Affiliation:
1. Sun Yat-sen University Cancer Center
Abstract
Abstract
Purpose
Prostate-specific membrane antigen (PSMA) positron emission tomography/magnetic resonance (PET/MR) is a novel imaging technology in neuro-oncology. This study aimed to investigate whether PET/MR-derived parameters could provide clinical characteristics and tumor heterogeneity in post-treatment glioma. The ability of PSMA PET/MR to distinguishing recurrence from treatment-related changes (TRCs) was evaluated.
Methods
Twenty-four glioma patients (fourteen males and ten females) suspected of recurrence who underwent PSMA PET/MR were included in this study. The results of PET/MR were evaluated qualitatively and quantitatively, including the visual intensity, classical metabolic and metric parameters, and primary tumor texture features.
Results
Classical metabolic parameters of PET including tumor-to-brain ratio (TBR max and mean), standardized uptake value (SUVmax and SUVpeak) showed differences in Karnofsky score (KPS) and metric parameters (area of lesion derived from MR) in 1p19q deletion status (P < 0.05). For the texture features, the shape-based, first-order and high-order primary tumor texture features of MR suggested the power to differentiate tumor grade and gene status (All P < 0.05). The sensitivity, specificity, and positive and negative predictive values of PSMA PET/MR in identifying recurrence were 64.29% (9/14), 80% (8/10), 81.82% (9/11), and 61.54% (8/13), respectively.
Conclusion
This work highlights the role of postoperative PSMA PET/MR in tumor-targeted imaging and differentiating recurrence in glioma, especially glioblastoma. PSMA PET/MR-derived parameters especially textural features provide additional value for characterizing glioma patient status and tumor heterogeneity. Our results indicate the significance of the hybrid PSMA PET/MR system in providing non-invasive glioma biological features, guiding precise surgical resection, and stratifying patients with PSMA targeted therapy.
Publisher
Research Square Platform LLC