Identifying metabolic networks using inter-organ analysis of whole-body [18F]FDG-PET imaging data – INTER-ORGAN PET

Context. Our understanding of diagnosis and effective treatment of complex diseases is limited. Recent studies point towards inter-organ axes communication and network effects that underpin numerous pathologies. Diagnostic molecular imaging methods, including whole-body positron emission tomography (WB-PET), could support the non-invasive as-sessment of metabolic and signaling pathways on cellular, organ or cross-organ levels. WB-PET using [18F]-labelled glucose (FDG) as the tracer-of-choice offers unique opportunities to assess glycolytic pathways and to map inter-organ signalling. However, the appreciation of wider network effects requires an architecture of molec-ular and phenotypic net-works for an in-depth understanding of disease expression and response to therapies.
Hypotheses. Our hypothesis is that we can assess these network effects through the use of whole-body [18F]-FDG PET, which allows simultaneous and quantitative characterisation of glycolic activity throughout the body. First, we will build an atlas of FDG-PET signals that are associated with non-pathological homeostatic metabolic networks in healthy subjects. Next, we will consider breast cancer as a perturbation of these networks. Thus, we hypothesize that breast cancer is associated with inter-organ aberrations or altered network effects. We will com-pare these aberrations as a function of the disease stages. In summary, we plan to utilize WB-FDG-PET infor-mation and tailored data analytics pipelines to describe inter-organ effects in breast cancer patients in comparison to normal, homeostatic pathways. Our assumption is that such information will ultimately be useful for personalized patient management.
Approach. We take advantage of our large databases of FDG-PET scans acquired in more than 1,000 subjects (healthy controls and cancer patients). Through the power of deep learning, we seek to 1) extract metabolic signals from multiple key organs following automated segmentation, 2) build an inter-organ association network with dense-ly connected nodes, and 3) categorize subjects based on these network nodes and their health status.
Added value. This project can add to our understanding of diseases (namely breast cancer) and help personalize medicine using WB-FDG-PET data. We will generate the first maps of inter-organ metabolic interactions in living organisms. We will effectively combine the extensive expertise of the two partners in molecular imaging, advanced quantitative biomarker analysis and data science. This project leverages a critical mass within complementary research teams to generate fundamental insights into the pathomechanisms.
Researchers. All coopted experts come with extensive complementary skills and knowhow and will interact close-ly. Crucial core competencies include experimental and clinical research in molecular imaging, the study of patho-physiological processes on a molecular level, as well as computational biology and multi-parametric data analysis.