To evaluate the effectiveness of polarization-sensitive optical coherence tomography (PS-OCT) combined with artificial intelligence (AI) in diagnosing subclinical keratoconus and aiding refractive surgery decisions.
Approach:
Key Findings:
Overall accuracy for PS-OCT was 82%, while Pentacam and MS-39 achieved 86%.
PS-OCT classified 43 of 109 subclinical keratoconus eyes as healthy, compared to 30 for Pentacam and 33 for MS-39.
In highly asymmetric keratoconus cases, PS-OCT classified 21 of 33 fellow eyes as healthy, versus 12 for Pentacam and 16 for MS-39.
PS-OCT identified subtle keratoconus cases that were classified as subclinical or uncertain by tomography-based models.
Interpretation:
Limitations:
The study is retrospective and requires prospective longitudinal studies to validate the predictive value of PS-OCT reclassifications.