Serum protein triplet TGF-β1, TIMP-1, and YKL-40 serve as diagnostic and prognostic profile for astrocytoma

• Published in: Scientific reports 2021-06, Vol.11 (1), p.13100-13100, Article 13100
• Main Authors: Urbanavičiūtė, Rūta, Zabitaitė, Rūta, Kriščiukaitis, Algimantas, Deltuva, Vytenis-Pranas, Skiriutė, Daina
• Format: Article
• Language: English
• Subjects: 631/67/1922; 692/53; Algorithms; Astrocytoma; Astrocytoma - blood; Astrocytoma - diagnosis; Astrocytoma - metabolism; Astrocytoma - mortality; Biomarkers, Tumor - blood; Brain Neoplasms - blood; Brain Neoplasms - diagnosis; Brain Neoplasms - metabolism; Brain Neoplasms - mortality; Case-Control Studies; Chitinase-3-Like Protein 1 - blood; Decision trees; Enzyme-Linked Immunosorbent Assay; Gene Expression Regulation, Neoplastic; Glioblastoma; Glioma; Heterogeneity; Humanities and Social Sciences; Humans; IP-10 protein; Kaplan-Meier Estimate; Middle Aged; multidisciplinary; Prediction models; Prognosis; Proteins; Science; Science (multidisciplinary); Serum proteins; Survival; Survival Analysis; Tissue inhibitor of metalloproteinase 1; Tissue Inhibitor of Metalloproteinase-1 - blood; Transforming Growth Factor beta1 - blood; Transforming growth factor-b1; Tumors
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-021-92328-3

Astrocytoma is the most common glial tumour of the CNS. The most malignant form is grade IV Astrocytoma, also called Glioblastoma. Due to its heterogeneity, aggressiveness and lethal nature scientists are trying to find less invasive methods for early prediction of tumour onset, recurrence, response to therapy and patients’ survival. Here, applying decision tree classification algorithm we performed astrocytoma specific protein profile analysis on serum proteins TIMP-1, active and latent form of TGF-β1, IP-10, ANGPT-1, OPN, and YKL-40 using enzyme-linked immunosorbent detection assay (ELISA). Results have demonstrated that astrocytoma specific profile consisted of three proteins—active form of TGF-β1, TIMP-1 and YKL-40 and was able to correctly classify 78.0% (103/132) of sample and 83.3% (60/72) of astrocytoma sample. Calculating decision tree algorithm associated with astrocytoma patient survival, prediction model reached an accuracy of 83.3% (60/72). All together these results indicate that glioma detection and prediction from patient serum using glioma associated proteins and applying mathematical classification tools could be achieved, and applying more comprehensive research further could be implemented in clinic.