Comparison Of Zeiss Ai Calculator To Other Contemporary Intraocular Lens (Iol) Power Calculation Formulas
Published 2025 - 43rd Congress of the ESCRS
Reference: PPE.04 | Type: ESONT Abstract | DOI: 10.82333/m9h2-b758
Authors: Kathleen Scruggs* 1 , Nathan Cannon 1 , Seth Pantanelli 1 , Liam Redden 2 , Kamran Riaz 2
1Ophthalmology,Penn State Health Milton S. Hershey Medical Center,Hershey,United States, 2Ophthalmology,Oklahoma Health Center Dean McGee Eye Institute,Oklahoma City,United States
Purpose
To compare the accuracy of the novel Zeiss AI Calculator (ZAIC) to two established formulas, Barrett Universal II (BU2) and Kane.
Setting
Two academic centers (Penn State Eye Center, Hershey, PA and Dean McGee Eye Institute, Oklahoma City, OK)
Methods
This multicenter retrospective consecutive case series included 246 surgery naive eyes of 246 patients with IOLMaster 700 biometry, cataract extraction, monofocal IOL (MX60E) implantation, and refraction ≥ 21 days after surgery. For each eye, actual postoperative spherical equivalent (SE) was compared to predicted SE to each formula. Outcome measures included mean prediction error (MPE), mean absolute error (MAE), standard deviation (SD) of MPE, and proportion of eyes ≤ 0.5D of predicted. Differences were analyzed with Wilcox-Holladay-Wang-Koch heteroscedastic testing and Holm correction.
Results
There was no significant difference (all p=1.00) in the MPEs of ZAIC (0.099D), BU2 (0.74D), or Kane (0.010D). However, the SD of ZAIC (0.381D) was lower than that of BU2 (0.411D; p=0.001) but not lower than that of Kane (0.38 D; p=0.542). Likewise, the MAE of ZAIC (0.296D) was lower than that of BU2 (0.315 D; p= 0.044) but not lower than that of Kane (0.298D; p= 0.649. Finally, the proportion of predictions that were accurate to within 0.5 D was higher for ZAIC (82.1%) compared to BU2 (76.4%; p=0.11) but not compared to Kane (82.1%; 9=1.00).
Conclusions
ZAIC outperformed BU2 in terms of MAE, SD of the MPE, and proportion of predictions accurate to within 0.5 D. However, ZAIC and BU2 performed sumilarly in terms of MPE. There was no difference in performance between ZAIC and Kane across all outcome measures.