Create AI models for prostate cancer and kidney cancer that incorporate different types of health data, including image data, medical health records, and biomarkers.
Overcome problems of data sparsity through privacy-protecting methods merging multilingual inputs, improving access to data, and facilitating strong, multinational cooperation.
Use AI to group patients with suspected prostate cancer and kidney cancer into different categories based on their clinical characteristics, prognosis, and best treatment options. This will be measured by changes in diagnostic and treatment accuracy.
Enable meaningful patient stratification based on an augmented assessment of clinical phenotype, individual prognosis, and optimal treatment strategies. In this context, COMFORT aims to explore novel and advanced stratification concepts, e.g., based on an AI risk score calculated from the input data.
Implement the developed models into hospital IT infrastructure to verify and measure their performance in a large, prospective multicentre international study and to enable continuous learning approaches.
Provide evidence on how the models help improve patient care, by directly comparing them to standard-of-care approaches in a real-world clinical environment. This will be completed post-project.
Assess the trust of healthcare providers and patients in AI and investigate how it can be increased.
Present insights from AI models to patients and healthcare providers to maximise usability and societal acceptance.
Create a plan for getting approval to make the AI models available as medical devices in the European Union. This will be done after the project is finished.