Scientists using an artificial intelligence learning model can spot type 2 diabetes early on an abdominal CT scan, academics have said.
In the US, around 13 per cent of adults have type 2 diabetes and more than 34 per cent have pre-diabetes.
According to the researchers, it is vital to detect pre-diabetes early to combat the development of type 2 diabetes by making healthy lifestyle changes.
- High rate of type 2 diabetes and high blood pressure associated with brain changes
- Cardiovascular disease triggered by type 2 diabetes and cognitive impairment, research shows
The findings of this study have revealed that abdominal CT scans can identify the early signs of type 2 diabetes.
Prior research has identified that individuals with type 2 diabetes have more visceral fat stored in their pancreas compared to people without the condition.
Top author, Dr Hima Tallam said: “The analysis of both pancreatic and extra-pancreatic features is a novel approach and has not been shown in previous work to our knowledge.”
During the study, the team of scientists analysed the data of more than 8,000 individuals who had been screened between 2004 and 2016. More than 500 of the participants had type 2 diabetes and more than 1,800 has dysglycemia.
In addition, the team used a deep learning model to assess images from CT scans. By using the model, the researchers can examine the participant’s pancreatic features, visceral fat, density and volumes of the surrounding abdominal muscles and organs.
They found that pancreas density was lower amongst people with type 2 diabetes compared to those without.
People with the condition also had a higher amount of visceral fat compared to individuals without type 2 diabetes, the findings have reported.
Fellow author, Dr Ronald M. Summer said: “We found that diabetes was associated with the amount of fat within the pancreas and inside the patients’ abdomens.
“The more fat in those two locations, the more likely the patients were to have diabetes for a longer period of time.”
The authors added: “This study is a step towards the wider use of automated methods to address clinical challenges.
“It may also inform future work investigating the reason for pancreatic changes that occur in people with diabetes.”
The study was published in the journal Radiology.
