Blood-Based Biomarkers in High Grade Gliomas: a Systematic Review

High-grade gliomas (HGG) are the most common malignant primary brain tumor in adults. During the course of disease, several challenges occur, like measuring tumor burden, monitoring of treatment response, estimating the patient's prognosis, and distinguishing between true progression and pseudo-progression. So far, no blood-based biomarker has been established in the clinical routine to address these challenges. The aim of this systematic review was to analyze the present evidence on blood-based biomarkers for HGG. We systematically searched in PubMed, Web of Sciences, Scopus, and Cochrane Library databases for publications before 30th of March 2018 reporting on associations of blood-based biomarkers in HGG patients with different endpoints as overall survival, progression-free survival, and postoperative monitoring. Quality assessment of the studies according to QUIPS and STARD guidelines was performed. In accordance with the GRADE guidelines, level of evidence (I-IV) for each of the tested biomarkers was assessed. One thousand six hundred eighty unique records were identified. Of these, 170 original articles were included to this review. Four hundred fifteen different blood-based biomarkers analyzed in 15.041 patients with HGG as also their corresponding recurrent tumors. Ten predictive biomarkers reached level II of evidence. No biomarker achieved level I of evidence. In this review, 10 blood-based biomarkers were selected as most promising biomarkers for HGG: alpha 2-Heremans-Schmid glycoprotein (AHSG), albumin, glucose, insulin-like growth factor- binding protein 2 (IGFBP-2), macrophage inflammatory protein 1 delta (MIP-1 delta), macrophage inflammatory protein 3 ss (MIP-3 ss), neutrophil-lymphocyte ratio (NLR), red blood cell distribution width (RDW), soluble glycoprotein 130 (Sgp130), and chitinase-3-like protein 1 (YKL-40). To further assess the clinical significance of these biomarkers, the evaluation in a larger cohort of HGG and their corresponding subgroups would be necessary.