Acta Ophthalmologica 2016
Glaucoma diagnostic capacity of optic nerve head haemoglobin measures compared with spectral domain OCT and HRT III confocal tomography Carmen Mendez-Hernandez,1,2 Ignacio Rodriguez-U~ na,1 Manuel Gonzalez-de-la Rosa,3 Paula Arribas-Pardo1 and Julian Garcia-Feijoo1,2 1 Ophthalmology Department, Hospital Clinico San Carlos, Institute of Health Care Research of the Hospital Clinico San Carlos (IdISSC), University Complutense of Madrid, Madrid, Spain 2 Cooperative Research Network in Ophthalmology (RETICS), Institute of Salud Carlos III, Madrid, Spain 3 University of La Laguna, La Laguna, Spain
ABSTRACT. Purpose: The computer program LAGUNA ONHE determines optic nerve head haemoglobin (ONH Hb) on retinal photographs based on detecting colour differences. This study compares the diagnostic capacity of Laguna ONhE with that of spectral domain optical coherence tomography (OCT) and confocal tomography (HRT III). Methods: In a prospective, observational, cross-sectional study, glaucomatous (n = 66) and healthy (n = 52) eyes were examined by Spectralis OCT, HRT III and Laguna ONhE. The following Laguna ONhE variables were determined: ONH Hb across the vertical disc diameter (8&20 Hb), estimated cup–disc ratio (C/D) and the glaucoma discriminant function (GDF), which combines the slope of Hb amount with the mean in 8&20 Hb. The three diagnostic methods were compared by calculating areas under ROC curves (AUCs). Correlations between variables were assessed through Spearman’s rho coefficient. Results: Areas under ROC curves (AUCs) were 0.785 (95% CI: 0.700–0.863) for GDF, 0.807 (95% CI: 0.730–0.883) for OCT retinal nerve fibre layer thickness (OCT-RNFL) and 0.714 (95% CI: 0.618–0.810) and 0.721 (95% CI: 0.628–0.815) for the HRT III variable GPS (glaucoma probability score) and vertical C/D ratio, respectively. Glaucoma discriminant function (GDF) was correlated with OCT-RNFL (0.587, p 0.001; 0.507, p 0.045; and 0.119, p 0.713 for mild, moderate and advanced glaucoma, respectively), mostly so with inferior OCT-RNFL (0.622; p < 0.001). Glaucoma discriminant function (GDF)-HRT III correlations were lower (rim area 0.471, p < 0.0001; rim/disc area 0.426, p < 0.0001; vertical C/D 0.413, p < 0.0001; GPS 0.408, p < 0.0001; rim volume 0.341, p < 0.0001). Conclusion: Similar diagnostic power was observed for Laguna ONhE, Spectralis OCT and HRT III. Key words: glaucoma – haemoglobin – HRT – OCT – optic – retinography
nerve
perfusion – perimetry
Acta Ophthalmol. 2016: 94: 697–704 ª 2016 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd
doi: 10.1111/aos.13050
Introduction High intraocular pressure (IOP) is one of the most important and the only treatable risk factor identified so far for the development of glaucoma. Among other factors implicated both in the development and in the progression of this disease are blood flow dysregulation and an abnormal haemoglobin (Hb) content of the optic nerve head (ONH) (Flammer et al. 2002; Moore et al. 2008; Gonzalez de la Rosa et al. 2013). Despite the findings of large clinical trials (Gordon et al. 2002, 2007; Leske et al. 2003; Chauhan et al. 2008) indicating that vascular diseases are not proved risk factors of glaucoma onset or progression, some patients with glaucoma show vascular alterations such as vasospasm, systemic hypotension, angiographic vascular perfusion defects and modified blood flow, which could lead to a decreased blood supply to the ONH and retina (Tezel et al. 2010). Our group has recently proposed the use of software to interpret retinal photographs in terms of the amount of Hb present in the optic nerve disc (ONH Hb) (Gonzalez de la Rosa et al. 2013). This computer program, LAGUNA ONHE (optic nerve head haemoglobin), can be used to colorimetrically analyse ONH photographs to estimate the amount of Hb present at the disc, and this is relatively simple and inexpensive, as only a non-mydriatic fundus
697
Acta Ophthalmologica 2016
camera is needed. Our studies have shown its high reproducibility, good glaucoma diagnostic capacity and good correlation with most morphologic and functional indices of primary openangle glaucoma (POAG) (Gonzalez de la Rosa et al. 2013; Mendez-Hernandez et al. 2014). The method also shows excellent intraobserver, interobserver, within-session and between-session reproducibility in pseudophakic patients with and without glaucoma along with high interanalysis reproducibility when performed by a masked examiner (Mendez-Hernandez et al. 2014). This study was designed to compare the capacity of the Laguna ONhE method to diagnose glaucoma with that of optical coherence tomography (OCT) and heidelberg retina tomograph (HRT).
Patients and Methods Subjects
The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the local ethics committee. Informed consent was obtained from each participant. The study sample was comprised of 118 eyes of 66 patients with POAG and 52 healthy subjects. Only one eye was chosen randomly in each participant. Patients with POAG were screened by an independent observer who examined the subjects including visual field tests and funduscopy. Patients with POAG were consecutively recruited from the glaucoma department of our hospital. Inclusion criteria were focal (localized notching) or diffuse neuroretinal rim narrowing, concentric enlargement of the optic cup, or both, and/or reproducible abnormal visual field test results. The MD visual field index (OCTOPUS G1-program) was used to define the patient subgroups: mild (MD ≤ 6 dB), moderate (MD > 6–12 dB) and advanced (MD ≥ 12 dB) glaucoma. Patients with a spherical equivalent greater than 5D or 3D of astigmatism, significant lens opacity, or with an oblique or tilted ONH, a narrow angle, suspected ocular hypertension (OHT) or secondary glaucoma were excluded. The healthy age-matched subjects were recruited among persons under-
698
going a routine eye test at our hospital, the relatives of the patients enroled and hospital staff. The inclusion criteria for this study group were as follows: IOP <18 mmHg measured by Goldmann applanation tonometry (GAT); an intact neuroretinal rim with no signs of haemorrhage, notches, excessive cupping defined as concentric enlargement of the optic cup (Jonas et al. 1993; Jonas & Dichtl 1995) or retinal nerve fibre layer (RNFL) defects; normal white-on-white automated perimetry variables (mean deviation (MD) < 2 dB and loss variance (LV) < 7 dB; OCTOPUS G1-program). Individuals with a history of diabetes, moderate or severe arterial hypertension (patients treated with two or more drugs), cardiovascular or haematologic diseases as well as systemic diseases or ocular diseases that could affect vision or a history of neurologic disease were excluded from the study. Age, and previous cataract and glaucoma surgery were not criteria for exclusion. Study protocol
In an initial visit, all participants were subjected to a complete ophthalmologic examination including complete clinical history, white-on-white perimetry using the Octopus perimeter TOP G1 program (Haag-Streit AG, Bern, Switzerland), visual acuity measurement and refraction, slit lamp biomicroscopy, Goldmann applanation tonometry (GAT), gonioscopy, fundus examination and retinography (Canon non-mydriatic retinal camera CR-DGi; Canon Inc., Tokyo, Japan). All participants also underwent optical coherence tomography (Spectralis OCT; Heidelberg Engineering, Heidelberg, Germany), confocal scanning laser tomography (HRT III, Heidelberg Retina Tomograph Model 3; Heidelberg Engineering) and ONH Hb determination on the retinal photographs taken in the examination. The most relevant parameters of each diagnostic procedure, Spectralis OCT and HRT3, were selected for data analysis. Image quality of each procedure was evaluated. All Spectralis OCT and HRT3 images were acquired with a quality greater than 7/10. Optic nerve Hb contents were measured using the Laguna ONhE procedure as described in our previous studies (Gonzalez de la
Rosa et al. 2013; Mendez-Hernandez et al. 2014). ONH Hb measurements
Zones of the ONH with a high Hb content mainly reflect red light while areas of low Hb reflect higher proportions of green and blue light. The LAGUNA ONHE program uses mathematical algorithms for automatic component segmentation to delimit the ONH border and identify the central retinal blood vessels. Two areas of the ONH were defined: the central retinal vessels and the ONH tissue itself. The formulas were then calculated at those pixels corresponding to the vessels as a whole and for every isolated pixel of tissue. The ONH areas with high Hb content mainly reflect red light. In contrast, areas with low Hb content reflect a lower proportion of the red component compared with the green and blue light. The blue, green and red components of the picture were assessed with an image analysis program using MATLAB image processing toolbox (MathWorks Inc., Natick, MA, USA). The method whereby analyses Hb amount has been previously described (Gonzalez de la Rosa et al. 2013). Figure 1 shows how the LAGUNA ONHE program automatically divides the image of the papilla into 24 sectors shared across three concentric rings. The outer ring mostly aligns with the neuroretinal rim, the central ring corresponds to the transition area, which may include the neuroretinal rim and cupping, and the inner rim mainly comprises the cup area. As with other imaging techniques used in glaucoma diagnosis, this alignment can be more difficult in tilted discs. On the basis of colorimetric analysis of the papilla, it is known that both in control individuals and in patients with glaucoma, Hb levels are greater in the more peripheral sectors, that is, in the outer ring. The middle ring tends to show a lower amount of Hb than the outer ring, and the innermost ring features the least amount of Hb. In other words, tissue Hb levels diminish from the disc periphery towards the centre such that the difference between peripheral and central Hb levels can be calculated. By determining colour changes in the central disc zone, the software estimates the cup and cup–disc ratio. Laguna ONhE also uses an index called the glaucoma discriminant function Glaucoma discriminant function
Acta Ophthalmologica 2016
Fig. 1. Image of a glaucomatous papilla. Upper images are the colour fundus photograph of the optic disc (left) and the disc automatically divided in radial sectors (right) whereas the lower images are the disc automatically divided into eight 45-degree radial sectors and three concentric rings (left) and the corresponding pseudo-image representing the amount of haemoglobin (right). A colorimetric scale is shown at the top of the lower left image to assess the amount of haemoglobin.
(GDF), which combines the Hb slope, obtained through multiple regression analysis of Hb estimated for each sector with the mean Hb amount determined in the vertical disc diameter, in sectors 8 and 20 (8&20 Hb). ONH Hb is expressed as a percentage (%). Glaucoma discriminant function (GDF) is expressed as a whole number from +75 to 90. It describes the probability that the optic nerve disc shows an Hb amount within the normal range (positive values) or in the range consistent with glaucoma (negative values). The relationship is therefore direct such that the further the GDF from zero in either direction, the greater the likelihood will be that the disc belongs to one or the other diagnostic groups. Data were compared between patients with POAG and healthy subjects using the Student’s t-test. The chisquared test was used to compare data among the patient subgroups. It was first checked that all data were normally distributed (Kolmogorov–Smir-
nov test). Data are presented as means SD. Relationships between structural parameters were examined through Spearman correlation. Areas under receiver operating characteristic curves (AUCs) were calculated for all parameters determined in each test. To compare the AUCs obtained with the different devices, the Hanley/McNeil method was used (Hanley & McNeil 1983). All statistical tests were performed using the software package IBM SPSS (version 17.0; IBM Corp., Somers, NY, USA). Significance was set at p < 0.05.
Results The study sample was comprised of 118 eyes of 118 subjects: 52 healthy control eyes and 66 eyes with POAG. All the data analysed were normally distributed. Table 1 provides the main demographic and clinical characteristics of the study subjects, and the variables determined using Spectralis
OCT, HRT3 and Laguna ONhE (minimum, maximum, mean and standard deviation). This table also shows comparisons of means between the glaucoma subgroups (Student t-test) and levels of statistical significance (p). There was a slight predominance of women, 68 women (57.6%) versus 50 men (42.4%). No significant differences were detected in age or gender between the study groups including the glaucoma subgroups (chi-squared test). OCT-measured RNFL thickness (OCT-RNFL) was significantly lower in the patients with glaucoma than in controls (p < 0.001). No significant differences between groups were detected for the HRT3 variables such as disc area, rim area, cup/disc area ratio or rim volume. However, the patients with glaucoma showed abnormal vertical cup/disc ratios and glaucoma probability scores (GPS) (p < 0.0001). The amount of Hb observed in the vertical disc diameter (sectors 8 and 20) (8&20 Hb) was significantly lower in the patients with glaucoma (p = 0.001) and in all the glaucoma subgroups (p = 0.018 mild, p = 0.002 moderate, p < 0.001 advanced). The mean GDF value in the control group was positive. In the mild glaucoma subgroup, GDF was negative (p = 0.001) and acquired a more negative value with glaucoma severity (p < 0.001 for moderate and advanced glaucoma). Estimated C/D ratios did not differ significantly among groups. Table 2 provides the correlations detected between GDF and the OCT, HRT3 or perimetry data adjusted by age, sex and disc size in the patients with glaucoma and glaucoma patient subgroups. Glaucoma discriminant function (GDF) was well correlated with disc area determined by HRT3 in moderate glaucoma (r = 0.587; p = 0.008). Glaucoma discriminant function (GDF) showed greater correlation with OCTRNFL in normal subjects (r = 0.322; p = 0.024) and in both mild and moderate glaucoma subgroups (r = 0.587; p = 0.001 and r = 0.507; p = 0.045, respectively), and also showed correlation with Octopus G1 MD in moderate glaucoma (r = 0.609; p = 0.012). Receiver operating characteristic (ROC) curves were constructed for all the variables determined using each diagnostic procedure, and the cut-off
699
700
32 70.20; 8.35 (50–82)
52 67.80; 10.13 (52–84)
27/25 (51.91/48.10)
0.05; 1.38 ( 2.60–0.30)
92.79; 14.11 (68.00–134.00)
1.84; 0.67 (0.65–3.33)
1.19; 0.52 (0.35–3.09)
0.46; 0.94 (0.06–0.72)
0.65; 0.16 (0.30–0.94)
0.41; 0.97 ( 0.25–0.70)
0.42; 0.25 (0.00–0.94)
0.42; 0.32 (0.08–0.93)
n Age (yr)
Sex, M/F (%)
MD TOP G1, (dB)
RNFL thickness OCT (lm)
Disc area HRT III (mm2)
Rim area HRT III (mm2)
Cup/disc area ratio HRT III
Rim/disc area ratio HRT III
Rim volume HRT III (mm3)
Vertical C/D ratio HRT III
GPS HRT III
0.60; 0.31 (0.08–0.92)
0.56; 0.24 (0.01–0.98)
0.23; 0.19 ( 0.18–0.65)
0.56; 0.21 (0.05–0.95)
0.43; 0.21 (0.05–0.95)
1.07; 0.44 (0.09–1.84)
2.02; 0.62 (0.52–3.31)
81.59; 18.51 (33.00–117.00)
3.54; 1.55 (3.02–5.96)
11/21 (34.40/65.64)
Mild glaucoma (2) Mean; SD (range)
Control (1) Mean; SD (range)
8.67; 1.74 (3.10–11.90)
10/9 (52.63/47.37)
69.20; 1.11 (50–85)
0.73; 0.19 (0.23–0.93)
0.63; 0.22 (0.05–0.91)
0.24; 0.22 ( 0.15–0.94)
0.56; 0.18 (0.14–0.93)
0.79; 1.51 (0.12–0.72)
1.06; 0.46 (0.17–1.85)
1.95; 0.52 (0.64–3.09)
70.05; 12.75 (52.00–92.00)
19
Moderate glaucoma (3) Mean; SD (range)
15.53; 5.32 (7.00–26.50)
9/6 (60.00/40.00)
68.67; 9.10 (50–83)
0.75; 0.23 (0.10–0.93)
0.70; 0.21 (0.23–0.98)
0.16; 0.12 (0.01–0.47)
0.45; 0.26 (0.12–0.88)
0.55; 0.26 (0.12–0.88)
0.89; 0.49 (0.24–1.84)
2.13; 0.48 (1.58–2.98)
51.33; 14.00 (36.00–82.00)
15
Advanced glaucoma (4) Mean; SD (range)
Table 1. Demographic/clinical characteristics and the disc and visual field variables (mean; SD) recorded in the study groups.
7.74; 5.57 (3.02–26.50)
23/43 (34.80/65.10)
69.36; 9.13 (50–85)
0.67; 0.27 (0.08–0.93)
0.61; 0.23 (0.01–0.98)
0.22; 0.19 ( 0.18–0.94)
0.54; 0.22 (0.05–0.95)
0.56; 0.83 (0.05–0.95)
1.15; 0.47 (0.09–3.09)
2.02; 0.56 (0.52–3.31)
71.39; 19.88 (33.00–134.00)
66
POAG (2,3,4) Mean; SD (range)
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus versus
t-test (p)
2: 0.986 3: 0.875 4: 0.876 POAG: 0.275 2: 0.226† 3: 0.536† 4: 0.224† POAG: 0.916† 2: <0.001* 3: <0.001* 4: <0.001* POAG: <0.001* 2: 0.001* 3: <0.001* 4: <0.001* POAG: <0.001* 2: 0.227 3: 0.509 4, 0.124 POAG: 0.105 2: 0.291 3: 0.355 4: 0.050 POAG: 0.076 2: 0.830 3: 0.271 4: 0.709 POAG: 0.546 2: 0.041* 3: 0.053 4: <0.001* POAG: 0.002* 2: 0.288 3: 0.428 4: 0.305 POAG: 0.102 2: 0.019* 3: 0.003* 4: <0.001* POAG: <0.001* 2: 0.014* 3: <0.001* 4: <0.001* POAG: <0.001*
Acta Ophthalmologica 2016
56.91 11.55 (25.00–78.00) 49.60; 11.07 (25.00–66.30) 57.36; 9.62 (32.00–72.50) 66.46; 11.91 (40.46–84.70) 8&20 Hb Laguna ONhE
60.06; 11.60 (29.40–78.00)
0.65; 0.14 (0.38–1.00) 0.75; 0.11 (0.54–1.00) 0.65; 0.10 (0.49–0.88) 0.60; 0.14 (0.38–0.91) 0.52; 0.10 (0.27–0.73) Estimated C/D ratio Laguna ONhE
† Chi-square test. * difference significant at p < 0.05. C/D = cup/disc, GDF = glaucoma discriminant function, GPS = glaucoma probability score, MD = mean defect, M/F = male/female, N = number of cases, POAG = primary open-angle glaucoma.
versus versus versus versus versus versus versus versus versus versus versus versus 1 1 1 1 1 1 1 1 1 1 1 1 14.14; 26.13 ( 89.00–37.00) 32.60; 24.03 ( 89.00–8.00) 14.05; 20.38 ( 64.00–18.00) 5.53; 26.23 ( 69.00–37.00) 11.08; 17.68 ( 30.00–51.00) GDF Laguna ONhE
Table 1. (Continued).
Control (1) Mean; SD (range)
Mild glaucoma (2) Mean; SD (range)
Moderate glaucoma (3) Mean; SD (range)
Advanced glaucoma (4) Mean; SD (range)
POAG (2,3,4) Mean; SD (range)
t-test (p)
2: 0.001* 3: <0.001* 4: <0.001* POAG: <0.001* 2: 0.467 3: 0.376 4: 0.419 POAG: 0.315 2: 0.018* 3: 0.002* 4: <0.001* POAG: 0.001*
Acta Ophthalmologica 2016
providing the best balance between sensitivity and specificity was determined in each case (Tables 3 and 4). OCT-RFNL and GDF had the greatest AUCs. AUC for OCT-RNFL was 0.81, p = 0.001 with a sensitivity of 83.3%, a positive predictive power of 74.3%, a specificity of 63.5% and a negative predictive power of 75.0%. Regarding GDF, the AUC obtained was 0.78, p < 0.001, with a sensitivity of 72.7%, a positive predictive power of 76.2%, a specificity of 71.2% and a negative predictive power of 67.3%. In the advanced glaucoma group, AUCs obtained were 0.97, p < 0.01 for OCTRNFL and 0.94, p < 0.01 for GDF.
Discussion The role of ocular blood flow to the optic nerve in glaucoma has been a controversial field partly due to limitations in assessing ocular blood flow accurately (Hayreh 1997; Harris et al. 2001). The supply of oxygen of the ONH depends in part on various factors, such as blood flow, blood oxygen saturation and amount of circulating blood. Blood flow may be indirectly estimated from velocity measured in neighbouring arteries by Doppler or measured locally using complex methods of laser Doppler (Michelson et al. 1995) or laser speckle flowgraphy (Yaoeda et al. 2000). Blood oxygen saturation determination for the study of glaucoma has proven to be difficult (Michelson & Scibor 2006) because of instrument difficulties and contradictory results (Hardarson et al. 2009; Traustason et al. 2009). The third factor, amount of circulating blood, is what the Laguna ONhE method attempts to quantify. In a recent study, this factor was measured by OCT angiography as ONH vascularization (Jia et al. 2012). The method proposed here differs from spectrophotometric analysis of the ONH, targeted at measuring Hb oxygen saturation. Laguna ONhE tries to quantify Hb itself as an indirect measure of the volume of blood that irrigates the papilla. In clinical practice, methods that are efficient and accessible to most users are necessary. Laguna ONhE is a noninvasive simple method that is not expensive as it is performed on retinal photographs and thus only a fundus camera is needed. Many diagnostic devices such as OCT angiography are
701
Acta Ophthalmologica 2016
Table 2. Correlations adjusted by age, gender and disc size between GDF (Laguna ONhE) and OCT, HRT III or perimetry variables. GDF Control
Mild glaucoma p
r Octopus TOP G1 MD LV Spectralis OCT Average RNFL thickness HRT III Disc area Rim area Rim volume Rim/disc area ratio Cup/disc area ratio Vertical cup/disc ratio GPS
Moderate glaucoma p
r
p
r
Advanced glaucoma p
r
0.263 0.213
0.067 0.142
0.314 0.096
0.097 0.620
0.609* 0.236
0.012 0.378
0.194 0.057
0.546 0.859
0.322*
0.024
0.587§
0.001
0.507*
0.045
0.119
0.713
0.214 0.368§ 0.158 0.326* 0.158 0.344* 0.181
0.127 0.009 0.203 0.022 0.278 0.016 0.214
0.325 0.266 0.363 0.208 0.215 0.213 0.073
0.069 0.163 0.053 0.279 0.263 0.266 0.705
0.587§ 0.200 0.277 0.159 0.032 0.123 0.227
0.008 0.459 0.299 0.557 0.905 0.650 0.398
0.400 0.254 0.290 0.233 0.233 0.172 0.298
0.140 0.425 0.361 0.466 0.466 0.564 0.346
Correlation at a significance level below p = 0.01. * Correlation at a significance level below p = 0.05. Statistical significant correlations are shown in bold values.
§
Table 3. AUCs for the main variables measured using the different methods.
OCT: average RNFL thickness Laguna ONhE: GDF HRT III: vertical cup/disc ratio GPS Cup/disc area ratio
AUC
SE
Cut-off
S (%)
PPP (%)
Sp (%)
NPP (%)
p value
Comparison between AUCs (p)†
0.81 0.78 0.72 0.71 0.68
0.04 0.04 0.05 0.05 0.05
88.50 2.50 0.52 0.67 0.36
83.3 72.7 75.8 63.6 62.1
74.3 76.2 71.2 75.0 71.9
63.5 71.2 61.5 73.1 69.2
75.0 67.3 63.5 61.3 59.0
0.001* <0.001* <0.001* <0.001* 0.001*
0.59 – 0.22 0.15 0.05
S and Sp = Indicated are the best cut-offs for sensitivity and specificity, PPP and NPP = positive predictive power and negative predictive power. SE = standard error of the AUC. * p < 0.05. † Compared with Laguna ONhE-GDF AUC (Hanley/McNeil method). Table 4. AUCs for the main variables measured using the different methods in the patient subgroups.
HRT III vert. cup/disc ratio HRT III GPS HRT III cup/disc area ratio OCT average RNFL thickness Laguna ONhE-GDF
Mild glaucoma
Moderate glaucoma
Advanced glaucoma
AUC
p
AUC
p
AUC
p
0.66 0.66 0.63 0.69 0.68
0.02* 0.02* 0.05 0.01* 0.01*
0.76 0.76 0.71 0.88 0.83
0.01* 0.01* 0.01* <0.01* <0.01*
0.80 0.77 0.73 0.97 0.94
<0.01* 0.01* 0.01* <0.01* <0.01*
* p < 0.05.
inaccessible to many ophthalmologists. In addition, our software only requires a non-mydriatic retinography while mydriatic drops are needed to obtain high-resolution OCT angiography images. It also has been suggested that changes in ONH reflectance could help identify differences in Hb levels (Crittin & Riva 2004), but no practical methods to quantify such changes have been proposed. Further, the Laguna ONhE
702
method appears to show a greater relationship with other morphological and functional indices for glaucoma diagnosis than the recently described OCT angiograph (Jia et al. 2014). It is known that in patients with advanced glaucoma, besides the disc showing greater cupping, it also appears pale in colour and this pallor is more pronounced as glaucoma damage advances due to the increased excava-
tion. Our clinical experience indicates that the subjective colour of the disc may guide a diagnosis of glaucoma (Gonzalez de la Rosa et al. 2013) and may be useful in other diseases such as optic neuritis where reduced perfusion may be found in different regions from those affected in glaucoma. In multiple sclerosis, reduced perfusion is most frequently observed in the temporal rim (Bambo et al. 2013). In turn, the colour of the papilla is determined by the amount of Hb present relative to that observed on the constant white background provided by myelin. The red colour of Hb is the result of its great capacity to absorb short wavelengths and lower capacity to absorb long wavelengths of the visible light spectrum. Thus, the greater the absorption of short wavelengths, the greater the amount of Hb present. Also, by adequately selecting the wavelengths examined, the degree of tissue oxygenation can also be assessed.
Acta Ophthalmologica 2016
The glaucoma discriminant function, GDF, was able to discriminate between healthy eyes and eyes with POAG, and differences between the two subject groups were greater for an increasing severity of glaucoma. In contrast, the estimated C/D ratio provided by the program, though higher in the patients with glaucoma, failed to differ between the two study groups. This may be attributed to the fact that Laguna ONhE provides only an estimate of this ratio (Rodriguez Una et al. 2014). C/D ratios differed significantly among the three methods tested in our study. The Laguna ONhE C/D ratio is a rough estimate because it is based on 2D photographs. HRT3 provides a more accurate C/D ratio as it performs a 3D analysis of the ONH. Interestingly, the Laguna ONhE C/D ratio was more similar to the HRT3 vertical C/D ratio than the C/D area ratio. Thus, Laguna ONhE could be better at estimating the vertical C/D ratio, which becomes more evident in advanced glaucoma. We are presently working on a new version of Laguna ONhE targeted at improving its measure of the C/D ratio (Pena-Betancor et al. 2015). As in the present study, most investigations examining correlations between different structural tests and perimetry have mainly focused on overall indices provided by the tests compared (Lan et al. 2003; DaneshMeyer et al. 2006; Kawano et al. 2006; Yalvac et al. 2009; Zhong et al. 2009; Kaushik et al. 2010; Cvenkel & Sket Kontestabile 2011). In our patients with glaucoma, correlations between GDF and TOP G1 indices or the other structural variables examined were moderate. Greatest correlation was observed for OCT-RNFL, and slightly lower correlation was detected for the HRT3 variables. Correlations between GDF and OCTRNFL values for the nasal and temporal quadrants were low while higher correlation was observed at the vertical meridian (GDF showed greatest correlation with OCT-RNFL in the inferior sector, 0.622; p < 0.001). This could be due to the important role played by the vertical disc quadrants in glaucoma as the GDF is based on Hb levels in these sectors (Gonzalez de la Rosa et al. 2013; Rodriguez Una et al. 2014). Nouri-Mahdavi et al. (2004) reported that the factor that best distinguished
normal eyes from glaucomatous eyes was RNFL thickness in the superior and inferior quadrants. In the LAGUNA ONHE program, the vertical disc quadrants (sectors 8 and 20) are the most discriminant of glaucoma in that they show the greatest difference in Hb between healthy and glaucoma eyes (Gonzalez de la Rosa et al. 2013). This is consistent with the data in the literature, which indicates that the first fibres of the disc rim to undergo damage are those located at the superior and inferior poles such that the cup stretches vertically (Quigley & Addicks 1981). This is why Laguna ONhE uses the mean Hb value for these two sectors and the Hb gradient to construct the factor GDF (Gonzalez de la Rosa et al. 2013). The variable featuring the largest AUC was OCT-RFNL (AUC = 0.807; p = 0.001; S = 83.3%; Sp = 63.5%), followed by GDF (AUC = 0.781; p < 0.001; S = 72.7%; Sp = 71.2%) and HRT III vertical C/D ratio (AUC = 0.721; p < 0.001; S = 75.8%; Sp = 61.5%). Thus, the AUCs obtained were fairly similar for the different structural diagnostic markers. These values may be classified as moderate, although 95% confidence intervals in some of the ROC curves were analysed and sensitivity and specificity values were within the medium range. In a prior study (Gonzalez de la Rosa et al. 2013), we observed high correlation and a similar diagnostic capacity of LAGUNA ONHE to other procedures. Here we have compared the LAGUNA ONHE program with Spectralis OCT instead of OCT Cirrus and also stratified the patients with glaucoma by disease stage according to the MD value obtained. The non-mydriatic retinal camera used in this study also differs from the camera used in our earlier study. The findings of both studies indicate that Laguna ONhE offers results that are reproducible and independent of the diagnostic device used as the standard reference or POAG severity. Our study has limitations. The high variability in normal optic disc morphology, as well as refractive errors, poor fixation and eye movements, may affect measurement accuracy. Thus, to avoid the influence of some of these factors, patients with a spherical equivalent greater than 5D or 3D of astigmatism, or with an oblique or tilted ONH,
were excluded. In addition, this is a clinical-based, case–control study, so it may not be appropriate to generalize results to all patients with glaucoma. In conclusion, Laguna ONhE shows a good diagnostic capacity and diagnostic agreement with other procedures based on morphology. The GDF index was found to correlate well with TOPG1 perimetry indices and the most clinically meaningful Spectralis OCT and HRT3 variables. Future studies to compare colorimetry with spectrometry and retinal vessel oximetry as well as other automated software which also uses retinal photographs for glaucoma diagnosis may be necessary. This new photographic colorimetric procedure could be useful for an early structural diagnosis of glaucoma.
References Bambo MP, Garcia-Martin E, Perez-Olivan S et al. (2013): Diagnostic ability of a new method for measuring haemoglobin levels in the optic nerve head in multiple sclerosis patients. Br J Ophthalmol 97: 1543–1548. Chauhan BC, Mikelberg FS, Balaszi AG, LeBlanc RP, Lesk MR & Trope GE; Canadian Glaucoma Study Group (2008): Canadian Glaucoma Study: 2. risk factors for the progression of open-angle glaucoma. Arch Ophthalmol 126: 1030–1036. Crittin M & Riva CE (2004): Functional imaging of the human papilla and peripapillary region based on flicker-induced reflectance changes. Neurosci Lett 360: 141–144. Cvenkel B & Sket Kontestabile A (2011): Correlation between nerve fibre layer thickness measured with spectral domain OCT and visual field in patients with different stages of glaucoma. Graefes Arch Clin Exp Ophthalmol 249: 575–584. Danesh-Meyer HV, Ku JY, Papchenko TL, Jayasundera T, Hsiang JC & Gamble GD (2006): Regional correlation of structure and function in glaucoma, using the Disc Damage Likelihood Scale, Heidelberg Retina Tomograph, and visual fields. Ophthalmology 113: 603–611. Flammer J, Org€ ul S, Costa VP, Orzalesi N, Krieglstein GK, Serra LM, Renard JP & Stef ansson E (2002): The impact of ocular blood flow in glaucoma. Prog Retin Eye Res 21: 359–393. Gonzalez de la Rosa M, Gonzalez-Hernandez M, Sigut J et al. (2013): Measuring hemoglobin levels in the optic nerve head: comparisons with other structural and functional parameters of glaucoma. Invest Ophthalmol Vis Sci 54: 482–489. Gordon MO, Beiser JA, Brandt JD et al. (2002): The Ocular Hypertension Treatment Study: baseline factors that predict the onset
703
Acta Ophthalmologica 2016
of primary open-angle glaucoma. Arch Ophthalmol 120: 714–720. Gordon MO, Torri V, Miglior S et al. (2007): Validated prediction model for the development of primary open-angle glaucoma in individuals with ocular hypertension. Ophthalmology 114: 10–19. Hanley JA & McNeil BJ (1983): A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 148: 839–843. Hardarson SH, Gottfredsdottir MS, Halldorsson GH et al. (2009): Glaucoma filtration surgery and retinal oxygen saturation. Invest Ophthalmol Vis Sci 50: 5247–5250. Harris A, Jonescu-Cuypers C, Martin B, Kagermann L, Zalish M & Garzozi HJ (2001): Simultaneous management of blood flow and IOP in glaucoma. Acta Ophthalmol Scand 79: 336–341. Hayreh SS (1997): Factors influencing blood flow in the optic nerve head. J Glaucoma 6: 412–425. Jia Y, Tan O & Tokayer J (2012): Splitspectrum amplitude-decorrelation angiography with optical coherence tomography. Opt Express 20: 4710–4725. Jia Y, Wei E, Wang X et al. (2014): Optical coherence tomography angiography of optic disc perfusion in glaucoma. Ophthalmology 121: 1322–1332. Jonas JB & Dichtl A (1995): Advances in the assessment of optic disc changes in early glaucoma. Curr Opin Ophthalmol 6: 61–66. Jonas JB, Fernandez MC & St€ urmer J (1993): Pattern of glaucomatous neuroretinal rim loss. Ophthalmology 100: 63–68. Kaushik S, Pandav SS, Ichhpujani P & Gupta A (2010): Correlation of frequency-doubling perimetry with retinal nerve fiber layer thickness and optic disc size in ocular hypertensives and glaucoma suspects. J Glaucoma 20: 366–370. Kawano J, Tomidokoro A, Mayama C, Kunimatsu S, Tomita G & Araje M (2006): Correlation between hemifield visual field damage and corresponding parapapillary atrophy in normal-tension glaucoma. Am J Ophthalmol 142: 40–45. Lan YW, Henson DB & Kwartz AJ (2003): The correlation between optic nerve head topographic measurements, peripapillary nerve fibre layer thickness, and visual field indices in glaucoma. Br J Ophthalmol 87: 1135–1141.
704
Leske MC, Heijl A, Hussein M, Bengtsson B, Hyman L & Komaroff E; Early Manifest Glaucoma Trial Group (2003): Factors for glaucoma progression and the effect of treatment: the early manifest glaucoma trial. Arch Ophthalmol 121: 48–56. Mendez-Hernandez C, Garcia-Feijoo J, Arribas-Pardo P, Saenz-Frances F, RodriguezU~ na I, Fernandez-Perez C & de la Rosa MG (2014): Reproducibility of optic nerve head hemoglobin measures. J Glaucoma Oct 2: 7. Michelson G & Scibor M (2006): Intravascular oxygen saturation in retinal vessels in normal subjects and open-angle glaucoma subjects. Acta Ophthalmol Scand 84: 289– 295. Michelson G, Langhans MJ & Groh MJ (1995): Clinical investigation of the combination of a scanning laser ophthalmoscope and laser Doppler flowmeter. Ger J Ophthalmol 4: 342–349. Moore D, Harris A, Wudunn D, Kheradiya N & Siesky B (2008): Dysfunctional regulation of ocular blood flow: a risk factor for glaucoma? Clin Ophthalmol 2: 849–861. Nouri-Mahdavi K, Hoffman D, Tannenbaum DP, Law SK & Caprioli J (2004): Identifying early glaucoma with optical coherence tomography. Am J Ophthalmol 137: 228– 235. Pena-Betancor C, Gonzalez-Hernandez M, Fumero-Batista F, Sigut J, Medina-Mesa E, Alayon S & Gonzalez de la Rosa M (2015): Estimation of the relative amount of hemoglobin in the cup and neuro-retinal rim using stereoscopic color fundus images. Invest Ophthalmol Vis Sci 56: 1562–1568. Quigley HA & Addicks EM (1981): Regional differences in the structure of the lamina cribrosa and their relation to glaucomatous optic nerve damage. Arch Ophthalmol 99: 137–143. Rodriguez Una I, Mendez Hernandez CD, Saenz-Frances F & Garcia Feijoo J (2014): Correlating cup-to-disc ratios measured by HRT-III, SD-OCT and the new color imaging Laguna ONhE procedure. Arch Soc Esp Oftalmol 90: 212–219. Tezel G, Yang X, Luo C, Cai J, Kain AD, Powell DW, Kuehn MH & Pierce WM (2010): Hemoglobin expression and regulation in glaucoma: insights into retinal ganglion cell oxygenation. Invest Ophthalmol Vis Sci 51: 907–919.
Traustason S, Hardarson SH, Gottfredsdottir MS, Eysteinsson T, Karlsson RA, Stef ansson E & Harris A (2009): Dorzolamide-timolol combination and retinal vessel oxygen saturation in patients with glaucoma or ocular hypertension. Br J Ophthalmol 93: 1064–1067. Yalvac IS, Altunsoy M, Cansever S, Satana B, Eksioglu U & Duman S (2009): The correlation between visual field defects and focal nerve fiber layer thickness measured with optical coherence tomography in the evaluation of glaucoma. J Glaucoma 18: 53–61. Yaoeda K, Shirakashi M, Funaki S, Funaki H, Nakatsue T & Abe H (2000): Measurement of microcirculation in the optic nerve head by laser speckle flowgraphy and scanning laser Doppler flowmetry. Am J Ophthalmol 129: 734–739. Zhong Y, Chen L, Cheng Y & Huang P (2009): Correlation between blue-on-yellow perimetry and scanning laser polarimetry with variable corneal compensation measurements in primary open-angle glaucoma. Jpn J Ophthalmol 53: 574–579.
Received on March 30th, 2015. Accepted on February 4th, 2016. Correspondence: Carmen Mendez-Hernandez, MD, PhD Department of Ophthalmology Hospital Clinico San Carlos Prof. Martin Lagos Av. s/n 28040 Madrid Spain Tel: +34 913303000 Fax: +34 913303138 Email: cdmendezh@gmail.com This study was supported in part by the Instituto de Salud Carlos III grant PI09/90933. Some of the contents of this study were presented at the 2013 ARVO Annual Meeting, Seattle, FL, USA, May 2013. Some of the contents of this study were described as a poster in World Glaucoma Congress, Vancouver, Canada, July 2013. GDLR. M. has a proprietary interest in the LAGUNA ONHE program. The remaining authors declare no conflict of interest.