Prediction of effective Lens position using anterior segment optical coherence tomography in Chinese subjects with angle closure

Purpose To assess the accuracy of biometric parameters measured by anterior segment optical coherence tomography (AS-OCT) and partial coherence interferometry (PCI) in prediction of effective lens position (ELP) compared with previous formulas in PACG patients. Methods 121 PACG eyes were randomly divided into training set (85 eyes) and validation set (36 eyes) with same procedure including AS-OCT, PCI, phacoemulsification and IOL implantation surgery. Preoperative anterior chamber depth (pre-ACD), scleral spur depth (SSD), scleral spur width (SSW), lens vault (LV) and cornea thickness (CT) were measured from AS-OCT image. Axial length (AL) and corneal power (K) were measured by PCI. All the 7 parameters were analyzed by multiple linear regression in training set and a statistic regression formula was developed. In validation set, one-way ANOVA was applied to compare the new regression formula with Sanders-Retzlaff-Kraff theoretic (SRK/T), Holladay 1, Haigis, and a regression formula developed in previous study. Results The coefficient of determination (R 2 ) of different parameter combinations are 0.19 (pre-ACD, AL), 0.25 (AL, K) and 0.49 (SSD, AL, SSW) in training set. In validation set, the correlation between predicted and measured ELP are: new formula (R 2 = 0.50, P = 0.9947) Holladay 1 (R 2 = 0.12, P < 0.0001), SRK/T (R 2 = 0.11, P < 0.0001) and Haigis (R 2 = 0.06, P < 0.0001). Conclusion Among 7 tested parameters, pre-ACD contribute little in ELP prediction. Formula consist of SSD, AL and SSW showed better accuracy than other formulas tested.