Imaging: using AI to warn of osteoporosis

The bone substance of many people deteriorates with age. This process, known as osteoporosis, often goes unnoticed, even if there are fractures in the vertebrae. The vertebral fracture could be detected by X-ray examination or computer tomography (CT), but this does not always happen. For example, because the CT scan is performed for other reasons and a vertebral fracture is overlooked in the stress of everyday life in the clinic.

In order to improve osteoporosis diagnostics, researchers led by Prof Claus-Christian Glüer, Department of Radiology and Neuroradiology in Kiel at the University Medical Centre Schleswig-Holstein (UKSH), have now developed software. The programme uses methods of artificial intelligence (AI)  and can therefore automatically detect indications of osteoporosis and prognostically unfavourable vertebral fractures on computer tomographies taken for a variety of reasons.

KI reliably detects osteoporosis in CT images

CT images of the chest are often taken to look at the lungs, for example. Although the spine can be seen on the image, it is not examined because there may be another problem in the foreground. „Our programme can run in the background during such examinations. It automatically looks at the spine and gives an indication of fractures of the vertebral bodies that might otherwise not have been detected," explains Eren Yilmaz, computer scientist and first author of the study. This is important because the presence of vertebral fractures significantly increases the risk of further fractures.

The software works with so-called neural networks. These are algorithms that are modelled on the way the human brain works and are often used to recognise patterns.

The software works with so-called neural networks.

The researchers tested the AI on 159 CT images of the spine from seven hospitals in Germany. Experienced radiologists examined the images beforehand and detected 170 fractures. The neural network correctly classified 90 per cent of the cases with fractures and 87 per cent of the vertebrae without fractures,

KI recognises mild and severe grade

In addition, the programme can not only detect fractures, but also differentiate between mild fractures (grade 1) and more severe ones (grade 2 or higher). „This diagnosis is crucial for assessing the future fracture risk“, says Yilmaz. This also applies in particular to hip fractures, which are associated with a high reduction in quality of life and increased mortality, especially in old age.

„We are therefore developing an early warning system to prevent the serious consequences of osteoporosis“, says the computer scientist. The technology is not yet available for general use in hospitals. However, it should at least be possible to use it for research purposes in the foreseeable future.

Article from "medizin & technik" from 10 May 2023