Specialists from ITMO University and the V. A. Almazov National Medical Research Center have created an innovative algorithm based on artificial intelligence that allows for the rapid detection of cardiac fibrosis. The development significantly speeds up the processing of MRI scans, reducing it from 1-2 hours to several minutes.
The new algorithm uses a semi-automatic deep learning model. First, the neural network determines the area of the heart, then identifies the presence of fibrosis, recognizes 17 segments of the heart, and assesses the volume of fibrosis in each of them.
Unlike a specialist, who spends about 1-2 hours on processing, our model copes in a couple of minutes
Fibrosis is scar tissue that can occur after a heart attack or infectious diseases. Timely diagnosis with MRI is important for choosing effective treatment.
The user only needs to mark a few points on the image of the heart, and tissue segmentation and diagram creation are fully automated
Scientists continue to work on improving the method and plan to create a fully automatic algorithm that can instantly analyze images without user intervention.
Read more on the topic:
Project for analysis tomography for cancer diagnosis based on AI launched at NSU
NtechLab started creating a neural network for the diagnosis of Alzheimer's disease