Source/Disclosure
Issuer:
Koch DD et al. Artificial intelligence formula for intraocular lens power calculation for short eyes. Presented at: American Society of Cataract and Refractive Surgery. May 5-8, 2023. San Diego.
Disclosure:
Koch reports on consulting Carl Zeiss Meditec.
Important points:
- The Zeiss AI IOL Calculator had a smaller mean absolute error than the other IOL formulas.
- This model is based on 500,000 simulated eyes and real data.
SAN DIEGO — Research shows that artificial intelligence-based IOL calculators can help optimize models for different types of eye parameters, including short eyes.
When performing IOL calculations, it is important to optimize the IOL constant if the study cohort is representative of the entire population. Douglas D. Koch, MD, Speaking at the American Society for Cataract and Refractive Surgery’s DOS Digital Day, he said:
“However, optimizing for a constant can be misleading or unhelpful if the cohort is not representative of the population as a whole, for example with short eyes,” he said. The optimized constants may not be useful to other surgeons or colleagues.”
Koch presents data on the Zeiss Artificial Intelligence IOL Calculator, the first AI-based, IOL-specific formula.
According to Koch, in addition to biometric measurements and target refraction, the Zeiss AI IOL calculator uses information optimized for each unique IOL model rather than using IOL constants.
According to him, the model starts with paraxial ray tracing and a database of 500,000 simulated model eyes that are used to predict IOL power. From there, pre-operative biometry and post-operative anterior chamber depth are used to predict his post-operative IOL position. It also includes curated data from his Veracity surgery planning data collection at Zeiss.
Koch said that in a study consisting of 155 single and 123 paired eyes, the Zeiss AI IOL calculator had lower mean absolute errors than other calculators such as Kane, Barrett and Pearl-DGS. (P. .05). It also reduced the root-mean-square absolute error.
“The future of this is that the eye can build this database from a wide range of parameters, such as anterior chamber depth, lens thickness, and corneal diameter, and optimize it for each model,” Koch said. said.
