Українська правда

Artificial intelligence has been trained to detect tiny brain tumors that previously went unnoticed by MRI scans

- 3 July, 08:59 AM

As reported by BGR, researchers from the Netherlands Cancer Institute, together with Robovision Healthcare, have presented a new development — BrainMets.ai, a system that uses artificial intelligence (AI) to detect metastases in the brain. According to the results of a study published in the journal Radiology, the technology showed a high level of accuracy, especially in cases of microtumors, which are often overlooked during standard MRI.

Brain metastases are the most common type of central nervous system tumor, occurring in approximately 17% of adult cancer patients. The problem is that these tumors are mostly less than 3 mm in diameter, making them difficult to identify even with multiple MRI scans.

BrainMets.ai achieved a sensitivity of 97.4% for overall lesion detection. For tumors larger than 12 mm, accuracy was 100%, for sizes 6–12 mm — 98%, 3–6 mm — 97.9%, and for tumors smaller than 3 mm — 93%.

"This tool is a true leap forward. It increases our confidence in detecting even the smallest lesions, supporting more timely and accurate treatment decisions," said Dr. Laurens Topff, a neuroradiologist at the Netherlands Cancer Institute and clinical leader of the project.

It is noted that this technology is not intended to replace doctors, but is intended to speed up the diagnostic process and reduce the risk of missed cases. According to the developers, an important factor in accuracy was the quality of the training data on which the AI is trained. Thanks to carefully selected data, accurate annotations and an optimized neural network architecture, it was possible to obtain such highly effective results.

Load more