Quality-adjusted survival after tumor resection and/or radiation therapy for elderly patients with glioblastoma multiforme

Abstract

Objective: Prognostic factors are poorly defined for the elderly subpopulation with glioblastoma multiforme and have been exclusively related to conventional survival analysis. In this study an additional quality adjusted survival analysis (QAS) was performed. The prognostic evaluation of both survival- and QAS data after standard treatment were checked for concordant/discordant findings. Their usefulness for estimation of treatment effects and treatment strategies was then evaluated.

Methods: 123 patients >or= 65 years of age with a supratentorial, de novo glioblastoma were included in the current retrospective report. Microsurgery plus radiation therapy (planned tumor dose: 60 Gy) was performed in 58 patients, and radiation therapy alone after stereotactic biopsy (planned tumor dose: 60 Gy) in 65 patients. The functional status of each patient was scored when joining the study and at every follow-up using 15 selected neurological signs and symptoms (NSSs). Gradation of severity of each NSS was performed with subjective weights. Survival time of each patient was adjusted according to any changes in these NSSs to become the Quality Time (Q-TIME). Time intervals spent with side effects of the treatment (TOX) were subtracted from Q-TIME to become the patient's QAS (QAS = Q-TIME-TOX). Prognostic factors for both survival and QAS were obtained from the Cox model.

Results: Overall survival and QAS were 24 weeks and 10.5 weeks, respectively. Perioperative morbidity and mortality were 5.2 % and 1.7 % in the surgery group and 1.5 % and 1.5 % in the biopsy group, respectively (p > 0.05). Tumor resection gained favorable prognostic importance for patients with midline shift in terms of both survival and QAS (p < 0.0001). Otherwise, radiation therapy alone was as effective as surgery plus radiation therapy (concordant finding). A pretreatment Karnofsky Score (KPS) < 70 was an unfavorable predictor for QAS (p < 0.002) but not for survival (discordant finding). Median QAS for patients with a pretreatment KPS < 70 was only 10 weeks. Age did not reach prognostic relevance.

Conclusion: The dramatic decrease of QAS as compared with survival indicates extremely limited posttreatment improvement and/or rapid deterioration of the neurological score after standard treatment for the older subpopulation with glioblastoma multiforme. Supportive treatment should be considered for patients with a pretreatment KPS < 70.