目的：比较角膜生物力学分析仪(Corvis ST)、A型超声测厚仪(A超)、Lenstar LS 900(Lenstar)在测量近视眼角膜厚度的差异，为屈光术前筛查提供参考。方法：纳入2020年3至12月佛山市第二人民医院收治的屈光术前近视患者244例(244只右眼)，并按中央角膜厚度(central corneal thickness，CCT)将患者分为3组：薄角膜组(CCT≤520 μm)、正常角膜组(520 μm570 μm)，分别采用Corvis ST、A超、Lenstar三种仪器测量3组患者的中央角膜厚度，并纳入对比指标。结果：Corvis ST与A超、Lenstar在薄角膜组、正常角膜组、厚角膜组的测量值对比差异均有统计学意义(均P<0.05)；在相关性方面，Corvis ST与A超、Lenstar在薄角膜组、正常角膜组、厚角膜组呈正相关(r=0.841，P<0.001 vs r=0.832，P<0.001；r=0.866，P<0.001 vs r=0.918，P<0.001；r=0.497，P<0.001 vs r=0.801，P<0.001)；在一致性方面，Corvis ST与A超、Lenstar在薄角膜组、正常角膜组、厚角膜组上一致性系数ICC及95%置信区间分别为0.602(0.091，0.857)，0.758(0.221，0.901)、0.741(0.207，0.890)，0.845(0.396，0.938)、0.417(0.005，0.679)，0.667(0.172，0.851)，95%一致性区间范围分别是(-2.2~24.2 μm，-6.7~19.5 μm；-8.4~24.6 μm，-6.8~19.4 μm；-8.7~23.8 μm，-7.0~18.3 μm)。结论：通过对3组数据统计分析得出，Corvis ST与A超、Lenstar的测量值相比，差异均有统计学意义，且Corvis ST的测量值均高于后两种仪器；Corvis ST与后两者均有着较高的相关性，且在正常角膜厚度测量上相关性最高；在一致性方面，从ICC点估计及95%置信区间数值分析，一致性较差；从Bland-Altman角度分析，Corvis ST与后两者的绝对值最大差值超过10 μm，一致性较差，但在与Lenstar的绝对值最大差值均在20 μm内。因此，在测量精度要求较高的屈光手术筛查方面，Corvis ST与A超、Lenstar之间不能相互替代，而对于大规模临床筛查，其角膜厚度测量值对眼压的影响有一定的临床参考意义。

Objective: To compare the differences of corneal thickness measured by corneal biomechanical analyzer (Corvis ST), A-mode ultrasonic thickness meter (A-ultrasound) and Lenstar LS 900 (Lenstar) in myopia, so as to provide reference for preoperative screening of refractive surgery. Methods: A total of 244 patients (244 right eyes) with myopia before refractive surgery in Foshan Second People’s Hospital from March to December 2020 were included in the study, and the patients were divided into three groups according to the central corneal thickness (CCT): the thin cornea group (CCT≤520 μm), the normal cornea group (520 μm570 μm). The central corneal thickness (CCT) of the three groups was measured by Corvis ST, A-type ultrasonic thickness gauge and Lenstar LS 900, respectively, and included in the comparison index. Results: The measured values of Corvis ST, A-ultrasound and Lenstar in the thin cornea group, the normal cornea group and the thick cornea group were statistically significant (all P<0.05). Corvis ST was positively correlated with A-ultrasound and Lenstar in the thin cornea group, the normal cornea group and the thick cornea group (r=0.841, P<0.001 vs r=0.832, P<0.001; r=0.866, P<0.001 vs r=0.918, P<0.001; r=0.497, P<0.001 vs r=0.801, P<0.001). In terms of consistency, the consistency coefficients ICC of Corvis ST, A-ultrasound and Lenstar in the thin cornea group, the normal cornea group and the thick cornea group were 0.602 (0.091, 0.857), 0.758 (0.221, 0.901); 0.741 (0.207, 0.890), 0.845 (0.396, 0.938); 0.417 (0.005, 0.679), 0.667 (0.172, 0.851), and the range of 95% consistency interval were -2.2–24.2 μm, -6.7–19.5 μm; -8.4–24.6 μm, -6.8–19.4 μm; -8.7–23.8 μm, -7.0–18.3 μm. Conclusion: Through the statistical analysis of the data from the three groups, it is concluded that the measured values of Corvis ST are statistically significant compared with those of A-ultrasound and Lenstar, and the measured values of Corvis ST are higher than those of the latter two instruments. Corvis ST has a high correlation with the latter two, and the correlation is the highest in the measurement of normal corneal thickness. In terms of consistency, from the ICC point estimation and 95% confidence interval numerical analysis, the consistency was poor. From the perspective of Bland Altman, the maximum absolute value difference between Corvis ST and Lenstar was more than 10 μm, indicating poor consistency, but the maximum absolute value difference between Corvis ST and Lenstar was within 20 μm. Therefore, in refractive surgery screening with high measurement accuracy, Corvis ST, A-ultrasound and Lenstar cannot replace each other. However, in large-scale clinical screening, the measured value of corneal thickness has a certain clinical reference significance on the impact of intraocular pressure.