IMG-65. Temporal muscle thickness as an independent prognostic marker in newly diagnosed glioblastoma patients

Abstract BACKGROUND Sarcopenia is associated with increased mortality, and it is desirable to measure muscle function in its evaluation. However, this is difficult in patients with glioblastoma due to neurological symptoms. In recent years, temporal muscle thickness (TMT) on MRI has been used as an alternative indicator of sarcopenia, and in this study, we investigated the significance of temporalis muscle thickness as a prognostic predictor in patients with glioblastoma. METHODS The subjects were 225 cases of primary glioblastoma treated at our hospital between July 2006 and September 2024. TMT was measured from MRI data, and overall survival (OS) was retrospectively examined. In addition, the association of TMT with age and gender was also compared. RESULTS The median age was 68 years (30-87 years), and the median TMT was 8.11 mm (4.43-12.51 mm). When the patients were divided into two groups based on the median TMT (109 high, 116 low), the median OS was 24.7 months and 12.0 months, respectively, which was significantly longer in the high group (p=0.0001). When the patients were further divided into three groups (74 high, 76 middle, 75 low), the median OS was 26.2 months, 19.6 months, and 10.5 months, respectively, with OS being significantly longer in the order of TMT (p<0.0001). Based on age, there were 111 aged patients and 114 young patients, and the TMT was 7.6±1.6mm and 8.2±1.3mm, respectively, which was significantly thicker in the young group (p=0.002), and median OS was also significantly longer in the young group at 23.8 months and 15.8 months (p<0.0001), showing an association with age. However, there were 129 males and 96 females, and TMT was significantly thicker in males at 8.2±1.5mm and 7.6±1.4mm (p=0.004), but median OS was significantly longer in females at 27.6 months and 16.2 months (p<0.0001), with no association. CONCLUSION In patients with glioblastoma, TMT was a useful prognostic predictor.