Main menu


Scientists develop a new way to use magnetic resonance to detect brain tumors

University of Washington researchers have developed a new model and method for identifying and classifying brain tumors as one of six common types of tumors, using a single 3D MRI scan, a new study published in Radiology: Artificial Intelligence shows.

"This is the first study to treat the most common intracranial tumors and determine the tumor class or absence of a tumor from a 3D MRI volume," said University of Washington Ph.D. Satrajit Chakrabarti.

The six most common types of intracranial tumors are high-grade glioma, low-grade glioma, brain metastases, meningioma, pituitary adenoma and acoustic neuroma.

Each of these types has been documented through histopathology, which requires tissue to be surgically removed from a suspected cancer site and examined under a microscope. According to the researchers, machine and deep learning approaches using MRI data could make it easier to detect and classify brain tumors.

To build their machine learning model, called a convolutional neural network, the researchers developed a large, multi-institutional data set for 3D intracranial MRI scans from four publicly available sources.

"These results indicate that deep learning is a promising approach to mechanistically classify and evaluate brain tumors," Chakrabarti said. "The model achieved high accuracy on a heterogeneous data set and demonstrated excellent generalization capabilities on unseen test data."

Dr Soteras, one of the model developers, added that these models can be extended to other types of brain tumors or neurological disorders, which may provide a path to augment much of the neuroradiology workflow, as the new technology is the first step towards developing an AI-enhanced radiology workflow that It can support the interpretation of images by providing quantitative information and statistics.