Temporospatial changes of TSPO expression in glioblastoma in response to therapy and during disease progression – a multimodal TSPO PET, amino acid PET, and MRI study

我们之前的研究和其他人的研究表明，TSPO PET和FET PET在恶性胶质瘤中表现出不同的摄取模式，并为对比增强MRI提供了补充信息。TSPO PET信号升高与肿瘤分级和预后不良相关，特别是在FET增强相对缺乏的情况下，可能反映了治疗引起的肿瘤及其炎症微环境的变化。然而，迄今尚未对TSPO信号在治疗反应和疾病进展过程中的纵向评估及其对胶质母细胞瘤患者临床结局和个体预后的影响进行评估。因此，本项目旨在通过单臂观察研究，利用纵向多模态TSPO / FET PET和ceMRI，观察胶质母细胞瘤患者TSPO表达的时空变化，并将成像结果与细胞和分子水平的变化以及临床结果联系起来。在工作包中，将招募1例新诊断为胶质母细胞瘤的脑肿瘤患者（n=58），并进行前瞻性研究。招募期为18个月，随访期为18个月。所有从临床检查、临床随访、影像学和神经病理分析中检索到的数据将存入电子数据库。在肿瘤切除和连续立体定向活检期间收集的肿瘤标本将使用图像引导神经导航系统进行。组织样本的定位将分配给PET和MRI数据，这使得PET信息与A2/3项目的组织病理学发现和通路分析具有精确的空间相关性。与A2/A3项目的合作允许深入分析肿瘤微环境的变化，细胞对TSPO信号的贡献（髓细胞与肿瘤细胞）和分子改变，在一个独特的纵向研究队列中，在TSPO PET图像指导下获得的组织样本在初次诊断和复发时。工作包2侧重于多模态成像数据的高级图像分析及其与工作包1中获得的临床和神经病理数据的相关性。分析包括药代动力学建模、机器学习和深度学习方法，以进行非侵入性组织分类，在早期时间点区分治疗应答者和无应答者，预测肿瘤复发的时间和位置，并最终将成像特征与总体生存期联系起来。

Our previous work and that of others has shown that TSPO PET and FET PET demonstrate diverging uptake patterns in malignant glioma and provide complementary information to contrast enhanced MRI. Increased TSPO PET signal correlates with tumor grade and poor prognosis and, especially in the relative absence of FET enhancement, may reflect therapy induced changes in the tumor and its inflammatory microenvironment. However, longitudinal assessment of TSPO signaling in response to therapy and during disease progression and its impact on the clinical outcome and individual prognosis of glioblastoma patients has not been assessed so far. Therefore, this project aims to visualize temporal and spatial changes of TSPO expression by means of a single arm observational study using longitudinal multimodal TSPO / FET PET and ceMRI for patients with glioblastoma, and to correlate imaging findings with changes on the cellular and molecular level as well with the clinical outcome. In work package 1 patients with newly diagnosed brain tumors suspicious for glioblastoma (n=58) will be recruited and prospectively enrolled. Recruitment phase of the study will be 18 months with follow-up of another 18 months. All data retrieved from clinical workup, clinical-follow-up, imaging and neuropathological analysis will be put into an electronic database. Tumor specimen collected during tumor resection and serial stereotactic biopsies will be performed using an image guided neuronavigation system. Localizations of tissue samples will be assigned to PET and MRI data, which enables a precise spatial correlation of PET information with histopathological findings and pathway analyses for project A2/3. This collaboration with project A2/A3 allows in-depth analyses of changes of the tumor microenvironment, cellular contributions to the TSPO signal (myeloid cells vs. tumor cells) and molecular alterations in a unique longitudinal study cohort with tissue samples obtained with TSPO PET image guidance at primary diagnosis and at relapse. Work package 2 focusses on advanced image analysis of the multimodal imaging data and its correlation with clinical and neuropathological data obtained in work package 1. The analyses include pharmacokinetic modelling, machine learning and deep learning approaches to perform non-invasive tissue classification, distinguish therapy responders from non-responders at an early time point and to predict the time and localization of tumor recurrence and ultimately, to link imaging features with overall survival.