Current applications of deep-learning in neuro-oncological MRI

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Neuro-oncology

Background:

• Magnetic Resonance Imaging (MRI) is crucial for managing neurological neoplasms.
• Deep learning (DL), a subset of artificial intelligence, offers advanced capabilities for MRI analysis.

Purpose of the Study:

• To provide an overview of the current applications of DL in neuro-oncology using MRI.
• To summarize the state-of-the-art DL techniques applied to neuro-oncology MRI.

Main Methods:

• A systematic review of the PubMed database was conducted.
• Search terms included MRI, DL, and neuro-oncology, focusing on MeSH terms and title/abstract.
• Eligible studies were categorized into technological innovation, diagnosis, and follow-up.

Main Results:

• Forty-one publications post-2016 were reviewed.
• Most studies focused on technological innovation (22), followed by diagnosis (12) and follow-up (7).
• Applications included image acquisition improvement, synthetic CT generation, auto-segmentation, tumor classification, outcome prediction, and response assessment, primarily using standard MRI sequences and Convolutional Neural Networks (CNNs).

Conclusions:

• DL in neuro-oncology MRI is an emerging research area with diverse applications.
• Key challenges include acquiring large datasets, ensuring cross-institutional/vendor applicability, and clinical validation for practice integration.