Relevance. Despite all the treatment glioblastoma recurs as an aggressive and therapy-resistant tumor, and patients quickly die from these neoplasms. The study of glioblastoma recurrence processes and search for prognostic factors of the disease should lead to the improvement of survival rates of patients with this pathology. Purpose of the study. To study the influence of clinical and molecular-genetic factors on the median second recurrence-free period. Materials and methods. Progression-free survival after first recurrence in 34 patients aged 28 to 81 years with recurrent glioblastoma was analyzed. The diagnosis was established according to the WHO 2021 classification of CNS tumors. In each observation we studied such clinical parameters as patient’s age, functional status according to the Karnovsky scale pre- and postoperatively, peculiarities of neuroimaging picture (prevalence of tumor process, localization of recurrence, tumor volume), conducted treatment and molecular-genetic characteristics of the tumor (determination of mRNA expression level of genes: MGMT, VEGF, PDGFRA, β-tubulin III, ERCC-1, TOP2A). Results. Among the clinical and demographic characteristics, the median of the survival was influenced by the patients’ age and functional status after surgery. The median of the survival was more than 2 times higher in the group of patients under 50 years old, compared to patients over 50 years old (18.5 vs 8 weeks). The dependence of the median of the survival on the post- operative functional status (according to the Karnovsky scale) was determined (p = 0.001). The median of the survival in case of a single brain lobe lesion was more than 5 times higher than in case of widespread tumor process, though without statistical reliability (p = 0.09, 21.5 vs 4 weeks). Survival rates were higher when recurrence was localized within 2 cm of the area of removal of the primary neoplasm. After disease progression, the MGMT gene lost its predictive value. Patients with low expression of the TOR2A gene had a higher survival rate than those with medium and high expression (47.5 vs 3 weeks, p = 0.001; 47.5 vs 22.5 weeks, p = 0.06). The median of survival was higher than at high levels at low and medium PDGFRA gene expression levels (29 vs 0 weeks, p = 0.04; 21 vs 0 weeks; p = 0.05, respectively). Maximum survival rates were recorded in the group of patients after total and subtotal removal of tumor recurrence (22 and 18.5 weeks, p = 0.05). Administration of second-line chemotherapy with temozolomide statistically significantly increased the median of the second BRS (p = 0.01). Conclusion. Recurrent glioblastomas are characterized by an extremely aggressive course. Therefore, such prognostic factors as patient age, degree of tumor resection, tumor process prevalence, degree of tumor resection and 2nd line chemotherapy come to the forefront. It should be noted that the MGMT gene loses its predictive value during disease progression, while the TOR2A gene and PDGFRA gene become prognostic markers.