析评估模型在预测 OS 和 PFS 方面的性能。
评估了一组 178 例患者（平均年龄，56 岁±13[SD]；116 例男性，62 例女性）的 243 次 MRI 时间点，以及分别包含 55、70、610 和 419 次 MRI 时间点的四个外部数据集。中位 Dice 评分 0.79（IQR，0.53-0.89），检测残留或复发性肿瘤的 AUC 为 0.84（95%CI：0.79，0.89）。在内部测试集中，估计的细胞肿瘤体积与 OS（风险比[HR]每毫升 1.04；<0.001）和 PFS（HR 每毫升 1.04；<0.001）显著相关，调整年龄、性别和大体全切除（GTR）状态后。在外部测试集中，调整年龄、性别和 GTR 状态后，估计的细胞肿瘤体积与 OS 显著相关（HR 每毫升 1.01；<0.001）。
一种包含先进成像的 DL 模型可以准确分割增强和非增强的细胞肿瘤，区分复发性或残留肿瘤与治疗后变化，并预测胶质母细胞瘤患者的 OS 和 PFS。
分段，胶质母细胞瘤，多壳扩散 MRI
© 2024 RSNA
Zotero 插件