October 13, 2021
Type 2 diabetes can be diagnosed with a whole-body magnetic resonance imaging (MRI) scan. This is the result of a current study by researchers from the German Center for Diabetes Research, the Institute for Diabetes Research and Metabolic Diseases of the Helmholtz Center Munich at the University of Tübingen, the Max Planck Institute for Intelligent Systems and the University Clinic Tübingen. They used deep learning methods (machine learning) and data from more than 2000 MRIs to identify patients with (pre-) diabetes.
Being overweight and having a lot of body fat increases the risk of diabetes. But not every overweight person also gets it. What matters is where the fat is stored in the body. If fat accumulates under the skin, it is more harmless than fat in deeper areas of the abdomen (so-called visceral fat). How the fat is distributed in the body can be shown well with full-body magnetic resonance imaging. “We have now investigated whether type 2 diabetes could also be diagnosed on the basis of certain patterns of body fat distribution in the MRI,” said last author Prof. Robert Wagner, explaining the researchers’ approach.
Deep learning trained with over 2,000 MRI images
To recognize such patterns, the researchers used artificial intelligence (AI). They trained deep learning networks (machine learning) with full-body MRI images of 2,000 people who had also undergone screening with an oral glucose tolerance test (abbreviated oGTT). With the oGTT, also known as the sugar stress test, a disturbed glucose metabolism can be detected and diabetes diagnosed. This is how the AI learned to detect diabetes.
Fat accumulation in the lower abdomen is an important indication of the development of diabetes
“An analysis of the model results showed that fat accumulation in the lower abdomen plays a decisive role in the detection of diabetes,” reports Wagner. Further additional analyzes also showed that some of the people with a preliminary stage of diabetes (prediabetes) and people with a diabetes subtype that can lead to kidney disease can also be identified via MRI scans.
The researchers are now working on deciphering the biological control of body fat distribution. One goal is to use new methods such as the use of AI to identify the causes of diabetes in order to find better preventive and therapeutic options.
Deep learning is a special method from the field of machine learning with artificial neural networks (ANN) and thus also a sub-area of artificial intelligence (AI). Deep learning is particularly suitable when there is a lot of unstructured data – such as images and recordings. In order to teach deep learning algorithms to correctly evaluate images and predict diagnoses, they are trained on annotated (information provided) data.
The results were published in the journal JCI Insight: https://insight.jci.org/articles/view/146999/pdf