Long Term Survival in Patients Suffering from Glioblastoma Multiforme: A Single-Centre Observational Cohort Study

Background: Glioblastomas (GBM) are generally burdened, to date, by a dismal prognosis, although long term survivors have a relatively significant incidence. Our specific aim was to determine the exact impact of many surgery-, patient- and tumor-related variables on survival parameters. Methods: The surgical, radiological and clinical outcomes of patients have been retrospectively reviewed for the present study. All the patients have been operated on in our institution and classified according their overall survival in long term survivors (LTS) and short term survivors (STS). A thorough review of our surgical series was conducted to compare the oncologic results of the patients in regard to: (1) surgical-(2) molecular and (3) treatment-related features. Results: A total of 177 patients were included in the final cohort. Extensive statistical analysis by means of univariate, multivariate and survival analyses disclosed a survival advantage for patients presenting a younger age, a smaller lesion and a better functional status at presentation. From the histochemical point of view, Ki67 (%) was the strongest predictor of better oncologic outcomes. A stepwise analysis of variance outlines the existence of eight prognostic subgroups according to the molecular patterns of Ki67 overexpression and epidermal growth factor receptor (EGFR), p53 and isocitrate dehydrogenase (IDH) mutations. Conclusions: On the grounds of our statistical analyses we can arm that the following factors were significant predictors of survival advantage: Karnofsky performance status (KPS), age, volume of the lesion, motor disorder at presentation and/or a Ki67 overexpression. In our experience, LTS is associated with a gross total resection (GTR) of tumor correlated with EGFR and p53 mutations with regardless of localization, and poorly correlated to dimension. We suppose that performing a standard molecular analysis (IDH, EGFR, p53 and Ki67) is not sucient to predict the behavior of a GBM in regards to overall survival (OS), nor to provide a deeper understanding of the meaning of the di↵erent genetic alterations in the DNA of cancer cells. A fine molecular profiling is feasible to precisely stratify the prognosis of GBM patients.