Clinical and Industry Partners Sought to Validate Synthetic Medical Imaging AI

Healthcare AI needs large, diverse and complete datasets, but medical-imaging data are frequently limited by missing examinations, incomplete clinical information, underrepresented patient groups and differences between scanners or imaging protocols.

Prodigy AI Solutions is developing a generative machine-learning model designed to complement existing medical datasets with high-quality synthetic data.

Starting with thoracic CT imaging, the model is designed to:

• Expand underrepresented demographic and clinical groups
• Generate missing CT acquisition types, such as non-contrast, contrast-enhanced and low-dose scans
• Reconstruct selected missing clinical variables
• Reduce technical differences caused by scanners and acquisition protocols
• Preserve clinically relevant anatomical and radiomic characteristics

The solution is currently at the MVP and experimental-validation stage. It is intended for medical research and AI-model development—not as a substitute for real patient data, clinical evidence or medical diagnosis.

We are looking for:

• Hospitals, radiology departments and oncology centres
• Radiologists, oncologists and radiomics specialists
• Universities and medical-imaging research groups
• Medical-device and healthcare AI companies
• Pharmaceutical and life-sciences organisations
• European research consortia and innovation programmes

Possible collaboration activities include clinical and technical validation, evaluation on authorised datasets, joint research projects, scientific publications, proof-of-concept studies and commercial pilots.

Prodigy contributes the ML architecture, generative-AI development, cloud and container deployment capabilities, and previous experience evaluating AI infrastructure on CINECA’s LEONARDO supercomputer. Our team is also the developer of Verbis Graph, an enterprise knowledge-retrieval platform available through major cloud marketplaces.

We welcome partners who can contribute clinical expertise, validation methodologies, authorised research data or real-world use cases. Together, we aim to make healthcare AI datasets more complete, representative and useful while maintaining rigorous clinical, privacy and fairness evaluation.