Computer-Aided Detection (CADe) and Segmentation Methods for Breast Cancer Using Magnetic Resonance Imaging (MRI)

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Oncology

Background:

• Breast cancer remains a significant health issue, underscoring the need for effective early detection methods.
• Magnetic resonance imaging (MRI) offers high sensitivity for detecting invasive breast cancers, crucial for improving patient outcomes.
• Computer-aided detection (CADe) systems enhance MRI efficacy by identifying suspicious lesions and assisting radiologists.

Purpose of the Study:

• To provide a comprehensive review of current computer-aided detection (CADe) systems in breast MRI.
• To analyze the technical aspects of CADe pipelines and segmentation models, from classical methods to deep learning.
• To identify challenges and future directions for CADe systems in clinical breast MRI practice.

Main Methods:

• Review of technical pipelines and segmentation models in breast MRI CADe.
• Analysis of classical intensity-based methods, machine learning (ML), and deep learning (DL) architectures.
• Examination of CADe implementation with multi-parametric MRI acquisitions.

Main Results:

• Advancements range from traditional algorithms to sophisticated DL models like U-Nets.
• Current CADe systems face challenges including variable accuracy rates, data interpretation complexity, and performance variability.
• Technical hurdles include image artifacts and the need for explainable detection algorithms.

Conclusions:

• Despite progress, CADe systems require more robust, generalizable, and interpretable algorithms for clinical adoption.
• Future research should focus on explainable AI, multi-purpose AI, and integration with large language models for enhanced reporting.
• Standardizing MRI protocols and reducing costs are essential for optimizing CADe system accessibility and clinical utility.