gms | German Medical Science

Objective: Standardized treatment algorithms for recurrent glioblastoma multiforme (rGBM) are still lacking. The neurosurgical armamentarium for circumscribed rGBMs comprises resection and stereotactic brachytherapy (SBT) as possible local treatment concepts. In here, we report our experiences of both treatment options.

Methods: We retrospectively identified patients (KPS ≥70) with circumscribed rGBM after multimodal therapy who were either treated by resection or SBT at two centers. Resections were performed using 5-aminolevulinic acid fluorescence guidance; for SBT, exclusively temporary I-125 seeds (reference dose: 50 Gy, dose rate: <15 cGy/h) were utilized. Study endpoints were overall survival (OS) and postrecurrence survival (PRS). Survival was assessed with the Kaplan-Meier method.

Results: 81 patients (median age: 58.6 years, median KPS: 80) were identified; 39 patients allotted to the SBT and 42 patients to the resection cohort. All surgical cases and 30/39 (76.9%) SBT patients were included at first recurrence, the remaining SBT cases at later stages. SBT patients were younger (56.9 vs. 64.6 years; p=0.02); rGBMs treated by SBT were smaller (median volume: 2.1 vs. 31.5 cm3; p<0.01), more commonly deep-seated (p=0.03) and on the dominant hemisphere (p<0.01). Treatment-associated morbidity was recorded at 6.2%, with wound healing deficits being the most common complication. The perioperative hospitalization was significantly shorter in the SBT cohort (median: 6.0 vs. 12.0 days; p<0.05). The cohorts did not differ in terms of their respective treatment-free intervals after SBT/resection. Salvage therapy included chemotherapy in 57/81 (70.4%) patients; no difference was seen between both groups. Over the course of follow-up 5/39 (12.8%) patients received multiple SBTs and 6/42 (14.3%) patients underwent multiple resections. At the end of follow-up 71/78 (91.0%) patients had died, three patients were lost to long-term follow-up. In the SBT versus resection cohorts, OS were 30.2 months (CI95%: 19.5-39.7 months) versus 25.5 months (CI95%: 16.2-30.6; p=0.49) and PRS 9.8 months (CI95%: 8.0-14.3) versus 9.3 months (CI95%: 6.0-15.3; p=0.34), respectively. Multivariate statistical analyses revealed tumor volume to be the strongest prognostic factor for favorable PRS for the total study population (p=0.007) and the resection cohort (p=0.01); no correlation between extent of resection and tumor volume was recorded.

Conclusion: Due to the differences between the two patient cohorts comparative outcome analyses are per se difficult; however, PRS was equivalent irrespective of the applied local treatment concept. Large tumor volume was identified as the strongest predictive factor for worse PRS in the resection cohort. SBT may be considered a favorable approach for small tumors due to its lesser burden of treatment.