Advances in the Qualitative Diagnosis of Glioma：Correlation between Radiological Images and Genetic Alterations

Abstract

Radiological imaging plays a pivotal role in glioma patient care. It provides qualitative information about the tumor, such as the presumed pathological diagnosis and molecular status. In addition, it can pro- vide anatomical information necessary for surgery and is helpful for monitoring treatment response. In this review, we discuss the following topics • 1. The progress in radiomics in the field of glioma. 2. Detailed analysis of the T2~FLAIR mismatch sign. 3. The potential of quantitative magnetic resonance (MR) imaging in the realm of qualitative glioma imaging. Radiomics "focuses on improvements in image analysis, using an automated high-throughput extraction of large amounts (200+) of quantitative features of medical images". Despite extensive research, its diagnostic accuracy for detecting IDH mutations is limited to approximately 85% sensitivity and specificity. The diagnosis of lp/19q co-deletion is 10% less accurate than that of the IDH mutation. The accuracy for diagnosing MGMT promoter methylation is still uncertain. Furthermore, the generalization of diagnostic algorithms derived from machine learning is another critical issue. While many researchers in the community have pushed radiomic research to the limit, a conventional qualitative imaging feature, namely,"the T2-FLAIR mismatch sign," was discovered. This imaging feature is able to identify IDH-mutant, lp/19q non-codeleted astrocytomas with a sensitivity of 20-50% and a specificity of almost 100%. Through radiomic research of gliomas, the authors noticed potential effects of differences in image acquisition parameters between different institutions on the low sensitivity of the T2-FLAIR mismatch sign for detecting IDH-mutant and lp/19q non-codeleted astrocytomas. Indeed, tuning the image acquisition parameters for FLAIR significantly improved the sensitivity of the T2-FLAIR mismatch sign. Finally, the future of MR-based glioma imaging relies on quantitative MR acquisition. This technique directly measures the tissue's Tl- and T2~relaxation times, which provides valuable information for cancer tissue characterization. For example, we found that IDH-mutant, lp/19q non-codeleted astrocytomas contain tissues with very long Tl- and T2~relaxation times (longer than 3,000 ms in Tl_relaxation time). The commercialization of rapid quantitative MR acquisition technology could further boost the capability of radiomics.