ObjectiveTo compare the biometric parameters provided by A-scan ultrasonography and the Lenstar optical biometer in guinea pig eyes, including anterior segment depth (ASD), lens thickness (LT), vitreous chamber depth (VCD), and axial length (AL), and differences of them between treated form deprivation (FD) eyes and untreated fellow eyes after 4 weeks of FD. MethodsThree-week-old guinea pigs (N = 41) were subjected to biometric measurements before monocular FD (baseline) and after a 4-week FD. Statistical analyses including within-subject standard deviation (SDwithin), coefficient of variation (CV), and intraclass correlation coefficient (ICC), used to evaluate repeatability for both the A-scan ultrasonography and the Lenstar individually, and correlation and Bland-Altman analyses were used to assess agreement between the two methods. The absolute values of ASD, LT, VCD and AL as measured by the two devices were compared, and the differences of them between treated (T) and untreated fellow (F) eyes (ΔASD, ΔLT, ΔVCD and ΔAL) (Δ = T-F) were compared between the two devices after 4 weeks of FD. ResultsMeasurements by the Lenstar (ICC: 0.923–0.994) were more repeatable than A-scan ultrasonography (ICC: 0.825–0.870). There was a high correlation for AL (r = 0.851, P < 0.001), a moderate correlation for VCD (r = 0.571, P < 0.001) and LT (r = 0.423, P < 0.001), and a low correlation for ASD (r = 0.230, P < 0.01) between the two devices. The values for ASD, VCD and AL measured by A-scan ultrasonography were larger than those measured by the Lenstar (all, P < 0.001), while LT provided by A-scan ultrasonography was much smaller than that of the Lenstar (P < 0.001). Bland-Altman plots showed poor agreement of absolute values of the four parameters between the two devices. Moreover, there was a high correlation between both methods for ΔAL (r = 0.704, P < 0.001), a moderate correlation for ΔVCD (r = 0.534, P < 0.001) and ΔASD (r = 0.574, P < 0.001), and no correlation for ΔLT (r = 0.303, P = 0.054). The ΔASD, ΔLT, and ΔAL measurements obtained by A-scan ultrasonography were greater than those obtained by the Lenstar (all, P < 0.001), while ΔVCD was mildly smaller using A-scan ultrasonography (P < 0.05). Bland-Altman plots illustrated there is good agreement of ΔAL, ΔVCD, ΔASD, and ΔLT between the two devices. ConclusionsThe Lenstar exhibited better repeatability and provided smaller measurements for AL, VCD and ASD than A-scan ultrasonography. Furthermore, a high correlation and a good agreement for the ΔAL was observed between the two devices after a period of FD. In summary, the two devices cannot replace each other directly to obtain absolute values of ASD, LT, VCD and AL, but the Lenstar still can serve as an option in measuring ΔAL between eyes in guinea pig myopia model.
Read full abstract