Comparisons of 13-lined ground squirrel ocular biometry derived from multiple modalities

Abstract

Purpose : To assess inter-modality agreement of axial length measurements in the 13-lined ground squirrel (13-LGS)

Methods : Both eyes of four euthermic 13-LGS (1M, 3F) were imaged in this study. All data were collected between the hours of 10am and 5pm within a 1-week timeframe in October. T₂-weighted coronal slices (TR=4026ms, TE_eff=33ms, ETL=8, NA=4, slices=8 contiguous, slice thickness=0.3mm) were acquired on a 9.4T Bruker BioSpec 94/20 small-animal MRI scanner with a custom-made surface coil. In 2D MRI images, axial length was measured as the maximal distance between the anterior surface of the cornea to the sclera using the plot profile tool in ImageJ. Whole-globe OCT volume scans (7mm; 650 A-scans/B-scan, 300 B-scans) were collected in each eye concurrently using a Proveo surgical scope equipped with a Leica EnFocus Spectral Domain OCT system. Within an OCT volume scan, the B-scan identified as containing the maximal distance between the anterior cornea to the retina was analyzed using the plot profile tool in ImageJ. For each eye, two OCT measurements were obtained, measuring to either the inner limiting membrane (OCT-ILM) or to the choroid (OCT-Ch). Finally, an Accutome ophthalmic ultrasound system was used to collect contact A-scans, and a total of 10 single acquisitions were averaged to produce axial length measurements in each eye.

Results : Data from all three methods were normally distributed (Shapiro-Wilk test). Using a one-way ANOVA, we found a difference in axial length between at least two different methods (F(1.356, 9.490) = 8.542), p=0.012). Post-hoc testing with Tukey’s HSD Test for multiple comparisons showed that the mean axial length was significantly different between OCT-ILM and OCT-Ch (p<0.0001) and between OCT-Ch and A-scan (p=0.0042). No significant differences in axial length values were observed between any other modalities.

Conclusions : On average, axial length measurements of the 13-LGS eye differed by up to 0.5mm between methods. Such errors affect the accuracy of quantitative measures extracted from retinal images. For example, errors of this magnitude would result in errors of cone density (per mm²) of up to 11%. Device settings for the OCT and A-scan are tuned for the human eye, so adjustment may be necessary to improve their accuracy in the 13-LGS eye.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Table 1: Measurements (in mm) of 13-lined ground squirrel axial length across multiple modalities.