Thomas Olsen1, Martin Thorwest. 1. University Eye Clinic, Aarhus Kommunehospital, Aarhus, Denmark. tkolson@dadlnet.dk
Abstract
PURPOSE: To study the conditions for consistent axial length measurements with partial coherence interferometry (PCI) performed with the Zeiss IOLMaster. SETTING: University Eye Clinic, Aarhus, Denmark. METHODS: A consecutive, unselected series of 1289 cataractous eyes were measured with the optical technique of PCI according to the IOLMaster as well as with conventional (contact) A-scan ultrasound (US) for the measurement of axial length. For each PCI reading, the signal-to-noise ratio (SNR) was recorded and used for comparison with the US measurement. All patients had routine phacoemulsification with implantation of an intraocular lens (IOL). In 284 cases, the patients were reexamined 2 to 3 months after surgery and the axial length was again measured using PCI. The readings of the IOLMaster, which had been calibrated against immersion US from the manufacturer, were recalculated to represent the true optical length and used in the analysis of the consistency of the measurements. RESULTS: Not all readings obtainable with the IOLMaster were of good quality, and large differences with conventional US were found. The error between US and PCI decreased significantly with increasing SNR, showing a minimum error at an SNR value above 2.1. The SNR correlated significantly with the visual acuity with considerable scatter, however. Excluding readings with a poor quality (SNR <2.1), the postoperative PCI measurements showed a high correlation with the preoperative measurement (r = 0.99), showing a mean difference of 0.08 mm +/- 0.12 (SD). The difference was highly significantly different than zero (P < .001) and may be explained by a higher refractive index of the biological lens than assumed in the original calibration of the IOLMaster. CONCLUSIONS: The quality of the axial length readings of the IOLMaster was influenced by the SNR value. However, with proper SNR evaluation and recalibration of the PCI measurements, it is possible to achieve consistent PCI readings with little variation between preoperative and postoperative measurements. These results are promising for a higher accuracy of the IOL power calculation.
PURPOSE: To study the conditions for consistent axial length measurements with partial coherence interferometry (PCI) performed with the Zeiss IOLMaster. SETTING: University Eye Clinic, Aarhus, Denmark. METHODS: A consecutive, unselected series of 1289 cataractous eyes were measured with the optical technique of PCI according to the IOLMaster as well as with conventional (contact) A-scan ultrasound (US) for the measurement of axial length. For each PCI reading, the signal-to-noise ratio (SNR) was recorded and used for comparison with the US measurement. All patients had routine phacoemulsification with implantation of an intraocular lens (IOL). In 284 cases, the patients were reexamined 2 to 3 months after surgery and the axial length was again measured using PCI. The readings of the IOLMaster, which had been calibrated against immersion US from the manufacturer, were recalculated to represent the true optical length and used in the analysis of the consistency of the measurements. RESULTS: Not all readings obtainable with the IOLMaster were of good quality, and large differences with conventional US were found. The error between US and PCI decreased significantly with increasing SNR, showing a minimum error at an SNR value above 2.1. The SNR correlated significantly with the visual acuity with considerable scatter, however. Excluding readings with a poor quality (SNR <2.1), the postoperative PCI measurements showed a high correlation with the preoperative measurement (r = 0.99), showing a mean difference of 0.08 mm +/- 0.12 (SD). The difference was highly significantly different than zero (P < .001) and may be explained by a higher refractive index of the biological lens than assumed in the original calibration of the IOLMaster. CONCLUSIONS: The quality of the axial length readings of the IOLMaster was influenced by the SNR value. However, with proper SNR evaluation and recalibration of the PCI measurements, it is possible to achieve consistent PCI readings with little variation between preoperative and postoperative measurements. These results are promising for a higher accuracy of the IOL power calculation.