Investigation of cystine as differential diagnostic biomarker between astrocytomas and oligodendrogliomas based on global- and targeted analysis using liquid chromatography/tandem mass spectrometric analysis

Abstract

Astrocytoma and oligodendroglioma are primary brain tumors classified as gliomas. Because there is difference in the prognostic significance of the extent of resection between IDH-mutant astrocytoma and oligodendroglioma, intraoperative differential diagnosis between them provides important information for optimal extent of resection. Although the characteristics of genetic mutations and chromosomal aberrations in both tumors have been reported, there is no molecular diagnostic methods that is able to be used quickly and simply for differentiate the two tumors. Therefore, we aimed to search for biomarker candidates for differentiating them with metabolome analysis using liquid chromatography/tandem mass spectrometry and develop a molecular diagnostic method based on quantitative analysis.

We searched for peaks that differed in two types of gliomas using global metabolomics. Next, we identified five biomarker candidates as hypoxanthine, inosine, cystine, proline and uric acid, respectively. Next, we developed an LC/MS/MS analytical method for five biomarker candidates and quantified them in brain tumors. Cystine had significantly lower amounts in astrocytomas than in oligodendrogliomas. We developed two prediction models for differentiation of the two gliomas and validated them using the separated two dataset. The logistic regression model with only cystine provided the best prediction performance. It was suggested that mass spectrometric analysis of cystine in surgery might be useful for differentiating astrocytoma and oligodendroglioma with 91.7% positive prediction value and 80.0% specificity whereas negative predictive value and sensitivity was lesser than 70%, so that further exploration for unknown metabolite is mandatory.