Multi-sequence MRI radiomics ML model predicts Ki-67 in glioblastoma

Retrospective radiomics model using multi-sequence MRI achieved AUC 0.746 (95% CI 0.614–0.879) for predicting Ki-67 expression in glioblastoma; external validation is needed.

• The combined model integrated T1WI, T2WI, FLAIR, and CET1WI features, outperforming single-sequence models.
• SHAP analysis was used to interpret feature contributions to predictions.
• Study limitation: retrospective single-center design without external validation.

Automated summary

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