A multicenter analysis of prognostic factors for patients undergoing re-irradiation for recurrent glioblastoma

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Background Re-irradiation (re-RT) has gained popularity in the treatment of recurrent or progressive glioblastoma (rGBM) after first-line treatment. However, uncertainties regarding optimal treatment dose, fraction size and target volume definition remain. In this multicenter analysis of the Western German Cancer Center (WTZ), we analyze prognostic factors for patients treated with re-RT for rGBM with different fraction regimens. Methods We analyzed patients with rGBM (CNS-WHO °4, IDH-wildtype) who received re-RT after standard first-line treatment (±salvage resection/chemotherapy) between 01/2010 and 12/2021. Clinical, histopathological, and radiological data were evaluated to identify prognostic factors influencing event-free and overall survival, as well as treatment tolerability. Results We identified 118 patients of whom 87 received normofractionated (1.8-2 Gy/fx) re-RT with a mean dose of 40.8 Gy using a mean GTV-PTV margin of 8.8 mm. 31 patients were treated with moderate hypofractionation (2.4-3.5 Gy/fx) up to a mean dose of 36 Gy with a mean GTV-PTV margin of 4.1 mm. Propensity score weighting was used to compensate prognostic factors between these two cohorts. Re-resection classification (no, STR, GTR, p < 0.0001), age at diagnosis (p = 0.0082), unifocal vs. multifocal progression (p = 0.021) and time between first and second RT (p = 0.0109) were identified as significant prognostic factors using stepwise parameter selection. The propensity score weighted survival was 8.8 months for normofractionation and 9.9 months for hypofractionation (p = 0.63). Conclusion Age at initial diagnosis, progression pattern and time between first and second RT are prognostic for patients undergoing re-irradiation. No significant survival difference between normofractionated and hypofractionated approaches was detected, which should be further investigated.