Identifying predictors of glioma evolution from longitudinal sequencing.

Area of Science:

• Oncology
• Genomics
• Cancer Evolution

Background:

• Clonal evolution is a key driver of cancer progression and therapeutic resistance in gliomas.
• Understanding early molecular features that guide post-treatment cancer evolution is crucial but remains unclear.

Purpose of the Study:

• To identify early molecular predictors of tumor evolution in adult diffuse gliomas.
• To develop a predictive model for glioma evolution and patient stratification.
• To elucidate the mechanism by which MYC influences treatment resistance and hypermutation.

Main Methods:

• Multivariate analysis of sequencing and clinical data from 544 initial-recurrent adult diffuse glioma pairs.
• Development and validation of the machine learning model CELLO2 (Cancer EvoLution for LOngitudinal data version 2).
• In vitro experiments using glioma cell lines and patient-derived gliomaspheres to study temozolomide resistance and MYC function.

Main Results:

• CDKN2A deletion predicted later tumor necrosis in IDH-mutant gliomas; Ki67 predicted hypermutation in IDH-wild-type gliomas.
• MYC gain or MYC-target activation at diagnosis predicted treatment-induced hypermutation in all glioma subtypes.
• CELLO2 model stratified patients, identifying a high-risk group with MYC gain and poor survival.
• MYC drives temozolomide resistance by promoting hypermutation via increased vulnerability of mismatch repair genes.

Conclusions:

• Early molecular events, particularly MYC alterations, are significant predictors of diffuse glioma evolution and therapeutic resistance.
• The CELLO2 model offers a tool for predicting glioma progression and stratifying patients for precision oncology.
• Targeting MYC-driven hypermutation presents a potential strategy to overcome temozolomide resistance in gliomas.