The illustration depicts a brain with a glioma and its digital twin. Image credit: Baharan Meghdadi, Systems Biology of Human Disease Group, University of Michigan
A new machine learning-based approach to mapping real-time tumor metabolism in brain cancer patients could help doctors discover which treatment strategies are most likely to be effective against individual cases of glioma.
To overcome challenges in mapping tumor metabolism inside the brain, University of Michigan researchers developed a computer-based “digital twin” that can predict how an individual patient's brain tumor will react to each treatment.
The researchers built a type of deep learning model called a convolutional neural network and trained it on synthetic patient data, generated based on known biology and chemistry and constrained by measurements from eight patients with glioma who were infused with labeled glucose during surgery. By comparing their computer models with different data from six of those patients, they found the digital twins could predict metabolic activity with high accuracy.
In experiments conducted on mice, the team confirmed that the diet only slowed tumor growth in mice that the digital twin had identified as good candidates for the treatment.
The digital twin also predicted how tumors would respond to the drug mycophenolate mofetil, which targets how cancer cells build DNA. It correctly identified that some tumors could bypass the drug's effects by using a “salvage pathway” to grab nutrients from their surroundings. Again, the team confirmed the predictions with mouse experiments.
Thus, a doctor could use a patient’s digital twin to test whether a specific diet or drug would starve the cancer before the patient changes their meal plan or starts a new medication.
“This work moves us closer to truly personalized cancer care—not just for brain cancer, but eventually for a variety of tumors. By simulating different therapies virtually, we hope to spare patients from unnecessary treatments and focus on those likely to help,” said co-corresponding author of the study Costas Lyssiotis, the professor of oncology at the University of Michigan.
The team has applied for patent protection with the assistance of U-M Innovation Partnerships and is seeking partners to bring the technology to market.
Data from University of Michigan