A research team from Germany has used artificial intelligence to improve the assessment of treatment response in brain tumours
Gliomas are the most common and most malignant brain tumours in adults. The current standard of treatment consists of surgery to remove as much of the tumour as possible, followed with chemotherapy and radiotherapy. The response to these treatments and tumour regrowth is assessed by Magnetic Research Imaging (MRI) scans.
However, measuring changes happening to the tumour manually, using MRI scans, is prone to error and may lead to slightly different results.
This can have a negative effect on the assessment of the tumour’s response to therapy and consequently deciding the course of future treatments.
Researchers from the University Hospital of Heidelberg the German Cancer Research Centre, set out to change this by using the huge potential of artificial intelligence.
In a study published in Lancet Oncology, the research team developed a framework for different algorithms to work together and process complex data. The data in question was MRI scans of people affected by gliomas.
The team used a reference database of approximately 500 MRI scans to ‘train’ the programme to automatically recognise and localise the brain tumours.
In addition, the programme was also trained to recognise and measure individual areas of the tumour. The combined information then allowed researchers to accurately assess the tumour’s response to therapy.
The researchers then tested this technology using a different set of MRI scans. They found that the technology improved the assessment of therapy response by 36%.
Furthermore, the researchers designed and evaluated a software that enables them to integrate this programme into existing radiological infrastructure.
This would help to immensly improve the assessment of therapy response in real time in clinical trials.
The impact of artificial intelligence in the field of brain tumour research is critical and as a charity, we recognise this and have invested in quality research that is harnessing the power of artificial intelligence to help improve the diagnosis of different types of tumours.
You can read more about this research here.