Posters

New method for assessing the accuracy of formulas for intraocular lens power calculation according to the normal eye biometric parameters

Poster Details

First Author: M.Rodriguez-Vallejo SPAIN

Co Author(s):    N. Garzon   J. Martinez   Y. Zhou   F. Poyales   J. Fernandez        

Abstract Details

Purpose:

To introduce a new graphical method for assessing the accuracy of intraocular lens power calculation depending on the deviation from the normality for several eye biometric parameters obtained from a population of subjects candidates to cataract surgery, including: Axial Length (AXL), Anterior Chamber Depth (ACD), Lens Thickness (LT), White-to-White (WTW) and mean Anterior Corneal Radius (Rm)

Setting:

IOA Madrid, Innova Ocular; Qvision, Vithas Virgen del Mar Hospital

Methods:

A population of 3519 subjects, older than 50, were measured with the IOL Master 700. Percentiles at 5%, 25%, 75% and 95% for each variable were calculated categorizing the eyes in 5 groups that contained the 50% central eyes, and 20% and 5% of eyes at each side according to the cut-off points from the percentiles. These previously defined ranges were used to classify 142 operated eyes for which the Barrett Universal II was used to calculate the IOL power. The performance of the formula (% outside 0.50 D and 1.00 D) was evaluated for each parameter.

Results:

Mean ± SD of the population was 23.78 ± 1.58 mm for AXL, 3.12 ± 0.41 mm for ACD, 4.63 ± 0.40 mm for LT, 11.92 ± 0.41 mm for WTW and 7.70 ± 0.27 mm for Rm. All variables were approximately normal distributed except AXL with asymmetry of 1.68 and kurtosis of 5.95. Rm was close to normal but with kurtosis of 3.36. The general predictability from the operated eyes was 2% outside +/- 1.00 D and 16% outside +/- 0.50 D. The accuracy decreased more markedly for eyes with ACD and LT at percentile from 0 to 25.

Conclusions:

We developed a new graphical method that allows to evaluate the accuracy of a formula for intraocular lens power calculation depending on the eye characteristics. The eyes were classified according to their location in the normal population distribution for each parameter. This method allows to evaluate not only the accuracy depending on axial length but in all the parameters that are usually used by the formulas allowing to detect the weakness and strengths of a particular formula in comparison to others.

Financial Disclosure:

None