The Reliability and Reproducibility of Corneal Confocal Microscopy in Children

Area of Science:

• Ophthalmology
• Neurology
• Pediatrics

Background:

• Diabetic neuropathy can affect children, necessitating early detection methods.
• In vivo corneal confocal microscopy (IVCCM) offers a noninvasive approach to assess corneal nerves.

Purpose of the Study:

• To evaluate the interrater agreement and 1-month repeatability of corneal nerve fiber length (CNFL) measurements in children using IVCCM.
• To compare manual and automated image analysis for CNFL quantification in pediatric subjects.

Main Methods:

• Seventy-one children (8-18 years), including 44 with type 1 diabetes, underwent IVCCM.
• Manual analysis of 547 corneal images by two masked observers; 1-month repeat assessment in 21 patients.
• Automated image analysis using specialized software (ACCMetrics).

Main Results:

• Excellent image-level (ICC 0.94) and good patient-level (ICC 0.86) reproducibility for CNFL.
• High agreement between manual and automated analysis (ICC 0.89), with manual measurements yielding higher values.
• Minimal bias observed between observers in Bland-Altman analysis.

Conclusions:

• IVCCM demonstrates good reproducibility for CNFL assessment in pediatric subjects.
• Image-level reproducibility surpasses patient-level, suggesting image selection refinement is key.
• IVCCM is a promising tool for early, noninvasive detection of neuropathy in children, particularly those with diabetes.