Using neural networks to recognize arthritis in an MRI



Credit: Pixabay/CC0 Public domain

Arthritis is a widespread condition that affects hundreds of thousands of people and results in inflammation of the joints. It has many different causes, and if doctors want to treat the disease properly, it is important that they can determine exactly what type of arthritis the patient has. It is often not an easy undertaking. A number of different parameters need to be considered and a definitive diagnosis is often only possible as the disease progresses.

Computer scientists from the 5 Pattern Recognition Chair of Computer Science at the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) and physicians from the 3 Department of Medicine – Rheumatology and Immunology and the Institute of Radiology at the Universitätsklinikum Erlangen conducted a study to determine if neural networks can determine whether a patient has rheumatoid arthritis (RA) or psoriatic arthritis (PS). Result: the AI ​​was able to differentiate between the two types in 75% of cases.

The team had only recently investigated whether neural networks could determine the type of arthritis using high-resolution computed tomography images. They succeeded. According to Professor Frank Roemer of the Institute of Radiology, “The advantage of MRI over CT is that an MRI gives a more accurate picture of the extent of inflammation and the joint structures affected.”

For the study, the team led by computer scientist Lukas Folle used five different MRI sequences from 649 patients to train and test an innovative neural network. The network was able to classify the type of arthritis patients suffered from in 75% of cases based on the MRI images.

Additionally, the team tested how the neural network classified cases of psoriasis, which can often progress to PSA. The network correctly classified most cases of psoriasis that later developed into PsA as PsA.

Folle thinks “it is possible that the neural network picks up on early changes or other structural features in patients with psoriasis and uses them to classify them accordingly.”

“Our results indicate that MRI scans can show changes that the neural network has identified as relevant to classifying different forms of arthritis and that have not been described to date,” adds PD. Dr. David Simon, a physician involved in the study. “We now aim to continue training and improving the neural network, with a view to its eventual use in clinical practice,” explains Lukas Folle.

The team has just published its results in Rheumatology.

Neural network learns to differentiate healthy bones from inflamed bones using finger joints

More information:
Lukas Folle et al, Advanced neural networks for MRI classification in psoriatic arthritis, seronegative and seropositive rheumatoid arthritis, Rheumatology (2022). DOI: 10.1093/rheumatology/keac197

Provided by Friedrich-Alexander-University Erlangen-Nurnberg

Quote: Using Neural Networks to Recognize Arthritis in an MRI (2022, September 23) Retrieved September 23, 2022 from html

This document is subject to copyright. Except for fair use for purposes of private study or research, no part may be reproduced without written permission. The content is provided for information only.

Source link


Comments are closed.