Abstract

Purpose: Two new post-op Anterior Chamber Depth (ACD) prediction calculation models have been developed. A retrospective data analysis of cataract patients, measured on Lenstar LS900, Haag-Streit, was devised to investigate the position prediction accuracy of the new models. The accuracy of the two models is also assessed for a separate clinical study of a novel ocular measurement device.

Methods: For the retrospective data analysis, a total of 306 eyes scheduled for cataract surgery, have been measured preoperatively and postoperatively at 4-8 weeks by means of a standard-of-care (SOC) ocular biometer (Lenstar LS900, Haag-Streit). Two new prediction models (Model A and B) for post-op IOL position, have been developed. For both models, the raw data of the SOC device were used to obtain true optical path properties and multi-regression analysis with preoperative clinical factors was performed to derive Models A and B.<br /> Model A depends on the pre-op ACD, Crystalline Lens Thickness (LT) and IOL Thickness variables and Model B depends on the pre-op ACD, LT, IOL Thickness, WTW, Gender and Anterior Corneal Axis.<br /> Models A and B have been tested on a second prospective cataract patients study, where a new ocular biometry system that consists of a Purkinje imaging method and an optical coherence device was used to measure the ocular parameters for IOL power calculation. The post-op ACD measured with the new device and the calculated predicted ACD values are compared with the corrected Lenstar measured ACD parameters.

Results: In the retrospective data analysis, the Model A mean position error between the ACD optical path and predicted IOL position was 0.1mm ±0.07mm for monofocal IOL’s. The Model B mean position error was 0.09 ±0.07 mm. No statistical significant difference was found between the measured and calculated means for both models (p = 0.1; paired t-test).<br /> The eyes measured on the new ocular biometer have shown a mean position error of 0.13 ±0.03 mm for Model A (p = 0.2; paired t-test) and 0.17 ±0.04 mm for Model B (p = 0.3; paired t-test). The difference between the post-op ACD measured on the two devices is 0.1±0.03mm (p = 0.54; paired t-test).

Conclusions: Two new ACD prediction models have been shown to deliver good prediction results compared with the existing Lenstar LS900. The models also provide the basis for accurate IOL power calculation with a new ocular biometer.