Garcia Vicente Ana María, Pérez-Beteta Julián, Amo-Salas Mariano, Bosque Jesús J, Noriega-Álvarez Edel, Soriano Castrejon Ángel María, Pérez-García Víctor M
Nuclear Medicine Department, University General Hospital, 13005 Ciudad Real, Spain.
Mathematical Oncology Laboratory (MOLAB), Department of Mathematics, Faculty of Medicine, Castilla-La Mancha University, 13005 Ciudad Real, Spain.
J Clin Med. 2022 Oct 14;11(20):6065. doi: 10.3390/jcm11206065.
(1) Aim: To study the associations between imaging parameters derived from contrast-enhanced MRI (CE-MRI) and 18F-fluorocholine PET/CT and their performance as prognostic predictors in isocitrate dehydrogenase wild-type (IDH-wt) high-grade gliomas. (2) Methods: A prospective, multicenter study (FuMeGA: Functional and Metabolic Glioma Analysis) including patients with baseline CE-MRI and 18F-fluorocholine PET/CT and IDH wild-type high-grade gliomas. Clinical variables such as performance status, extent of surgery and adjuvant treatments (Stupp protocol vs others) were obtained and used to discriminate overall survival (OS) and progression-free survival (PFS) as end points. Multilesionality was assessed on the visual analysis of PET/CT and CE-MRI images. After tumor segmentation, standardized uptake value (SUV)-based variables for PET/CT and volume-based and geometrical variables for PET/CT and CE-MRI were calculated. The relationships among imaging techniques variables and their association with prognosis were evaluated using Pearson’s chi-square test and the t-test. Receiver operator characteristic, Kaplan−Meier and Cox regression were used for the survival analysis. (3) Results: 54 patients were assessed. The median PFS and OS were 5 and 11 months, respectively. Significant strong relationships between volume-dependent variables obtained from PET/CT and CE-MRI were found (r > 0.750, p < 0.05). For OS, significant associations were found with SUVmax, SUVpeak, SUVmean and sphericity (HR: 1.17, p = 0.035; HR: 1.24, p = 0.042; HR: 1.62, p = 0.040 and HR: 0.8, p = 0.022, respectively). Among clinical variables, only Stupp protocol and age showed significant associations with OS and PFS. No CE-MRI derived variables showed significant association with prognosis. In multivariate analysis, age (HR: 1.04, p = 0.002), Stupp protocol (HR: 2.81, p = 0.001), multilesionality (HR: 2.20, p = 0.013) and sphericity (HR: 0.79, p = 0.027) derived from PET/CT showed independent associations with OS. For PFS, only age (HR: 1.03, p = 0.021) and treatment protocol (HR: 2.20, p = 0.008) were significant predictors. (4) Conclusions: 18F-fluorocholine PET/CT metabolic and radiomic variables were robust prognostic predictors in patients with IDH-wt high-grade gliomas, outperforming CE-MRI derived variables.
（1）目的：研究对比增强磁共振成像（CE-MRI）和18F-氟胆碱PET/CT得出的影像参数之间的关联，以及它们在异柠檬酸脱氢酶野生型（IDH-wt）高级别胶质瘤中作为预后预测指标的性能。（2）方法：一项前瞻性多中心研究（FuMeGA：功能性和代谢性胶质瘤分析），纳入基线有CE-MRI和18F-氟胆碱PET/CT以及IDH野生型高级别胶质瘤的患者。获取诸如功能状态、手术范围和辅助治疗（Stupp方案与其他方案）等临床变量，并将总生存期（OS）和无进展生存期（PFS）作为终点进行区分。通过对PET/CT和CE-MRI图像的视觉分析评估多灶性。肿瘤分割后，计算基于PET/CT标准化摄取值（SUV）的变量以及基于PET/CT和CE-MRI的体积和几何变量。使用Pearson卡方检验和t检验评估影像技术变量之间的关系及其与预后的关联。采用受试者工作特征曲线、Kaplan-Meier法和Cox回归进行生存分析。（3）结果：评估了54例患者。PFS和OS的中位数分别为5个月和11个月。发现从PET/CT和CE-MRI获得的体积依赖性变量之间存在显著的强相关性（r>0.750，p<0.05）。对于OS，发现与SUVmax、SUVpeak、SUVmean和球形度存在显著关联（HR分别为：1.17，p = 0.035；1.24，p = 0.042；1.62，p = 0.040和0.8，p = 0.022）。在临床变量中，只有Stupp方案和年龄与OS和PFS存在显著关联。没有CE-MRI得出的变量与预后存在显著关联。在多变量分析中，PET/CT得出的年龄（HR：1.04，p = 0.002）、Stupp方案（HR：2.81，p = 0.001）、多灶性（HR：2.20，p = 0.013）和球形度（HR：0.79，p = 0.027）与OS存在独立关联。对于PFS，只有年龄（HR：1.03，p = 0.021）和治疗方案（HR：2.20，p = 0.008）是显著的预测指标。（4）结论：18F-氟胆碱PET/CT代谢和影像组学变量是IDH-wt高级别胶质瘤患者强有力的预后预测指标，优于CE-MRI得出的变量。
