Enhancing corneal ectasia susceptibility detection: analysis of a new algorithm (BAD-D v4)

Area of Science:

• Ophthalmology
• Corneal Imaging
• Refractive Surgery

Background:

• Accurate detection of post-refractive ectasia susceptibility is crucial for preoperative evaluation in laser vision correction (LVC).
• Iatrogenic ectasia, though reduced, can cause severe vision loss, necessitating improved diagnostic tools.
• Current screening methods require enhancement for reliable identification of ectasia risks.

Purpose of the Study:

• To develop and validate an optimized version of the Belin/Ambrósio Enhanced Ectasia Display (BAD-D v4).
• To enhance the detection of keratoconus and very asymmetric ectasia (VAE).
• To assess the risk of post-refractive ectasia more accurately.

Main Methods:

• A new optimized logistic regression algorithm was utilized.
• The study analyzed a dataset of 3,886 eyes from 3,351 patients, including normal, keratoconus (KC), and VAE categories.
• Validation was performed across 26 international centers.

Main Results:

• BAD-D v4 demonstrated superior efficacy in differentiating normal eyes from ectatic conditions (AUROC 0.997-0.998).
• It showed high accuracy in Normal vs. Disease (KC + VAE) classification (AUROC 0.974-0.966).
• Performance in the challenging Normal vs. VAE-NT group was also significant (AUROC 0.905-0.858), outperforming BAD-D v3.

Conclusions:

• BAD-D v4 offers enhanced accuracy for early ectasia detection.
• The tool effectively identifies risks for post-refractive ectasia.
• BAD-D v4 maintains the existing index scale and user experience while improving diagnostic capabilities.