Work

10133-NQPC-6 Can THK5351 PET Predict Cognitive Function and Feasibility of Home-Based Care in Elderly Glioblastoma Patients?

Year: 2025

Type: article

Abstract: Background With the aging of glioblastoma (GBM) patients, maintaining QOL and higher-order cognitive function has become increasingly important. However, evaluating these aspects and providing ongoing support is often challenging. This study investigated whether serial changes in tau PET imaging using THK5351 could serve as a surrogate marker for cognitive function during treatment in elderly GBM patients, and whether PET uptake patterns might influence prognosis and the need for care support. Methods We retrospectively analyzed 14 patients aged ≥75 years who received bevacizumab (Bev) as initial therapy between February 2020 and March 2025. All patients were diagnosed with GBM via biopsy and were classified into two groups based on KPS: ≥60 and <60. Patients with KPS <60 who underwent tumor resection after Bev were defined as the neoadjuvant group. Cognitive function was evaluated after Bev and before discharge using the MMSE, TMT, and FAB, and findings were correlated with clinical course and support systems. Results The KPS ≥60 group included 2 patients (mean age 77.5), the KPS <60 group 9 patients (82.6), and the neoadjuvant group 3 patients (76.3). No post-Bev KPS deterioration was observed. Among KPS <60 patients, 4 with THK5351 uptake in the corpus callosum or hippocampus showed declines in MMSE. In contrast, the neoadjuvant group showed no hippocampal uptake and MMSE improvement. TMT-A was improved in 2 neoadjuvant cases. All KPS <60 patients failed TMT-B. FAB showed impairment in 2 KPS <60 cases, while the neoadjuvant case improved in all domains. Seven KPS <60 patients died in long-term care; others died at home. Conclusion THK5351 PET uptake in the corpus callosum or hippocampus may reflect cognitive decline and predict difficulty with home discharge, offering a useful tool for planning treatment and supportive care.

Source: Neuro-Oncology Advances

DOI: 10.1093/noajnl/vdaf236.086