Tariciotti Leonardo, Caccavella Valerio M, Fiore Giorgio, Schisano Luigi, Carrabba Giorgio, Borsa Stefano, Giordano Martina, Palmisciano Paolo, Remoli Giulia, Remore Luigi Gianmaria, Pluderi Mauro, Caroli Manuela, Conte Giorgio, Triulzi Fabio, Locatelli Marco, Bertani Giulio
Unit of Neurosurgery, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Milan, Italy.
Department of Oncology and Hemato-Oncology, University of Milan, Milan, Italy.
Front Oncol. 2022 Feb 24;12:816638. doi: 10.3389/fonc.2022.816638. eCollection 2022.
Neuroimaging differentiation of glioblastoma, primary central nervous system lymphoma (PCNSL) and solitary brain metastasis (BM) remains challenging in specific cases showing similar appearances or atypical features. Overall, advanced MRI protocols have high diagnostic reliability, but their limited worldwide availability, coupled with the overlapping of specific neuroimaging features among tumor subgroups, represent significant drawbacks and entail disparities in the planning and management of these oncological patients.
To evaluate the classification performance metrics of a deep learning algorithm trained on T1-weighted gadolinium-enhanced (T1Gd) MRI scans of glioblastomas, atypical PCNSLs and BMs.
We enrolled 121 patients (glioblastoma: n=47; PCNSL: n=37; BM: n=37) who had undergone preoperative T1Gd-MRI and histopathological confirmation. Each lesion was segmented, and all ROIs were exported in a DICOM dataset. The patient cohort was then split in a training and hold-out test sets following a 70/30 ratio. A Resnet101 model, a deep neural network (DNN), was trained on the training set and validated on the hold-out test set to differentiate glioblastomas, PCNSLs and BMs on T1Gd-MRI scans.
The DNN achieved optimal classification performance in distinguishing PCNSLs (AUC: 0.98; 95%CI: 0.95 - 1.00) and glioblastomas (AUC: 0.90; 95%CI: 0.81 - 0.97) and moderate ability in differentiating BMs (AUC: 0.81; 95%CI: 0.70 - 0.95). This performance may allow clinicians to correctly identify patients eligible for lesion biopsy or surgical resection.
We trained and internally validated a deep learning model able to reliably differentiate ambiguous cases of PCNSLs, glioblastoma and BMs by means of T1Gd-MRI. The proposed predictive model may provide a low-cost, easily-accessible and high-speed decision-making support for eligibility to diagnostic brain biopsy or maximal tumor resection in atypical cases.
在表现出相似外观或非典型特征的特定病例中，胶质母细胞瘤、原发性中枢神经系统淋巴瘤（PCNSL）和孤立性脑转移瘤（BM）的神经影像学鉴别仍然具有挑战性。总体而言，先进的MRI方案具有较高的诊断可靠性，但其在全球范围内的可用性有限，再加上肿瘤亚组之间特定神经影像学特征的重叠，这是显著的缺点，并且在这些肿瘤患者的治疗规划和管理中存在差异。
评估一种基于胶质母细胞瘤、非典型PCNSL和BM的T1加权钆增强（T1Gd）MRI扫描训练的深度学习算法的分类性能指标。
我们纳入了121例患者（胶质母细胞瘤：n = 47；PCNSL：n = 37；BM：n = 37），这些患者均接受了术前T1Gd - MRI检查及组织病理学确诊。对每个病灶进行分割，并将所有感兴趣区域（ROI）导出到DICOM数据集中。然后按照70/30的比例将患者队列分为训练集和保留测试集。使用一个Resnet101模型（一种深度神经网络（DNN））在训练集上进行训练，并在保留测试集上进行验证，以在T1Gd - MRI扫描上区分胶质母细胞瘤、PCNSL和BM。
DNN在区分PCNSL（曲线下面积（AUC）：0.98；95%置信区间（CI）：0.95 - 1.00）和胶质母细胞瘤（AUC：0.90；95%CI：0.81 - 0.97）方面取得了最佳分类性能，在区分BM方面能力中等（AUC：0.81；95%CI：0.70 - 0.95）。这种性能可能使临床医生能够正确识别适合进行病灶活检或手术切除的患者。
我们训练并在内部验证了一种深度学习模型，该模型能够通过T1Gd - MRI可靠地区分PCNSL、胶质母细胞瘤和BM的疑难病例。所提出的预测模型可为非典型病例中诊断性脑活检或最大程度肿瘤切除的适用性提供低成本、易于获取且高速的决策支持。
