My SciELO
Services on Demand
Article
Indicators
- Cited by SciELO
Related links
- Similars in SciELO
Share
Revista Cubana de Oftalmología
On-line version ISSN 1561-3070
Abstract
HERNANDEZ LOPEZ, Iván; CARDENAS DIAZ, Taimi and BATISTA LEYVA, Alfo José. Optimization of constants for accurate calculation of five intraocular lens models. Rev Cubana Oftalmol [online]. 2022, vol.35, n.2 Epub June 01, 2022. ISSN 1561-3070.
Objective:
To optimize the constants used by the SRK/T, SRK/T2, Holladay 1 and Hoffer Q formulas for five intraocular lens (IOL) models implanted during cataract surgery at the Cuban Institute of Ophthalmology (January/2006-October/2019).
Methods:
47341 patients were retrospectively studied. The constants were optimized by adjusting the average prediction error (EPm) to zero and obtaining the value that showed the maximum number of eyes with absolute prediction error |EP| ≤ 0.25, for both ultrasonic and optical biometry. The effectiveness of the optimized constants was verified by analyzing the lowest average and median absolute errors (EAM/EAMed) and the percentage of eyes in a given range of prediction error as a function of the value of the constants. The Haigis formula performance index and a newly introduced index were also analyzed.
Results:
The optimized constants were significantly different from the manufacturer's values. The values obtained for EPm = 0 were different from the EAM and EAMed minima. The percentage of eyes with |EP| ≤ 0.25 and 0.50 D differed when the optimization criteria were different. SRK/T and SRK/T2 showed the best performances, according to both indexes.
Conclusions:
Optimized constants improve postoperative outcomes for each IOL-formula combination. Different optimization criteria lead to different results. The EAM, the EAMed and the percentage of eyes with |EP| ≤ 0.50 D are valid parameters to check the performance of the formulas, until a single, reliable and consensual index is available.
Keywords : intraocular lens calculation; optimization of constants; cataract surgery.