Specialists from the Netherlands Cancer Institute together with the Robovision Healthcare company have developed the BrainMets.ai system, which uses artificial intelligence to detect metastases in the brain. This technology has shown high accuracy, especially in cases of microtumors that are often missed during traditional MRI examination, informs BGR.
Metastatic brain lesions are the most common tumors of the central nervous system, occurring in approximately 17% of adult cancer patients. The main difficulty is that most of these tumors have a diameter of less than 3 mm, which makes their detection much more difficult even with numerous MRI images.
The BrainMets.ai system showed a sensitivity of 97.4% in the overall detection of lesions. For tumors larger than 12 mm, the accuracy was 100%, for sizes 6–12 mm — 98%, 3–6 mm — 97.9%, and less than 3 mm — 93%.
“This technology is an important step forward. It increases confidence in detecting even the smallest lesions, helping doctors make more timely and accurate treatment decisions.” said Dr. Laurens Topff, a neuroradiologist from the Netherlands Cancer Institute and the clinical leader of the project.
The developers emphasize that the goal of this system is not to replace doctors, but to optimize the diagnostic process and reduce the likelihood of missed pathologies. They also emphasized that one of the key performance factors was the high quality of the training data: thanks to carefully selected examples, accurate annotations and advanced neural network architecture, it was possible to achieve impressive results.
