Non-invasive classification of gliomas using multi-nucleus MRI spectroscopy: a comparative study with histological biopsy

Le résumé fourni par la source

One of the main challenges in brain cancer diagnosis is the need for an invasive procedure, a histological biopsy, which confirms the presence of a brain tumor and classifies it, that is, determines its grade and oncotype, in order to guide the treatment strategy. At the same time, MRI is routinely used to visualize and characterize the tumor non-invasively, that is, to describe its characteristics such as size, location, and appearance. This raises the question: why not use only non-invasive MRI to confirm and classify, in addition to characterizing, brain tumors? This retrospective study included 25 patients with gliomas of various oncotypes and grades. The examinations were performed using a 3T high-field MRI scanner with a semi-LASER multi-nuclear spectroscopy MRI sequence. Tumor spectroscopic profiles, including glycolytic, lactate, and acid-base parameters, as well as energy balance and proliferation markers, were then analyzed to classify brain tumors virtually and non-invasively. Furthermore, the results of these virtual biopsies were validated by histological biopsies performed on the patients, by comparing the tumor classifications obtained via spectroscopy with those derived from histological analysis, particularly through the expression of lactate monocarboxylate transporters. A comparative statistical analysis was then performed to evaluate the diagnostic performance of virtual biopsy compared to histological biopsy. Our results show that virtual MRI biopsy can classify the oncotype and grade of certain brain tumors with performance comparable to that of histological biopsy. However, future studies involving a larger number of patients and including a wider range of brain tumors are needed to validate these results. In addition, another line of research aimed at evaluating the ability of virtual MRI biopsy to confirm the presence of brain tumors is being considered. This study will include patients with non-tumor conditions who have undergone multi-nuclear spectroscopy semi-LASER MRI as well as histological biopsy. MRI is a promising method for non-invasively classifying and characterizing brain tumors. It could thus reduce the need for invasive biopsy in brain diagnostics.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.