DR. RESHMA RANADE

Dr. ROHIT SHETTY, Dr. POOJA KHAMAR, Dr. Gairik Kundu

Abstract

Purpose: To study performance of AI based smartphone application in identifying keratoconus progressors&suggest appropriate management.

Methods: Good quality 2500 scans of 200 eyes were exported from PentacamHR&were classified into: Stable&Progressing.Keratometry parameters,KC indices&Zernike wavefront aberrations were given to AI. Machine learning was used to teach AI algorithm for management options like Crosslinking, ICRS, TCAT, DALK, Penetrating keratoplasty&Femtosecond laser assisted Lamellar Keratoplasty.Progression&management derived data was used to build smartphone application.

Results: Random forest classifier-based AI model predicted disease progression with area under curve at 0.92, sensitivity&specificity as 0.8&0.87 respectively. Sensitivity &specificity of KC app in identifying progressors & treatment was 96.40%.

Conclusion:KC app works excellently to sort, organise&display relevant information in accessible manner & serves as one stop solution to practitioner.

Full Text