Automated assessment of rheumatoid arthritis symptoms in mice

A machine learning approach to the assessment of rheumatoid arthritis: design and implementation

© Fraunhofer ITMP | IGD
Ein Foto von Mauspfoten von unten und die ersten Schritte zur Generierung eines Algorithmus zur Automatisierung der Wertung
© Fraunhofer ITMP | IGD
Unterseite der Mäusepfote mit Merkmalspunkterkennung

Rheumatoid arthritis (RA) is a chronic autoimmune disease that primarily affects the joints, leading to inflammation, pain, stiffness and progressive joint damage. Rheumatoid arthritis is one of the most common autoimmune diseases, affecting approximately 1% of the world's population. The social relevance of RA is significant due to its impact on quality of life, ability to work, and overall well-being of sufferers. In addition, RA is often associated with comorbidities such as cardiovascular disease, osteoporosis, and depression, which further increases the social and economic burden.

 

Reliable, reproducible data and time savings

In the context of research on rheumatoid arthritis, disease symptoms in the mouse model can be evaluated by observing the paws. This evaluation has been done manually so far. Automation of this paw evaluation system would enable increased accuracy, efficiency and consistency. This project is developing this automated assessment of rheumatoid arthritis symptoms in the mouse model. During the course of the disease, photos of the mouse paws are taken and used to develop an algorithm.

Such software can help ensure that results from RA research are reliable and reproducible and that data collection is less time-intensive. It would help make the results of paw tests in RA research more comparable, and thus better evaluate new treatments and pave the way for new insights and discoveries. We believe that this tool will not only save hours of work but also provide more accurate and reproducible data.

 

Outlook

The rheumatoid arthritis mouse model has been implemented and comprehensive photos from different perspectives, along with accompanying videos, have already been created. We are currently working on developing a suitable algorithm that closely examines important features of the paw for evaluation. We are also planning further animal studies to collect image and video data to train the algorithm and test its accuracy. 

After the validation of the software, it will be made available to other research groups for use and evaluation in similar mouse studies. In addition, we want to further optimize the test environment. To do this, the software will be integrated into a device that simplifies the positioning of the mouse paws and optimizes the image and video data collection. Our goal is to use the algorithm to automatically evaluate each paw and to create an overall evaluation of all four paws.