Comparison Of 11 Intraocular Lens Power Calculation Formulas In Eyes Undergoing Combined Descemet Membrane Endothelial Keratoplasty And Cataract Surgery
Published 2024 - 42nd Congress of the ESCRS
Reference: FP06.05 | Type: Free paper | DOI: 10.82333/3sdt-jj97
Authors: Eva Sapok* 1 , Klemens Paul Kaiser 1 , Thomas Kohnen 1 , Ingo Schmack 1
1Department of Ophthalmology,Goethe University,Frankfurt,Germany
Purpose
To evaluate the accuracy of 11 intraocular lens (IOL) calculation formulas: Barrett Universal II, Emmetropia Verifying Optical (EVO), Haigis, Hill-Radial Basis Function (Hill-RBF), Hoffer Q Savini/Taroni (Hoffer QST), Holladay 1, Holladay 2, Kane, Postoperative spherical Equivalent prediction using ARtificial Intelligence and Linear algorithms, by DebellemanieĢre, Gatinel, and Saad (PEARL DGS), Sanders-Retzlaff-Kraff/theoretical (SRK/T) and T2 in eyes undergoing “Triple DMEK” (Descemet membrane endothelial keratoplasty combined with cataract surgery).
Setting
Department of Ophthalmology, Goethe University, Frankfurt, Germany. Retrospective case series.
Methods
The retrospective case series included 80 eyes of 80 patients (52 females, 28 males) with a mean age of 67.08 ± 7.64 years (from 45 to 83 years old) having Descemet membrane endothelial keratoplasty combined with cataract surgery at the Department of Ophthalmology, Goethe University, Frankfurt, Germany between 2016 and 2023. Preoperative biometry measurements were obtained from an IOLMaster 700; Carl Zeiss Meditec. Evaluation of the measurements was performed by comparison of the mean prediction error and the mean and median absolute prediction error. Percentages of eyes within prediction errors of ±0.5 diopters (D), ±1.0 D and ±2.0 D were evaluated as well.
Results
The median absolute refractive prediction error was lowest in the Hill-RBF (0.76 D) followed by Kane and SRK/T (0.80 D each), Hoffer QST (0.81 D), EVO and T2 (0,82 D each), Barrett Universal II and PEARL DGS (0.83 D each), Haigis (0.85 D) and Holladay 1 and Holladay 2 (0.88 D each). The lowest maximum absolute prediction error was obtained with SRK/T (3.97 diopters). The hyperopic shift was lowest in the Holladay 1 formula with a mean prediction error of -0.41 D. Regarding eyes within a predicition error of ±0.5 D Hoffer QST (37.5%) performed best, followed by Haigis and Holladay 1 (36.3% each). All formulas lead to more than 78% of eyes within a predicted refraction error of ±2.0 D.
Conclusions
Prediction of the actual postoperative refraction error in eyes that need both Descemet membrane endothelial keratoplasty (DMEK) and cataract surgery is still a challenge and not as accurate, as in healthy eyes without corneal changes. In our case series modern formulas for IOL power calculation as the Hill-Radial Basis Function and the Kane formula achieved the best results in IOL power calculation accuracy, although the refractive prediction with the Holladay 1 formula showed the lowest hyperopic shift. This should be taken into account when calculating and choosing the IOL power in patients undergoing DMEK combined with cataract surgery.