investigator_user investigator user funding collaborators pending menu bell message arrow_up arrow_down filter layers globe marker add arrow close download edit facebook info linkedin minus plus save share search sort twitter remove user-plus user-minus
  • Project leads
  • Collaborators

Machine Learning Comparison of Brain Anatomy in Cancer and Alzheimers Disease

Shelli R. Kesler

0 Collaborator(s)

Funding source

National Institutes of Health (NIH)
Difficulties with cognitive skills such as memory, thinking and attention are very common after breast cancer chemotherapy. These cognitive difficulties have been associated with chemotherapy-related injury to several brain regions, particularly those regions that also tend to be affected by brain aging. There is concern that chemotherapy may accelerate brain aging and increase patients' risk for Alzheimer's disease. Certain previous studies support this concern while others do not and therefore the effect of breast cancer chemotherapy on risk for Alzheimer's disease remains unclear. The proposed research will compare brain anatomy of breast cancer survivors to that of women who later developed Alzheimer's disease. Using machine learning, a form of artificial intelligence, we will determine a pattern of brain anatomy that indicates the probability of developing Alzheimer's disease. This pattern, or classifier, will be determined using magnetic resonance imaging (MRI) scans that were obtained for approximately 100 women with Alzheimer's disease compared to MRI scans from a group of approximately 100 healthy, unaffected women. We will then apply the classifier to MRI scans from 108 breast cancer survivors, 67 who received chemotherapy treatment and 41 who did not. Using the machine learning classifier, we will calculate a score for each breast cancer survivor that indicates her probability of developing Alzheimer's disease based on her brain anatomy. We believe that probability scores will be significantly higher in the chemotherapy group compared to the no-chemotherapy group. We will also explore possible predictors of probability score such as demographic, disease, genetic and treatment factors. This project has the potential to improve our ability to identify patients at risk for persistent and/or progressive chemotherapy-related brain injury using a simple, non-invasive five minute MRI scan. This information could potentially inform treatment decision-making and prioritize patients for early intervention.

Related projects