Aim: IDH-mutant lower grade glioma (LGG) has been proven to have a good prognosis. However, its high recurrence rate has become a major therapeutic difficulty. Materials & methods: We combined epigenomic deconvolution and a network analysis on The Cancer Genome Atlas IDH-mutant LGG data. Results: Cell type compositions between recurrent and primary gliomas are significantly different, and the key cell type that determines the prognosis and recurrence risk was identified. A scoring model consisting of four gene expression levels predicts the recurrence risk (area under the receiver operating characteristic curve = 0.84). Transcription factor PPAR-α explains the difference between recurrent and primary gliomas. A cell cycle-related module controls prognosis in recurrent tumors. Conclusion: Comprehensive deconvolution and network analysis predict the recurrence risk and reveal therapeutic targets for recurrent IDH-mutant LGG.

For a long time, the guidance for adjuvant chemoradiotherapy for lower grade glioma (LGG) lacks instructions on the application timing and order of radiotherapy (RT) and chemotherapy. We, therefore, aimed to develop indicators to distinguish between the different beneficiaries of RT and chemotherapy, which would provide more accurate guidance for combined chemoradiotherapy. By analysing 942 primary LGG samples from The Cancer Genome Atlas (TCGA) and the Chinese Glioma Genome Atlas (CGGA) databases, we trained and validated two gene signatures (Rscore and Cscore) that independently predicted the responsiveness to RT and chemotherapy (Rscore AUC = 0.84, Cscore AUC = 0.79) and performed better than a previous signature. When the two scores were combined, we divided patients into four groups with different prognosis after adjuvant chemoradiotherapy: RSCS (RT‐sensitive and chemotherapy‐sensitive), RSCR (RT‐sensitive and chemotherapy‐resistant), RRCS (RT‐resistant and chemotherapy‐sensitive) and RRCR (RT‐resistant and chemotherapy‐resistant). The order and dose of RT and chemotherapy can be adjusted more precisely based on this patient stratification. We further found that the RRCR group exhibited a microenvironment with significantly increased T cell inflammation. In silico analyses predicted that patients in the RRCR group would show a stronger response to checkpoint blockade immunotherapy than other patients.

No abstract