Journal Article

Biological tumor volume predicts survival in recurrent High-Grade glioma: A multiparametric 18FFET PET/MRI study

Henssen, Dylan, Scherlach, Cordula, Arens, Anne I J, Schildan, Andreas, Jähne, Katja, Seidel, Clemens, Rullmann, Michael, Schürer, Matti, Striepe, Stephan, Stassart, Ruth, Sabri, Osama, Hesse, Swen

2025

Overview

Single-session, multiparametric [¹⁸F]FET PET/MRI is used to detect tumor recurrence in high-grade glioma, but its prognostic value for overall survival remains uncertain. This study evaluated whether biological tumor volume, tumor-to-background ratio (TBRmax), cerebral blood volume (rCBVmax), and choline/NAA ratio (Cho/NAA) could predict survival in recurrent high-grade glioma.BACKGROUND AND PURPOSESingle-session, multiparametric [¹⁸F]FET PET/MRI is used to detect tumor recurrence in high-grade glioma, but its prognostic value for overall survival remains uncertain. This study evaluated whether biological tumor volume, tumor-to-background ratio (TBRmax), cerebral blood volume (rCBVmax), and choline/NAA ratio (Cho/NAA) could predict survival in recurrent high-grade glioma.Twenty-six patients with histopathologically confirmed tumor progression underwent simultaneous [¹⁸F]FET PET/MRI. PET-derived biological tumor volume and TBRmax, MRI-derived rCBVmax, and Cho/NAA ratio were analyzed. A Cox proportional hazards model assessed associations with overall survival, adjusting for the number of lesions and treatment strategy.MATERIALS AND METHODSTwenty-six patients with histopathologically confirmed tumor progression underwent simultaneous [¹⁸F]FET PET/MRI. PET-derived biological tumor volume and TBRmax, MRI-derived rCBVmax, and Cho/NAA ratio were analyzed. A Cox proportional hazards model assessed associations with overall survival, adjusting for the number of lesions and treatment strategy.Biological tumor volume (hazard ratio = 2.22, 95%-CI: 1.035-4.762, p = 0.041) and the number of lesions (hazard ratio = 1.03, 95%-CI 1.00-1.06, p = 0.036) were significantly associated with survival. TBRmax (p = 0.089), rCBVmax (p = 0.088), and Cho/NAA ratio (p = 0.734) were not predictive. Treatment strategy after tumor recurrence diagnosis did not significantly impact overall-survival (HR = 0.208, p = 0.649). PET/MRI interaction terms did not enhance survival prediction.RESULTSBiological tumor volume (hazard ratio = 2.22, 95%-CI: 1.035-4.762, p = 0.041) and the number of lesions (hazard ratio = 1.03, 95%-CI 1.00-1.06, p = 0.036) were significantly associated with survival. TBRmax (p = 0.089), rCBVmax (p = 0.088), and Cho/NAA ratio (p = 0.734) were not predictive. Treatment strategy after tumor recurrence diagnosis did not significantly impact overall-survival (HR = 0.208, p = 0.649). PET/MRI interaction terms did not enhance survival prediction.Biological tumor volume is a significant prognostic imaging biomarker in recurrent high-grade glioma, emphasizing tumor burden over metabolic activity or perfusion of individual lesions. Volume-based PET metrics may offer better survival prediction than traditional PET or MRI parameters. Prospective multicenter studies are needed to validate these findings and explore automated segmentation and machine learning approaches for improved prognostication.CONCLUSIONBiological tumor volume is a significant prognostic imaging biomarker in recurrent high-grade glioma, emphasizing tumor burden over metabolic activity or perfusion of individual lesions. Volume-based PET metrics may offer better survival prediction than traditional PET or MRI parameters. Prospective multicenter studies are needed to validate these findings and explore automated segmentation and machine learning approaches for improved prognostication.