Abstract
Purpose: :
The purpose of the software algorithms being developed is to test feasibility of measuring peripheral neurons in the cornea. The ultimate aim is a clinical tool for following progression of neuropathy and to thereby provide a test that corroborates, confirms or helps determine the cause of neuropathic symptoms.
Methods: :
Algorithms were designed and prototyped in software programs. The algorithms automatically find and mark peripheral nerve fibers that are seen in scanning laser ophthalmoscope images of the cornea. They also calculate morphometry and intensity measures of the nerve fibers. A database of subjects that were categorized with levels of neuropathy through electromyography was tested to determine trends in morphometry. The categories were: healthy (no-diabetes), diabetic but no neuropathy, mild, moderate and severe.
Results: :
Statistical hypothesis testing, using standard linear regression analysis, showed a statistically significant trend in several morphometry and intensity measures.
Conclusions: :
The trends found in the statisical significance testing provide a promise of finding objective and quantitative means for following progression of peripheral neuropathy.
Keywords: imaging/image analysis: clinical • diabetes • cornea: stroma and keratocytes