ASLNet: an explainable deep learning framework for glioma grading and survival prediction

1

ASLNet is a deep learning model developed to predict glioma grade and overall survival using noninvasive ASL MRI data.
2

The study included 471 patients with diffuse glioma, utilizing ASL MRI volumes for model training and evaluation.
3

ASLNet achieved an AUC of 0.79 for glioma grading and 0.70 for survival prediction, demonstrating its predictive capabilities.
4

Saliency analysis identified hyperperfused tumor cores as key regions for grade prediction, highlighting ASL's clinical relevance.
5

The findings support ASL-based deep learning as a valuable tool for noninvasive glioma risk stratification and prognosis.

Frontiers In Oncology

by Rafail C. Christodoulou, Georgios Vamvouras, Platon S. Papageorgiou, Evros Vassiliou, Sokratis G. Papageorgiou, Christina Kalogeropoulou, Peter Zampakis, Elena E. Solomou, Michalis F. Georgiou
May 18, 2026

1