Oncology researcher (scientific coordination by Prof. Armando Santoro)
📍 Humanitas Research Hospital, Milano
Grant received: €3.000Developing artificial intelligence models (GANs) to improve diagnostic imaging in lung cancer, focusing on early prediction and risk stratification. Project coordinated by Prof. Armando Santoro at Humanitas.
The project is in the phase of collecting and harmonising clinical data from two retrospective cohorts of NSCLC patients treated with chemo-immunotherapy and immunotherapy. The database will soon be shared with the AI Center to generate and validate synthetic data using artificial intelligence models; in parallel, patient enrolment in the prospective validation cohort and coordination of the liquid biopsy have begun. Data cleaning and the structuring of the final database are currently being completed before the transfer to the AI Center; the development of generative models and the comparison between real and synthetic data for statistical reliability and clinical usefulness will follow.
«This funding is a crucial step in turning real-world clinical data into innovative tools for cancer research. It will allow us to develop the project and to share future results in international scientific settings.»
— Silvia Masini
To generate adequate synthetic data from a real-world cohort of well-annotated consecutive lung cancer patients using GANs and other generative models, and to validate the generated synthetic data with a validation framework in terms of statistical fidelity, clinical utility and privacy preservability.
We expect to generate high-fidelity synthetic datasets that reliably reproduce the statistical distribution and clinical complexity of real-world NSCLC cohorts. Validated through the SAFE and MOSAIC frameworks, the models will demonstrate an improved ability to predict outcomes in first-line immunotherapy, while ensuring robust privacy preservation and enabling secure data sharing.
Every €3,000 raised = one more grant in the next call. You run, you donate, they do research.
Donate now See the open call