An estimated 18,000 people in the United States die of spinal cord and brain tumors each year. To help doctors differentiate between the severity of cancers in the brain, an international team of researchers led by Murat Günel, MD, chair of neurosurgery at Yale School of Medicine, and Nixdorff-German Professor of Neurosurgery, built machine learning models that use complex mathematics to learn how various types of brain tumors look in the brain.
The models are designed to “learn” from this gathered data to make predictions and help doctors diagnose the stage of brain cancers faster and more accurately. The study was published in European Radiology in February 2020.
Slow-growing to Highly Aggressive Tumors Were Studied
The team began by compiling data from a public tumor machine resonance imaging (MRI) database called The Cancer Imaging Archive. Board-certified neuro-radiologists then identified and selected 229 patients with brain tumors, from which the researchers collected clinically relevant data, such as tumor volume, location, and features. These patients had brain tumors representing a spectrum of how likely they were to become malignant—from lower-grade gliomas, which are relatively slow-growing tumors that originate from glial cells of the brain, to glioblastomas, the highly aggressive counterpart to gliomas.
Using this data, the researchers then developed nine machine learning models and evaluated each one based on its tumor classification accuracy, precision, and sensitivity. When taken together, the machine learning models could predict which tumors were lower-grade gliomas or glioblastomas with a high degree of accuracy.
Machine Learning Models Are Highly Accurate
“Our machine learning models used to differentiate the tumor types were very accurate,” said Hang Cao, a medical student from Xiangya Hospital who worked with Gunel and is the lead author of the study.
This work is fundamentally important to our understanding of brain tumors and a great example of the collaborative, multidisciplinary effort we use to advance the field and provide the best care to brain tumor patients.
Dr. Jennifer Moliterno
However, two models outperformed the rest, achieving a classification accuracy of at least 79%. In addition, the team also found significant differences in how the cancers looked, their volumes in various regions of the brain, and their locations. Their results also suggest that clinical brain tumor features collected from MRI scans are a useful tool for non-invasive glioma grading and assessment.
Standards Will Be Needed Before Implementing These Models
The timeline for using machine learning models such as these in a clinical setting is not yet known. Although it would be possible to implement now as a stand-alone evaluation, the process is not yet integrated into the clinical evaluation of the patient. A clear set of standards will need to be established by the scientific community and then embraced by the manufacturers of software and hardware used in radiology departments.
“This work is fundamentally important to our understanding of brain tumors and a great example of the collaborative, multidisciplinary effort we use to advance the field and provide the best care to brain tumor patients,” said co-author Jennifer Moliterno, MD, associate professor of neurosurgery at Yale School of Medicine and clinical program leader of the Brain Tumor Program.
Originally published March 9, 2020; updated May 10, 2022.