Abstract
Purpose: To compare routine clinical examination with optical coherence tomography angiography (OCTA) for the assessment of limbal conjunctival ischemia following a chemical burn. Setting: Validity analysis. Methods: We assessed 10 participants (15 eyes) with an acute chemical injury. Clinical photographs were used to determine the extent of any limbal conjunctival epithelial defect and ischemia. These were compared with the extent of limbal ischemia identified on OCTA images of the ocular surface. Quantitative and longitudinal analysis using the OCTA software were also performed. Correlations with visual outcome were sought using clinical and OCTA-derived variables. Results: The extent of clinically determined limbal ischemia was less than that identified with OCTA (2.3±3.6 clock hours vs 5.1±4.2 clock hours, P =.003), which in turn was less than the size of limbal conjunctival epithelial defect (7.3±5.1 clock hours, P =.03). Longitudinal OCTA analysis showed that mean vessel area increased by 0.2%±0.1% during the study, corresponding to a rate of vascular recovery of 0.9 mm2/d. Significant correlations were found between visual outcome at 3 months and limbal conjunctival fluorescein staining (r = 0.67, P =.006), and limbal conjunctival ischemia on OCTA (r = 0.76, P =.001). Conclusions: OCTA can objectively identify and monitor the recovery of limbal ischemia following an acute ocular chemical injury. OCTA confirms that limbal ischemia is usually more extensive than is suggested by clinical examination, and the former is highly correlated with visual outcome. OCTA therefore is a useful tool in the management of ocular chemical injury.
| Original language | English |
|---|---|
| Pages (from-to) | 165-174 |
| Number of pages | 10 |
| Journal | American Journal of Ophthalmology |
| Volume | 205 |
| Early online date | 10 May 2019 |
| DOIs | |
| Publication status | Published - 1 Sept 2019 |
Funding
In this study, we have confirmed that OCTA can reliably image blood flow within the conjunctival vessels, permitting the documentation of areas of ischemia following chemical injury. In a proportion of eyes, OCTA could also detect ischemia of the intrascleral vessels, giving an estimate of the depth of the injury. The extent of limbal conjunctival ischemia determined by OCTA often differed from other clinical estimates; however, the former was found to best correlate with the final visual outcome after chemical injury. Chemical or thermal burns cause rapid denaturation of the tissue of the ocular surface. Epithelial loss and vascular closure (ischemia) are both important clinical signs that reflect the severity of the injury. Both are associated with irreversible loss of LESC, which can lead to limbal stem cell deficiency, conjunctival overgrowth on the cornea, and visual loss. It is not known whether the extent of an epithelial defect or ischemia at the corneoscleral limbus best reflects permanent LESC damage. An epithelial defect may only involve the superficial tissue and not include the deeper limbal stem cell crypt, in which case normal reepithelialization will occur. 14,15 Clinical estimation of the extent of limbal ischemia is largely subjective and could be highly variable between different examiners. 16 Blood vessels may also have a relatively normal appearance despite vascular stasis, which can only be distinguished by dynamic studies. Therefore, an objective method to accurately delineate the full extent and depth of limbal conjunctival vascular nonperfusion may provide a more accurate prognosis. 2,4,6 Dye-based angiography has been used to image the vasculature of the anterior segment 17–20 and identify conjunctival ischemia after chemical burns. 21 Kuckelkorn and associates used fluorescein angiography (FA) to delineate the area and depth of tissue ischemia, and they also observed that the extent of injury could be greater than that suggested by the clinical appearance. 21 However, owing to the rapid fluorescein transit time, booster injections were needed to fully assess a region of interest, with image degradation from leakage of fluorescein into the tissue. 21 Low-dose fluorescein injection, in which the dye is bound to albumin, reduces vascular leakage, but this has not been used to assess chemical burns. 18 Indocyanine green angiography has also been used to image the anterior segment, although these studies were directed to the effect of inflammation on the marginal corneal vascular arcades 20 or pathologic corneal neovascularization rather than chemical injury. 11,19 High levels of agreement between OCTA and FA have been demonstrated in studies assessing macular vasculature. 8,22 Advantages of OCTA include its ability in generating high-contrast, well-defined images of the retinal microvasculature without any obscuration from dye leakage–related hyperfluorescence. 7 Furthermore, images obtained by OCTA could be segmented and quantitatively analyzed so that individual layers of vasculature could be separately assessed. OCTA could also be performed more rapidly with no systemic risks or side effects compared with dye-based angiography. Limitations of OCTA include the fact that it could not assess vessel permeability and leakage, and that OCTA could be affected by shadow and motion artifacts similar to dye-based angiography. 7,23 Nevertheless, there is now a growing body of evidence demonstrating the usefulness of this technology. A number of reports have described the use of OCTA in the anterior segment. 9–12 The appearance of normal conjunctival vasculature has been described by our group previously. 9 Recently, Akagi and associates used swept-source OCTA to identify both the normal conjunctival and intrascleral vasculature. 12 In this study, we were able to identify the normal vasculature in unaffected areas of limbal conjunctiva (shown in Figures 1 and 2 ), as well as areas of ischemia. The latter was similar in appearance to previous reports using fluorescein angiography to investigate chemical burns on the ocular suface. 21 We found other advantages of OCTA in the assessment of ocular chemical burn. In this study, OCTA imaging was able to detect vascular flow despite the presence of conjunctival edema and hemorrhages, as shown in Figures 1 and 3 . This contrasts with OCTA studies in the posterior segment, in which masking effects by retinal edema and hemorrhages were reported. 24 The reason behind the discrepancy is uncertain. The severity of chemical injuries suffered by our consecutive cohort of patient ranged from mild to severe, and therefore selection bias toward less edema was unlikely. Otherwise, the differences may be because the conjunctival vascular flow is higher, the caliber of the vasculature is larger, or that the conjunctival vessels are more superficial in relation to surrounding tissues compared to retinal vessels. Future comparative studies using dye-based angiography could help confirm the presence of absence of masking effects. Excellent agreement was found between limbal ischemia determined by OCTA and conjunctival epithelial defect identified by fluorescein staining clinically; however, the OCTA findings did not agree with clinical estimates of limbal ischemia. The latter could be because we used clinical photographs to assess clinical signs, thus unable to dynamically assess vascular flow, and may have underestimated the full extent of limbal conjunctival ischemia. Furthermore, the detection of vascular pattern by OCTA is dependent on vascular flow (or more precisely, erythrocyte movement within the blood vessels). Areas with very slow blood flow, for example, in the setting of conjunctival vasospasm, therefore may not be detected by OCTA and appear as ischemic instead. 8 However, we did not observe dramatic changes of conjunctival vasculature on OCTA after a relatively short follow-up, suggesting that reversible vasospasm did not exert significant influence on our data. It has been shown that the extent of limbal epithelial defect could predict clinical outcomes after chemical burn better than relying on signs of limbal ischemia. 4,25 However, although we found that visual outcome at 3 months correlates well with limbal conjunctival fluorescein staining ( r = 0.67), it was superseded by OCTA estimation of limbal ischemia ( r = 0.76). We also noted that areas of limbal stem cell failure corresponded to OCTA-determined areas of severe limbal ischemia. OCTA therefore improves visualization of the limbal conjunctival vasculature, making it a more reliable sign useful for prognostication and treatment. It is interesting that the severity grading of either Roper-Hall or Dua classifications did not correlate with visual outcome in this study, suggesting that precise measures of the extent of limbal damage may be more informative instead. Gupta and associates have previously shown that the Dua classification is better associated with the clinical outcome at 1 year than the traditional Roper-Hall classification. 25 They noted that there was a positive correlation between the Dua classification and the formation of symblepharon, but a similar relationship was not found with visual outcome. Future comparison with dye-based angiography with longer follow-up would help to determine the precision of OCTA and its ability in prognostication after chemical burn injuries. Repeat OCTA examinations were easy to perform, suggesting that this is a suitable noninvasive method for longitudinal monitoring of vascular recovery. Indeed, we were able to provide an estimate area of vascular reperfusion after an ocular chemical burn using OCTA. However, we are cautious in our interpretation, because the lack of image registration would negatively influence the precision of our assessment. We nevertheless believe it is a development that should be further explored. En face OCTA potentially provides an additional dimension to the assessment of a chemical injury as, in a minority of eyes, we could distinguish the deeper intrascleral vessels. Although this may not be a more sensitive marker of LESC loss than clinical estimates of superficial limbal ischemia, it may be an index of more severe damage leading to iris atrophy, cataract, secondary ocular hypertension, or hypotony. Future studies are required to confirm our findings. The limitations of this study include the small patient cohort, the limited follow-up, and the lack of correlation of the acute clinical signs to long-term ocular surface changes, although the latter may also reflect the inadequacies of current grading systems. In addition, the study has only 1 image grader assessing all the images, and therefore an element of subjective bias may have influenced our results. Further studies on OCTA-outcome correlation and intergrader agreement would help to clarify these issues. The majority of the images were captured with the retinal mode of OCTA, so some estimates of vessel area and density may be inaccurate because of software calibration. Nevertheless, we found that the difference between the AngioRetina and the cornea modes of OCTA was primarily the signal strength index. Although the cornea mode could provide better-quality images, the retinal mode OCTA provides quantitative data that was unavailable in the cornea mode. As such, AngioRetina is currently our preferred mode of image acquisition. In summary, we report the application of OCTA to assess ischemia following an acute ocular chemical injury. Our results show that OCTA may give a more reliable and objective record of areas of vascular nonperfusion compared to clinical assessment, as well giving an estimate of the depth of the vascular damage. As such, the use of OCTA may help refine the prognosis after injury, enable monitoring of revascularization during healing and the documentation of the effect of treatment. All authors have completed and submitted the ICMJE form for disclosure of potential conflicts of interest and none were reported. The research was supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology , London, United Kingdom. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health, United Kingdom. All authors attest that they meet the current ICMJE requirements to qualify as authors.
Fingerprint
Dive into the research topics of 'Anterior Segment Optical Coherence Tomographic Angiography Assessment of Acute Chemical Injury'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver