Approximating sum-of-segments axial length from a traditional optical low-coherence reflectometry measurement

Purpose: To present the Cooke-modified axial length (CMAL(Initial)) method, which closely approximates sum-of-segments AL. Notably, sum-of-segments AL has been shown to improve predictions of many intraocular lens (IOL) power formulas; however, calculating this AL requires information that is not readily available. Design: Comparative case series. Patients: Distinct datasets of 215 eyes and 1442 eyes, which were measured before cataract surgery with a commercially available optical biometer (Lenstar LS 900), were identified. The AL measured by this machine was labeled "traditional AL." Main Outcome Measure: Prediction Error. Methods: The CMAL(Initial), sum-of-segments AL, and traditional AL methods with Bland-Altman plots and r(2) values were compared, along with graphs of prediction errors. Results: The CMAL(Initial) was developed from 215 eyes and evaluated in the 1422-eye validation dataset. The r(2) for CMAL(Initial) versus the sum-of-segments AL was 0.99983. The predictions based on CMAL(Initial) were compared with those based on traditional AL using the Hoffer Q, Holladay 1, SFIK/T, and Holladay 2 IOL formulas. The CMAL(Initial) produced more accurate predictions in all four formulas (P<.001). Eyes in all datasets were then combined to create the final recommendation: CMAL(Final) = 1.23853 + 0.95855 x traditional AL - 0.05467 x lens thickness (all measurements in millimeters). Conclusions: A modified AL method (CMAL) was easy to calculate. Using CMAL improved predictions for at least four IOL power prediction formulas, especially at extreme ALs. Caution is advised if using CMAL with other formulas. (C) 2019 ASCRS and ESCRS