Quantitative Analysis of the Corneal Collagen Distribution after In Vivo Cross-Linking with Second Harmonic Microscopy

Abstract

Corneal cross-linking (CXL) is a surgical procedure able to modify corneal biomechanics and stabilize keratoconus progression. Although it is known that CXL produces changes in corneal collagen distribution, these are still a topic of discussion. Here we quantitatively compare the corneal stroma architecture between two animal models four weeks after in vivo conventional CXL treatment, with second harmonic generation (SHG) imaging microscopy and the structure tensor (ST). The healing stage and the stroma recovery were also analyzed by means of histological sections. Results show that the CXL effects depend on the initial arrangement of the corneal collagen. While the treatment increases the order in corneas with a low level of initial organization, corneas presenting a fairly regular pattern are hardly affected. Histological samples showed active keratocytes in anterior and middle stroma, what means that the recovery is still in progress. The combination of SHG imaging and the ST is able to objectively discriminate the changes suffered by the collagen arrangement after the CXL treatment, whose effectiveness depends on the initial organization of the collagen fibers within the corneal stroma.

Figure 1

Example of the use of the ST in an artificial image presenting a structural organization with two preferential directions. The PO histogram represents the frequency of appearance of a certain orientation within the fibers (i.e., the closer to 1, the higher the presence of that direction in the image).

Figure 2

SHG signal from the corneal stroma in untreated rabbit (left) and adult chicken (right) corneas. Depth locations were 40 and 120 μm, respectively. Bar length: 50 μm.

Figure 3

SHG images of a control rabbit cornea (left panels) and a post-CXL cornea (right panels). Depth positions correspond to the anterior (40 μm, top), mid (120 μm, middle), and posterior (240 μm, bottom) locations. The size of the images is the same as in the previous figure. Scale bar: 50 μm.

Figure 4

Histograms of PO distribution for a control and a post-CXL rabbit cornea at three depth locations. These were computed by applying the ST to the SHG images in Figure 3. The corresponding DoI mean values are also included for direct comparisons.

Figure 5

SHG intensity profiles as a function of depth for control (blue dots) and post-CXL (red dots) rabbit corneas. The data of each line are the average for all the specimens. The values have been normalized for a direct comparison.

Figure 6

SHG images of a control adult chicken cornea (left panels) and a post-CXL cornea CXL (right panels). Depth positions (from top to bottom) correspond to the anterior, mid, and posterior locations. The size of the images is the same as in Figure 3.

Figure 7

Histograms of PO distribution for an untreated (left) and a post-CXL chicken cornea (right) at three depth locations. The corresponding averaged DoI values are also included.

Figure 8

Comparison of SHG signal profiles as a function of corneal depth in control (blue dots) and post-CXL (red dots) chicken eyes. Each line corresponds to the mean values across all the specimens for each experimental condition. Data have been normalized for direct comparisons.

Figure 9

Representative DoI maps for post-CXL and control corneas (anterior stroma) for a rabbit (top) and an adult chicken (bottom) specimens. The averaged DoI value across each map is also given.

Figure 10

Comparison of averaged DoI values for control (red bars) and post-CXL (blue bars) corneas for three corneal locations. Data correspond to the mean values across all specimens (rabbits (a) and chickens (b)). Error bars indicate the standard deviation.

Figure 11

Histological sections of rabbit and chicken corneas (control vs. post-CXL) stained with Masson's trichrome. Dark stained cells are keratocytes. Activated keratocytes present a bigger size (see text for further information).