To get the highest possible performance in image recognition tasks, deep learning often requires tens of thousands of labeled images. This is a stumbling block in the development of deep learning systems for complex image datasets, especially MRI, which is critical for detecting neurological abnormalities.
By deriving key labels from radiology reports and accurately assigning them to the corresponding MRI examinations, researchers from King's College London's School of Biomedical Engineering & Imaging Sciences have automated brain MRI image labeling, which is required to teach machine learning image recognition models. In less than half an hour, more than 100,000 MRI examinations may now be labeled. This is the first study to allow researchers to label intricate MRI image collections at scale, and it was published in European Radiology. Manually labeling more than 100,000 MRI examinations, according to the researchers, would take years.
"By overcoming this bottleneck, we have massively facilitated future deep learning image recognition tasks, and this will almost certainly accelerate the arrival into the clinic of automated brain MRI readers. The potential for patient benefit through, ultimately, timely diagnosis, is enormous," said senior author, Tom Booth, Ph.D., from the School of Biomedical Engineering & Imaging Sciences at King's College London.
Model performance was evaluated on unseen images, rather than just evaluating the performance of unseen radiology reports.
The future challenge lies in completing the deep learning image recognition tasks, which present multiple technical challenges. Secondly, ensuring that the developed models continue to perform accurately across different hospitals using different scanners will be another hurdle.
"This study was possible thanks to a very broad team of experts who are working on these challenges. There is a huge base of supporting organisers and facilitators who are equally important in delivering this research. Obtaining clean data from multiple hospitals across the UK is an important step to overcome the next challenges. We are running an NIHR portfolio adopted study across the UK to prospectively collect brain MRI data for this purpose,” said Tom Booth.