Machine learning applications in prostate cancer magnetic resonance imaging

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence

Background:

• Machine learning (ML) and deep learning (DL) are increasingly applied in medical imaging.
• Prostate magnetic resonance imaging (MRI) offers rich data for ML analysis.
• Distinguishing between clinically significant and indolent prostate cancers is crucial for patient management.

Purpose of the Study:

• To review recent applications of ML in prostate MRI.
• To explain ML and DL concepts relevant to radiology.
• To outline potential clinical uses of ML in prostate cancer assessment.

Main Methods:

• Review of recent literature on ML applications in prostate MRI.
• Explanation of ML/DL principles and their differences from classical algorithms.
• Highlighting ML/DL pipeline differences for prostate MRI.

Main Results:

• ML applications cover gland segmentation, lesion aggressiveness assessment, cancer detection/diagnosis, localization, staging, and recurrence prediction.
• Many applications utilize unenhanced MRI sequences.
• Promising results are reported for various prostate MRI tasks.

Conclusions:

• ML demonstrates significant potential in enhancing prostate MRI interpretation.
• Clinical validation across diverse settings, vendors, and field strengths is essential.
• Further research is needed to bridge the gap between promising results and routine clinical practice.