常量(计算机编程)
多中心研究
超声波
医学
数学
核医学
眼科
外科
随机对照试验
计算机科学
放射科
程序设计语言
作者
David L. Cooke,Rhonda Waldron,Giacomo Savini,Kamran M. Riaz,Leonardo Taroni,David A. Murphy,Fabio Guaraldi
标识
DOI:10.1097/j.jcrs.0000000000000865
摘要
To compare axial length (AL) measurements obtained by optical biometry (OB) and immersion ultrasound (iUS) to investigate the agreement between the 2 techniques and how to use OB constants for eyes with AL by iUS.Multicenter study.Retrospective observational case series.Agreement between OB and iUS ALs was investigated in 4 subsets. Also, in a test database, the prediction error (PE) for iUS AL was assessed with 4 methods: (1) data-optimized constants; (2) user group for laser interference biometry (ULIB) constants with iUS biometry; (3) with recalibrated AL; and (4) ULIB A-constant - 0.23.A Combined 1970 eyes were measured with both OB and iUS biometry. OB mean AL was 0.0873 mm longer than iUS AL. The latter was made equivalent to OB ALs with this equation: Recalibrated iUS AL = 1.0228 × iUS AL - 0.4556. In a fifth database (n = 1079) with OB AL measurements only, after AL was artificially shortened by 0.0873 mm, the original A-constant had to be reduced by 0.23 to maintain a zero PE. In a sixth database (n = 127) with iUS AL, the original ULIB A-constant provided the poorest outcomes. Using either recalibrated iUS AL or ULIB A-constant - 0.23 zeroed out the mean PE and achieved the lowest median absolute error.AL measurements by iUS can be used with ULIB constants for OB by subtracting 0.23 from the A-constant; alternatively, the iUS AL may be recalibrated. The recalibrated iUS AL should be treated as AL measurements obtained by OB. It is longer than iUS AL in long eyes.
科研通智能强力驱动
Strongly Powered by AbleSci AI