Abstract:

Objective  To assess the accuracy of Sirius ray-tracing method for intraocular lens power calculation in post-corneal -refractive-surgery eyes. Design Comparative case series. Participants 21 post-LASIK eyes and 2 post-PRK eyes undergone phacoemulsification in Beijing Aier eye Hospital from June 2015 to December 2017. Methods The axial length, keratometry and anterior chamber depth was measured with IOL Master. The data of anterior segment including corneal power, anterior chamber depth and the angle of anterior chamber was measured with the Sirius three-dimensional corneal topography. The intraocular lens power was calculated using Sirius ray-tracing method, Haigis-L formula and Barrett True K No History formula. The implanting IOL power was determined mainly according to the result of Haigis-L formula, also referring to Barrett True K No History formula and Sirius ray-tracing method. All phacoemulsification operation were performed by the same operator. Postoperative refractions were measured with subjective refraction one month postoperatively. The predictive error （the difference between the actual postoperative refractive error and that predicted by the formula）, the absolutely predictive error （the absolute value of the prediction error）and the absolutely predictive error distribution（the percentage of absolute prediction error within 0.50 D  and the percentage of absolute prediction error within 1.0 D） of the four methods were calculated. The absolutely predictive error and the absolutely predictive error distribution of different methods were compared. Main Outcome Measures The predictive error, the absolutely predictive error and the absolutely predictive error distribution of the four methods. Results The predictive error of the four methods was (+0.22±0.57)D, (-0.07±0.73) D, (-0.18±0.81) D, (-0.00±0.57) D respectively. The absolutely predictive error was (0.47±0.38)D, (0.57±0.44) D, (0.64±0.52) D, (0.44±0.34) D respectively. The absolutely predictive error of the mean value method was the lowest, but there was no statistical difference in all the formulas (P=0.364). The absolutely predictive error of Sirius ray-tracing method was no more than 0.5 D in 16 eyes (69.57%), and no more than 1.0 D in 18 eyes (91.30%). The absolutely predictive error of Haigis-L formula was no more than 0.5 D in 11 eyes (47.83%), and no more than 1.0 D in 20 eyes (86.96%). The absolutely predictive error of Barrett True K No History formula was no more than 0.5 D in 12 eyes (52.17%), and no more than 1.0 D in 18 eyes (78.26%). The absolutely predictive error of the mean value method was no more than 0.5 D in 14 eyes (60.87%), and no more than 1.0 D in 20 eyes (86.96%). There was no statistical difference in all the formulas in the number of eyes in which the absolutely predictive error was no more than 0.50 D (P=0.453), and there was also no statistical difference in the number of eyes no more than 1.0 D (P=0.636). Conclusions  When used in IOL power calculation in post-LASIK or post-PRK eyes, the range of variation of Sirius ray-tracing method was narrower than Haigis-L formula and Barrett True K No History formula  which were commonly used in clinically. Sirius ray-tracing method could result mild hyperopia. The mean value method should be used in the IOL power for post-corneal-refractive-surgery eyes. (Ophthalmol CHN, 2019, 28: 98-103)

Key words: ray-tracing, Haigis-L formula, Barrett True K No history formula, intraocular lens, excimer laser, Sirius