Functional, Structural, and AI-Based MRI Analysis: A Comprehensive Review of Recent Advances

Area of Science:

• Advancements in medical imaging analysis
• Integration of artificial intelligence in radiology

Background:

• Magnetic Resonance Imaging (MRI) analytical methods have evolved significantly since its invention.
• Artificial intelligence (AI) has recently transformed MRI data analysis, impacting functional MRI (fMRI), image segmentation, and radiomics.

Purpose of the Study:

• To review recent progress in functional MRI (fMRI) and structural MRI analysis.
• To summarize AI-driven techniques including deep learning, machine learning, and radiomics in MRI.

Main Methods:

• Review of fMRI and quantitative MRI (qMRI) techniques (T1/T2 mapping, proton density).
• Description of diffusion, perfusion, and Dixon method approaches.
• Analysis of MRI radiomics studies (2012-2025) and deep learning segmentation architectures.

Main Results:

• Increasing adoption of quantitative MRI methods for enhanced cross-study comparison and reproducibility.
• Rapid advancements in deep learning for MRI segmentation, with novel architectures and loss functions.
• AI-driven MRI analysis shows promise for broader clinical practice integration.

Conclusions:

• Significant progress in AI-enhanced MRI analysis, particularly in quantitative methods and deep learning.
• Challenges persist in standardization, validation, and clinical translation of advanced MRI techniques.
• Further research and development are needed for routine integration into medical practice.