Streamlining X-ray Processing with AI


Over the past two decades, the number of data points per patient has increased exponentially, with a thousand more images per patient being produced today compared to the turn of the millennium. This surge in data has created a significant challenge for radiologists who are struggling to keep up with the volume of examinations. The ability to effectively diagnose and monitor diseases through X-rays was becoming increasingly difficult due to the sheer volume of data. Emids was approached to find a way to manage this data effectively and efficiently without compromising on the quality of patient care.


In response to this challenge, a collaborative project was launched with Emids (formerly Macadamian), Radiobotics, and Bispebjerg Hospital in Copenhagen. The project, known as X-AID, aimed to streamline the workflow of processing X-rays, particularly for non-acute patients. The solution involved the integration of an AI algorithm into the workflow to provide an automated assessment of musculoskeletal X-rays.

The AI technology was trained on unlabeled data from experts within the field to predict unseen instances. This allowed for the automation of text reports, making the report creation process faster and enhancing the quality by identifying things that might have been missed by human radiologists.

The project leveraged Emids’ HealthConnect platform-as-a-service, to accelerate the development and integration of this AI algorithm into clinical workflows. Once commercialized, the solution could be scaled up at other hospitals using the existing integration tools built into HealthConnect.


The implementation of the AI solution led to significant improvements in the diagnostic quality of reading X-rays. It reduced the variation between different radiologists’ interpretations, leading to a more uniform description of the diseases that the machine learning algorithms were trained to find.

This innovative solution demonstrated the power of collaboration, bringing together the clinical and domain expertise of the team at Bispebjerg Hospital, the specialists in machine learning and AI algorithm development at Radiobotics, and Emids’ domain expertise in the healthcare space.

By leveraging the strengths of these three organizations, the project was able to tackle a significant challenge in healthcare, driving a higher likelihood of success and demonstrating the potential for innovation in the field.

Sign up to receive our latest insights