|
Researchers at the
Shiley Eye Center at UCSD
Advance Glaucoma Diagnosis
Using Machine Learning
Everyday, each of us individually
learns from our own experience. Physicians, researchers and
computers do as well. Michael Goldbaum, MD, and his research
team are talking this learning experience to entirely new levels.
Do you or does someone you
know, have glaucoma? Did you know that glaucoma is hereditary?
Researchers
at the Shiley Eye Center are analyzing new technology that
can diagnose glaucoma earlier and also detect more quickly
whether or not it is progressing. This new technology, Machine
Learning Classifiers, is a set of new sophisticated mathematical
processors that can equal or surpass the diagnostic work
of glaucoma experts or the current methods of processing
data. The result could be the diagnosis of glaucoma, the
identification of glaucoma onset, or an indication that
glaucoma is progressing. According to Michael H.
Goldbaum M.D., Professor of Ophthalmology at UCSD,
"These new sophisticated mathematical techniques permit
us to improve patient management.”
With more than 3 million Americans diagnosed
and another 10 million at risk, glaucoma is the leading
cause of treatable blindness within the United States. Early
in the course of the disease, you may not be aware that
you are losing your vision. By the time you notice a change
in your eyesight, the glaucoma may be advanced and difficult
to treat effectively. Early diagnosis and treatment of the
disease enhances the likelihood of maintaining lifelong
vision.
|
Using computational power and math modeling,
Machine Learning Classifiers (MLCs) help the physicians
at Shiley Eye Center diagnose patients, predict outcomes
and see progression of disease. MLCs can do this in a short
period of time, which hastens the prcess of finding information
that is useful for treating patients. MLCs take in data,
which could be visual fields or images of the optic disk
and the surrounding retina, process it and provide a result.
Using sophisticated mathematics, MLCs learn to recognize
patterns and separate data into classes, for instance eyes
with glaucoma versus healthy eyes. The physicians and clinicians
at Shiley hope to use MLCs to determine what information
is important for:
• detecting glaucoma;
• predicting which eyes, suspected of having
glaucoma, will develop it
• predicting in which eyes glaucoma-related visual
defects will worsen.
According to Robert N. Weinreb
M.D., Director of the Hamilton Glaucoma Center
at UCSD, “These new technologies can assist us to
more sensitively determine which patients to treat and which
patients to monitor. They also enable us to ascertain the
type and level of recommended medical or surgical treatment.”
|
 
|
