| MS Seminar

Name of the Speaker: Ms. Divya Bharti (EE22S018)
Guide: Dr. Mohanasankar S
Online meeting link: https://meet.google.com/nap-gswh-nqc
Date/Time: 29th April 2025 (Tuesday), 4 PM
Title: AAD-DCE: An Aggregated Multimodal Attention Mechanism for Early and Late Dynamic Contrast Enhanced Prostate MRI Synthesis.

Abstract :

Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) is a widely adopted imaging technique for detailed assessment of tissue perfusion and the detection of suspicious lesions, playing a crucial role in guiding clinical decisions, including targeted biopsies in prostate cancer. Despite its diagnostic value, the requirement for Gadolinium-based (Gad) contrast agents raises concerns due to potential toxicity, especially in patients with renal impairment or those requiring repeated imaging. Recent deep learning-based approaches have explored the synthesis of DCE-MRI to eliminate the need for contrast administration. However, these methods predominantly rely on unimodal inputs, typically either non-contrast or low-dose contrast images, limiting their ability to accurately model localized perfusion dynamics within critical anatomical regions.

To address these challenges, we propose AAD-DCE, a novel multimodal deep learning framework for DCE-MRI synthesis. Our approach is built upon a Generative Adversarial Network (GAN) architecture, augmented with an Aggregated Attention Discriminator that integrates both global and local discriminators. These discriminators generate spatially-aware attention maps, which guide the generator in synthesizing both early and late-phase DCE-MRI with high fidelity. Our findings highlight the potential of attention ensembling and multimodal integration in advancing contrast-free prostate cancer imaging, offering a safer and more accessible alternative to traditional DCE-MRI protocols.