[Advance of screening keratoconus before refractive surgery with machine learning]

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Imaging

Background:

• Traditional keratoconus screening methods lack accuracy, particularly for subclinical cases.
• Subclinical keratoconus poses a risk for postoperative corneal ectasia after refractive surgery.
• The expanding application of AI in ophthalmology offers potential for improved diagnostic capabilities.

Purpose of the Study:

• To review artificial intelligence algorithms for keratoconus screening in refractive surgery.
• To discuss corneal feature extraction techniques for AI-based diagnostics.
• To evaluate the predictive accuracy of machine learning models in keratoconus detection.

Main Methods:

• Review of current literature on AI and machine learning applications in ophthalmology.
• Analysis of algorithms used for keratoconus screening.
• Examination of feature extraction methods from corneal imaging data.

Main Results:

• Machine learning shows promise in improving the accuracy of keratoconus screening.
• AI algorithms can enhance the detection of subclinical keratoconus.
• Various algorithms and feature extraction techniques are being explored for diagnostic accuracy.

Conclusions:

• AI, particularly machine learning, is a developing area for accurate keratoconus screening.
• Improved screening can better determine refractive surgery indications and mitigate risks.
• Further research into AI models is essential for clinical application in ophthalmology.