A comparison of two methods to calculate axial length

Abstract

Purpose: To compare prediction accuracy with the axial length (AL) calculation method of the Lenstar biometer (traditional AL) and that of the ARGOS biometer (sum-of-segments AL).

Setting: Private practice clinic.

Design: Comparative case series.

Main outcome measure: Mean absolute error (MAE).

Methods: Predictions were developed for nine formulas, grouping them into those derived with ultrasound (US) (SRK/T, Holladay 1 and 2, Hoffer Q, Haigis) and those derived with optical biometry (Barrett, OKULIX, Olsen from PhacoOptics, and Olsen from Lenstar). Formulas were ranked by MAE using sum-of-segments AL and traditional AL, in short eyes (traditional AL <22.0 mm), long eyes (traditional AL >26.0 mm), and all eyes.

Results: The study comprised 1442 eyes (54 short eyes and 67 long eyes) of 1070 patients. The best-ranking formula for long eyes was Haigis using sum-of-segments AL. For short eyes and for all eyes, OKULIX using sum-of-segments AL was best. Using sum-of-segments AL instead of traditional AL, Holladay 2 improved the most; Olsen from PhacoOptics worsened the most.

Conclusions: Some biometers used traditional AL, and at least one used sum-of-segments AL. Formula accuracy varied depending on how various commercial biometers internally calculate AL. Using sum-of-segments AL instead of traditional AL improved predictions for formulas designed on US data (SRK/T, Holladay 1, Holladay 2, Hoffer Q, and Haigis), although it worsened the Barrett and Olsen formulas. OKULIX was generally improved with sum-of-segments AL. When ranking by MAE, OKULIX ranked first.