Combining Spectral-Domain OCT and Air-Puff Tonometry Analysis to Diagnose Keratoconus.

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Keratoconus is a progressive corneal ectasia.
• Early diagnosis is crucial for effective management.
• Current diagnostic methods may struggle with early or forme fruste stages.

Purpose of the Study:

• To evaluate the diagnostic capability of AI using spectral-domain optical coherence tomography (SD-OCT) and air-puff tonometry.
• To differentiate between normal and various stages of keratoconus (forme fruste, early, and advanced).

Main Methods:

• Patients with normal eyes, laser vision correction, or keratoconus (FFKC, EKC, AKC) were included.
• AI models (random forest, neural networks) were trained using SD-OCT and air-puff tonometry data.
• Model performance was assessed using receiver operating characteristic (ROC) curves, calculating area under the curve (AUC), sensitivity, and specificity.

Main Results:

• AI models demonstrated high diagnostic accuracy for advanced (AKC) and early (EKC) keratoconus.
• The combined SD-OCT and air-puff tonometry AI model achieved 99% AUC for differentiating forme fruste keratoconus (FFKC) from normal eyes.
• Specificities reached 94.74% for FFKC detection.

Conclusions:

• AI parameters accurately diagnose advanced and early keratoconus.
• AI-assisted diagnosis using SD-OCT and air-puff tonometry shows promise for improving FFKC detection.
• This technology could enhance patient treatment outcomes for keratoconus.