2024 Alzheimer's Association Research Grant (AARG)

Increasing Neuroimaging Access of Dementia Through Low-Field Portable MRI

Can a flexible, low-cost form of MRI make access to dementia diagnosis more available to underserved communities?

William Taylor Kimberly, M.D., Ph.D.
Massachusetts General Hospital
Boston, MA - United States

Background

Brain scanning (or neuroimaging) is an important technique that physicians use to assess the structural brain health of people at risk of dementia. Neuroimaging includes several different technologies, including magnetic resonance imaging (MRI). The most widely used MRI method today, known as 3T-MRI, takes high-resolution images that can be analyzed quickly by physicians. However, 3T-MRI machines are heavy and expensive, and they must be used in specially designed spaces within clinics. The stationary nature of these machines presents a limitation for evaluating people at risk of dementia, as these individuals usually need to undergo multiple scans at different times and places. Moreover, the high cost of 3T-MRI limits its use for underserved populations.

Dr. William Taylor Kimberly and colleagues have been studying an alternate MRI technology called low-field portable MRI (LF-MRI). This technique uses machines that are much lighter and less costly than 3T-MRI, and they can be placed in a variety of locations, including outpatient clinics, enabling people at risk of dementia to more easily undertake multiple images over time. Traditionally, however, LF-MRI machines have been discouraged because the images they produce have been less clear – unable to visualize subtle brain changes linked to dementia. To overcome this problem. Dr. Kimberly’s team has developed a type of machine learning (or advanced computer science) technique that can process “low-resolution” LF-MRI images to make them more detailed and more able to detect dementia-related brain changes. The researchers call this machine learning technique “LF-SynthSR.”

Research Plan

Dr. Kimberly and the team will now test and refine their technique. First, male and female participants of diverse backgrounds (White, Asian, Hispanic, and Black American) will undertake LF-MRI scans and standard 3T-MRI scans. Studies have found that brain scan technology does not always detect Alzheimer’s accurately in underrepresented groups. Thus, the researchers will assess how well their procedure, when compared with 3T-MRI, can image changes in brain structure across different sexes and ethnicities. 

Next, they will enroll a second group of diverse participants with Alzheimer’s, mild cognitive impairment (MCI, a state of subtle memory loss that may precede Alzheimer’s), or no cognitive impairment. These individuals will receive LF-MRI as well as blood exams and cognitive tests. The researchers will then test whether analyzing dementia-related proteins in the blood combined with MRI-visualized brain changes can accurately distinguish people with MCI or Alzheimer’s from those without cognitive impairment.
 
Lastly, Dr. Kimberly’s group will conduct another brain scan experiment with participants who have MCI or Alzheimer’s. These participants will be receiving drugs or other therapies designed to reduce their brain levels of beta-amyloid (a protein fragment that accumulates into hallmark plaques in the Alzheimer’s brain). It will be used LF-MRI to monitor the individuals for amyloid-related imaging abnormalities (ARIA), which are side effects (such as brain blood vessel damage) that can occur due to anti-amyloid therapy. The team will compare how well LF-MRI can monitor for ARIA, compared with 3T-MRI.  

Impact

Dr. Kimberly’s project could promote a flexible, cost-efficient tool for diagnosing Alzheimer’s and cognitive decline in diverse populations. Such work could help reduce the disparity in Alzheimer’s diagnosis and treatment in underrepresented groups.