Novel Data-Driven Models for Robust Ultra-Low-Dose PET Image Reconstruction and CT-Free PET Image Synthesis (PETGEN)

This project develops advanced AI models to improve ultra-low-dose PET imaging and enable CT-free PET image synthesis. The goal is to reduce radiation exposure while maintaining high diagnostic image quality for safer and more efficient medical imaging.

Key data

Description

Positron Emission Tomography (PET) plays a key role in diagnosing cancer. However, PET imaging exposes patients to radiation from injected tracers and accompanying CT scans. Reducing this radiation while maintaining diagnostic image quality remains a major clinical challenge.

This project develops advanced AI-based methods for ultra-low-dose PET imaging and CT-free PET image synthesis. Using modern data-driven and generative models, we aim to reconstruct high-quality PET images from extremely low-dose acquisitions and explore imaging workflows that reduce the need for CT scans.

The project is conducted in close collaboration with leading clinical and academic partners, supporting the development of clinically robust and generalisable AI solutions for next-generation PET imaging. A strong focus is placed on translational research, bringing data-driven innovations in medical imaging closer to clinical application. We welcome collaborations with clinical institutions interested in advancing multimodal imaging workflows.