Glioblastoma (GB) is one of the most aggressive and treatment-resistant cancers due to its complex tumor microenvironment (TME). We previously showed that GB progression is dependent on the aberrant induction of chaperone-mediated autophagy (CMA) in pericytes (PCs), which promotes TME immunosuppression through the PC secretome. The secretion of extracellular matrix (ECM) proteins with anti-tumor (Lumican) and pro-tumoral (Osteopontin, OPN) properties was shown to be dependent on the regulation of GB-induced CMA in PCs. As biomarkers are rarely studied in TME, in this work, we aimed to validate Lumican and OPN as prognostic markers in the perivascular areas of the peritumoral niche of a cohort of GB patients. Previously, we had validated their expression in GB xenografted mice presenting GB infiltration (OPN) or GB elimination (Lumican) dependent on competent or deficient CMA PCs, respectively. Then, patient sample classification by GB infiltration into the peritumoral brain parenchyma was related to GB-induced CMA in microvasculature PCs, analyzing the expression of the lysosomal receptor, LAMP-2A. Our results revealed a correlation between GB-induced CMA activity in peritumoral PCs and GB patients' outcomes, identifying three degrees of severity. The perivascular expression of both immune activation markers, Iba1 and CD68, was related to CMA-dependent PC immune function and determined as useful for efficient GB prognosis. Lumican expression was identified in perivascular areas of patients with less severe outcome and partially co-localizing with PCs presenting low CMA activity, while OPN was primarily found in perivascular areas of patients with poor outcome and partially co-localizing with PCs presenting high CMA activity. Importantly, we found sex differences in the incidence of middle-aged patients, being significantly higher in men but with worse prognosis in women. Our results confirmed that Lumican and OPN in perivascular areas of the GB peritumoral niche are effective predictive biomarkers for evaluating prognosis and monitoring possible therapeutic immune responses dependent on PCs in tumor progression.