We all know that many types of accidents can lead to head injuries. The brain can thus be damaged to different extent, and this damage can eventually lead to physical, cognitive, and emotional symptoms. Traumatic axonal injury (TAI) is such an important type of brain injury, and MRI is the best way to show such damage.
In the TAI-MRI project, we have therefore aimed to develop a classification system that can describe the severity of TAI and predict the outcome of the brain injury for the patient. We have also aimed to develop automated methods, using artificial intelligence, to detect and quantify TAI in MRI images.
Researchers and clinicians from four different countries (Norway, the UK, the Netherlands, and Belgium) have collaborated to develop a reliable classification system. Data from several studies (patients from approximately 12 hospitals in Europe) with a total of 1250 patients with traumatic brain injury have been used in the project. Extensive descriptions and post-processing of MRI images have been performed, and data have been analyzed using advanced statistical methods. The preliminary results from the first 500 Norwegian patients show, among other things, that bilateral damage in the brainstem and thalamus seems to indicate particularly great severity in the acute phase. The use of advanced MRI shows some correlation with measured changes in white matter (area with axons) and later symptoms in mild traumatic brain injury. These results will therefore have an impact on an improved classification system, which will be further tested in the large EU study CENTER-TBI and other studies. However, it has proved challenging to obtain sufficiently high accuracy for the automatic methods that have been developed in the project. Although the methods have taken major steps forward, there seems to be some way to go until they can be used in both research and clinical work with patients with traumatic brain injury.
Data analyses are still ongoing, however, results so far indicate that the project has contributed (and will further contribute) with new knowledge about the clinical significance of MRI biomarkers. In the long term, this can have a major impact on the management of traumatic brain injury.
Traumatic injuries to the head can cause differing degrees of damage to the brain, ranging from none to mild, moderate or severe traumatic brain injury (TBI). The management of patients with TBI depends on the severity of the injury and the primary imaging modality is still CT in the acute phase. But today we know that CT may only show the "tip of the iceberg" of the actual injuries to the brain, and in some instances the scan miss injuries altogether. In particular, traumatic axonal injury (TAI) is difficult to detect by CT. In the last decade, different MRI techniques therefore have been increasingly used. MRI can detect visible TAI lesions, but also other more subtle brain injuries, with a much higher sensitivity than that observed for CT.
In TAI-MRI, we aim to develop a classification system that can better describe severity of TAI and predict the outcome of injuries. The current classification system is based on neuropathological studies from 1980s and has been extrapolated to classify also injuries on MRI in surviving patients. This classification system has shown limitation in reflecting the actual burden of axonal injuries. A classification system that better reflects the distribution of axonal injuries and that also takes into account the prognostic significance of the different TAI lesions would help both doctors and health care professionals as well as patients and their families to understand the effects of brain injury and also what prognosis can be expected during the first year. To develop such an important and reliable classification system, researchers from four different countries will collaborate and data from three different studies including almost 1400 patients will be included and analyzed. MRI methods have been continuously developed during the last couple of decades, and new promising technological advances enable us to analyze MRI data in a more automated way. The results of this project should improve the care of patients with TBI.