Pilocytic astrocytoma (PA) is the most common type of brain tumour in paediatric population connected with favourable prognosis although in numbered cases recurrence or dissemination could be observed. PA accounts for 30% of all brain tumours occurring in children. The tumours affect various anatomical structures and show different radiological appearance. Genetics of this tumour as well as the plausible correlations between molecular profile and clinical course of the disease and/or radiological features are still undefined. The purpose of our research was the identification of gene expression profiles related to localization and radiological features of pilocytic astrocytomas and clinical course of the disease. Eighty six children with PAs, operated on in the Department of Neurosurgery, Polish Mother’s Memorial Hospital Research Institute, were included in this study. The group was comprised of 55 males and 31 females. The mean age of patients at the time of diagnosis was 7 years (ranging from 1 to 17 years). All specimens were diagnosed at the Department of Molecular Pathology and Neuropathology Medical University of Lodz, according to the WHO criteria. The analysis was done by high density oligonucleotide microarrays (GeneChip Human Genome U133 Plus 2.0) in 50 snap‑frozen tissue samples diversified in terms of localization (28 cerebellar tumours, 11 optic tracts and hypothalamic tumours, 9 hemispheric tumours, 2 brain stem tumours), radiological appearance (27 solid or mainly solid tumours, 10 cystic tumours where the mural nodule and the cyst wall were enhanced, 8 cystic tumours where only the mural nodule was enhanced, 5 largely necrotic tumours) and clinical course (5 cases of progressive disease after subtotal resection, 2 cases connected with neurofibromatosis type 1). Bioinformatic analyses with using Bioconductor packages were done after normalization of data with using GC‑RMA algorithm. Gene expression profile of pilocytic astrocytomas highly depends on the tumour localization. This correlation reach statistical significance (p=0.001). Eight hundred sixty‑two probesets differentiated tumours of different localization with high significance of global test. Most prominent differences were noted for IRX2, PAX3, CXCL14, LHX2, SIX6, CNTN1 and SIX1 genes. Analysis of the genes differentiating between radiological features showed much weaker transcriptome differences, with the borderline significance in the global test of association (p=0.88). No genes showed significant association with the tendency to progression in univariate analysis (p=0.83). The results of microarray analysis were confirmed by QRT‑PCR. In the conclusion we showed that gene expression profile in pilocytic astrocytomas is connected with tumour localization and such relationship suggest different origin of PA arising within various anatomical brain structures. Morphological and radiological features as well as clinical course of disease seem not to be associated with different gene expression pattern.