In the national press again: Tackling osteoarthritis with AI

The use of artificial intelligence in medicine is on the rise, with ImageBiopsy Lab at the forefront in the fight against osteoarthritis (OA). The aim is to extract information from an X-ray in order to enable an early prediction of the disease and therefore optimize patient outcomes. Armed with a Deep Learning algorithm trained on over 150 000 X-ray images, the ImageBiopsy Lab solution proposes almost real-time diagnoses to orthopedists and radiologists. Up to 85% of these diagnoses are accepted by physicians. “We want to exceed 90% and this will require further work” says CEO Richard Ljuhar.

Saving time and cost for patient benefit

The issues with the current diagnostic workflow can be tackled with an automated software solution. High variation rates within diagnoses demand a standardized method. Furthermore, studies conducted within the clinical validation process of the software have shown that physicians have spent up to half an hour per X-ray image. In contrast, ImageBiopsy Lab’s software delivers an almost instantaneous result.

The future is present in X-rays

However, research at ImageBiopsy Lab is not restricted to simply diagnosing OA in X-rays. “We have developed models which enable a prediction of the risk. In studies, we were able to predict with an accuracy of 80% whether the severity of the condition will deteriorate within the next four years” Ljuhar explains. This is already being used in clinical studies.

ImageBiopsy Lab originated from Braincon, an Austrian company. Over €1 million worth of funding from aws, INiTS and the Wirtschaftsagentur Wien enabled a rapid growth. With collaborators like Prof. Stefan Nehrer, head of the Center for Regenerative Medicine and Orthopedics of the Danube University Krems, ImageBiopsy Lab quickly spun off on its own. With an expected €200 000 in revenues this year, we are taking the healthcare sector by storm.

Please see the national press releases below:

Link (de): (24.1.2018)

Link (de): (24.1.2018)

Link (de): (24.1.2018)