CVas is a non-specific response to a variety of stimuli wherein abnormal blood vessels grow into the normally clear cornea. CVas is a double-edged sword, which on the one had is the body’s response to insult and helps in wound healing, combing invading organisms and eliminating noxious stimulants. On the other hand, CVas results in oedema, leakage of protein and lipids, perpetuate inflammation and affect the transparency of the cornea, affecting sight. Common causes of CVas include inflammatory conditions of the skin and connective tissues, infections, trauma, and certain ocular surface disorders [5,6,7, 16].

Various techniques are used in the evaluation and assessment of the extent, depth and severity of CVas. These include slit lamp biomicroscopy and photomicroscopy, videography, fluorescein angiography, OCT and optical coherence tomography angiography (OCTA), and corneal topography [12, 16, 17]. These modalities of assessment have several limitations such as imprecision, lack of quantification, difficulty of vessel delineation in areas of scarring, and evaluation of the blood flow within the vessels [18].

More recently, optical coherence tomography angiography (OCTA) has been established as a method for imaging retinal and corneal vessels. OCTA takes advantage of the high scan speed of Fourier-domain OCT for angiographic imaging and uses motion contrast to detect in vivo flow within blood vessels, without the need for contrast dye. This technology has been commercially available since 2014, however, much of its use has been limited to the research setting [19,20,21]. Furthermore, with the development of anti-angiogenic therapy, non-invasive techniques are needed for quantitative analysis in anterior segment vasculature. Binotti et al. and Ang et al. have used OCTA for the quantitative analysis of corneal neovascularisation to evaluate the extent of corneal NV and evaluate the severity of the condition [21,22,23,24]. However, anterior segment optical coherence tomography angiography (ASOCTA) evaluates a limited area and depth compared to ASOCT, which is designed for the external eye. To meet this challenge, ASOCT with angiography technology needs to be developed to help diagnosis and assessment of anterior segment diseases [24].

Unlike most studies on the use of ASOCT in corneal examination, our study was unique in that it monitored CVas through correlating ASOCT with the slit lamp examination and images. CVas presented as a hypo-reflective, dark back-shadow, which arises from depth of the location of major vascular trunks and extends throughout the entire corneal thickness. We have also established that the width of the shadow corresponded to the width of the vessel that casts the shadow. In multiple corneal NV, sequential vertical lines of back-shadows were clearly visible in a single scan, giving it the appearance of a bar code, previously described as the ‘bar code sign of corneal OCT’ [2].

Moreover, have been able to establish that the obstruction of the passage of light in ASOCT was mainly due to the circulating blood column (active vessels) rather than the vessel wall, as both arteries (afferent) and veins (efferent) have produced an equally dense back shadow despite their different vessel-wall structures. Our study also showed that partially regressed vessels as determined clinically by slit lamp imaging, cast a less dense back-shadow on ASOCT, and ghost vessels in cases of resolved CVas, which did not contain a circulating blood column, the back shadow was absent. Further evidence to demonstrate that the ‘regressed/closed’ vessels were not circulating blood comes from the observation that these vessels do not bleed during trephination for subsequent keratoplasty and that lipid keratopathy clears following fine needle diathermy occlusion of the vessels [14, 25].

We quantified this observation by measuring the integrated density in 26 ASOCT scans by measuring the density of the back-shadow cast by (active vessels) and (regressed vessels) [2]. Many studies have reported the attempt to clinically assess corneal NV in a qualitative or semi-quantitative manner based on the number of quadrants involved, depth of penetration and the centripetal progression. Vassileva et al., who determined the state of neovascularisation before and after intervention by comparing the size, number and centricity of vessels as well as the number of affected quadrants [26]. Similarly, Pillai et al. recorded the extent of corneal vascularisation with respect to the number of quadrants involved, the depth of vessels and whether they were active or quiescent [14].

Faraj et al. have also described the various clinical characteristics of corneal NV to standardise the clinical grading of corneal NV severity. Through their analysis of CVas images based on different parameters such as source, location and depth; number of quadrants affected; length (peripheral, mid-peripheral and central); branching pattern, leakage and nature of the blood flow. They also classified CVas into active young, active old, mature, partially regressed and regressed [10, 25].

In order to achieve a more accurate and quantitative measure of the extent of vascularisation, digital images provided the best way forward. In 1994, an in vivo automated method previously described the quantification of CVas based on contrast enhancement, followed by density threshold identification for the blood vessels and, finally, pixel measurement [27].

The same concept was adopted using ASOCT, by AlMaazmi et al. who studied aqueous leak in non-traumatic corneal perforations using ASOCT. The glue and corneal blood vessels consistently cast a dense shadow posteriorly. Relying on ASOCT they were able to differentiate between the back shadow cast by an active vessel and tissue glue through monitoring the density, width and depth of the shadow [28].

Varma et al. have used image processing software on ASOCT images in eyes with limbal stem cell deficiency (LSCD). They have described a method of quantifying the reflectivity patterns of the line scans of ASOCT and studies both epithelial and stromal reflectivities cast by scar tissue and/or corneal NV. However, as several corneal disorders with vascularization may mimic LSCD, they suggested that vessels in LSCD were mainly superficial and showed segmentation [29].

This study establishes a simple, semi-quantitative method using slit-lamp and ASOCT, to assess and monitor vessel activity. We have noted that all produced an equally dense back shadow, by comparing the integrated density in active (pre-treatment) and regressed (post-treatment) vessels. We inferred that the impediment to the passage of light was by the column of circulating blood in the vessels rather than the vessel wall. Hence, the presence of the ‘barcode sign’ was consistent with vessels with active circulation, and the clinical resolution, demonstrated by the reducing intensity or absence of the back shadow, corresponds to the low or absent blood flow in regressing and regressed vessels. This highlights the use of ASOCT in assessing CVas and their state of activity. In our personal experience, we have observed the back shadowing produced by corneal vessels with the Topcon OCT device (3D-OCT-2000) system (Topcon Corporation) and can also be seen in images in the published literature with the RTVue (Optovue, Inc., CA, USA) device [30].

Supplemental material is available at Eye’s website.