Lens Thickness– Driven Refractive Modulation After Segmented Optical Biometry: Challenging The Composite Index Assumption In Intraocular Lens Calculation

To determine whether true geometric axial length (AL) calculation using parameter-specific refractive indices improves intraocular lens (IOL) power prediction compared with conventional composite-index biometry, and to evaluate the modulatory effect of lens thickness (LT) on refractive shift and predictive error across contemporary IOL formulas.

Single-center clinical study including cataract eyes implanted with TECNIS Eyhance™  (Johnson & Johnson Vision) or Clareon® (Alcon) intraocular lenses.

One hundred eighty-eight eyes of one hundred thirty-three patients were retrospectively analyzed. Optical biometry was performed using the Lenstar LS 900 (Haag-Streit), applying a composite refractive index. Standard measurements were compared with segmented calculations using parameter-specific refractive indices for axial length (AL), central corneal thickness (CCT), anterior chamber depth (ACD), and lens thickness (LT) to derive true geometric distances. Predictive errors of Barrett, Hill-RBF, SRK/T, Haigis, Hoffer Q, and Holladay formulas were calculated using both standard and segmented AL. Hyperopic and myopic shifts, vergence adjustments required to neutralize prediction error, and ROC-defined LT thresholds were determined.

Lens thickness (LT) emerged as the dominant determinant of refractive modulation. ROC analysis demonstrated strong discrimination for prediction shift (AUC 0.753–0.820, p<0.001), with segmented correction inducing a consistent myopic shift above approximately 4.4–4.7 mm LT. For error correction, LT also showed significant predictive capacity (AUC up to 0.723, p<0.001). In eyes exceeding approximately 4.5 mm LT, segmentation reduced predictive error for all formulas except Haigis. Conversely, Haigis accuracy worsened beyond this threshold, demonstrating an inverse response pattern.

Composite-index optical biometry may obscure clinically relevant geometric variations. True geometric axial length calculation significantly alters refractive behavior in thicker lenses, improving prediction accuracy for most vergence-based formulas while paradoxically degrading Haigis performance. Lens thickness appears to be the critical biometric trigger for segmentation benefit. LT-guided biometry refinement may therefore represent a targeted strategy to enhance refractive precision in selected eyes.