Precision medicine

by integrating Multiscale Functional Imaging and Advanced Machine Learning

In the coming decades, our aging population and the expected increase of people with chronic disease will pose an increasing burden on society. There is therefore an urgent need to develop new preventive and new therapeutic strategies for common age-related diseases. Tailored prevention requires an understanding of the earliest stages of disease in combination with the identification of persons who are at risk. Tailored therapy relies on methods for accurate diagnosis and prognosis. Recent research has shown that image-based techniques are among the most promising of the available technologies to

improve diagnosis, prognosis, and treatment selection. On the one hand, the application of sophisticated imaging technology in the context of large studies, combined with the use of advanced data-driven approaches including deep learning holds the current promise of precision medicine, i.e. individualized prognosis, and treatment tailored to the individual patient, compare the right-hand side of the diagram below. On the other hand, state-of-the-art physics-driven clinical imaging techniques, like CT, MRI, optical and ultrasound imaging, can provide different structural and functional information at high

resolution. Integrating all the information from multiple modalities at multiple spatial and temporal scales for a patient-specific assessment is an extremely challenging problem, compare the left-hand side of the diagram below. 

PRECISION MEDICINE entails a paradigm shift from a one-size-fits-all to a tailored and personalized approach in medicine. This is relevant for early diagnosis and to deliver the right treatment in the right way at the right moment to enable a better prognosis for patients. It holds the promise of improving quality of life and, simultaneously achieving a reduction of health care costs.

The aim of our consortium is to realize a breakthrough in medical imaging technology to facilitate precision medicine: the optimal treatment for individual patients. We can achieve this by a unique combination of expertise in model-driven imaging science (physics-based) on the one hand, and in data-driven imaging science (data science-based) on other hand, compare the dashed block in the diagram below. Most important, at each TU we have identified the experts on the left- and right-hand side, as well as those acting at the interplay of both worlds.

Our consortium aims to develop novel imaging tools for patient-specific diagnosis, prognosis, and precision therapy, by leveraging imaging across different modalities, different scales and state-of-art machine learning techniques.

This consortium consists of: