Determination of Surgical Strategies for Burn-Induced Conjunctivalized Corneas Using Optical Coherence Tomography

CLINICAL SCIENCE

Chunxiao Wang, PhD, MD,* Xueying Xia, MD,† Feipeng Wang, MBBS,* Shaobi Ye, MBBS,* and Shiyou Zhou, MD, PhD*

onjunctivalization of corneas is a severe complication of Purpose: To evaluate the clinical use of optical coherence Climbal stem cell deficiency. It mostly follows ocular tomography (OCT) in optimizing surgical treatment strategies for burns in some developing countries, leading to significant conjunctivalized corneas secondary to ocular burns. visual impairment.1–3 Surgical treatment for conjunctivalized Methods: This noncomparative observational study included 25 corneas is a great challenge for corneal surgeons. In this patients with stable ocular burns and conjunctivalized corneas. OCT condition, corneal opacity obscures the clinical differentiation of deeper corneal structures. It is difficult to determine the was performed on each eye. The thickness of corneal opacity or fi pseudopterygium and the underlying healthy stroma were measured. speci c layers of the diseased tissues and whether limbal stem Individual surgical strategies were performed based on clinical cell transplantation (LSCT), lamellar keratoplasty (LKP), or examination and OCT images. penetrating keratoplasty (PKP) should be applied appropriately before surgery.4–6 Inappropriate surgical procedures may Results: Three types of conjunctivalized corneas were evaluated, increase the risk of transplant rejection, aggrevate visual including conjunctival pannus (4 of 25), pseudopterygium (10 of impairment, and cause wasting of donor corneas. Knowledge 25), and a white fibrovascular membrane (11 of 25). All 25 patients on whether the corneal stroma is damaged and the involved received a procedure of allograft limbal stem cell transplantation. In depth is especially useful in the determination of specific addition, with information provided in OCT images, 8 patients had surgical strategies. combined lamellar keratoplasties; 3 patients had deep anterior The development of optical coherence tomography lamellar keratoplasties, and 2 patients received penetrating kerato- (OCT) technology makes it possible to define the depth of plasties. The remaining 12 patients received limbal stem cell corneal pathologic abnormalities in the presence of corneal transplantation alone. All fibrovascular tissues were successfully opacity, such as dystrophy or scarring.7–10 Considering that removed from the cornea in all patients. Fourier-domain OCT has weaker penetration, a commercial Conclusions: time-domain OCT system, Visante OCT (Carl Zeiss Meditec, OCT is a valuable method in the evaluation of Inc, Dublin, CA) was used to determine the depth of conjunctivalized corneas. This is helpful in determining the surgical conjunctival tissue or corneal opacity in burnt corneas in treatments for individual patients, allowing for less corneal graft formulating specific surgical management strategies for rejection and making good use of corneal donors. individual patients. Key Words: conjunctivalization of corneas, optical coherence tomography, lamellar keratoplasty and penetrating keratoplasty – PATIENTS AND METHODS (Cornea 2015;34:1233 1239) This study followed the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Zhongshan Ophthalmic Center. Informed consent was Received for publication March 3, 2015; revision received May 10, 2015; obtained from all patients. accepted May 14, 2015. Published online ahead of print July 6, 2015. From the *Cornea Department, State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China; and †Cosmetic and Plastic Surgery Department, Guangdong Study Population Women and Children Hospital, Guangzhou, China. Thirty-two patients, from July 2012 to June 2013, with “ fi ” Supported in part by Guangdong Scienti c Program 2012 in China (Grant burn-induced conjunctivalized corneas consulted our institute number: 2012B010300010) and the Fundamental Research Funds of the State Key Laboratory of Ophthalmology (2014QN10). The funding and asked for surgeries. After preliminary examination and organization had no role in the design or conduct of this research. screening, 25 eyes (25 patients) were included in this The authors have no conflicts of interest to disclose. prospective study. The inclusion criteria for the patients were Reprints: Shiyou Zhou, MD, PhD, State Key Laboratory of Ophthalmology, (1) a history of stable ocular burns after between 10 and 18 Zhongshan Ophthalmic Center, Sun Yat-sen University, 54 South Xianlie Rd, Guangzhou, 510060 Guangdong, China (e-mail: zhoushiy@mail. months, and (2) ocular burns of grades III and IV according to sysu.edu.cn). the Roper-Hall classification. Patients who had previous Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. corneal surgeries, recurrent corneal conjunctivalization, and

Cornea Volume 34, Number 10, October 2015 www.corneajrnl.com | 1233

a Schirmer II test score result less than 3 mm were excluded. for more than two thirds of the corneal thickness. While All possible comprehensive ophthalmic examinations were performing LKP or DALK, the impaired corneal tissue was performed on each participant, including visual acuity, dissected and peeled off layer by layer according to the depth intraocular pressure, slit-lamp biomicroscopy, Schirmer II information provided by the OCT images until the recipient test, and fundus examination. Visante OCT (Carl Zeiss bed was transparent. Fresh corneal donors were prepared Meditec Inc) and digital corneal photography were performed according to the shape of the recipient beds with a limbus of before and after surgeries. All patients received surgical 2-mm width. A procedure of PKP was performed if full treatments from the same surgeon (S.Z.). thickness corneal stroma was impaired. All corneal graft buttons were made from donor corneas acquired within 12 hours after their death (obtained from eyes of deceased donors OCT Examination and provided by Guangdong Provincial Eye Bank, China). During examination, a 360-degree scan was initially performed, and then the meridian presenting the deepest opacity point of the cornea was manually selected for RESULTS analysis. From the images, the specific corneal layer where Twenty-five eyes of 25 patients (age range, 4–43 years) the conjunctivalized tissue had reached was first determined. were evaluated in this study, including 4 eyes with thin With the caliper provided by the software, the thickness of the conjunctival pannus (Fig. 1A), 10 eyes with pseudopterygium superficial conjunctivalized tissue was measured in the (Fig. 2A), and 11 eyes with a thick full-circumference corneal center or 2.5 to 4.0 mm from the center (depending fibrovascular membrane (Fig. 4A). Cross-sectional OCT on the point of deepest corneal impairment in different images of corneas were acquired in 24 of 25 eyes. In 1 eye patients). The underlying cornea with normal light reflectivity with thick pseudopterygium, a shadow was present under was assumed to be healthy and measured whenever possible. the hyperreflectivity, with no view of the endothelium. For this patient, ultrasound biomicroscopy (UBM) was performed and cross-sectional images were acquired. All 25 Surgical Treatments patients received LSCT. In addition, with information pro- Surgical treatments were performed according to the vided in OCT images, 8 patients had combined LKP; 3 patients clinical examinations and OCT information. Conjunctival had DALK, and 2 patients received PKP. The remaining 12 peritomy and allograft LSCT was performed on each patient. patients received LSCT alone. Representative cases are Donor corneoscleral rims having stem cells were prepared illustrated below. alone or in combination with corneal donor buttons. Patients received LSCT alone if the corneal stroma was not damaged or the involved anterior stroma was less than one third of the Case 1 corneal thickness, based on the rules of phototherapeutic A 35-year-old man had an acid burn in his right eye 1 keratectomy (PTK).11 LKP and deep lamellar keratoplasty year before he presented to our institute. The best-corrected (DALK) were applied when the corneal stroma was damaged visual acuity was 20/40. Clinical examination disclosed

FIGURE 1. Anterior segment pho- tographs (A and C) and OCT images of case 1 (B and D). A, Preoperative photograph showing thin conjunctival pannus invading the cornea from the 2.30 to 8.30 position. B, Preoperative OCT image showing a slim band of hyperreﬂectivity limited between the epithelium and anterior stroma. C, Postoperative photograph 1 week after the procedure of LSCT. D, Postoperative OCT image, 1 week after the procedure of LSCT.

FIGURE 2. Anterior segment pho- tographs (A and C) and OCT images of case 2 (B and D). A, Preoperative photograph showing conjunctival pseudopterygium covering more than half the area of the cornea, with no view of deeper corneal structures. B, Preoperative OCT image showing hyperreﬂectivity on the surface, with a thickness of 440 mm of the underlying healthy corneal tissue in the center. C, Postoperative photograph 1 month after the procedure of LSCT. D, Postoperative OCT image 1 month after the procedure of LSCT.

conjunctival pannus invading the cornea from the 2.30 to 8.30 (Fig. 1C). Corneal transplantation was considered to be not position, with no involvement of the pupil zone. With the slit- applicable for this patient. Postoperative OCT performed 1 lamp microscope, we postulated that the pannus was deep to the week later revealed a homogeneous corneal stromal reflectivity Bowman membrane (Fig. 1A). Preoperative OCT confirmed and well-apposed limbal tissue (Fig. 1D). the postulation as a slim band of hyperreflectivity noted between the epithelium and the anterior stroma, with a thickness of 181 mm at a distance of 3.5 mm from the center. A clear demarcation Case 2 between the pannus and the underlying corneal stroma was well A 30-year-old man had a history of alkali burn in his delineated in the OCT images. The residual corneal thickness left eye, and the best-corrected visual acuity was 20/200. was approximately 502 mm (Fig. 1B). Considering no corneal Clinical examination showed that the conjunctival pseudop- stroma to be involved, we performed a conjunctival pannus terygium covered more than half of the area of the cornea and excision combined with allograft LSCT on this patient pupil with severe symblepharon on the inferior cornea. The

FIGURE 3. Anterior segment pho- tographs (A and C) and OCT images of case 3 (B and D). A, Preoperative photograph showing conjunctival pseudopterygium covering half the area of the cornea, with similar clinical impression with case 2. B, Preoperative OCT image showing the hyperreﬂectivity in deep corneal stroma, with a thickness of 208 mm of the residual healthy corneal bed in the pericenter. C, Postoperative photograph 1 month after partial LKP. D, Postoperative OCT image 1 month after a procedure of partial LKP, revealing a fully integrated graft with the recipient bed.

FIGURE 4. Anterior segment pho- tographs (A and C) and OCT images of case 4 (B and D). A, Preoperative photograph showing a white ﬁbro- vascular membrane covering the entire cornea and with no view of the corneal stroma. B, Preoperative OCT image showing a thickness of 410 mm of the abnormal hyper- reﬂectivity in the center. C, Post- operative photograph 2 weeks after a procedure of full LKP. D, Post- operative OCT image 2 weeks after a procedure of LKP, revealing a fully integrated graft to the bed.

visualization of the deeper stroma and anterior chamber was repeated and it revealed a recipient bed with a fully structures was limited by pseudopterygium (Fig. 2A). From integrated graft. The reflectivity of the graft and the bed was the Visante OCT images, hyperreflectivity, corresponding to homogenous (Fig. 3D). the thick fibrovascular tissue, could be well visualized. With the measurement tool, it was found that the relatively healthy corneal stroma underlying pseudopterygium was Case 4 440 mm in the center and 525 mm peripherally (Fig. 2B). A 20-year-old man had thermal burns in his right eye This thickness accounted for more than four-fifth of the full 15 months ago, with a visual acuity of hand motions at 50 cm. corneal thickness, thus illustrating a nonrequirement of Slit-lamp biomicroscopy showed that total limbal stem cell corneal transplantation in this case. After symblepharon lysis, deficiency occurred and the cornea was covered with a dense the patient underwent a pseudopterygium excision and LSCT, fibrovascular membrane (Fig. 4A). Preoperative OCT showed as described in case 1. One month after surgery, the visual that the fibrovascular membrane, represented by the abnormal acuity increased to 30/60, and clinical examination showed hyperreflectivity, was approximately 410 mm thick in the a clear cornea (Fig. 2C). Postoperative OCT revealed center. A relatively healthy corneal stromal bed was identified complete apposition of the donor limbus to the recipient in OCT images with a thickness of 128 mm, suggesting stroma (Fig. 2D). a healthy Descemet membrane and endothelium (Fig. 4B). As thus, we proposed a procedure of LKP with a full limbus for this patient. The OCT was repeated 2 weeks later, and it Case 3 revealed a fully integrated graft with restoration of corneal A 22-year-old man with dense conjunctival pseudoptery- stroma, having only a moderate change in optical density gium in his right eye had an alkali burn 1 year ago and had a identifying a perfectly aligned graft–host junction character- visual acuity of 20/100. Similar to case 2, pseudopterygium istic of this form of surgery. The stroma was homogenous involved the corneal area from the 2.30 to 9-o’clock position, with reflectivity throughout. Clinical examinations showed and the underlying corneal stroma could not be viewed that the cornea was transparent and the postoperative visual (Fig. 3A). Preoperative OCT showed that the corneal scar, acuity had improved to 20/100 (Fig. 4C, D). represented by the hyperreflectivity was 516 mm deep in the pericenter, accounting for three-fourth thickness of the full stroma (Fig. 3B). Thereby, a procedure of “D-shape” partial Case 5 LKP was designed for this patient. After a 180-degree A 33-year-old woman had a history of ocular thermal conjunctival peritomy and removal of the pathologic tissue, burn with a visual acuity of hand motions at 30 cm. The the diseased corneal stroma was dissected and peeled away clinical impression of this patient was the same as that of case 4 according to the depth information provided by the OCT (Fig. 5A). Preoperative OCT depicted that the stromal hyper- images. A fresh donor cornea was prepared according to the reflectivity was heterogeneous and close to Descemet mem- shape of the recipient bed. One month later, In clinical brane. The deepest point was 427 mm in the center. The examinations, the cornea was transparent and the best-corrected contour of the corneal posterior surface was smooth and visual acuity had increased to 20/60 (Fig. 3C). Visante OCT regular (Fig. 5B). We considered the endothelium to be

FIGURE 5. Anterior segment pho- tographs (A and C) and OCT images of case 5 (B and D). A, Preoperative photograph showing similar clinical manifestation with case 4, with a white ﬁbrovascular membrane and no view of the corneal stroma. B, Preoperative OCT image showing inhomogeneous stromal hyper- reﬂectivity and smooth and regular corneal posterior surface. C, Post- operative photograph of case 5, 2 weeks after a procedure of deep anterior LKP. D, Postoperative OCT image, 2 weeks later, revealing complete graft apposition to Descemet membrane and the endothelium.

functional, and a procedure of LKP or DALK to be applied for best-corrected visual acuity after the surgery improved to this patient. During the operation, the opaque corneal tissue 20/60. was peeled off layer by layer until a small amount of stroma was left. The posterior stroma was still cloudy. After a small exposure of Descemet membrane, it was found transparent, and Case 6 the patient eventually received a procedure of DALK. Post- A 38-year-old man had a history of liquid iron burns in operative OCT showed complete graft apposition to Descemet his right eye 14 months ago. He had a visual acuity of hand membrane, and no double anterior chamber was formed. The motions at 30 cm and similar clinical impression as that of endothelium was identified by a smooth hyperreflective line on the previous 2 patients (Fig. 6A). The OCT image showed the posterior surface. Slit-lamp examination showed hyperreflectivity of full corneal layers homogeneously in the a well-apposed graft and a clear cornea (Fig. 5C, D). The center. The contour of the endothelium was irregular, with

FIGURE 6. Anterior segment pho- tographs (A and C) and OCT images of case 6 (B and D). A, Preoperative photograph showing similar clinical manifestation with cases 4 and 5. B, Preoperative OCT image showing homogeneous hyperreﬂectivity of full corneal stroma and irregular endothelial surface with some scat- tering dots adhered. C, Post- operative photograph 2 weeks after a procedure of PKP. D, Postoperative OCT image 2 weeks after a procedure of PKP.

some scattered dots adhered to the endothelial surface (Fig. indicating impairment of the endothelium. We performed PKP 6B). With this information, a procedure of PKP was for this patient. To confirm the endothelium condition we had considered for this patient. Postoperative clinical examina- presumed, the recipient corneal button was stained with trypan tion showed a transparent corneal graft, and the OCT images blue and alizarin red, a traditional method to detect the activity exhibited homogeneous reflectivity (Figs. 6C, D). The visual of corneal endothelial cells. We found that the nuclei were acuity of this patient increased to 20/100. stained by trypan blue, indicating that the cell membrane was impaired or the cells were dead. One limitation of time-domain OCT is that the penetration DISCUSSION is not yet deep enough in some cases. In 1 eye of our study, the In this study, we report the value of OCT in imaging hyperreflectivity corresponding to conjunctival pseudopterygium burn-induced conjunctivalized corneas. The cross-sectional caused signal shadowing and the view of the underlying stroma images afforded by OCT provided useful information for was not acquired. UBM was then performed on this patient and making individual surgical strategies and provided guidelines revealed a healthy deeper stroma. This patient eventually for surgical treatments. With an axial resolution of 5 to 18 mm, received partial LKP, as in case 3. Even so, UBM is not the the cross-sectional images of OCT can differentiate different optimal method for our study because of its lower resolution, corneal layers.12 OCT has been widely used in various corneal requirement of coupling fluid immersion of the examined eye, surgeries, including LASIK, PTK, PKP, DALK, and endothelial and inconvenient operation methods.21 For corneas with thick keratoplasty.13,14 It can evaluate the thickness of the anterior pseudopterygium, a more suitable imaging technique should be opacity in diseased corneas, such as corneal scarring, trauma, in discussion for whole cross-sectional images besides UBM. and stromal dystrophies.8,10,15–17 Most of these studies used In conclusion, we have reported the application of OCT Fourier-domain OCT. However, even with high resolution and in imaging the structures of conjunctivalized corneas due to fast scan speed, images generated from Fourier-domain OCT are severe ocular burns. This may provide clinicians with useful somewhat limited by severe corneal opacities or conjunctival information for the determination of surgical treatment pseudopterygium. Time-domain OCT is slower but its penetra- strategies for individual patients. tion is much stronger than Fourier-domain OCT because of its longer wavelength (1310 vs. 830 nm).18 We had previously demonstrated that Visante OCT (time-domain OCT) had good correlation and agreement with the traditional UBM system in REFERENCES imaging corneas with dense scars.9 In this study, we found it 1. Spector J, Fernandez WG. Chemical, thermal, and biological ocular exposures. Emerg Med Clin North Am. 2008;26:125–136. especially useful in making appropriate clinical decisions for 2. Saini JS, Sharma A. Ocular chemical burns–clinical and demographic surgical intervention, specifically, in differentiating the condi- profile. Burns. 1993;19:67–69. tions like in cases 2 and 3. Similar clinical impression of these 2 3. Hong J, Qiu T, Wei A, et al. Clinical characteristics and visual outcome patients interfered with the surgeon’s choice of specific of severe ocular chemical injuries in Shanghai. Ophthalmology. 2010; – surgeries. Under the guidance of OCT, a procedure of LKP 117:2268 2272. 4. Tseng SC. Concept and application of limbal stem cells. Eye (Lond). was accurately applied in case 3. Thus, donor corneas could be 1989;3:141–157. accurately prepared before surgeries. 5. Tan DT, Dart JK, Holland EJ, et al. Corneal transplantation. Lancet. As a method of corneal transplantation, PKP has 2012;379:1749–1761. traditionally been regarded as the gold standard in the manage- 6. Wiley LA, Joseph MA, Pemberton JD. Microkeratome-assisted anterior lamellar keratoplasty. Arch Ophthalmol. 2008;126:404–408. ment of corneal scars or opacities. In these conditions, the 7. Majander AS, Lindahl PM, Vasara LK, et al. Anterior segment optical assessment of endothelium function is often overlooked. coherence tomography in congenital corneal opacities. Ophthalmology. Replacement of potential healthy endothelium may increase 2012;119:2450–2457. risks of graft rejection. This is essential in burn-induced corneal 8. Hong JP, Kim TI, Chung JL, et al. Analysis of deposit depth and injuries that face high risks of graft rejection in vascularized morphology in granular corneal dystrophy type 2 using fourier domain optical coherence tomography. Cornea. 2011;30:729–738. corneas. The application of OCT is especially useful in these 9. Zhou SY, Wang CX, Cai XY, et al. Optical coherence tomography and conditions. Although it cannot provide cellular level resolution ultrasound biomicroscopy imaging of opaque corneas. Cornea. 2013;32: like confocal microscopy, it can provide useful information e25–e30. suggesting the condition of the endothelium. Imaging of the 10. Ma JJ, Tseng SS, Yarascavitch BA. Anterior segment optical coherence tomography for transepithelial phototherapeutic keratectomy in central deep stroma and posterior surface underlying the dense corneal stromal scarring. Cornea. 2009;28:927–929. fibrovascular membrane aids in the determination of general 11. Stasi K, Chuck RS. Update on phototherapeutic keratectomy. Curr Opin suitability for LKP and DALK surgeries, as in cases 4 and 5 in Ophthalmol. 2009;20:272–275. our study. One advantage of this application is that the corneal 12. Huang D, Li Y, Radhakrishnan S. Optical coherence tomography of the – donors, especially those preserved for long periods of time or anterior segment of the eye. Ophthalmol Clin North Am. 2004;17:1 6. 13. Nesi TT, Leite DA, Rocha FM, et al. Indications of optical coherence have unsatisfactory endothelium, can be used effectively and tomography in keratoplasties: literature review. J Ophthalmol. 2012; 19,20 appropriately. It has been recently reported that cryopre- 2012:989063. served corneal tissue has a comparable surgical result with fresh 14. Ventura BV, Moraes HV Jr, Kara-Junior N, et al. Role of optical corneal tissue in the procedure of DALK.19,20 This is especially coherence tomography on corneal surface laser ablation. J Ophthalmol. fi 2012;2012:676740. signi cant in countries with a shortage of donor corneas. For 15. Kim TI, Hong JP, Ha BJ, et al. Determination of treatment strategies for case 6, the OCT images demonstrated full thickness hyper- granular corneal dystrophy type 2 using Fourier-domain optical coher- reflectivity of corneal stroma and nonsmooth posterior surface, ence tomography. Br J Ophthalmol. 2010;94:341–345.

16. Wylegala E, Dobrowolski D, Nowinska A, et al. Anterior segment 19. Chen W, Lin Y, Zhang X, et al. Comparison of fresh corneal tissue optical coherence tomography in eye injuries. Graefes Arch Clin Exp versus glycerin-cryopreserved corneal tissue in deep anterior lamellar Ophthalmol. 2009;247:451–455. keratoplasty. Invest Ophthalmol Vis Sci. 2010;51:775–781. 17. Mori H, Miura M, Iwasaki T, et al. Three-dimensional optical coherence 20. Javadi MA, Feizi S, Javadi F, et al. Deep anterior lamellar keratoplasty tomography-guided phototherapeutic keratectomy for granular corneal using fresh versus cryopreserved corneas. Ophthalmology. 2014;121: dystrophy. Cornea. 2009;28:944–947. 610–611. 18. Ramos JL, Li Y, Huang D. Clinical and research applications of anterior 21. Dada T, Sihota R, Gadia R, et al. Comparison of anterior segment optical segment optical coherence tomography—a review. Clin Experiment coherence tomography and ultrasound biomicroscopy for assessment of Ophthalmol. 2009;37:81–89. the anterior segment. J Cataract Refract Surg. 2007;33:837–840.