According to the study presented at the annual meeting of the Radiological Society of North America (RSNA), artificial intelligence can be used as an automated and accurate measurement instrument marker common heart disease in patients undergoing chest CT screening for lung cancer.
Low-dose CT of the chest is approved for screening of lung cancer in people at high risk, such as smokers, "with the experience." Although CT is designed for the diagnosis of lung cancer, it also displays the coronary artery calcium - a recognized index that helps doctors decide which patients should take preventive medicines to lower cholesterol (statins).
"The new guidelines for the control of cholesterol encourage the use of calcium indicator when doctors decide whether to prescribe statins - the scientists noted -. For some patients, with an average risk for cardiovascular diseases and calcium indicator equal to 0, statins can be postponed . If the calcium component is high, the statin is needed. "
Despite the predictive value of coronary calcium aretrii usually not measured in screening lung CT in low doses, as measurements require special software and increase the time of interpretation.
The research group recently developed and tested technique which uses a deep training, complex AI type for automatic measurement of calcium in the coronary artery on the CT images of the thorax. They trained the system to measure the calcium during cardiac CT and CT of the chest. They then tested the system on the results of computed tomography of thousands of smokers aged 55-74 years, who were part of the National Lung Screening studies (NLST).
It was found that calcium indicators derived from AI were close to those who have identified experts. In addition, there was a significant relationship between calcium indicators over the next 6.5 years, calculated by the system and cardiovascular death.
Scientists have noted that the automated method of extracting information about the level of calcium can help patients and doctors make decisions about preventive therapy, as well as be used to divide people into groups of high and low risk.
deep learning system works in the background without affecting the time of screening.
In the future, this tool can be used in almost every scan of the chest to generate clinically relevant information for a large number of patients.