En face optical coherence tomography: a new method to analyse structural changes of the optic nerve head in rat glaucoma

Abstract

Aim: To investigate en face optical coherence tomography (eOCT) and its use as an effective objective technique for assessing changes in the glaucomatous rat optic nerve head (ONH) in vivo, and compare it with confocal scanning laser ophthalmoscopy (cSLO).

Methods: 18 Dark Agouti (DA) rats with surgically induced ocular hypertension were imaged with eOCT and cSLO at regular intervals. Assessment included three dimensional (3D) topographic reconstructions, intensity z-profile plots, a new method of depth analysis to define a "multilayered" structure, and scleral canal measurements, in relation to the degree of intraocular pressure (IOP) exposure.

Results: The increased depth resolution of the eOCT compared to the cSLO was apparent in all methods of analysis, with better discrimination of tissue planes. This was validated histologically. eOCT demonstrated several significant changes in imaged rat ONH which correlated with IOP exposure, including the area of ONH (p<0.01), separation between retinal vessel and scleral layers (p<0.05), and anterior scleral canal opening expansion (p<0.05).

Conclusion: eOCT appears to be effective in assessing rat ONH, allowing detailed structural analysis of the multilayered ONH structure. As far as the authors are aware, this is the first report of scleral canal expansion in a rat model. They suggest eOCT as a novel method for the detection of early changes in the ONH in glaucoma.

Figure 1

Image analysis method 1. En face OCT (eOCT) imaging of optic nerve head (ONH) in glaucomatous rats with 3D reconstructive blocks of a series of eOCT sections. The images consist of an eOCT image above and a confocal scanning laser ophthalmoscopy (cSLO) image below. Retinal blood vessel layers (RBVL) and scleral layer are easily demonstrated as indicated on the eOCT 3D reconstructive image (A). The edges of the ONH are clearly visible in the cross section through the ONH taken from A (B), and validated by confocal 3D reconstruction of immunohistochemistry (C, colour inset). Progressive widening of the scleral canal was seen in all glaucomatous rats (arrows in D–F). Typical examples of eOCT images (3D reconstructions D–F, XY projections G–I) are shown in the same glaucomatous animal at baseline (D, G), 2 months (E, H) and 3 months (F, I) with IOP integral of 0, 542.65, and 665.85 mm Hg days, respectively.

Figure 2

Image analysis method 2 of eOCT (A–F) and cSLO (G–K) images. Reflectance maps (akin to conventional surface topography) show the same eye of the same animal analysed with eOCT (A) and cSLO (G). Corresponding points, selected as shown by the cross intersection in A and G, are further analysed using longitudinal slice reconstructions in the y–z axis (B, H) and x–z axis (C, I). The corresponding confocal z-profile plots are shown in (D) and (J). A double peak z-profile was obtained with eOCT (D) but only a single peak K obtained with cSLO (J). Transversal optical slices corresponding to the peaks, as indicated in (D) and (J), are shown with eOCT (E, F) and cSLO (K).

Figure 3

Image analysis method 2. Multisurface depth analysis generating histograms for eOCT (A) and cSLO (B) of the same eye in the same animal. Histograms represent the number of (x,y) locations of the same depth of the reflecting surface (vertical axis) as a function of its depth (horizontal axis). All histograms are centred at the minimum between two peaks and normalised to the highest peak. Compared to cSLO, eOCT produced sharper and narrower peaks. For the analysis, we only included eyes with a double peak histogram for both eOCT and cSLO, where each peak could be approximated by a Gaussian curve. A typical Gaussian curve fit is shown in (C), from which the full width at the half maximum of the first (FWHM1) and second (FWHM2) peaks were calculated, as well as the distance between the two peaks (separation).

Figure 4

Image analysis method 3. A typical longitudinal slice from a 3D eOCT image in which the anterior scleral canal opening (ASCO), posterior scleral canal opening (PSCO) and scleral canal thickness (“t”) are marked as indicated (A). Corresponding measurements in the paraffin cross section through the ONH of a normal rat eye stained with haematoxylin and eosin, are shown for comparison (B).

Figure 5

Comparison of eOCT and cSLO using data acquired in methods 1 (A) and 2 (B) with corresponding box plots (A, B) and descriptive statistics (C). The eOCT demonstrates a narrow range for all three parameters compared to the larger spread and high variance of the same parameters assessed by the cSLO. Whereas no correlation was found between eOCT and cSLO in any parameter measured (Pearson’s correlation coefficient), there was a significant difference between the two imaging techniques for each parameter (ANOVA, p<0.05).

Figure 6

Analysis of changes in ONH structure with IOP exposure using eOCT. The change in disc area was found to be significantly correlated with ΔIOP integral, (A, Method 1, Pearson’s r = 0.546, p<0.01), as was the degree of separation between the two anatomical layers, corresponding to the retinal blood vessel and the scleral layers (B, Method 2, Pearson’s r = 0.456, p<0.05). Furthermore, we found a positive correlation between anterior scleral canal opening (ASCO) expansion and ΔIOP integral (C, Method 3, Pearson’s r = 0.462, p<0.05). All these results suggest that the eOCT can be used to objectively assess anatomical changes in ONH structure over time in glaucomatous rats.