PurposeTo evaluate the refractive predictability of a partial coherence interferometry (PCI) biometry device (IOL Master®) for cataract surgery and to investigate factors that may affect it.MethodsRetrospective review of 209 eyes from 151 patients that had undergone preoperative PCI biometry and an uneventful phacoemulsification cataract surgery with posterior chamber intraocular lens (IOL) implantation was conducted. Prediction error defined as the intended refraction minus the postoperative refraction in spherical equivalent (SE) and the absolute error were analyzed according to IOL calculation formulas, patient characteristics, preoperative visual acuity (VA) and refraction, posterior subcapsular cataract (PSC), signal-to-noise ratio (SNR), and axial length (AL).ResultsThe overall refractive predictability of the PCI device was good. Generally, the SRK/T formula performed better than the SRK-II formula. Refractive predictability was slightly worse in eyes with ≥+2.0 diopters (D) of preoperative SE (with both SRK-II and SRK/T) and in eyes with an AL≤23.0 mm (only with SRK-II. No other factors significantly affected the refractive predictability of the PCI, although poor VA, dense PSC, and poor SNR were closely interrelated.ConclusionsThe SRK/T formula performed significantly better than the SRK-II formula. Eyes with an AL≤23.0 mm were associated with significantly greater hyperopic shifts in postoperative refraction with the SRK-II formula, but not with the SRK/T formula. A preoperative SE≥+2.0D was related to a significantly greater hyperopic shift in postoperative refraction. With proper verification of measured data and a suitable IOL calculation formula, good refractive predictability is expected from PCI biometry regardless of patient characteristics, preoperative VA, SNR, PSC, and AL.

Purpose: To assess the refractive outcome of cataract surgery employing IOL master and A-scans in diabetic and non-diabetic patients Methods: The retrospective comparative study included 205 eyes of consecutive patients who had uneventful cataract surgery implanting I-Flex IOL. Axial length was measured with IOL master and A-scan and IOL power was calculated using various formulas (SRK II, SRK/T, Haigis, Holladay). Subjects was separated into five groups according to axial length, and then the groups were divided into diabetic and non-diabetic subgroups. Differences between the predicted refraction and the actual refraction were compared and analyzed at two months after the operation. Results: The mean absolute errors (MAE) of ten groups showed no significant differences. Comparing diabetic groups and non-diabetic groups, there were no statistically significant differences. Also the result of the two modalities, IOL master and A-scan, were not different in statistical analysis. Conclusions: In diabetic and non-diabetic patients, IOL master and A-scan may be the accurate methods for calculating IOL power regardless of the axial length.