Lipid metabolic reprogramming represents a fundamental oncogenic mechanism. However, clinical relevance of lipid metabolism (LM) alterations in gliomas remain to be fully elucidated. LM-based classification and subsequent correlative analyses were performed using glioma bulk RNA-seq datasets retrieved from The Cancer Genome Atlas (TCGA). Radiomic models were constructed and validated using magnetic resonance imaging (MRI) datasets from The Cancer Imaging Archive (TCIA) and our in-house glioma cohort. Immunohistochemical (IHC) staining was employed to detect the protein expression of the key gene in glioma tissue samples. Single cell RNA-seq datasets from the GBMap database were used to characterize the distribution and functional roles of the key gene in the tumor microenvironment (TME) of gliomas. Consensus clustering based on LM pathways identified three distinct subtypes, respectively dominated by steroid metabolism (ST-type), triglyceride metabolism (TC-type), and sphingolipid metabolism (SP-type). The SP-type was independently associated with poorer prognosis and displayed enhanced activity in pathways linked to aggressive tumor phenotypes and radiotherapy resistance. Radiomic features enabled accurate identification of SP-type gliomas, thereby offering a non-invasive strategy for predicting this subtype. Genes GLA, GBL1, and HSD3B7 were identified as signature genes of the SP-type. Higher expression of HSD3B7 was associated with poor prognosis and exerted functional effects on multiple signaling pathways across various cellular components of the TME, which may contribute to glioma progression. Gliomas exhibited marked heterogeneity in LM and can be classified into three subtypes, with the SP-type exhibiting the most aggressive clinical behavior.
Tu et al. (Wed,) studied this question.