1.5 year experience with swept source OCT biometry

Session Details

Session Title: Cataract Surgery Equipment/Instrumentation/Surgical Devices II

Session Date/Time: Monday 12/09/2016 | 17:00-18:30

Paper Time: 17:40

Venue: Hall C1

First Author: : P.Hoffmann GERMANY

Co Author(s): :    M. Abraham   L. Zumhagen                 

Abstract Details

Purpose:

To evaluate biometric measurements generated with a swept source OCT biometer (Zeiss IOLMaster 700) for the purpose of IOL calculation.

Setting:

Private eye clinic in Germany

Methods:

2548 data sets were evaluated (ongoing study). The following parameters were examined: success rate, axial length (AL), central corneal thickness (CCT), anterior chamber depth (ACD), lens thickness (LT), corneal radii (r), anterior corneal astigmatism (ACA). As plug-in raytracing software cannot be installed on the machine, a simple thin-lens formula was developed uses AL, ACD and LT data to predict the position of the IOL. As a reference, the Holladay I formulae was used. SF constant was optimized to give a mean error of zero for any of 5 different IOL models used in a sub-set of 281 eyes (ongoing study).

Results:

98.8% of consecutive cataractous eyes could be successfully measured. AL was (minimum, median, maximum, mean, standard deviation) 20.01, 23.41, 32.55, 23.60, 1.47 mm. Mean corneal radius was 5.36, 7.72, 9.10, 7.73, 0.28 mm. ACD was 1.38, 3.15, 4.84, 3.12, 0.42 mm. Lens thickness was 3.13, 4.66, 5.99, 4.66, 0.44 mm. ACA was 0.0, 0.76, 12.88, 0.98, 1.21 D. Using the Holladay I formula, a mean prediction error of zero with a standard deviation of 0.446 D and a median / mean absolute error of 0.292 / 0.348 D could be achieved. Our proprietary formulae yielded a standard deviation of 0.397 D and median / mean absolute error of 0.260 / 0.310 D. This is statistically significant (T-test, P ≈ 0.02).

Conclusions:

Generated data sets are similar to the Haag-Streit Lenstar biometer but the rate of unmeasurable cataracts is considerably lower. Unfortunately, built in software still relies on outdated formulae. Plug-ins like PhacoOptics or Okulix cannot be used. Our very simple formulae improves IOL calculation precision by ≈ 10-12% over the best classical formula. Unlike classical formulae, no systematic errors do exist. The refractive improvement will be even more pronounced in very short eyes or highly aberration correcting IOLs. It would be desirable to have modern IOL calculation software integrated information about asphericity and posterior corneal curvature provided.

Financial Disclosure:

NONE