Abstract
Various hematological markers are associated with survival in patients with glioblastomas (GBMs), as they reflect inflammation and nutrition status. However, single markers are insufficient for predicting prognosis in GBM, and a comprehensive scoring system is needed. In this study, we developed a simple, inexpensive, and non-invasive scoring system, referred to as the Sanbo Scoring System (SSS), to predict survival in patients with GBMs. Patients with GBM were retrospectively assigned to two independent cohorts at Sanbo Brain Hospital and National Cancer Center/Cancer Hospital. Clinical records, including age, routine blood tests, biochemistry and coagulation examinations, and IDH-1 status, were collected. In total, 274 and 87 patients with GBMs at Sanbo Brain Hospital and National Cancer Center/Cancer Hospital were included as derivation and validation cohorts, retrospectively. We developed the SSS based on data for the derivation cohort, i.e., age, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), albumin-to-globulin ratio (AGR), and fibrinogen levels. These patients were divided into three groups that differed with respect to age, inflammation-nutrition status, and overall survival (p < 0.001), i.e., SSS 0, 1, and 2. NLR, PLR, and fibrinogen levels were lower and AGR was higher in the SSS 2 group than in the other groups, indicating better inflammation and nutrition statuses. Additionally, the longest overall survival was observed in this group. A multivariate analysis showed that SSS was an independent prognostic factor. The validation cohort supported all the results. SSS was a simple, non-invasive, and effective scoring system, and independently predicted survival in GBMs.