Predictability of pseudophakic refraction using patient-customized paraxial eye models

To determine whether patient-customized paraxial eye models that do not rely on exact ray tracing and do not consider aberrations can accurately predict pseudophakic refraction.Bascom Palmer Eye Institute, Miami, Florida.Prospective study.Cataract surgery patients with and without a history of refractive surgery were included. Manifest refraction, corneal biometry, and extended-depth optical coherence tomography (OCT) imaging were performed at least 1 month postoperatively. Corneal and OCT biometry were used to create paraxial eye models. The pseudophakic refraction simulated using the eye model was compared with measured refraction to calculate prediction error.49 eyes of 33 patients were analyzed, of which 12 eyes from 9 patients had previous refractive surgery. In eyes without a history of refractive surgery, the mean prediction error was 0.08 ± 0.33 diopters (D), ranging from -0.56 to 0.79 D, and the mean absolute error was 0.27 ± 0.21 D. 31 eyes were within ±0.5 D, and 36 eyes were within ±0.75 D. In eyes with previous refractive surgery, the mean prediction error was -0.44 ± 0.58 D, ranging from -1.42 to 0.32 D, and the mean absolute error was 0.56 ± 0.46 D. 7 of 12 eyes were within ±0.5 D, 8 within ±0.75 D, and 10 within ±1 D. All eyes were within ±1.5 D.Accurate calculation of refraction in postcataract surgery patients can be performed using paraxial optics. Measurement uncertainties in ocular biometry are a primary source of residual prediction error.