AI diagnoses the condition of blood vessels
The new software developed by researchers at King's College London and the University of Edinburgh, working under the guidance of Dr. Paul Bentley, allows you to accurately identify and assess the severity of microangiopathy – one of the most common causes of stroke and dementia (dementia).
Microangiopathy is a very common disease among the elderly, which is characterized by a decrease in the amount of blood flowing to the interneuronal contacts in the deep layers of the white matter of the brain, which leads to damage and, ultimately, to the death of nerve cells. As a result, a person may develop mental balance disorders, as well as stroke or dementia. The severity of microangiopathy increases with age and worsens against the background of hypertension and diabetes mellitus.
To date, doctors diagnose microangiopathy by detecting changes in the white matter recorded in images obtained using magnetic resonance imaging or computed tomography. In fact, the diagnosis consists in the fact that the doctor measures the size of the lesion on the image. In the images obtained using computed tomography, the boundaries of this focus are often very poorly distinguishable, which makes it difficult to determine the severity of the disease. Magnetic resonance imaging in this case is more sensitive, but its application is limited by the availability of equipment, as well as the impossibility of its use during emergency care and when working with patients of very advanced age.
The authors explain that the importance of the approach developed by them lies in the fact that the software allows you to automatically determine the severity of the disease with high accuracy based on the results of computed tomography. It can be used both for screening diagnostics and for monitoring dementia, as well as when it is necessary to make emergency decisions about stroke therapy.
According to Dr. Bentley, the software will allow you to personalize the treatment of acute neurological conditions. For example, in case of a stroke, blood clot-dissolving drugs can be administered to quickly release the lumen of the artery. However, such approaches are fraught with serious side effects, such as bleeding. In the future, the program can be used to assess the risk of bleeding and doctors will be able to make a decision on the administration of thrombolytic drugs based on individual patient data.
The program will also help identify patients at risk for developing dementia or loss of mobility due to slowly progressing microangiopathy, drawing the attention of attending physicians to potentially avoidable causes of its development, such as high blood pressure or diabetes mellitus.
To test their software, the authors used 1082 archival images of the brain of stroke patients obtained using computed tomography in 70 UK clinics in the period 2000-2014. The program identified and measured a marker of microangiopathy and calculated an indicator corresponding to the severity of the disease. The results obtained were compared with the results provided by a commission of highly qualified doctors who visually assessed the severity of the disease using the same images. The degree of conformity for the results obtained through the program and the results presented by the commission was comparable with the degree of conformity of the opinions of different experts.
Additionally, the authors obtained magnetic resonance and computed tomography data from 60 patients and used the results of magnetic resonance examination to determine the exact size of the microangiopathic injury zone. Subsequent comparison of the results demonstrated that the accuracy of diagnostics using software is 85%.
Currently, researchers are practicing the use of similar techniques to assess the extent of brain tissue atrophy and other conditions traditionally diagnosed by computed tomography.
Article by Liang Chen et al. Rapid Automated Quantification of Cerebral Leukoaraiosis on CT Images: A Multicenter Validation Study published in the journal Radiology.
Evgenia Ryabtseva, portal "Eternal Youth" http://vechnayamolodost.ru based on the materials of Imperial College London: Artificial Intelligence improves stroke and dementia diagnosis in brain scans.