In a recent study published in the Journal of the American Society for Mass Spectrometry, researchers from the University of Florida (UF) Health revealed an innovative approach that combines machine learning (ML) with liquid chromatography-high resolution mass spectrometry (LC-HRMS) to enhance the characterization of brain tumors, specifically meningioma tumors.
Meningioma tumors are classified into three grades: grade I, grade II, and grade III, with grade II tumors posing a particular challenge to clinicians. As Dr. Jesse L. Kresak, a clinical associate professor at UF College of Medicine, explains, “Do we take [the tumor] out and watch to see if it comes back? Or do we also irradiate the area with the idea of preventing a recurrence?”
To tackle this dilemma, the research team analyzed 85 meningioma samples, generating detailed chemical profiles of the small molecules and fats present in each tumor. This characterization enabled the researchers to identify unique biomarkers and more accurately differentiate between tumor grades. Initially, ML was not part of the plan. However, the team recognized the potential for ML to uncover hidden patterns and insights beyond their manual analysis capabilities.
By incorporating ML into the evaluation process, the researchers significantly improved efficiency without compromising accuracy. While a clinician typically assesses around 20 data points in ten minutes, ML allowed the analysis of 17,000 data points in less than a second. Moreover, one ML model achieved 87 percent initial accuracy in classifying tumor grades, with the potential for further improvement through the examination of additional samples.
What is meningioma?
Meningioma is a type of brain tumor that arises from the meninges, the protective membranes surrounding the brain and spinal cord.
What are the different grades of meningioma tumors?
Meningioma tumors are classified into three grades: grade I, grade II, and grade III. Grade I tumors are slow-growing and less threatening, while grade III tumors are more aggressive and require a combination of tumor removal and radiation treatment.
How does machine learning improve brain tumor evaluations?
Machine learning enhances brain tumor evaluations by analyzing large amounts of data and identifying patterns and biomarkers that may be difficult to detect manually. This technology allows for more efficient and accurate characterization of tumors, aiding in treatment decisions.
What other advancements are being made in oncology using data analytics and artificial intelligence?
Health systems, including the University of Texas MD Anderson Cancer Center, are exploring the potential of data analytics and artificial intelligence to revolutionize oncology. These technologies can support personalized care, enhance data management, and improve patient outcomes.