Objective:

To explore the use of artificial intelligence (AI) in predicting rigid lens parameters for keratoconus patients and its implications for improving scleral lens fittings in clinical practice.

Key Findings:

• CNN demonstrated the highest accuracy in predicting lens parameters compared to the mean K method, which is a traditional fitting approach, and commercial GP guidelines.
• The CNN approach combined three topographic maps, significantly outperforming traditional fitting methods.
• AI has the potential to reduce fitting time and follow-ups for keratoconus patients, streamlining the fitting process.

Interpretation:

The study indicates that AI, particularly CNNs, can significantly enhance the fitting process for keratoconus patients by providing more accurate predictions, paving the way for future advancements in scleral lens fittings.

Limitations:

• Limited sample size and scope of lens parameters considered, which may affect the generalizability of the findings.
• Comparison was made with only one GP lens design, limiting the applicability of results to other designs.

Conclusion:

AI is poised to transform the fitting process for keratoconus and scleral lenses, necessitating further research to integrate these findings into clinical practice and improve patient outcomes.