Brain tumors are a major source of disease burden in pediatric population, with the most common tumor types being pilocytic astrocytoma, ependymoma and medulloblastoma. In every tumor entity, surgery is the cornerstone of treatment, but the importance of gross-total resection and the corresponding patient prognosis is highly variant. However, real-time identification of pediatric CNS malignancies based on the histology of the frozen sections alone is especially troublesome. We propose a novel method based on differential mobility spectrometry (DMS) analysis for rapid identification of pediatric brain tumors.
We prospectively obtained tumor samples from 15 pediatric patients (5 pilocytic astrocytomas, 5 ependymomas and 5 medulloblastomas). The samples were cut into 36 smaller specimens that were analyzed with the DMS.
With linear discriminant analysis algorithm, a classification accuracy (CA) of 70% was reached. Additionally, a 75% CA was achieved in a pooled analysis of medulloblastoma vs. gliomas.
Our results show that the DMS is able to differentiate most common pediatric brain tumor samples, thus making it a promising additional instrument for real-time brain tumor diagnostics.