LOOBesity: Closing the loop towards precision medicine in obesity
Problem: Elevated stress hormone metabolism (glucocorticoid) exacerbates health risks in obesity
Goal: Optimization of treatment for obese patients with altered metabolism
Obesity is a widespread disease that increases the risk of secondary conditions such as cardiovascular (e.g. high blood pressure) and metabolic (e.g. diabetes) complications. Diets and exercise often fail to achieve lasting success. Moreover, the choice of therapy depends on the personal experience of the treating physicians, and there is a lack of concrete indicators for which therapy is most suitable for which individual. This is precisely where the LOOBesity research project, led by Felix Beuschlein, comes in: the goal is to develop individualized and therefore more effective therapies against obesity for a selected group of affected individuals.
Particularly affected are people with elevated glucocorticoid metabolism, which is typically associated with stress. This group exhibits a higher risk for secondary diseases. For this patient group, molecular and radiological criteria are used to derive insights into the composition and distribution of body fat. Using artificial intelligence (AI), the individual risk for secondary diseases is assessed, and based on this, a personalized therapy recommendation is developed. Individuals at increased risk should then receive optimized treatment that leads to long-term weight reduction.
LOOBesity combines various data sources such as cohort studies, molecular signature analyses, imaging techniques, and computer-assisted decision support systems to develop targeted personalized treatment strategies for obese patients. In 2024, the project made significant progress and largely achieved its set goals.
Selected research findings
An important milestone was the expansion of the Zurich obesity cohort to 437 individuals. In addition, the study protocol was extended to include microbiome analysis. This added dimension helps to better understand the different manifestations of obesity and to predict their response to various treatments.
Personalized pharmacotherapy
Progress has also been made in the field of personalized pharmacotherapy. A functional prototype of a decision support system was developed that evaluates the benefits and risks of treatment with so-called GLP-1 receptor agonists in real time (graphic). These agents (GLP-1 receptor agonists) are used to treat type 2 diabetes, and some of these substances can lead to a reduction in body weight. Based on successful pilot experiments, this decision support system will be integrated into clinical processes over the course of 2025.
Another central field of research is the molecular analysis of adipose tissue. Significant progress has been made in the molecular characterization of adipose tissue samples. The methods for sample collection and RNA sequencing were optimized, which has significantly improved the quality of the data. The goal is to identify correlations between gene expression data and individual responses to therapies.
Imaging techniques for the characterization of body fat
In addition to molecular analysis, phenotyping using non-invasive imaging techniques also plays a crucial role. For this purpose, special magnetic resonance imaging (MRI) methods were developed to quantitatively measure fat distribution in muscle tissue. In parallel, machine learning-based methods were developed for the automatic segmentation of body fat. An important application of artificial intelligence (AI) in image analysis is the quantitative determination of metabolically active brown adipose tissue.
The current standard method for assessing the metabolic activity of this tissue is to measure local glucose consumption using positron emission tomography (PET) – a costly method that requires radioactive substances ([18F]-FDG, a short-lived radioisotope fluorine-18) and is therefore not suitable for use in large groups. However, the LOOBesity project was able to demonstrate that comparable results can be obtained through the analysis of X-ray computed tomography (CT) images. Using neural network analyses, it was shown that glucose consumption data correlate with X-ray attenuation in CT images. Since CT is more cost-effective and also involves significantly lower radiation exposure, this method could potentially be used more broadly in the future.
Outlook
In summary, LOOBesity developed key tools in its first year to successfully advance the project. The combination of microbiome, molecular, imaging, and clinical data forms the foundation for a precise characterization of different types of obesity and the development of tailored therapies. These approaches will be further refined in the next project phase in order to sustainably improve the treatment of obesity.
Groups involved:
- Felix Beuschlein: Clinic for Endocrinology, Diabetology, and Clinical Nutrition,University Hospital Zurich
- Thomas Frauenfelder: Institute for Diagnostic and Interventional Radiology, University Hospital Zurich
- Ender Konukoglu: Department of Information Technology and Electrical Engineering, ETH Zurich
- Milo Puhan: Epidemiology, Biostatistics and Prevention Institute, University of Zurich
- Christian Wolfrum: Department of Health Sciences and Technology, ETH Zurich
Featured in:
Effectively Combat Obesity
↑ Interview with Prof. Beuschlein (in German)
Project Overview
Lead:
Prof. Dr. Felix Beuschlein
Director Department of Endocrinology, Diabetology and Clinical Nutrition and
Director Center for Obesity and Metabolic Surgery at the University Hospital Zurich
Start: February 2023
Duration: 2023 – 2028
Universities: ETH Zürich, University of Zurich
Hospitals: University Hospital Zurich
Researchers: 20 – 25
Partners: TU München
Patients: 500