CEST MRI and ML-based segmentation predict MGMT methylation in high-grade gliomas

Fluid-suppressed CEST MRI and ML segmentation predict MGMT methylation in high-grade gliomas: max APTw in enhancing tumor and necrosis achieved AUC 0.76-0.82, sensitivity up to 100% in a 44-patient study.

• In 44 patients (24 MGMT methylated), fluid-suppressed CEST metrics differed significantly: higher APTw and CEST@2ppm values in non-methylated tumors, particularly in necrosis and enhancing tumor with ML-based segmentation.
• Best AUC for predicting MGMT methylation: max APTw in enhancing tumor and necrosis (0.76–0.82); max CEST@2ppm in enhancing tumor and necrosis (0.67–0.85). Sensitivity (69–100%) exceeded specificity (59–82%).
• Small single-center study; findings require validation in larger multicenter cohorts.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.