Year: 2025
Type: article
Abstract: - Prostate cancer is one of the leading causes of cancer-related mortality among men, especially in the aging population. Timely and accurate detection plays a pivotal role in improving survival rates and optimizing treatment strategies. Traditional diagnostic approaches relying on multi-parametric MRI (mp-MRI) are effective but highly dependent on manual interpretation, which is both time-consuming and vulnerable to human error. This study introduces an AI-powered diagnostic framework that combines 3D deep learning and feature fusion techniques to detect prostate cancer more accurately and efficiently. By integrating T2-weighted, diffusion-weighted (DWI), and dynamic contrast-enhanced (DCE) MRI modalities, we develop a comprehensive system that leverages a 3D U-Net for precise lesion segmentation and late fusion classification models to enhance decision-making. The system is trained and tested on public datasets like PROSTATEx and SPIE-AAPM-NCI, and its performance is evaluated using clinically relevant metrics such as Dice Score, AUC, sensitivity, and specificity. A user-friendly visualization dashboard is also developed to assist radiologists in interpreting the results. Our work demonstrates the potential of AI-driven tools to support clinicians in early prostate cancer diagnosis while maintaining patient data privacy through anonymized processing. Key Words: Prostate Cancer, Deep Learning, 3D U-Net, MRI, AI in Healthcare, Diagnosis.
DOI: 10.55041/ijsrem49647
Author: Dr. Priya Nandihal
Institution: Dayananda Sagar Academy of Technology and Management
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4411080361