Head MRI scans can contain information beyond the brain. A collaboration involving Hirosaki, Tohoku and Kyushu universities has developed AI that measures the masseter, a jaw muscle used in chewing, and explored its relationship with oral health. Hirosaki University highlighted the work on September 24.[1]
A regional cohort provides the evidence
The researchers trained the model using Tohoku University Hospital images with repeated expert input, then applied it to the Hirosaki cohort of the Japan Prospective Studies Collaboration for Aging and Dementia. The analysis included 2,077 people, averaging 69.9 years old.[2]
Larger muscle volume was associated with fewer oral-health problems. The strongest relationship concerned structural problems, followed by periodontal inflammation and declining oral function.[2]
Volume is not chewing strength
The software identifies muscle regions in three-dimensional T1-weighted head images. Its documentation requires scans that extend far enough down to include the masseter muscles. It can process multiple scans and produce left- and right-muscle measurements.[3]
A volume measurement describes size. It does not directly measure bite force or, by itself, establish a diagnosis.
The limits matter
The study used observations from one point in time. It cannot establish which came first: differences in muscle volume or oral health. Nor does it demonstrate that exercising the jaw prevents dental problems. The university identifies follow-up research and external validation as next steps.[1]
An association with cognitive frailty did not remain statistically significant after full adjustment. That finding does not establish a dementia test or a prevention strategy.[2]
A tool for reusing existing images
The publicly available software offers researchers a way to extract additional measurements from suitable MRI datasets.[3] Its promise is a broader use of existing research images. Whether those measurements can reliably guide decisions for individual patients requires further evidence.
