Retinal Layer Thickness Changes in Eyes With ...

5 downloads 9021 Views 1MB Size Report
built Optical Coherence Tomography Retinal Image. Analysis algorithm (OCTRIMA). ..... mentation results using our own custom-built software,. OCTRIMA, which ...
i



m

a

g

c l i n i c a l

i

n

g

s c i e n c e



Retinal Layer Thickness Changes in Eyes With Preserved Visual Acuity and Diffuse Diabetic Macular Edema on Optical Coherence Tomography Gábor Márk Somfai, MD, PhD; Erika Tátrai, MD; Mária Ferencz, MD, PhD; Carmen A. Puliafito, MD, MBA; Delia Cabrera DeBuc, PhD n BACKGROUND AND OBJECTIVE: Diabetic macular edema has several patterns on optical coherence tomography. This retrospective study aimed to assess which retinal layers show thickness changes in type 1 diffuse diabetic macular edema with preserved vision. n PATIENTS AND METHODS: Eleven eyes with diffuse diabetic macular edema on optical coherence tomography and eight control eyes with 20/20 bestcorrected visual acuity were enrolled. Optical coherence tomography images were segmented using an algorithm of the authors’ design (OCTRIMA): regional thickness data in the central area, pericentral rings, and peripheral rings were obtained for the retinal cellular layers. INTRODUCTION

Diabetes mellitus and consequent diabetic macular edema (DME) is the leading cause of legal blindness and impaired vision in the developed world. Focal

n RESULTS: The retinal nerve fiber layer showed no changes, the ganglion cell and inner plexiform composite layer and the ganglion cell complex were thicker only in the pericentral regions, and all other layers were thicker in all regions in diffuse diabetic macular edema. Macular thickness was normal in the peripheral region. n CONCLUSION: The results show that the outer retina in the foveal area is the most affected in diffuse diabetic macular edema. [Ophthalmic Surg Lasers Imaging 2010;41:593597.] macular edema is caused by foci of vascular abnormalities, primarily microaneurysms, which tend to leak fluid, whereas diffuse macular edema is caused by dilated retinal capillaries in the retina. Optical coherence tomography (OCT) is a novel

From the Department of Ophthalmology (GMS, ET, MF), Semmelweis University, Budapest, Hungary; Keck School of Medicine (CAP), University of Southern California, Los Angeles, California; and Bascom Palmer Eye Institute (DCD), University of Miami Miller School of Medicine, Miami, Florida. Originally submitted December 9, 2009. Accepted for publication June 10, 2010. Posted online August 30, 2010. Presented in part at the meeting of the European Association for Vision and Eye Research, October 4-7, 2007, Portoroz, Slovenia. Supported by a Juvenile Diabetes Research Foundation grant, an NIH center grant P30-EY014801, an unrestricted grant to the University of Miami from Research to Prevent Blindness, Inc., and the Zsigmond Diabetes Fund of the Hungarian Academy of Sciences. Dr. Puliafito receives a royalty payment from Massachusetts Eye & Ear Infirmary as a co-inventor of OCT and is a consultant to Carl Zeiss Meditec, Inc. The remaining authors have no financial or proprietary interest in the materials presented herein. Dr. Puliafito did not participate in the editorial review of this manuscript. Address correspondence to Gábor Márk Somfai, MD, PhD, Semmelweis University, Faculty of Medicine, Department of Ophthalmology, Mária str. 39., 1085 Budapest, Hungary. E-mail: [email protected] doi: 10.3928/15428877-20100830-04

Ophthalmic Surgery, Lasers & Imaging · Vol. 41, No. 6, 2010

593

i

m

a

Table 1

Inclusion Criteria for the Diffuse Diabetic Macular Edema Group • Preserved visual acuity (20/20) • Refractive error within -3.0 D and +3.0 D spherical and cylindrical correction • Lack of any other ophthalmic disease (including any other macular pathology or glaucoma) • Negative history of previous retinal laser therapy or surgery • No untreated systemic diseases • Signal strength greater than 6 on all scans • Presence of type 1 diffuse macular edema on optical coherence tomography (thickening with homogenous optical reflectivity) • Lack of any alterations at the vitreoretinal interface on all scans D = diopters.

imaging method enabling high-resolution cross-sectional visualization of retinal morphology. OCT not only facilitates a better description of pathological features in the retina, but also helps to monitor the progression of ocular diseases. Based on the classification of Otani et al.,1 four patterns of DME have previously been described on OCT by Kang et al.2: type 1, thickening with homogenous optical reflectivity; type 2, thickening with markedly decreased optical reflectivity in the outer retinal layer; type 3A, foveolar detachment without traction; and type 3B, foveolar detachment with apparent vitreofoveal traction.1,2 It can be presumed that type 1 diabetic diffuse macular edema (dDME) in eyes with preserved visual acuity is the preceding stage leading to more severe edema. Our aim was to assess which retinal layers show thickness changes at an early stage of dDME using our custombuilt Optical Coherence Tomography Retinal Image Analysis algorithm (OCTRIMA).3

g

i

n

g

OCT were included in our study. The performance of fluorescein angiography was not among the inclusion criteria (Table 1). Healthy control subjects underwent standard ophthalmic examination followed by OCT examination in one randomly chosen eye after informed consent was obtained. Macular mapping was performed with a Stratus OCT device using the standard Macular Map Protocol (software version 4.0.1; Carl Zeiss Meditec Inc., Dublin, CA). Images of a healthy subject’s fundus image, horizontal OCT scan, and macular map in comparison with those obtained in dDME can be seen on the figure. Fundus photography and fluorescein angiography images were graded according to the grading scale proposed by the American Academy of Ophthalmology.4 All OCT images were exported and segmented using OCTRIMA. We obtained average thickness data for the following cellular layers: the retinal nerve fiber layer (RNFL), the ganglion cell and inner plexiform composite layer (GCL+IPL), the ganglion cell complex (GCC, comprising the RNFL+GCL+IPL), the inner nuclear layer (INL), the outer plexiform layer (OPL), and the outer nuclear layer (ONL), along with the total retinal thickness (annotated as “Macula”). Thickness results were calculated for the nine Early Treatment Diabetic Retinopathy Study regions and the central area, pericentral rings, and peripheral rings (regions) were calculated separately with a radius of 1, 1 to 3, and 3 to 6 mm, respectively. Because there are no ganglion cells and only the ONL can be continuously detected in the central region, we only assessed the thickness of this layer and the total retina. Statistical comparison of the results was performed by pairwise Mann–Whitney U test, and the level of significance was set at a P value of less than .05. For statistical analysis, Statistica 8.0 software (Statsoft Inc., Tulsa, OK) was used. RESULTS

PATIENTS AND METHODS

A retrospective chart review was conducted among patients with diabetes mellitus undergoing fundus photography and OCT examination at the Diabetes Outpatient Clinic of the Department of Ophthalmology of Semmelweis University between January 1, 2006, and January 4, 2007. Patients with type 1 dDME on

594



A total of 11 eyes of 8 patients with dDME were selected (2 men and 6 women) and 8 age-matched healthy controls (3 men and 5 women) were recruited. Median age was 54 years (range: 44 to 69 years) and 56 years (range: 38 to 72 years), respectively, which showed no statistically significant difference. All eyes in the dDME group had moderate non-

Copyright © SLACK Incorporated

i

m

a

g

i

n

g

Figure. (A) Color fundus image, (B) horizontal optical coherence tomography (OCT) scan, and (C) thickness map of a healthy control eye and (D) color fundus image, (E) horizontal OCT scan, and (F) thickness map of an eye with preserved visual acuity and diffuse diabetic macular edema on OCT. Note that there are only a few microaneurysms and a few hard exudates in the macula in the diabetic eye, with no visible alterations in macular structure on scan E.

DISCUSSION

proliferative diabetic retinopathy except for one eye that had mild nonproliferative retinopathy. Nine eyes had no maculopathy and mild maculopathy was present in only two eyes. The ganglion cells showed a significant thickening only in the pericentral region, but without the involvement of the RNFL. In the peripheral region, the GCC layer showed no statistically significant thickness difference in the two groups. All other layers were significantly thicker in the dDME group, but total retinal thickness showed no statistically significant changes in the peripheral ring. The ONL was thicker in all regions (Table 2).

Ophthalmic Surgery, Lasers & Imaging · Vol. 41, No. 6, 2010

There are only few data about the prevalence of type 1 dDME. In the study by Otani et al.,1 it was present in 42% (25 of 59 eyes), but their study does not mention the level of diabetic retinopathy in the eyes with dDME. In the study by Kang et al.,2 a total of 145 eyes were involved, with 19 eyes showing type 1 dDME in the group of eyes with mild to moderate diabetic retinopathy (13% of the study population). The overall prevalence of type 1 dDME in that study was 55%.2 Unfortunately, visual acuities of the first two mentioned subgroups were not specified by the two studies, but we assume that this type of edema

595

596



NL

Mean ± SD

NL

Mean ± SD

NL

NL

Median [IQR]

Mean ± SD

112 ± 12

Mean ± SD

225 ± 15

Mean ± SD

278 ± 19

279 [258–289]

.000

.008

309 ± 6

309 [307–314]

82 ± 6

82 [78–85]

333 ± 21

337 [328–346]

93 ± 12

96 [82–102]

40 ± 2

40 [38–41]

.017

.048

.010

270 ± 11

273 [259–279]

57 ± 3

57 [56–59]

29 ± 2

28 [28–30]

28 ± 2

29 [27–30]

103 ± 5

102 [100–104]

52 ± 2

52 [50–53]

51 ± 4

51 [48–53]

Control

278 ± 16

272 [268–291]

63 ± 5

62 [60–65]

31 ± 1

31 [29–32]

30 ± 2

31 [29–32]

101 ± 8

103 [96–105]

53 ± 4

52 [50–57]

48 ± 6

48 [43–54]

dDME

Peripheral Region

.457

.002

.017

.026

.934

.563

.364

P

RNFL = retinal nerve fiber layer; IQR = interquartile range; NL = no layer; SD = standard deviation; GCL+ IPL = ganglion cell and inner plexiform layer; GCC = ganglion cell complex; INL = inner nuclear layer; OPL = outer plexiform layer; ONL = outer nuclear layer. a “Macula” is standing for the median and mean values of the total retinal thickness of the Early Treatment Diabetic Retinopathy Study regions involved. b P values in bold indicate significant differences among groups by the Mann-Whitney U test.

223 [214–231]

Median [IQR]

134 ± 19

140 [119–154]

36 ± 3

37 [34–39]

.002

.026

.032

.083

P

n

Macula

112 [105–115]



41 ± 2

40 [39–43]

125 ± 7

126 [119–129]

92 ± 4

90 [89–95]

33 ± 3

34 [30–36]

dDME

Pericentral Region

i

Median [IQR]

NL NL

36 ± 2

37 [34–39]

119 ± 4

118 [116–121]

87 ± 3

87 [86–90]

31 ± 2

31 [29–32]

Control

g

ONL

NL

NL

Mean ± SD









P

a

Median [IQR]

NL

NL

NL

NL

NL

NL

NL

NL

dDME

Central Region

m

OPL

INL

NL

NL

Median [IQR]

Mean ± SD

GCC

NL

Median [IQR]

GCL+IPL

NL

Control

Median [IQR]

RNFL

Cell Layer

Table 2

Thickness Results of the Control and Diffuse Diabetic Macular Edema (dDME) Groupsa,b

i g

Copyright © SLACK Incorporated

i

m

a

g

with no visual loss must correspond to an early stage of retinal alterations. It can be indirectly speculated from the study by Kang et al. that type 1 dDME in eyes with mild to moderate diabetic retinopathy with a visual acuity of 20/20 should be relatively rare.2 Indeed, our retrospective chart review of 15 months revealed only 11 eligible patients in whom both OCT and fundus photography or fluorescein angiography have been performed. It should also be of note that OCT is not necessarily performed routinely in eyes with preserved visual acuity, mild signs of retinopathy, and no signs of maculopathy because of the lack of suspected macular changes. An important feature of type 1 dDME on OCT in such an early stage is that there can be no obvious alterations in macular structure on single OCT scans because the foveal depression is preserved with normal-appearing thickness and thus the subtle changes can only be recognized by the macular thickness map. Funduscopy does not necessarily reveal dDME because the foveal reflex may be present. Although there are no longitudinal OCT studies available in patients with diabetes mellitus investigating the course of dDME with preserved vision, we believe it might be the first stage of macular edema formation; therefore, we aimed to locate the retinal layers playing a part in the pathological process of edema generation in diabetes mellitus. In our study, an increased thickness was observed for all segmented cellular layers except the RNFL and the peripheral ganglion cell bodies (GCL+IPL). Although the reason for this remains unclear, it is of interest that the RNFL was previously reported to be decreased in peripapillary measurements in diabetes mellitus.5 The most pronounced thickening was observed in the ONL, including the rod and cone granules along with the external limiting membrane on Stratus OCT images.6 A possible explanation for the observed changes could be the different blood supply of the inner and outer retinal layers or a more severe metabolic involvement of the photoreceptors in this early stage of edema formation. The regional total retinal thickness analyses showed that although seemingly only the central and pericentral areas are mostly involved in the thickening of the retina in dDME, there are still marked changes within the peripheral ring that can be observed affecting the cellular layers. These changes are possibly counteracting each other and thus retinal thickness seemingly remains stable in the peripheral macula.

Ophthalmic Surgery, Lasers & Imaging · Vol. 41, No. 6, 2010

i

n

g

Recently, the Diabetic Retinopathy Clinical Research Network reviewed the discrepancy between the various definitions used for diffuse macular edema based on four examination methods (funduscopy, color fundus photographs, fluorescein angiography, and OCT).7 In our study, we used an OCT-based classification previously described by Kang et al.2 because we aimed to assess the retinal layers that are specifically responsible for thickening at this presumably early stage of more severe macular edema formation. However, it is possible that there will be a more detailed, complex definition for diffuse macular edema in the future that will enable a more thorough patient selection for the understanding of the pathophysiology of diabetic macular complications. The segmentation of OCT scans is an important tool for the clinician in understanding the pathological processes within the retina, as in the example above. We have previously shown the high reproducibility of segmentation results using our own custom-built software, OCTRIMA, which underlines the reliability of this technique.8 We believe that the methodology presented could permit both better detection and follow-up of layer injury, as well as understanding of diabetic retinal changes. However, a further prospective study is needed with a larger sample size to validate our results. REFERENCES 1. Otani T, Kishi S, Maruyama Y. Patterns of diabetic macular edema with optical coherence tomography. Am J Ophthalmol. 1999;127:688-693. 2. Kang SW, Park CY, Ham DI. The correlation between fluorescein angiographic and optical coherence tomographic features in clinically significant diabetic macular edema. Am J Ophthalmol. 2004;137:313322. 3. Cabrera Fernández D, Salinas HM, Puliafito CA. Automated detection of retinal layer structures on optical coherence tomography images. Opt Express. 2005;13:10200-10216. 4. Wilkinson CP, Ferris FL 3rd, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology. 2003;110:1677-1682. 5. Sugimoto M, Sasoh M, Ido M, Wakitani Y, Takahashi C, Uji Y. Detection of early diabetic change with optical coherence tomography in type 2 diabetes mellitus patients without retinopathy. Ophthalmologica. 2005;219:379-385. 6. Chen TC, Cense B, Miller JW, et al. Histologic correlation of in vivo optical coherence tomography images of the human retina. Am J Ophthalmol. 2006;141:1165-1168. 7. Browning DJ, Altaweel MM, Bressler NM, Bressler SB, Scott IU, Diabetic Retinopathy Clinical Research Network. Diabetic macular edema: what is focal and what is diffuse? Am J Ophthalmol. 2008;146:649-655. 8. DeBuc DC, Somfai GM, Ranganathan S, Tatrai E, Ferencz M, Puliafito CA. Reliability and reproducibility of macular segmentation using a custom-built OCT retinal image analysis software. J Biomed Opt. 2009;14:064023.

597

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.