Radiomics-Based Visualization and Quantitative Validation of IDH1 Heterogeneity in Gliomas

The goal of this clinical trail is to non-invasively visualise and quantitatively validate an radiomics model of genetic heterogeneity in adult patients with diffuse glioma to help clinicians better guide surgical resection and treatment options. It aims to answer are: 1. To overcome the limitations of the existing genetic diagnostic process in terms of equipment and technology requirements, high costs and long timelines, and to enable quantitative studies of isocitrate dehydrogenase 1 (IDH1) mutations, thus allowing refined patient stratification and further exploration of the role of molecular markers in improving patient prognosis. 2. To achieve non-invasive diagnosis of gene mutations within tumours by taking advantage of artificial intelligence and medical images, and to test the clinical feasibility of the model through typical target puncture, gene sequencing and quantitative gene expression analysis. Participants will read an informed consent agreement before surgery and voluntarily decide whether or not to join the experimental group. They will undergo preoperative magnetic resonance imaging, intraoperative brain puncture of typical tumour sites, and postoperative genotype identification. Their imaging data, genotype data, clinical history data, and pathology data will be used for the experimental study.

BACKGROUND The WHO 2016 officially introduced molecular markers into the pathological diagnosis of gliomas, marking a step into the era of molecular diagnosis of gliomas. Among them, isocitrate dehydrogenase 1 (IDH1) mutation is considered to be the 'backbone' in the development of gliomas, and affects the treatment plan and prognosis of patients. However, the clinical use of this molecular biomarker is still controversial, which is rooted in the lack of quantitative studies on IDH1 mutations. The spatial heterogeneity of gliomas has been demonstrated in existing studies, i.e., tumor tissues in different parts of the same glioma belong to different genetic subtypes. This implies that IDH1-mutant tumors do not indicate the presence of mutations in all tumor cells, thus further exacerbating the problems in clinical genetic diagnosis. OBJECTIVE To quantify gene mutations in tumours, we plan to use radiomics model with artificial intelligence and clinical big data, and verify its accuracy by tissue puncture. In this way, we can overcome the challenges of multisite sampling and second-generation sequencing, such as high equipment and technology requirements, high cost and long time, and thus theoretically realise the visualisation and quantification of genetic heterogeneity within gliomas. PROCESS Participants will read an informed consent agreement before surgery and voluntarily decide whether or not to join the experimental group. 1. Modelling of visualisation of genetic heterogeneity Before surgery, participants first Routine imaging and the resulting images will be used to build a radiomics model. The model will non-invasively predict IDH1 mutations in gliomas. 2. Typical site puncture After the enrolled participants were anaesthetised and craniotomised, clinicians selected typical tumor sites for puncture based on the model outputs. 3. Histopathological diagnosis The specimen from the same puncture site is divided into two parts, and the first part is routinely formalin-fixed for paraffin embedding and finally H\&E-stained sections. The pathologist first reads the H\&E sections and makes a histological diagnosis, describing the pathological morphology and characteristics, especially the tumor cell content and distribution. 4. IDH1 single nucleotide sequencing Another part of the sample is used for liquid nitrogen preservation. The Qiagen DNA/RNA Extraction Kit is used to extract DNA from the liquid nitrogen preserved tumour tissue, which is purified and subjected to the IDH1 polymerase chain reaction (PCR). The PCR product is purified and subjected to sequencing, and the sequencing product is detected on an ABI 7200 sequencer to determine whether IDH1 is mutated or not. 5. Mass spectrometry analysis of 2-Hydroxyglutarate (2-HG) expression levels The presence and expression of 2-HG in glioma samples is detected and analysed by mass spectrometry. 6. Validation of the radiomics-based IDH1 mutation prediction model. The 2-HG detection results were numerically compared with the model results. This is a single centre validation study. Compared with the routine glioma surgical procedure, this study adds intraoperative tumor-typical sits puncture to validate the predictive accuracy of the radiomics model and collects corresponding MRI images, tumour histology diagnosis, molecular pathology diagnosis. The radiomics model is built based on preoperative clinical data and is a non-invasive and rapid tool for quantitative analysis and visualisation of tumor genes.