Scientists in the United States have been able to detect early signs of type 2 diabetes on abdominal CT scans using a fully automated artificial intelligence system.
Almost one in eight adults suffers from diabetes, while one in three meets the criteria for prediabetes. Due to the slow onset of symptoms, it is important to be diagnosed with diabetes early. Some cases of prediabetes can last up to eight years. The earlier the diagnosis is made, the sooner the patient will make the necessary changes in his lifestyle and the appropriate treatment to “slow down” the progression of the disease.
Abdominal tomography can be a promising tool for diagnosing type 2 diabetes, as it provides important information about the pancreas. Previous studies have shown that patients with diabetes tend to accumulate more abdominal fat and fat in the pancreas than non-diabetics.
So far, however, the liver, as well as the muscles and blood vessels around the pancreas, have not been studied extensively in relation to diabetes. The analysis – with the help of artificial intelligence – of the characteristics of both the pancreas itself and the surrounding areas is an innovative approach.
The researchers, led by Dr. Ronald Summers, a radiologist at the National Institutes of Health Clinic in Maryland, and Perry Pickhardt, a professor of radiology at the University of Wisconsin School of Medicine, conducted the , analyzed data for nearly 9,000 patients, of whom 572 were diagnosed with type 2 diabetes and 1,880 with dysglycaemia (very high or very low blood sugar levels).
The researchers have developed a model of artificial intelligence (deep learning) training it with 471 tomographies. The model was then able to diagnose diabetes on other CT scans, with virtually no difference in accuracy compared to doctors. In addition to the pancreas itself, the model was able to find traces of diabetes based on belly fat, as well as the density and volume of muscles and organs around the pancreas.
Patients with diabetes generally have lower pancreatic density and more abdominal fat than diabetics. “We found that diabetes was related to the amount of fat in patients’ pancreas and abdomen. “The more fat there is in these two parts of the body, the more likely patients are to have diabetes for a longer period of time,” Summers said.
The artificial intelligence system was able to utilize the above to accurately distinguish patients with or without diabetes.
Follow Skai.gr on Google News
and be the first to know all the news