Comprehensive Summary
This study, presented by Sapok et al. investigated the accuracy of 11 intraocular lens (IOL) power calculation formulas, which are used to determine the position of an IOL with respect to axial length and corneal curvature, in patients undergoing Descemet membrane endothelial keratoplasty (DMEK) combined with cataract surgery. Researchers designed a retrospective accuracy and validity analysis with a consecutive case series utilizing 80 eyes (52 females, 28 males) from the Department of Ophthalmology, Goethe University, Frankfurt, Germany. In order to determine the accuracy of these 11 IOL power calculation formulas, researchers compared the predicted postoperative refraction with the observed postoperative refraction using the IOL Master 700 to obtain baseline measurements. Researchers quantified the accuracy of the power formulas using D (diopters) along with statistical analysis carried out using Microsoft Excel and SPSS software. Statistical analysis yielded 5 main outcome measurements which were mean prediction error (ME), mean absolute error (MAE), median absolute error (MedAE), standard deviation (SD) along with percentage of eyes at set intervals of deviation from the observed postoperative refraction compared to the predicted. Researchers were able to determine that the Barrett Universal II (BUII) formula, an AI powered IOLup1D formula, performed the best demonstrating the lowest MedAE (0.64 D ± 0.80) making it the most accurate formula amongst the 11 while Hill-RPD, an AI-powered calculation, and SRK/T performed the worst (both 0.74 D ±0.79). Researchers were able to determine that the main challenge that comes with simultaneous DMEK and cataract surgery is hyperopic shift and the use of IOLup1D, a specific surgical technique used to compensate for DMEK can greatly improve final visual outcome. Alongside IOLup1D, researchers found that AI powered IOL formulas possess great potential but require extremely large, diverse datasets to be practical in a clinical setting.
Outcomes and Implications
IOL power calculation research is necessary for ensuring good patient outcomes for DMEK and cataract surgeries, however when performed simultaneously, standard IOL formulas become less accurate due to the complexity of the condition. As a result, this study, through the analysis of 11 different IOL power calculations, addresses a clinical gap in treating those with complex corneal conditions accurately and effectively. By showing additional adjustments to the IOL procedure such as IOLup1D can improve postoperative refractive outcomes and that AI-powered IOL calculations such as Hill-RPD can provide additional predictive value, Sapok et al. highlights the evolving role of AI in personalized ophthalmology especially with those with more complex conditions. Clinically, the adoption of adjusted IOL procedures and AI-powered calculations possess great upside potential in terms of improving postoperative outcomes and accuracy during procedures. As AI-powered formulas are accessible in modern biometers and the IOLup1D adjustment involves choosing an IOL 1 diopter higher than the predicted power, both of these can be readily applied in near-term clinical settings.