Predicting the postoperative effective lens position with anterior chamber depth, lens thickness, and angle-to-angle depth using a new anterior-segment optical coherence tomography

Session Details

Session Title: IOL Power Calculations

Session Date/Time: Sunday 11/09/2016 | 16:00-18:00

Paper Time: 16:30

Venue: Hall C2

First Author: : I.Iehisa JAPAN

Co Author(s): :    S. Goto   S. Koh   K. Ohnuma   N. Maeda   T. Noda        

Abstract Details

Purpose:

To evaluate the accuracy of a new algorithm for predicting the postoperative effective lens position (ELP) using the angle-to-angle (ATA) depth, preoperative anterior chamber depth (ACDpre), and lens thickness (LT) measured by Fourier-domain anterior-segment optical coherence tomography (AS-OCT) and compared the results with the SRK/T, Haigis, and OKULIX formulas.

Setting:

National Hospital Organization, Tokyo Medical Center, Tokyo, Japan.

Methods:

Sixty eyes randomized to a training set (TS, 30 eyes) and a validation set (VS, 30 eyes) underwent cataract surgery. Using the TS, a new regression algorithm for predicting the ELP was developed by multiple linear regression using two parameters. The combination of the ATA depth and ACDpre + LT/2 measured by AS-OCT (CASIA2, SS-2000, Tomey) showed the highest coefficient of determination (R2) values. In the VS, the median values of the absolute predicted ELP error and the R2 between the predicted and measured ELPs were compared with the new algorithm and the SRK/T, Haigis, and OKULIX formulas.

Results:

The results of multiple linear regression analysis in the TS using the new algorithm that combined the ATA depth and ACDpre + LT/2 showed the highest R2 value, 0.86, for predicting the ELP. In the VS, the median value of the absolute predicted ELP error of the new algorithm was significantly (p<0.01) smaller than the established formulas (new algorithm, 0.099 mm; SRK/T, 1.025 mm; Haigis, 0.336 mm; OKULIX, 0.193 mm). The new algorithm yielded the highest R2 among the four formulas (new algorithm, 0.830; SRK/T, 0.379; Haigis, 0.510; OKULIX, 0.285).

Conclusions:

The ATA depth and ACDpre + LT/2 measured using the new AS-OCT are reliable preoperative parameters for predicting the ELP. The new algorithm with two preoperative parameters (ATA depth and ACDpre + LT/2) predicted the ELP more accurately than the SRK/T, Haigis, and OKULIX formulas. This new algorithm might improve the accuracy of IOL power calculations.

Financial Disclosure:

NONE