Identification of astrocytomas through serum protein fingerprint using MALDI-TOF MS and machine learning

ABSTRACT: This study aimed to explore the potential of serum proteomics as a non-invasive diagnostic tool using MALDI-TOF mass spectrometry (MS) combined with machine learning techniques. We analyzed serum samples from 269 patients, employing machine learning models to differentiate between healthy individuals and astrocytoma patients. The MALDI-TOF MS approach achieved a balanced accuracy of 94.5% in distinguishing GBM patients from healthy controls. However, it showed limited efficacy in classifying tumor grades or determining IDH1 mutational status. Further investigation using bottom-up proteomics by GeLC-MS/MS identified potential biomarkers, including proteins from the APO family and transthyretin, previously associated with high-grade gliomas. These findings highlight the promise of MALDI-TOF MS in identifying serum-based biomarkers for astrocytoma diagnosis.

INSTRUMENT(S):

ORGANISM(S): Homo Sapiens (human)

TISSUE(S): Blood Serum

DISEASE(S): Glioblastoma

SUBMITTER: Lucas Cardoso Lazari  

LAB HEAD: Giuseppe Palmisano

PROVIDER: PXD060922 | Pride | 2026-09-07

REPOSITORIES: Pride