Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882 DOI 10.1007/s00417-006-0476-2

SHORT COMMUNICATION

A pilot study on slit lamp-adapted optical coherence tomography imaging of trabeculectomy filtering blebs

Thomas Theelen & Pieter Wesseling & Jan E. E. Keunen & B. Jeroen Klevering

Received: 14 June 2006 /Revised: 8 October 2006 /Accepted: 9 October 2006 / Published online: 22 November 2006 # Springer-Verlag 2006

Abstract Results The study included 20 of 20 patients. After Background Our study aims to identify anatomical charac- completion of our study, 15 eyes of 15 patients (mean teristics of filtering blebs by means of slit lamp- age±SD 67±16 years) were eligible for data analysis and 5 adapted optical coherence tomography (SL-OCT) and to eyes missed at least one follow-up visit. Filtering surgery identify new parameters for the functional prognosis of the was considered successful ( ≤ 21 mmHg filter in the early post-operative period. without antiglaucomatous medication) in 11 of 15 eyes. SL- Methods Patients with primary open-angle glaucoma, aged OCT frequently demonstrated multiple hypo-reflective 18 years and older, scheduled for primary trabeculectomy at layers within Tenon’s capsule (“striping” phenomenon) in the Department of Ophthalmology, Radboud University the first post-operative week. Presumably, these layers Nijmegen Medical Centre, were considered for our study. corresponded with drainage channels in the histological All patients underwent standardized trabeculectomy with specimen. These channels were present in functional filters intra-operative application of . The filtering but not in the failures. In addition, the visualisation of the blebs were evaluated clinically and with SL-OCT on day 1 below the filtering zone was better defined in failures and 1, 2, 4 and 12 weeks following surgery. The resulting compared with successful filtering blebs (“shading” phe- data were analysed and weighed against surgical success. nomenon). We observed no differences in the volume and To better understand the SL-OCT data a small comparative clinical aspect of the blebs in the successful group histologic study was performed. compared with the unsuccessful group. Conclusions Successful filtering blebs show characteristic optical properties on SL-OCT. These phenomena suggest a diffusely enhanced fluid content and the presence of intra- Meeting presentation Presented in part as a poster at the 1st World Glaucoma Congress, Vienna, Austria, July 2005. bleb drainage channels in functional filtering blebs.

Financial disclosure Medical Workshop, Groningen, Netherlands, . made the slit lamp-adapted optical coherence tomography device Keywords Trabeculectomy Optical coherence available to one of the authors (T. Theelen) with support from the tomography. Prospective study manufacturer, Heidelberg Engineering, Lübeck, Germany. There are no other competing interests. T. Theelen (*) : J. E. E. Keunen : B. J. Klevering Department of Ophthalmology, Introduction Radboud University Nijmegen Medical Centre, Philips van Leydenlaan 15, 6525 Nijmegen, The Netherlands The surgical success of a trabeculectomy strongly depends e-mail: [email protected] on the extent of the healing processes at the filtering site. Therefore, careful clinical examination of the filtering bleb P. Wesseling is necessary in the early post-operative period and attempts Department of Pathology, Radboud University Nijmegen Medical Centre, have been made to standardize clinical signs and to develop Nijmegen, The Netherlands risk scores [2, 6]. However, only the surface and superficial 878 Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882 layers of the bleb may effectively be examined by slit lamp biomicroscopy. Slit lamp-adapted optical coherence tomography (SL- OCT) is a novel, high-resolution imaging technique that has recently been introduced for assessment of the anterior segment [3]. The non-contact approach of SL-OCT makes it possible to monitor the anterior segment structures non- invasively, which particularly suits the early post-operative evaluation following intraocular surgery. In this study we employed the anterior segment SL-OCT in the early post-operative evaluation of filtering blebs. With this technique we attempted to identify new criteria to recognize failing filters at a very early stage.

Materials and methods

We performed a prospective, observational study on patients aged 18 years and older consecutively scheduled for primary trabeculectomy for primary open-angle glauco- ma (POAG) at the Department of Ophthalmology, Radboud University Nijmegen Medical Centre. This study was performed in accordance with the tenets of the Declaration of Helsinki and Ethics Committee approval was obtained. Fig. 1 Standardized bleb photographs to interpret the clinical aspects of the bleb in a masked fashion. An overview of the morphological The inclusion criterion was medically uncontrolled situation is acquired with diffuse illumination and ×16 magnification POAG necessitating filtering surgery. We excluded patients (a). A high magnification image (×40) allows the assessment of with other than POAG, previous glaucoma microcyst formation (arrowheads) within the epithelial layer of the surgery or any other type of ocular surgery that might have bleb (b). Direction of the vertical and horizontal SL-OCT scans (c) affected the conjunctival integrity, chronic or recurrent affections as well as scarring of the conjunctiva, poor patient cooperation and inability to complete all follow-up visits. study instead of direct clinical evaluation, functional tests All patients underwent a standardized trabeculectomy on bleb appearance were not applicable. procedure [9] performed by one surgeon (TT). However, no In our present study a 1,310 nm laser diode with a adjustable or releasable sutures were used. During surgery, coherence length of 20 μm was used. The same examiner a 2×2 mm sponge soaked with 0.2 mg/ml of mitomycin C (TT) performed two SL-OCT scans of 12.5 mm lateral (MMC) was applied to the determined filtration site for 2 width on each study . The first scan was directed min. After MMC application the area was irrigated with vertical and the second horizontal to the corneoscleral 60 ml of physiologic saline [Balanced Salt Solution (BSS), limbus. The scanning depth in air was 6.5 mm and the Alcon Laboratories, Fort Worth, TX, USA]. resolution in ocular tissue was approximately 10 μm All patients underwent a standard ophthalmologic exam- longitudinally and 16 μm laterally. All scans were directed ination including best-corrected visual acuity, slit lamp through the most prominent parts of the filtering blebs. examination, Goldmann applanation tonometry and binoc- Time delay and intensity of the reflected infrared light were ular fundus examination on day 1 and 1, 2, 4 and 12 weeks computed into logarithmic grey-scale images to allow post-operatively. We defined surgical success as intraocular tomographic bleb analysis. To estimate the bleb volume, pressure (IOP)≤21 mmHg without additional antiglauco- the tomograms enclosed the peripheral and the matous medication and no need for bleb needling or bleb clinically non-filtering conjunctiva. On both tomograms revision independent of IOP values. We documented the made, the peak distance between the most superior bleb trabeculectomy filtering blebs by slit lamp photography reflex (presumed epithelium) and increased reflex within (Imagenet, Topcon Corporation, Tokyo, Japan) and slit the depth of the filtering zone (presumed sclera) was lamp-adapted optical coherence tomography (SL-OCT, determined and taken for further analysis. The lateral Heidelberg Engineering, Lübeck, Germany). borders of the filtering zone, e.g. the beginning of oedema, Bleb photography included an overview as well as a were determined by comparison with scans of the conjunc- detail shot (Fig. 1). As we used photographs in our present tiva away from the bleb. Assuming that the filtering zone Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882 879

Table 1 Semi-quantitative conversion scales for clinical features and SL-OCT characteristics of filtering blebs after trabeculectomiesa

Evaluation by Qualities measured Estimation of bleb qualities

01 2 3

Clinical photograph Hyperaemia None Mild Moderate Strong Corkscrew vessels None Mild Moderate Strong Epithelial microcysts No Yes –– Demarcation (Tenon’s cyst) No Yes –– SL-OCT Mean reflectivity bleb wall – Low Moderate Strong Mean reflectivity deep bleb tissue – Low Moderate Strong Superficial cyst-like hypo-reflective spaces None Single small Multiple small Single large Deep cyst-like hypo-reflective spaces None Single small Multiple small Single large Hypo-reflective layer formation within bleb None Single Multiple (“striping”) – Visibility of sclera beneath the filtering zone – Poor Moderate Good a Conversion of data was repeated for all visits of all patients, e.g. for day 1 and weeks 1, 2, 4 and 12 after surgery. This resulted in 750 data points from 15 eyes available for analysis may be approximated by a bisected ellipsoid, we calculated Windows (SPSS Inc., Chicago, IL, USA). Differences the volume of a bleb as between successful and unsuccessful cases and sex-depen- dent divergences were put through Fisher’s exact test. 4= p x y z ð1Þ 3 Pearson correlation coefficients were evaluated to analyse with x=maximal width, y=maximal length and z=maximal relations between OCT bleb measurements and IOP. We height of the filtering zone on SL-OCT. considered a p value of less than 0.05 statistically significant. We performed a comparative histological study on a 39- year-old man who presented with persistent hypotonia in the left eye following trabeculectomy with MMC applica- Results tion. Clinically, a diffuse, large, non-leaking filtering bleb with moderate hyperaemia and some epithelial microcysts Fifteen patients (9 male and 6 female; 6 right and 9 left eyes) was observed. A surgical revision was carried out with with a mean age of 67 (±16) years completed the study. No partial resection of the bleb, including the central zone that surgical revision was necessary within the study period. The had previously been investigated by SL-OCT. The speci- mean preoperative IOP was 27.6 (±6.6) mmHg. At 12 weeks, men was subsequently prepared for histologic evaluation. 11 patients fulfilled the study criteria for successful surgery We analysed qualitative and quantitative aspects of the and 4 did not, because bleb needling was necessary in those filtering blebs on SL-OCT and compared these data to our eyes to achieve adequate filtering. At the endpoint, the mean clinical findings. The clinical photographs and the SL-OCT IOP was 10.3 (±3.8) mmHg in successful cases and 18.5 scans were mixed and were interpreted by two co-workers (±5.3) mmHg in failures. Comparison of clinical aspects with (JEEK, BJK) in a masked fashion. bleb photographs revealed no differences between successful All quantitative measurements were performed using the and unsuccessful cases. custom software provided by the manufacturer of the SL- The evolution of bleb volume and IOP over time is OCT. For our data analysis, we converted all qualitative data portrayed in Fig. 2. At day 1 and week 12, respectively, into a semi-quantitative scale (Table 1). Results were mean bleb volume was 41.2 (±28.4) mm3 and 59.0 (±30.0) statistically analysed using the SPSS 12.0.1 software for mm3 in successful cases and 37.7 (±18.9) mm3 and 83.1

Fig. 2 Illustration of IOP evo- lution (left) and bleb volume development (right) within the study period. Error bars show 95% confidence interval of mean (D=day, W=week) 880 Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882

Fig. 3 The different optical layers of a filtering bleb as perceived by dark yellow) with some pseudo-cysts (3, purple), a swollen Tenon’s the anterior segment SL-OCT and their presumed histologic analogues. layer (4, green) including a large, fluid-filled, pseudo-cystic space (5, Partial native grey-scaled (a) and colour-coded (b)scansshow dark purple) and the sclera (6, blue) conjunctival epithelium (1, red), the oedematous lamina propria (2,

(±48.0) mm3 in failed cases. These differences were not Histopathological examination of the excised filtration statistically significant. bleb revealed a normal-appearing conjunctival epithelium According to our supplementary histologic study (see with loose subepithelial connective tissue. Besides some below) we interpreted the different optical strata on SL- endothelialized vessels, dispersed non-endothelialized fluid OCT to be conjunctival epithelium, stroma, Tenon’s layer channels were present within the lamina propria and and sclera (Fig. 3). Within the first post-operative week, we Tenon’s capsule. The correlation of the ex vivo histologic observed multiple parallel hypo-reflective layers inside the findings with the pre-surgical SL-OCT scan is depicted in presumed Tenon’s layer (“striping”) in 9/11 of successful Fig. 5. cases (Fig. 4a). In contrast, we could not detect this optical phenomenon in any of the unsuccessful cases. Poor visibility of the sclera beneath the bleb (“shading”) was Discussion commonly observed in successful cases (7/11) in the first post-operative week (Fig. 4b). Again, this phenomenon was The most significant bleb-related remodelling and encapsu- not present in unsuccessful cases. Both the striping and the lating processes that may cause IOP elevations occur in the shading phenomena were not incidental findings, but were first weeks following glaucoma filtering surgery [5]. As repeatedly observed in individual patients. clinical bleb evaluation is limited to superficial tissue layers,

Fig. 4 Fraction of a vertical OCT scan showing a “striping” a horizontal OCT scan demonstrating “shading” of the sclera phenomenon (arrowheads) within Tenon’s layer of a successful (arrows) beneath a well-functioning trabeculectomy filtering bleb filtering bleb, suggesting the presence of fluid channels (a). Part of (b). Bars=500 μm Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882 881

Fig. 5 Correlation between the preoperative SL-OCT scan (logarith- and sclera (S) can be identified. Non-endothelialized fluid channels mic grey-scale, left) and the post-operative histological specimen (arrowheads) and vessels (arrows) are present. Note the distinct (haematoxylin-eosin, right,anti-CD31,inset) of a filtering bleb. optical properties of the tissue layers on SL-OCT Conjunctival epithelium (E), lamina propria (L), Tenon’s layer (T) important data about bleb development may be missed. increased and diffuse water content of a successful filter Consequently, attempts have been made to assess the deeper will decrease the visibility of lower structures such as the parts of filtering blebs using ultrasound biomicroscopy sclera in the mid-infrared. (UBM) [10]. However, UBM depends on an immersion Another phenomenon associated with good functionality technique, which may complicate the examination of newly of the filtration site was the “striping” phenomenon within operated eyes and, in addition, may alter bleb configuration Tenon’s layer. According to our own histological data as due to conjunctival stretching and massage effects. well as to previously published data about bleb histology A confocal laser scanning ophthalmoscope (Heidelberg [1], we propose that these thin stripes of decreased Retina Tomograph with Rostock Cornea Module) was also reflectivity surrounded by hyper-reflective areas represent used to examine the filtering zone of trabeculectomy blebs fluid-filled channels with fine connective tissue septa. [4]. High-resolution imaging of focal bleb areas was “Striping” may thus indicate the abundant presence of possible, but this technique could not give an overview drainage channels in successful filters. over the whole filtering zone. Since the Rostock Cornea The number of patients in this pilot study is relative Module depends on a contact method, the resulting tissue small, thereby limiting the general applicability of the compression could modify the original anatomy of the bleb. results. The follow-up period is short, but one has to keep Furthermore, Stratus OCT, originally developed for in mind that most encapsulation processes occur within the retinal imaging, was applied for retrospective evaluation first weeks following trabeculectomy [5]. Arguably, suc- of filtering blebs long after surgery [7]. In contrast to our cessful post-surgical intervention has to take place in this present study, the authors investigated a series of 24 blebs early post-operative period. at a single point in time within a widespread period after Anterior segment SL-OCT allows non-invasive, high- surgery (6 months to 12 years). In contrast, we examined resolution analysis of the clinically invisible deeper regions our patients after a standardized surgical procedure follow- of the anterior segment. In this study, we demonstrate the ing a specific study protocol with several predetermined anatomic correlations of SL-OCT images of the deep layers follow-up visits. We found optical strata corresponding to of trabeculectomy filtering blebs. Differences were ob- the conjunctival epithelium, lamina propria conjunctivae, served between successful filters and failures in the early Tenon’s layer and sclera. Larger comparative studies of post-operative period. The “striping” and “shading” phe- histologic and SL-OCT bleb appearance will be necessary nomena seem to be important indicators for surgical to demonstrate the consistency of the proposed relation success. Additional clinical research on SL-OCT should between the microscopic anatomy and the in vivo optical demonstrate whether this device is helpful in recognizing properties on SL-OCT. failing filters at an early stage, thereby increasing the Decreased visibility of the sclera underlying the bleb on likelihood of successful intervention strategies. In addition, the first post-operative day (“shading”) seems to indicate such investigations should study whether this new imaging good filtering qualities. At 1,310 nm, the absorption profile method is able to give essential supplementary information of water significantly worsens tissue transparency, and light to clinical examinations by highly experienced glaucoma backscatter of deeper structures will be limited [8]. Hence, surgeons. 882 Graefe’s Arch Clin Exp Ophthalmol (2007) 245:877–882

Acknowledgements The authors thank Carina F.M. Meulendijks, tomography of the anterior segment. Graefes Arch Clin Exp MS, for statistical calculations, and B. Frits T. Hogewind, MD, for Ophthalmol 238:8–18 patient recruitment. The slit lamp-adapted optical coherence tomogra- 4. Labbe A, Dupas B, Hamard P, Baudouin C (2005) In vivo phy device was provided to one of the authors (TT) by Medical confocal microscopy study of blebs after filtering surgery. Workshop, Groningen, Netherlands, under support from the manufac- Ophthalmology 112:1979 turer, Heidelberg Engineering, Lübeck, Germany. There are no other 5. Parrish RK II, Folberg R (1996) Wound healing in glaucoma competing interests of any author. surgery. In: Ritch R, Shields MB, Krupin T (eds) The glaucomas, 2nd edn. Mosby, St. Louis, pp 1636–1637 6. Picht G, Grehn F (1998) Classification of filtering blebs in trabeculectomy: biomicroscopy and functionality. Curr Opin – References Ophthalmol 9:2 8 7. Savini G, Zanini M, Barboni P (2005) Filtering blebs imaging by optical coherence tomography. Clin Experiment Ophthalmol 1. Addicks EM, Quigley HA, Green WR, Robin AL (1983) 33:483–489 Histologic characteristics of filtering blebs in glaucomatous eyes. 8. Schmitt JM, Knuttel A, Yadlowsky M, Eckhaus MA (1994) Arch Ophthalmol 101:795–798 Optical-coherence tomography of a dense tissue: statistics of 2. Cantor LB, Mantravadi A, WuDunn D, Swamynathan K, Cortes attenuation and backscattering. Phys Med Biol 39:1705–1720 A (2003) Morphologic classification of filtering blebs after 9. Stalmans I, Gillis A, Lafaut AS, Zeyen T (2006) Safe trabeculectomy glaucoma filtration surgery: the Indiana Bleb Appearance Grading technique: long term outcome. Br J Ophthalmol 90:44–47 Scale. J Glaucoma 12:266–271 10. Yamamoto T, Sakuma T, Kitazawa Y (1995) An ultrasound 3. Hoerauf H, Wirbelauer C, Scholz C, Engelhardt R, Koch P, Laqua biomicroscopic study of filtering blebs after mitomycin C H, Birngruber R (2000) Slit-lamp-adapted optical coherence trabeculectomy. Ophthalmology 102:1770–1776