Ability and Reproducibility of Fourier-Domain Optical Coherence Tomography toDetect Retinal Nerve Fiber Layer Atrophyin Parkinson’s Disease

Elena Garcia-Martin, PhD,1,2 Maria Satue, MD,1 Isabel Fuertes, PhD,1 Sofia Otin, MD,1,2

Raquel Alarcia, PhD,3 Raquel Herrero, MD,1,2 Maria P. Bambo, MD,1,2 Javier Fernandez, PhD,1,2

Luis E. Pablo, PhD1,2

Purpose: To evaluate and compare the ability of 3 protocols of Fourier-domain optical coherence tomog-raphy (OCT) to detect retinal thinning and retinal nerve fiber layer (RNFL) atrophy in patients with Parkinson’sdisease (PD) compared with healthy subjects. To test the intrasession reproducibility of RNFL thicknessmeasurements in patients with PD and healthy subjects using the Cirrus (Carl Zeiss Meditec Inc., Dublin, CA) andSpectralis (Heidelberg Engineering, Inc., Heidelberg, Germany) OCT devices.

Design: Observational, cross-sectional study.Participants: Patients with PD (n � 75) and age-matched healthy subjects (n � 75) were enrolled.Methods: All subjects underwent three 360-degree circular scans centered on the optic disc by the same

experienced examiner using the Cirrus OCT instrument, the classic glaucoma application, and the new NsiteAxonal Analytics of the Spectralis OCT instrument.

Main Outcome Measures: Differences between the eyes of healthy subjects and the eyes of patients withPD were compared using the 3 protocols. The relationship between measurements provided by each OCTprotocol was evaluated. Repeatability was studied by intraclass correlation coefficients and coefficients ofvariation.

Results: Retinal nerve fiber layer atrophy was detected in eyes of patients with PD (P � 0.025, P�0.042, andP � 0.001) with the 3 protocols used, but the Nsite Axonal Analytics of the Spectralis OCT device was the mostsensitive for detecting subclinical defects. In eyes of patients with PD, RNFL thickness measurements deter-mined by the OCT devices were correlated, but they were significantly different between the Cirrus and Spectralisdevices (P � 0.038). Reproducibility was good with all 3 protocols but better using the Glaucoma application ofthe Spectralis OCT device.

Conclusions: Fourier-domain OCT can be considered a valid and reproducible device for detecting sub-clinical RNFL atrophy in patients with PD, especially the Nsite Axonal Analytics of the Spectralis device. Retinalnerve fiber layer thickness measurements differed significantly between the Cirrus and Spectralis devices despitea high correlation of the measurements between the 2 instruments.

Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussedin this article. Ophthalmology 2012;119:2161–2167 © 2012 by the American Academy of Ophthalmology.

Pm

Papcc

ta

Parkinson’s disease (PD) is a degenerative disorder of thecentral nervous system caused by the accumulation of theprotein alpha-synuclein in neuronal Lewy bodies and frominsufficient formation and activity of dopamine produced incertain neurons in the substantia nigra of the midbrain.1

Diagnosis of typical cases is based mainly on symptoms,with tests such as neuroimaging used for confirmation.2

Early in the course of the disease, the most obvious symp-toms are movement related (shaking, rigidity, slowness ofmovement, and difficulty with walking and gait). Later,cognitive and behavioral problems may arise, with dementia

commonly occurring in the advanced stage of the disease. i

© 2012 by the American Academy of OphthalmologyPublished by Elsevier Inc.

arkinson’s disease is more common in the elderly, withost cases occurring after the age of 50 years.3

Foveal vision is one of the affected nonmotor systems inD. Previous authors reported decreased contrast sensitivitynd color vision and altered visual evoked potentials inatients with PD.4,5 The foveal vision alterations seem to beaused by dysfunction of the intraretinal dopaminergic cir-uitry and final retinal output to the brain.4

Loss of ganglion cells can be detected by ocular imagingechnologies, such as optical coherence tomography (OCT)nd scanning laser polarimetry, both of which provide non-

nvasive, objective, and reproducible methods for evaluat-

2161ISSN 0161-6420/12/$–see front matterhttp://dx.doi.org/10.1016/j.ophtha.2012.05.003

Hetlfafi3mctwr

(3t

iwatsppNtqsitmqwqraociare

tisvam

wspaaaoTpSrna

Ophthalmology Volume 119, Number 10, October 2012

ing the retinal nerve fiber layer (RNFL). Several studieshave reported the importance of assessing RNFL thicknessin the early diagnosis and management of optic nerve pa-thologies, such as glaucoma,6 band atrophy with or withoutchiasmal compression, demyelinating diseases,7–10 opticneuritis,11,12 and PD.13 These OCT-based findings can re-veal changes in RNFL thickness before visual field defectsappear.14,15 This is one of a few studies that has usedspectral-domain OCT to examine RNFL axonal loss.13,16

Given the value of RNFL examination as a method of detect-ing neurodegenerative disease progression and facilitating diag-nosis of diseases such as multiple sclerosis,8,9,15,17,18 the aim ofthis study was to evaluate the ability of 2 Fourier-domain OCTdevices to detect retinal thinning and RNFL atrophy in patientswith PD and to determine the reproducibility of these parameters.Our study assessed RNFL thickness measurements in healthysubjects and subjects with PD using the 2 most commonly avail-able Fourier-domain OCT machines (Cirrus [Carl Zeiss MeditecInc., Dublin, CA] and Spectralis [Heidelberg Engineering, Inc.,Heidelberg, Germany]) and compared the results between the 2machines and between the new software for neuro-ophthalmologyevaluations of the Spectralis OCT (Nsite Axonal Analytics) andthe classic software (Glaucoma application). The correlations be-tween RNFL measurements and the reproducibility of the 3 OCTapplications were also analyzed.

Materials and Methods

This was an observational, prospective, cross-sectional study.Seventy-five patients with a definite diagnosis of PD (45 men, 30women; aged 51–75 years; mean age, 64.4 years) and 75 healthycontrols (45 men, 30 women; aged 52–76 years; mean age, 64.2years) were enrolled.

The diagnosis of PD was based on standard clinical and neu-roimaging criteria.19 Related medical records were carefully re-viewed, including the duration of the disease, the Mini-MentalState Examination (MMSE), and the treatment. The MMSE is abrief 30-point questionnaire that is used to screen for cognitiveimpairment and includes simple questions and problems in anumber of areas (time and place, repeating lists of words, arith-metic, language use, comprehension, and basic motor skills).20

Any score �25 points (of 30) is effectively normal (intact). Belowthis, scores can indicate severe (�9 points), moderate (10–20points), or mild (21–24 points) cognitive impairment. The rawscore may also need to be corrected for educational level andage.21 The MMSE test was scored in all patients by a neurologistexperienced in the diagnosis and treatment of PD at the time of aroutine 6-month clinical visit.

The disease-free controls were age- and sex-matched to thepatients with PD and were recruited from hospital staff and familymembers of healthy patients with no evidence of disease, includingneurologic disorders of any nature. Controls had no history ofocular or neurologic disease; their best-corrected visual acuity was�20/30 based on the Snellen scale. One eye from each subject wasrandomly selected and included in the study.

Exclusion criteria were the presence of significant refractiveerrors (�5 diopters of spherical equivalent refraction or 3 dioptersof astigmatism); intraocular pressure of �21 mmHg; media opaci-fications; systemic conditions that could affect the visual system; ahistory of ocular trauma or concomitant ocular diseases, includinga history of retinal pathology; glaucoma; and laser therapy or

ocular pathologies affecting the cornea, lens, retina, or optic nerve. c

2162

All procedures adhered to the tenets of the Declaration ofelsinki, and the experimental protocol was approved by the local

thics committee. All subjects provided informed consent to par-icipate in the study and underwent a complete neuro-ophthalmo-ogic evaluation that included pupillary, anterior segment, andundoscopic examinations; assessment of best-corrected visualcuity relative to the Snellen scale; and assessment of the visualeld. Three repetitions of scans were performed using each of thefollowing RNFL analysis protocols: the Cirrus HD-OCT instru-ent (Carl Zeiss Meditec Inc.), the RNFL protocol of the Glau-

oma application, and the RNFL-N Axonal Analytics of the Spec-ralis OCT instrument (Heidelberg Engineering, Inc.). Each eyeas considered separately, and only 1 eye of each subject was

andomly selected and included in the study.The visual field was assessed using a Humphrey Field Analyzer

Carl-Zeiss Meditec Inc.). A SITA Standard Strategy (program0-2) was used, with the recorded parameters being mean devia-ion (dB), pattern standard deviation, and pattern of defect.

The OCT tests were performed to obtain measurements of peripap-llary RNFL using the Cirrus and Spectralis OCT devices, both of whichere used in random order to prevent any effect of fatigue bias. In

ddition, all subjects were evaluated using the macular cube 512�128 ofhe Cirrus OCT device. The same experienced operator performed allcans. Between scan acquisitions, there was a time delay and subjectosition and focus were randomly disrupted, meaning that alignmentarameters had to be newly adjusted at the start of each image acquisition.o manual correction was applied to the OCT output. An internal fixation

arget was used because it provides the highest reproducibility.22 Theuality of the scans was assessed before the analysis, and poor-qualitycans were rejected. The Cirrus OCT instrument determines the quality ofmages using the signal strength measurement that combines the signal-o-noise ratio with the uniformity of the signal within a scan and iseasured on a scale of 1 to 10, where 1 is categorized as poor image

uality and 10 as excellent image quality. Only images with a score �7ere evaluated in our study. The Spectralis OCT instrument uses the blueuality bar in the image to indicate the signal strength. The quality scoreanges from 0 (poor quality) to 40 (excellent quality). Only images with

score �25 were analyzed. Three series of good-quality scans werebtained for each option. Only 2 patients were excluded because aentered scan could not be acquired because of poor fixation. Fifteenmages with artifacts or missing parts, or showing seemingly distortednatomy, were excluded.23 To obtain good-quality and centered images,epeat scan acquisition using the Cirrus OCT device was required in 8yes and the Spectralis OCT device in 7 eyes.

By following the recommended procedure for scan acquisition,he subject’s pupil was first centered and focused in an Iris View-ng camera on the system data-acquisition screen, and then theystem’s line-scanning ophthalmoscope was used to optimize theiew of the retina. The OCT scan was aligned to the proper depthnd patient fixation, and system polarization was optimized toaximize the OCT signal.

Three repetitions of optic disc cube 200�200 scans in each eyeere performed using the Cirrus HD-OCT device. In each series of

cans, mean RNFL thickness and quadrant RNFL thickness (su-erior, inferior, temporal, and nasal) were analyzed. Three imagecquisitions using circular peripapillary Spectralis OCT scans (3cquisitions with the RNFL protocol of the classic Glaucomapplication and another 3 acquisitions with the RNFL-N protocolf Nsite Axonal Analytics) were performed for all subjects usingruTrack eye-tracking technology. The mean number of scans toroduce each circular B scan was 9. Both RNFL protocols of thepectralis generate a thickness map with mean thickness, 4 quad-ants (superior, nasal, inferior, and temporal), and 6 sector thick-esses (superonasal, nasal, inferonasal, inferotemporal, temporal,nd superotemporal in the clockwise direction for the right eye and

ounterclockwise for the left eye). The presence of defects in the

astpniTitt

class

Garcia-Martin et al � Fourier-Domain OCT with Parkinson’s Disease

RNFL is provided by the comparison of measurements from eachpatient with the normative database of each instrument.

In the new Nsite Axonal Analytics system, Heidelberg Engi-neering, Inc., has incorporated the fovea-to-disc technology thatcorrectly orients the anatomy for papillomacular bundle (PMB)measurement accuracy and minimizes variability due to patienthead orientation and comparison with the normative data (Table1). The RNFL-N system places the temporal region of the scan inthe center of the viewing window for better analysis of axonal lossin the PMB, as observed in patients with neurologic diseases such

Table 1. Classification Notation and Color SchCoherence

Condition

RNFL thickness � 1st percentile1st percentile � RNFL thickness � 5th percentile5th percentile � RNFL thickness � 95th percentile95th percentile � RNFL thickness � 99th percentile99th percentile � RNFL thickness

nth percentile � nth percentile of the normal distributThe section classification displays the classification of rehigh thickness values compared with the normative RNcolored bands of the thickness profile graph and in the

Figure 1. Representation and interpretation of an axonal report usingtomography (OCT) device (Heidelberg Engineering, Inc., Heidelberg, Germ

eye; OS � left eye; SUP � superior; TMP � temporal.

s multiple sclerosis. The RNFL thickness graph for the RNFL-Ncans displays the scan results in the order of nasal - inferior -emporal - superior - nasal sectors (Fig 1). This protocol alsorovides 2 new neuro-ophthalmologic parameters: the PMB thick-ess and the nasal/temporal (N/T) ratio. The additional PMB sectors displayed and classified in the pie chart of the thickness map.he parameter N/T ratio is defined as the mean RNFL thickness

n the nasal quadrant divided by the mean RNFL thickness in theemporal quadrant. The measured N/T ratio is classified accordingo the rules shown in Table 1. One acquisition of macular cube

of Nsite Axonal Analytics of Spectralis Opticalography

Classification Result Color Scheme

Below normal limits RedBorderline low YellowWithin normal limits GreenBorderline high BlueAbove normal limits Purple

nerve fiber layer (RNFL) thickness for too low and toohickness data. The color scheme above is used in theification pie chart in Fig 1.

site Axonal Analytics application of the Spectralis optical coherencewith fovea-to-disc technology. INF � inferior; NAS � nasal; OD � right

emeTom

ion.tinalFL t

the Nany)

2163

oSciNbptofne

mscrg

eblnaTaP(

wtCbt2ot

CMLTRStR

Ophthalmology Volume 119, Number 10, October 2012

512�128 scan in each eye was performed using the Cirrus OCTinstrument, and macular volume was registered.

Statistical analyses were performed using SPSS software (ver-sion 19.0, SPSS Inc., Chicago, IL). The Kolmogorov–Smirnov testwas used to assess sample distribution. The RNFL thicknesseswere compared between patients and healthy controls using aStudent t test when normally distributed. Differences betweenCirrus and Spectralis RNFL measurements in each group were alsocompared using a Student t test for paired data. P values less than0.05 were considered indicative of statistically significant differ-ences. The relationships between the measurements obtained usingeach OCT protocol and between RNFL average thickness and PDseverity (measured with MMSE punctuation) were evaluated usingthe Pearson correlation analysis test.

For each parameter, the coefficient of variation (COV) was calculatedas the standard deviation divided by the mean of the measurement valueand expressed as a percentage. Most authors consider that devices with aCOV less than 10% have high reproducibility, whereas a COV less than5% indicates very high reproducibility.17 To assess the reliability of therepeated measurements, the intraclass correlation coefficients for absoluteagreement were calculated. These are used to measure the concordancefor continuous variables and correct correlations for systematic bias.The intraclass correlation coefficient interpretation that we used was slightreliability (for values 0–0.2), fair reliability (for values 0.21–0.4), mod-erate reliability (for values 0.41–0.6), substantial reliability (for values0.61–0.8), and almost perfect reliability (values of �0.81). Bland–Altman plots were used to assess agreement.

Results

Epidemiologic and disease characteristics of patients with PD andhealthy subjects are shown in Table 2. Age, sex, and intraocularpressure were not significantly different between the 2 groups. Theduration of PD ranged from 6 months to 17 years, with a medianof 7.5 years since diagnosis. Mean MMSE score was 28.9�8.3.Visual functional parameters, best-corrected visual acuity, meandeviation, and pattern standard deviation of visual field differedsignificantly (P � 0.001) between patients with PD and healthycontrols. Macular volume did not differ between both groups(Table 3 available at http://aaojournal.org).

Retinal Nerve Fiber Layer Thickness Comparisonbetween Eyes of Patients with Parkinson’s Diseaseand Healthy SubjectsWe compared RNFL parameters between eyes of healthy subjects

Table 2. Epidemiologic and Disease CharacteriHealthy Subjects, an

Healthy S

No. 75Age (yrs), range 64.2 (52Male:female (% men) 30:20 (65Intraocular pressure 14.2 (1.BCVA (Snellen scale), mean (SD) 0.95 (0.MD visual field (dB), mean (SD) �0.8 (1.PSD visual field (dB), mean (SD) 2.2 (0.Disease duration (yrs), mean (SD) —MMSE score, mean (range) —

BCVA � best-corrected visual acuity; MD, mean devpattern standard deviation; SD � standard deviation.

and patients with PD. Mean RNFL thickness measurements, based t

2164

n 3 individual scans, were used for the analysis. The Cirrus andpectralis OCT measurements indicated that RNFL thinning oc-urred in eyes of patients with PD. Both devices detected signif-cant differences in RNFL thickness between the 2 groups, but thesite axonal application of the Spectralis OCT instrument showedetter sensitivity for detecting axonal degeneration in PD com-ared with the Cirrus OCT device and the Glaucoma application ofhe Spectralis OCT device (Table 3, available at http://aaojournal.rg). By using the Cirrus OCT instrument, statistical RNFL dif-erences were observed in the inferior quadrant and mean thick-ess, where the RNFL thickness was 9.51 and 6.6 �m lower inyes of patients with PD than in control eyes, respectively.

With the Glaucoma application of the Spectralis OCT instru-ent, mean thickness, inferotemporal, and superotemporal RNFL

ectors were significantly different in eyes of patients with PDompared with eyes of healthy subjects. The largest difference wasegistered in the inferotemporal sector (13.49 �m lower in the PDroup; P � 0.014).

By using the axonal RNFL-N application, all the parameters,xcept the inferonasal and nasal sectors, were significantly differentetween the healthy and PD groups. Mean thickness was 4.01 �mower in the PD group (P � 0.001). The differences in the 2 neweuro-ophthalmologic parameters provided by the axonal RNFL-Npplication were also statistically significant between both groups:he PMB thickness was 0.69 �m lower in the PD group (P � 0.042),nd the N/T index had a higher reduction of temporal sectors in theD group (1.11 in PD group vs. 1.01 in healthy group; P � 0.007)Table 3, available at http://aaojournal.org).

Figure 2A shows RNFL thickness differences between thoseith PD and disease-free controls for the mean thickness and the

hickness of every optic nerve quadrant as measured with theirrus OCT device. Figure 2B and C show differences betweenoth groups measured with the Glaucoma application of the Spec-ralis OCT device (Fig 2B) and the Nsite Axonal Analytics (FigC). Retinal nerve fiber layer atrophy in patients with PD can bebserved for all measurements, but the Nsite Axonal Analytics hadhe highest sensitivity (Table 3, available at http://aaojournal.org).

orrelation Analysis and Comparison betweeneasurements Provided by 3 Retinal Nerve Fiberayer Protocolshe RNFL Cirrus measurements had significant correlations with theNFL thicknesses provided by the Glaucoma application of thepectralis OCT device (r � 0.885, P � 0.001, for RNFL mean

hickness) and by Nsite Axonal Analytics (r � 0.878; P � 0.001, forNFL mean thickness). We also observed good correlations between

of 75 Patients with Parkinson’s Disease and 75tistical Significance

ts Patients with Parkinson’s P

75 —64.4 (51–75) 0.896

30:20 (66) 0.66213.9 (2.1) 0.4780.74 (0.18) �0.001

�5.5 (4.3) �0.0014.1 (2.3) �0.0017.5 (2.1) —

28.9 (8.3) —

; MMSE � Mini-Mental State Examination; PSD �

sticsd Sta

ubjec

–76).6)

7)08)5)1)

iation

he measurements of both the Spectralis applications (Fig 3, available

Op8tt

SrC(�a

StCct

psadAha(

D

SbrFmltgbp

cbAdtteaosSsid

sOric

Garcia-Martin et al � Fourier-Domain OCT with Parkinson’s Disease

at http://aaojournal.org). We did not find a significant correlationbetween RNFL average thicknesses, provided by 3 protocols, andMMSE punctuation (r � 0.646, P � 0.345 with the Cirrus OCTdevice; r � 0.707, P�0.569 with the Glaucoma application of theSpectralis OCT device; r � 0.690, P�0.209 with the Nsite AxonalAnalytics of the Spectralis OCT device). However, mean thicknessvalues obtained with the 3 tomography protocols differed significantly(analysis of variance, P � 0.038).

Repeatability of Cirrus and Spectralis OpticalCoherence Tomography in Eyes of Patients withParkinson’s Disease and Healthy Subjects

The RNFL thickness measurements showed good COV and intra-class correlation coefficients in patients with PD (Table 4, avail-

Figure 2. Box plots comparing retinal nerve fiber layer (RNFL) thick-nesses in eyes from patients with Parkinson’s disease (n � 75) and healthysubjects (n � 75) using the Cirrus optical coherence tomography (OCT)device (Carl Zeiss Meditec Inc., Dublin, CA) (A), RNFL Glaucomaanalytic of the Spectralis OCT device (B), and Nsite RNFL AxonalAnalytics of the Spectralis OCT device (C).

able at http://aaojournal.org). The results obtained using Cirrus i

CT were highly reproducible for all quadrants and sectors inatients with PD, with a mean COV of 5.38�1.6% (range, 2.10%–.34%) and intraclass correlation coefficients �0.807. The meanhickness value had the lowest variability (COV�2.10% and in-raclass correlation coefficients � 0.969).

Measurements performed using the Glaucoma application of thepectralis OCT device had the best reproducibility for all the quad-ants and sectors in healthy subjects and patients with PD. The meanOV using this application in patients with PD was 2.35�1.1%

range, 1.03%–2.95%), and the intraclass correlation coefficients were0.962. The mean thickness had the lowest variability (COV�1.03%

nd intraclass correlation coefficients � 0.987).The results obtained using the Nsite Axonal Analytics of the

pectralis OCT device were also highly reproducible but lowerhan those obtained with the Glaucoma application. The meanOV value was 4.20�2.5% (range, 1.84%–7.87%), and the intra-lass correlation coefficients were �0.820 in the PD group. Meanhickness had the lowest variability (COV�1.84%).

Figure 4 (available at http://aaojournal.org) shows Bland–Altmanlots of RNFL mean thickness reproducibility between different mea-urements using the Cirrus OCT device (Fig 4A; available at http://aojournal.org), the Glaucoma application of the Spectralis OCTevice (Fig 4B; available at http://aaojournal.org), and the Nsitexonal Analytics of the Spectralis OCT device (Fig 4C; available atttp://aaojournal.org). In almost all parameters, there was less vari-bility in the eyes of healthy subjects than in eyes of patients with PDTable 4, available at http://aaojournal.org).

iscussion

everal studies suggest that the Spectralis OCT device hasetter reproducibility and can detect retinal pathologies moreeadily than the Cirrus HD-OCT device.17,24 Various reports ofourier-domain OCT RNFL and retinal thickness measure-ents in healthy and pathologic subjects have been pub-

ished. Several studies have been published comparinghe Spectralis and Cirrus OCT devices in eyes of those withlaucoma25 or multiple sclerosis.17 However, a comparisonetween Fourier-domain OCT measurements on eyes fromatients with PD has not been reported.

To our knowledge, studies comparing axonal and glau-oma applications of the Spectralis OCT device have noteen published. The RNFL-N measurements, Nsite Axonalnalytics evaluations, and PMB thickness have not beenescribed in healthy or pathologic eyes. Our study evaluatedhe ability of the new Nsite Axonal Analytics application ofhe Spectralis OCT instrument to detect RNFL changes inyes of healthy subjects and patients with PD. The studylso compares measurements obtained using the Nsite Ax-nal Analytics application of the Spectralis with the mea-urements provided by the Glaucoma application of thepectralis and Cirrus OCT devices. Our findings demon-trate that the new Fourier-domain OCT technology, whichncludes the Cirrus and Spectralis OCT devices, is useful foretecting axonal defects in patients with PD.

Altintas et al26 demonstrated a relationship between PDeverity and fovea thickness alterations using time-domainCT, but RNFL measurements did not have a similar cor-

elation. Braak et al1 studied the distribution of Lewy bodiesn patients with PD and suggested that the disease begins inlearly defined induction sites and advances in a topograph-

cally predictable sequence, so during early stages the in-

2165

woiCbtmdpedtvlmtleoflosnsseuos

OdARmspnRt

R

Ophthalmology Volume 119, Number 10, October 2012

clusion body pathology remains confined to the medullaoblongata and olfactory bulb, but as the disease progresses,the substantia nigra and other nuclear gray areas of themidbrain and basal forebrain are the focus of severechanges. Nevertheless, the presence of Lewy bodies in theRNFL has not been histologically verified.

Parkinson’s disease causes the death of pigmented dopa-mine neurons in the substantia nigra, as well as neuronalloss in other dopaminergic neurons, such as retinal gangliacells and higher visual areas that contain dopaminergic cells(the lateral geniculate nucleus, the cholinergic nucleus basa-lis of Meynert, and the visual cortex).26 Progressive retinaldopaminergic deficiency causes loss of retinal amacrinecells, which provide input to retinal ganglion cells. Thisreduction of retinal cells causes the corresponding decreasein RNFL thickness that OCT detects in patients with PD.However, Jindahra et al27 and Reich et al28 found evidencethat RNFL loss can occur with retro-geniculate lesions,indicating that OCT measurements may reflect a mixture ofanterior and posterior visual pathway disease.

Because of the distribution of ganglionar cell fibers in theoptic nerve head, the inferior quadrant undergoes a slightphysiologic alterations with age, so a decrease in its thick-ness suggests an underlying pathology, such as chronicglaucoma.29 The temporal quadrant is the sector that isaffected the most in early neurodegenerative diseases14; thefibers of the temporal quadrant follow the PMB. Our resultsagree with these findings: The RNFL-N parameters showedthat the PMB thickness decreased in patients with PD andthat the temporal sector was more susceptible (N/T indexwas higher in the PD group because the decrease in theRNFL thickness affected the temporal quadrant more thanthe other quadrants). However, we found more affectationof the inferior quadrant in patients with PD when using theCirrus OCT system. The distribution of dopamine in retinalayers combined with the acquisition technology of theCirrus system may explicate this finding.

In recent years, many new instruments have been introduced toquantify retinal ganglion cells, leading some authors to suggestthat changes in the RNFL may reflect similar pathologic changesoccurring elsewhere in the brain.7,10 Ocular imaging technologies,such as OCT, scanning laser polarimetry, and confocal scanninglaser ophthalmoscopy, allow the axonal constituents of the ante-rior visual pathway to be observed, thereby allowing direct visu-alization of part of the central nervous system. Because the RNFLcomprises only unmyelinated axons, measuring the RNFL thick-ness may be a method of monitoring axonal loss in patients withPD.13,16 A good correlation between RNFL thickness and mag-netic resonance imaging measurements of the brain, such as theparenchymal fraction and brain volumes, has been described inpatients with multiple sclerosis, and mean RNFL thickness isstrongly associated with normalized brain volume.7,30 We foundRNFL thinning in patients with PD, and the Nsite Axonal Ana-lytics application of the Spectralis OCT device seems to be themost sensitive for detecting this subclinical atrophy, but prospec-tive studies are needed to analyze the capability of RNFL mea-surements to function as a progression biomarker in patientswith PD.

We found significant differences between OCT applica-

tions and devices, so these results should be considered

2166

hen following the changes in a patient because an increaser reduction in RNFL or retinal thickness may be due to thenstrument used instead of actual pathologic changes. Theirrus OCT device uses the internal limiting membrane toe the anterior limit of the RNFL and the posterior border ashe posterior RNFL limit. In contrast, thickness measure-ents using the Spectralis OCT device are derived from

elineation of the anterior (internal limiting membrane) andosterior borders along a single A-scan at the appropriateccentricity within each radial B-scan. This eccentricity wasetermined to be equivalent to 1400 �m from the center ofhe optic nerve head, measured using a ruler within the OCTisualization software. When converting angular span toinear distance, the Spectralis instrument assumes an em-etropic human eye with an average axial length.31 Oph-

halmologists are incorporating the new Nsite Axonal Ana-ytics application in their clinical practice, especially forvaluating neuro-ophthalmic patients, so reports such as thisne may help them to interpret results and changes. Weound significant differences between the standard Spectra-is Glaucoma application measurements and the NSite Ax-nal application, although both applications measure theame peripapillary area. Differences in acquisition tech-iques may be the cause of this disparity, although moretudies comparing populations with several pathologieshould be performed to analyze and interpret these differ-nces. Nevertheless, the same tomography device should besed to evaluate the RNFL of patients to detect progressionr changes in disease, as several authors have previouslyuggested for other tomography devices.32–34

In conclusion, our study suggests that Fourier-domainCT devices, such as the Cirrus and Spectralis, are able toetect axonal atrophy in the RNFL of patients with PD.lthough the OCT variability was low, any reduction inNFL thickness should be carefully evaluated because itay be caused by device variability rather than PD progres-

ion. This limitation may be reduced using serial studies anderiodic RNFL evaluations. Longer prospective studies areeeded using Fourier-domain OCT to analyze the ability ofNFL thickness measurements to detect axonal degenera-

ion caused by the progression of PD.

eferences

1. Braak H, Del Tredici K, Bratzke H, et al. Staging of theintracerebral inclusion body pathology associated with idio-pathic Parkinson’s disease (preclinical and clinical stages).J Neurol 2002;249(Suppl):III/1–5.

2. Jankovic J. Parkinson’s disease: clinical features and diagno-sis. J Neurol Neurosurg Psychiatry 2008;79:368–76.

3. Park A, Stacy M. Non-motor symptoms in Parkinson’s dis-ease. J Neurol 2009;256(Suppl):293–8.

4. Bodis-Wollner I. Retinopathy in Parkinson disease. J NeuralTransm 2009;116:1493–501.

5. Price MJ, Feldman RG, Adelberg D, Kayne H. Abnormalitiesin color vision and contrast sensitivity in Parkinson’s disease.Neurology 1992;42:887–90.

6. Budenz DL, Chang RT, Huang X, et al. Reproducibility ofretinal nerve fiber thickness measurements using the StratusOCT in normal and glaucomatous eyes. Invest Ophthalmol

Vis Sci 2005;46:2440–3.

2

2

2

2

2

2

2

3

3

3

3

3

Garcia-Martin et al � Fourier-Domain OCT with Parkinson’s Disease

7. Gordon-Lipkin E, Chodkowski B, Reich DS, et al. Retinalnerve fiber layer is associated with brain atrophy in multiplesclerosis. Neurology 2007;69:1603–9.

8. Garcia-Martin E, Pueyo V, Martin J, et al. Progressivechanges in the retinal nerve fiber layer in patients with mul-tiple sclerosis. Eur J Ophthalmol 2010;20:167–73.

9. Garcia-Martin E, Pablo LE, Herrero R, et al. Diagnostic abilityof a linear discriminant function for Spectral domain opticalcoherence tomography in multiple sclerosis patients. Ophthal-mology 2012;119:1705–11.

10. Pulicken M, Gordon-Lipkin E, Balcer LJ, et al. Optical co-herence tomography and disease subtype in multiple sclerosis.Neurology 2007;69:2085–92.

11. Garcia-Martin E, Pueyo V, Ara J, et al. Effect of optic neuritison progressive axonal damage in multiple sclerosis patients.Mult Scler 2011;17:830–7.

12. Ratchford JN, Quigg ME, Conger A, et al. Optical coherencetomography helps differentiate neuromyelitis optica and MSoptic neuropathies. Neurology 2009;73:302–8.

13. Inzelberg R, Ramirez JA, Nisipeanu P, Ophir A. Retinal nerve fiberlayer thinning in Parkinson disease. Vision Res 2004;44:2793–7.

14. Garcia-Martin E, Pueyo V, Almarcegui C, et al. Risk factors forprogressive axonal degeneration of the retinal nerve fibre layer inmultiple sclerosis patients. Br J Ophthalmol 2011;95:1577–82.

15. Garcia-Martin E, Fuertes-Lazaro I, Fernández Tirado FJ, Pablo-Julvez LE. Usefulness of the new spectral-domain optical coherencetomography (SD-OCT) devices in the study of degenerative demen-tias [in Spanish]. Arch Soc Esp Oftalmol 2011;86:347–50.

16. Hajee ME, March WF, Lazzaro DR, et al. Inner retinal layer thin-ning in Parkinson disease. Arch Ophthalmol 2009;127:737–41.

17. Garcia-Martin E, Pueyo V, Pinilla I, et al. Fourier-domainOCT in multiple sclerosis patients: reproducibility and abilityto detect retinal nerve fiber layer atrophy. Invest OphthalmolVis Sci 2011;52:4124–31.

18. Moschos MM, Tagaris G, Markopoulos I, et al. Morphologicchanges and functional retinal impairment in patients with Parkinsondisease without visual loss. Eur J Ophthalmol 2011;21:24–9.

19. Gelb DJ, Oliver E, Gilman S. Diagnostic criteria for Parkinsondisease. Arch Neurol 1999;56:33–9.

20. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: apractical method for grading the cognitive state of patients forthe clinician. J Psychiatric Res 1975;12:189–98.

21. Crum RM, Anthony JC, Bassett SS, Folstein MF. Population-based norms for the Mini-Mental State Examination by ageand educational level. JAMA 1993;269:2386–91.

22. Schuman JS, Pedut-Kloizman T, Hertzmark E, et al. Repro-

ducibility of nerve fiber layer thickness measurements using

Footnotes and Financial Disclosures

Spain.

FTd

AiM

Cey

optical coherence tomography. Ophthalmology 1996;103:1889 –98.

3. Wu Z, Huang J, Dustin L, Sadda SR. Signal strength is animportant determinant of accuracy of nerve fiber layer thick-ness measurement by optical coherence tomography. J Glau-coma 2009;18:213–6.

4. Garcia-Martin E, Pinilla I, Sancho E, et al. OCT in retinitispigmentosa: reproducibility and capacity to detect macularand retinal nerve fiber layer thickness alterations. Retina 2012Mar 29. [Epub ahead of print].

5. Vizzeri G, Balasubramanian M, Bowd C, et al. Spectral do-main-optical coherence tomography to detect localized retinalnerve fiber layer defects in glaucomatous eyes. Opt Express[serial online] 2009;17:4004–18. Available at: http://www.opticsinfobase.org/oe/abstract.cfm?uri�oe-17-5-4004. Ac-cessed April 14, 2012.

6. Altintas O, Iseri P, Ozkan B, Caglar Y. Correlation betweenretinal morphological and functional findings and clinical sever-ity in Parkinson’s disease. Doc Ophthalmol 2008;116:137–46.

7. Jindahra P, Petrie A, Plant GT. Retrograde trans-synapticretinal ganglion cell loss identified by optical coherence to-mography. Brain 2009;132:628–34.

8. Reich DS, Smith SA, Gordon-Lipkin EM, et al. Damage to theoptic radiation in multiple sclerosis is associated with retinalinjury and visual disability. Arch Neurol 2009;66:998–1006.

9. Parikh RS, Parikh SR, Sekhar GC, et al. Normal age-relateddecay of retinal nerve fiber layer thickness. Ophthalmology2007;114:921–6.

0. Grazioli E, Zivadinov R, Weinstock-Guttman B, et al. Retinalnerve fiber layer thickness is associated with brain MRI out-comes in multiple sclerosis. J Neurol Sci 2008;268:12–7.

1. Fortune B, Cull GA, Burgoyne CF. Relative course of retinalnerve fiber layer birefringence and thickness and retinal func-tion changes after optic nerve transection. Invest OphthalmolVis Sci 2008;49:4444–52.

2. Knight OJ, Chang RT, Feuer WJ, Budenz DL. Comparison ofretinal nerve fiber layer measurements using time domain andspectral domain optical coherent tomography. Ophthalmology2009;116:1271–7.

3. Leung CK, Cheung CY, Weinreb RN, et al. Retinal nerve fiberlayer imaging with spectral-domain optical coherencetomography: a variability and diagnostic performance study.Ophthalmology 2009;116:1257–63.

4. Sung KR, Kim DY, Park SB, Kook MS. Comparison of retinalnerve fiber layer thickness measured by Cirrus HD and Stratusoptical coherence tomography. Ophthalmology 2009;116:

1264–70.

Originally received: February 4, 2012.Final revision: May 1, 2012.Accepted: May 2, 2012.Available online: June 28, 2012. Manuscript no. 2012-166.1 Department of Ophthalmology, Miguel Servet University Hospital, Zara-goza, Spain.2 Aragonés Institute of Health Sciences, Zaragoza, Spain.3 Department of Neurology, Miguel Servet University Hospital, Zaragoza,

inancial Disclosure(s):he author(s) have no proprietary or commercial interest in any materialsiscussed in this article.

ll subjects gave detailed consent to participate in this study, which was conductedn accordance with the guidelines established by the Ethics Committee of the

iguel Servet Hospital and based on the principles of the Declaration of Helsinki.

orrespondence: Elena Garcia-Martin, PhD, C/Padre Arrupe, Consultasxternas de Oftalmología, 50009 Zaragoza, Spain. E-mail: egmvivax@

ahoo.com.

2167