0
0
January 25, 2025 – Brain tumor surgery, particularly for gliomas, a type of malignant brain tumor, can offer life-extending benefits. However, it often brings with it the potential for cognitive challenges, such as difficulty concentrating and performing complex tasks. A new AI model developed by Ph.D. researcher Lars Smolders aims to predict these cognitive consequences, giving patients and medical professionals a clearer picture of what to expect after surgery.
While the removal of malignant brain tumors, such as gliomas, can significantly improve life expectancy, it can also mark the beginning of a difficult health journey. Smolders, a researcher from the Department of Mathematics and Computer Science, notes that many patients face cognitive challenges following surgery, which can severely affect their quality of life. These challenges, which include trouble concentrating and difficulty with complex tasks, are not always predictable, and their true extent remains unclear.
Traditional methods have relied on the location of the tumor to predict cognitive outcomes. However, Smolders and his team discovered that tumor location alone is not a reliable predictor of cognitive problems post-surgery. In response, they developed a more sophisticated AI model that takes advantage of MRI images of a patient’s brain taken before surgery.
The Role of White Matter in the Model
The key to the model lies in the brain’s white matter – long-distance bundles of neurons that interconnect different regions of the brain. Smolders and his collaborators focused on structural details of these white-matter connections visible in pre-surgery MRI scans. The AI model uses this data to assess how vulnerable a patient’s brain is to potential damage caused by surgery, radiation, or chemotherapy.
Smolders explains, “This model helps us understand how resilient a patient’s brain is to the disruptions caused by the tumor removal process, providing us with valuable insights into potential cognitive impairments.”
Towards a Clinical Revolution
The significance of this model lies in its ability to predict the likelihood of cognitive issues post-surgery. This prediction could empower surgeons to make more informed decisions, potentially sparing patients from undergoing surgery if the risk of significant cognitive impairment is high. The model is still in its early stages and needs to be validated on a larger group of patients before it can be used clinically.
“This research highlights the importance of understanding brain structure in predicting cognitive outcomes after brain tumor surgery,” says Smolders. “It could eventually lead to personalized treatments that improve quality of life after surgery.”
Overcoming Early Challenges
The development of this model was not without its challenges. Smolders and his team initially applied traditional methods from network neuroscience to study MRI images, but they found that these techniques failed when dealing with the significant deformations caused by brain tumors. This setback led Smolders to create new approaches that more effectively accounted for the unique challenges posed by brain tumor patients.
“Many existing methods relied on assumptions about brain structure that were not applicable to patients with tumors,” Smolders reflects. “We had to rethink everything to create a model that works for this patient group.”
Looking Ahead
Smolders is excited about the future possibilities for this AI model. He envisions integrating real-time brain activity data into the model to enhance its predictive power, ultimately reducing the risks of neurological impairments and improving patient outcomes.
His work continues through collaborations with the neurosurgery department at Elisabeth-Tweesteden Hospital in Tilburg. Smolders hopes to secure funding to continue this vital research as a postdoctoral researcher, with plans to build on his groundbreaking work.
As Smolders continues to explore the complexities of the human brain, he remains passionate about advancing our understanding of its structure and function, particularly in patients with brain tumors. His work, he hopes, will help pave the way for more effective treatments and better quality of life for patients worldwide.
Disclaimer: This AI model is still in development and must undergo further clinical validation. It is not yet available for general use and should not be relied upon for medical decision-making without consultation from healthcare professionals.
