Chromatic Multifocal Pupillometer for Objective Perimetry and … · 2017-06-12 · white and chromatic stimuli arranged in a dartboard pattern was developed (nuCoria Field Analyzer;

Chromatic Multifocal Pupillometer forObjective Perimetry and Diagnosis ofPatients with Retinitis Pigmentosa

Ron Chibel, MSc,1,2 Ifat Sher, PhD,1,2 Daniel Ben Ner, BA,1,2 Mohamad O. Mhajna, MD,1 Asaf Achiron, MD,3

Soad Hajyahia, MD,1 Alon Skaat, MD,1,2 Yakir Berchenko, PhD,4 Bernice Oberman, MSc,4

Ofra Kalter-Leibovici, MD,2,5 Laurence Freedman, PhD,4 Ygal Rotenstreich, MD1,2

Purpose: To assess visual field (VF) defects and retinal function objectively in healthy participants and pa-tients with retinitis pigmentosa (RP) using a chromatic multifocal pupillometer.

Design: Cross-sectional study.Participants: The right eyes of 16 healthy participants and 13 RP patients.Methods: Pupil responses to red and blue light (peak, 485 and 625 nm, respectively) presented by 76

light-emitting diodes, 1.8-mm spot size at different locations of a 16.2� VF were recorded. Subjective VFs of RPpatients were determined using chromatic dark-adapted Goldmann VFs (CDA-GVFs). Six healthy participantsunderwent 2 pupillometer examinations to determine testeretest reliability.

Main Outcome Measures: Three parameters of pupil contraction were determined automatically: percent-age of change of pupil size (PPC), maximum contraction velocity (MCV; in pixels per second), and latency of MCV(LMCV; in seconds). The fraction of functional VF was determined by CDA-GVF.

Results: In healthy participants, higher PPC and MCV were measured in response to blue compared withred light. The LMCV in response to blue light was relatively constant throughout the VF. Healthy participantsdemonstrated higher PPC and MCV and shorter LMCV in central compared with peripheral test points inresponse to red light. Testeretest correlation coefficients were 0.7 for PPC and 0.5 for MCV. In RP patients, testpoint in which the PPC and MCV were lower than 4 standard errors from the mean of healthy participantscorrelated with areas that were indicated as nonseeing by CDA-GVF. The mean absolute deviation in LMCVparameter in response to the red light between different test point was significantly higher in RP patients (range,0.16e0.47) than in healthy participants (range, 0.02e0.16; P < 0.0001) and indicated its usefulness as adiagnostic tool with high sensitivity and specificity (area under the receiver operating characteristic curve (AUC),0.97, ManneWhitneyeWilcoxon analysis). Randomly reducing the number of test points to a total of 15 pointsdid not significantly reduce the AUC in RP diagnosis based on this parameter.

Conclusions: This study demonstrates the feasibility of using a chromatic multifocal pupillometer forobjective diagnosis of RP and assessment of VF defects. Ophthalmology 2016;123:1898-1911 ª 2016 by theAmerican Academy of Ophthalmology.

Visual field (VF) testing is part of the current clinicalstandard for evaluating retinal degeneration and optic nervedamage.1,2 Dark-adapted Goldmann perimetry and auto-mated perimetry are used most commonly for detectingand monitoring patients with retinitis pigmentosa (RP).2,3

These methods bear significant limitations because theyare subjective by nature and rely heavily on subject coop-eration and attention. Hence, testing of young children,the elderly, and individuals with impaired communicationskills is doomed to yield unreliable results. These testsalso may be stressful for patients because they needto make conscious decisions on identification of near-threshold stimuli that appear rapidly and disappear.4

Moreover, test results may be affected by the patient’sfatigue, wakefulness, and attentiveness during the long

1898 � 2016 by the American Academy of OphthalmologyPublished by Elsevier Inc.

Downloaded from ClinicalKey.com at MACHBAFor personal use only. No other uses without permission.

procedure. Therefore, constant monitoring and instructionof participants by qualified personnel are needed to obtainreliable results.4 Furthermore, testeretest variability, inparticular in peripheral locations and in regions of VFdeficits, makes it difficult to determine whether the VF isworsening over the course of serial examinations.5e8

Hence, frequent examinations are needed and misdiag-nosis of early stages is common.7,9e11

Several attempts have been made to establish objectiveperimetry based on pupil light response to perimetric lightstimuli. Harms12 was the first to report pupilloperimetry in1949. More studies followed using perimetric white-lightstimuli with some success. However, specificity and sensi-tivity were not sufficiently high to be of clinical use.13e23

More recently, multifocal pupillographic perimetry using

http://dx.doi.org/10.1016/j.ophtha.2016.05.038ISSN 0161-6420/16

-Tel Aviv University September 04, 2016. Copyright ©2016. Elsevier Inc. All rights reserved.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.ophtha.2016.05.038&domain=pdf

http://dx.doi.org/10.1016/j.ophtha.2016.05.038

Chibel et al � Chromatic Multifocal Pupillometry

white and chromatic stimuli arranged in a dartboard patternwas developed (nuCoria Field Analyzer; nuCoria Pty. Ltd.,Acton, Australia).24e27 This method analyzes both eyessimultaneously, and although it is promising, it cannotdifferentiate between the rod and cone systems.

Retinitis pigmentosa encompasses a group of progressiveretinal degeneration diseases that predominantly affect therod photoreceptor system, resulting in night blindness inthe early phase of the disease and loss of peripheral visionthat progresses to tunnel vision. In later stages of RP,degeneration of cone photoreceptors causes progressivedecline of visual acuity.28 Disease progression is monitoredby electroretinography and perimetry.3,29,30 However, poortesteretest repeatability in patients with RP, specifically inareas with VF deficits,5,6,30,31 limits the ability to assessdisease progression and particularly to design and interpretclinical trials of potential therapeutic agents.

We recently demonstrated a proof of concept for using achromatic multifocal pupillometer for detection of VFdefects in RP patients. The pupil responses of healthyvolunteers and RP patients were recorded at 13 differentlocations of the 30� VF in response to blue- and red-lightstimuli (peak, 485 and 640 nm, respectively; light in-tensity, 40 cd/m2; target size, 64 mm2).32 Retinitispigmentosa patients demonstrated a significantly reducedpercentage of pupil contraction (PPC) compared withhealthy participants in testing conditions that emphasizedrod contribution (blue light) in nearly all VF locations.By contrast, the PPC in responses to red light (whichemphasizes cone contribution) was reduced significantlyin RP patients compared with healthy participants, mostlyin peripheral locations. In central locations, there was nosignificant difference between the PPC of RP patients andhealthy participants in response to red light. In a secondstudy, we demonstrated that RP patients demonstratedsignificantly lower PPC in response to blue light inperipheral locations of the central VF than healthyparticipants.33 Furthermore, in both studies, minimal PPCwas recorded in RP patients in areas that were notdetected in dark-adapted chromatic Goldmann peri-metry.32,33 In these studies, we evaluated only a singleparameter of the pupil response, the PPC. These studiessuggested that VF defects as well as rod and cone functionmay be assessed in RP patients using a chromatic multi-focal pupillometer.

In the current study, we examined the dynamics of thepupil response in the central VF of RP patients and healthyparticipants. We analyzed additional parameters of thepupil light response, the maximal contraction velocity(MCV), and the latency of MCV (LMCV) to determinethe effect of retinal degeneration on these parameters. Tothe best of our knowledge, this is the first report ofevaluation of the LMCV parameter in pupillometry studies.The central VF of RP patients was assessed using a chro-matic multifocal pupillometer and was compared with thepatients’ chromatic dark-adapted Goldmann VF (CDA-GVF)2 results as well as the pupillometry results of healthyparticipants. We report that RP patients demonstratedsignificantly lower PPC and MCV in areas that werereported as nonseeing by CDA-GVF and that the mean


absolute deviation in the LMCV parameter betweendifferent test point locations was significantly higher in RPpatients and may present a valuable parameter as a diag-nostic tool for RP.

Methods

Participants

The Sheba Medical Center Institutional Review Board EthicsCommittee approved this trial. The study was conducted accordingto the tenets of the Declaration of Helsinki and was registered atwww.clinicaltrials.gov (identifier, NCT02014389). Informed writ-ten consent was obtained from all participants. Sixteen healthyvolunteers, age matched with patients (see below; 6 men and 10women; mean age � standard deviation, 38.4�15.6 years; range,26e77 years) were included in the study. Inclusion criteria werenormal eye examination results, best-corrected visual acuity of20/20, normal color vision, no history of or current ocular disease,no use of any topical or systemic medications that could adverselyinfluence efferent pupil movements, and normal 24-2 Swedishinteractive threshold algorithm results, developed for the Hum-phrey standard perimeter (Humphrey Field Analyser II, Swedishinteractive threshold algorithm 24-2; Carl Zeiss Meditec, Inc., Jena,Germany).

The study patient group comprised 13 patients with RP (3women and 10 men; mean age � standard deviation, 36.15�14.6years; range, 20e65 years). Inclusion criteria for RP patients weretypical abnormal fundus appearance and previously recordedelectroretinography results that were abnormal under scotopic orphotopic conditions or both (in compliance with the protocol of theInternational Society for Clinical Electrophysiology of Vision)34

and typical abnormal kinetic chromatic Goldmann test results(loss of VF that is either concentric or that began superiorly andsubsequently demonstrated an arcuate scotoma that progressedeither from the nasal or the temporal side; incompletemidperipheral ring scotoma that broke through into theperiphery; or a residual central VF, with blue and red isoptersthat were either superimposed or in which the isopter in responseto the red stimulus was larger than that in response to the bluestimulus2,3).

Exclusion criteria were a concurrent ocular disease and anyother condition affecting the pupil response to light. Data recordedfor all patients included gender, diagnosis, and electroretinographyresponses. Patients were tested for best-corrected visual acuity andfor color vision by the Farnsworth D15 test. The right eyes of bothhealthy and RP participants were examined.

Light Stimuli

Light stimuli were presented using a Ganzfeld dome apparatus(Accutome, Inc, Malvern, PA; Fig 1) placed 330 mm from thepatient’s eye. All tests were performed in a dark room. Theuntested eye was covered. Participants were asked to fixate on awhite-light fixator (0.9 cd/m2; Fig 1B, white arrow) at the centerof the dome. Stimuli were presented from 76 targets (light-emitting diodes) with diameter of 1.8 mm2 in a VF of 16.2�. Thewavelength and intensity of light stimuli selected for this studywere 625�5 nm and 1000 cd/m2 for long-wavelength stimuli(red light) and 485�5 nm and 200 cd/m2 for short-wavelengthstimuli (blue light). The light intensities were chosen after pre-liminary calibrations that enabled us to identify the minimalstimulus intensity that yielded a substantial pupil response in 5healthy participants. Background luminance was 0.05 cd/m2. Lightintensities were determined by measurement with the LS-100

1899


http://www.clinicaltrials.gov

Figure 1. Photographs of the chromatic multifocal pupillometer. A, Side view of the device. B, Front view of the device. The black arrow points to theinfrared camera pin hole, and the white arrow points to the fixator light-emitting diode. C, The user interface for the chromatic multifocal pupillometer. D,Pupillometer-measured parameters.

Ophthalmology Volume 123, Number 9, September 2016

luminance meter (Konica Minolta, Tokyo, Japan). Stimulus dura-tion was 1 second and the interstimulus interval was 4 seconds.

Pupil Measurement

Pupil diameter was recorded in real time by a computerizedinfrared high-resolution camera (the camera pinhole is marked witha black arrow in Fig 1B) that recorded the pupil diameter at afrequency of 30 Hz. The software (Accutome, Inc.) searched forthe pupil in every image. Participants were dark adapted for3 minutes before testing. Tests in which the subject blinkedduring the first 2.5 seconds after stimulus onset were excludedautomatically, and the targets were retested.

Analysis of Pupil Responses

During the recording of the pupil diameter in response to eachlight stimulus, 5 parameters were calculated within the softwareusing the change in pupil diameter over time: the initial diameterof the pupil (in pixels), the minimum pupil diameter (in pixels),PPC, MCV (in pixels per second), and LMCV (in seconds). ThePPC was determined using the following formula, as describedpreviously32:

PPC ¼ ðInitial Pupil Diameter�Minimum Pupil DiameterÞ=ðInitial Pupil DiameterÞ� 100

The MCV was determined by calculating the maximum rate atwhich the pupil contracted from the light stimulus between theinitial pupil diameter measurement and the minimum pupil diam-eter measurement. The LMCV was determined by calculating thetime point for each pupil response at which the maximum rate ofpupil contraction (MCV) occurred from each light stimulus.

1900


Chromatic Dark-Adapted Visual Field

Retinitis pigmentosa patients were tested for kinetic VF by CDA-GVF.2 Briefly, a Goldmann perimeter (940-ST; Haag-Streit AG,Liebefeld, Switzerland) was used to map patients’ conventionaland 2-color dark-adapted VFs. Patients were dark adapted for30 minutes before testing. The setting used for stimuli were II3c forthe long-wavelength stimulus and 2 log units lower in luminance(II3 eC) for the short-wavelength stimulus. For quantification offunctional CDA-GVF, a schematic representation of the pupill-ometer target points was overlaid on the CDA-GVF test record,and pupillometer targets that were within seeing areas by the CDA-GVF were scored as 1. Pupillometer targets that were in nonseeingareas by the CDA-GVF were scored as 0. The fraction of func-tioning VF was calculated as the sum of scores divided by 76.

Statistical Analysis

Statistical analyses were performed using Excel (Microsoft,Redmond, WA) and R software version 3.0.1 (The Free SoftwareFoundation [FSF]). Results were presented as mean � standarderror (SE). Student t test was used to evaluate demographic dif-ferences between patients and controls. Testeretest reliability ofthe pupil response measurements was calculated using Pearsoncorrelation, and the correlation between pupillometer recordingsand CDA-GVF testing were calculated with Spearman r test.Variability in LMCV recordings was measured by the meanabsolute deviation, and this measure was comparedbetween the study group and the control group using a 2-sidedManneWhitneyeWilcoxon test. The effects of using fewer testpoints and the robustness of the LMCV for discrimination wasexamined via simulation by randomly selecting n test points (n ¼5, 10, 15, . 75) and calculating the area under the receiveroperating characteristic curve (AUC) obtained using the LMCV


Figure 2. Visual field results showing the mean (A, B) percentage of change of pupil size (PPC), (C, D) maximum contraction velocity (MCV), and (E, F)latency of maximum contraction velocity (LMCV) parameters of healthy participants in response to (A, C, E) red and (B, D, F) blue light at each test targetlocation, indicted by number. Color coding for each parameter is shown on the right. deg ¼ degree.


based on these n points. This was done for 200 repeats, and themean AUC for each n thus was obtained.

Results

Characterization of Pupil Responses of HealthyParticipants

First we characterized the pupil response parameters in each testpoint for red- and blue-light stimuli in control participants. Figure 2demonstrates color-coded maps of mean PPC (Fig 2A, B), MCV(Fig 2C, D), and LMCV (Fig 2E, F) recorded from controlparticipants in each test point location in response to red-lightstimuli (Fig 2A, C, E) and in response to blue-light stimuli (Fig2B, D, F). The mean PPC recorded in response to blue-lightstimuli ranged from 13% to 28% at different test point locations


(mean � SE, 19.4�0.22%) and was significantly higher comparedwith the red-light stimuli (range, 7%e22%; mean � SE, 12�0.2%;P < 0.0001, t test), even though the red-light stimuli were pre-sented at a 5-fold higher intensity than the blue-light stimuli.Similarly, the mean MCV was significantly higher in response toblue-light stimuli (range, 28e43 pixels/second; mean � SE,37�0.3 pixels/second) compared with the red-light stimuli (range,18e36 pixels/second; mean � SE, 23.17�3 pixels/second; P ¼2�10�7, t test). There was no significant difference in the meanLMCV between the red and blue light (range, 0.6e0.8 seconds;P ¼ 0.11, t test).

Higher PPC was recorded in central locations of the VFcompared with peripheral locations in response to both red andblue light (Fig 2A, B). A similar pattern of faster mean MCV incentral locations compared with peripheral locations wasdemonstrated clearly in response to red-light stimuli (Fig 2C).The centraleperipheral gradient pattern was less evident in

1901


Figure 3. Scatterplots showing the test versus retest of (A, B) percentage of change of pupil size (PPC), (C, D) maximum contraction velocity (MCV), and(E, F) latency of maximum contraction velocity (LMCV) parameters in serial testing of 6 healthy participants in response to (A, C, E) blue and (B, D, F)red light. sec ¼ second.


response to blue light (Fig 2D). The LMCV parameter wasrelatively constant throughout the VF field in response to bothred and blue light, with an interquartile range of 0.6 to 0.7seconds. The longest LMCV values were measured at peripheraltest points in response to both wavelengths (Fig 2E, F).

Test reliability was assessed by retesting 6 healthy controls.Good linear correlation between the test and the retest wasdemonstrated for the parameter PPC for both colors (blue:R2 ¼ 0.762; P < 0.0001; red: R2 ¼ 0.721; P < 0.0001; Fig 3A, B).The parameter MCV demonstrated lower but still reasonablecorrelation between test and retest (blue: R2 ¼ 0.513;P < 0.0001; red: R2 ¼ 0.522; P < 0.0001; Fig 3C, D). Thelowest correlation between test and retest was recorded for theLMCV parameter (blue: R2 ¼ 0.419; P < 0.0001; red: R2 ¼0.208; P < 0.0001; Fig 3E, F).

Retinitis Pigmentosa Patients DemonstratedDiminished Pupil Responses in Correlation withDisease Severity

Thirteen RP patients were divided into 2 groups based on theirCDA-GVF testing results. Group A consisted of 5 patients with

1902


some functional CDA-GVF in response to red and blue light,whereas group B included 8 patients with a severe disease with nodetection of either blue or red light by CDA-GVF (Table 1).Figure 4 demonstrates grey-scale maps of mean PPC (Fig4AeC), MCV (Fig 4DeF), and LMCV (Fig 4GeI) recorded ineach test point location in response to blue light in healthyparticipants (Fig 4A, D, G) and in patients with intermediateretinal degeneration (group A; Fig 4B, E, H) and severe retinaldegeneration (group B; Fig 4C, F, I). Color coding was set toresemble the Humphrey perimeter’s output, with a white colorfor normal values and darker colors for values that were lessthan normal. Normal values were set as the mean of healthyparticipants in each test point location (Figs 2 and 4). Deviationfrom normal values was determined based on the SEs calculatedfor each parameter in each target point in the healthyparticipants. Thus, for PPC and MCV, the darkest color wasused for test points in which the mean of patients was less than5 SEs from the mean of healthy participants in those points. ForLMCV, the darkest color was used for test points in which themean of patients was more than 5 SEs from the mean of healthyparticipants in those points. In group B, mean PPC and MCV inresponse to the blue light were less than 5 SEs from the mean ofhealthy participants in nearly all test point locations (Fig 4C, F).


Table

1.Su

mmaryof

PatientCharacteristics

Patient

No.

Gender

Age

(yrs)

Diagnosis

Best-Corrected

VisualAcuity

Group

*

Electroretino

graphy

Results

Isolated

Rod

Response

Dark-Adapted

A-W

ave

Dark-Adapted

B-W

ave

Single-Flash

A-W

ave

Single-Flash

B-W

ave

Maximum

(mm)

Implicit

Time(m

s)Maximum

(mm)

Implicit

Time(m

s)Maximum

(mm)

Implicit

Time(m

s)Maximum

(mm)

Implicit

Time(m

s)Maximum

(mm)

Implicit

Time(m

s)

1M

34RP

20/40

BND

ND

ND

ND

ND

ND

ND

ND

ND

ND

2M

44RP

20/360

0A

7410

095

24.5

149

3916

1626

333

F34

RP

20/25

AND

ND

1517

1142

216

3323

4F

32RP

20/32

AND

ND

ND

ND

536

ND

ND

530

5M

20US

20/360

0A

4978

1630

3451

931

526

6F

31RP

20/20

AND

ND

1519

744

716

1346

7M

27RP

20/50

BND

ND

ND

ND

ND

ND

ND

ND

ND

ND

8M

55RP

20/63

BND

ND

ND

ND

ND

ND

ND

ND

ND

ND

24M

21RP

20/64

BND

ND

614

2657

926

2238

25M

22RP

20/640

BND

ND

1517

1744

1017

1026

26M

28RP

20/20

BND

ND

ND

ND

ND

ND

ND

ND

ND

ND

29M

65RP

20/30

BND

ND

ND

ND

ND

ND

ND

ND

ND

30M

58RP

20/50

BND

ND

ND

ND

ND

ND

ND

ND

ND

ND

F¼

female;

M¼

male;

ND

¼no

tdetected;R

P¼

retinitispigm

entosa;U

S¼

Usher

synd

rome.

*Group

A¼

somelight

detectionin

respon

seto

both

redandblue

bychromaticdark-adapted

Goldm

annvisualfield;

groupB¼

nolight

detectionin

respon

seto

atleast1colorby

chromaticdark-adapted

Goldm

annvisual

field.



Similarly, the mean LMCV was more than 5 SEs from the meanof healthy participants in 68 of 76 test points (89% of theVF; Fig 4I).

By contrast, the mean PPC recorded in patients in group A wasequal to or was lower by less than 2 SEs from the mean of healthyparticipants in most of the VF test points (50 test points, 66% of theVF; Fig 4B). In 26 test points (34% of the VF) that were locatedmostly in the periphery of the VF, the mean PPC of group Bpatients was less than 2 SEs from the mean of healthyparticipants. The MCV and LMCV parameters also demonstratedan intermediate defect in pupil response. Thus, in 33 test pointlocations (43% of the VF), the mean MCV was less than 3 SEsfrom the mean of healthy participants (Fig 4E), and in 25 testpoint locations (33% of the VF), the LMCV was less than 3 SEsfrom the mean of healthy participants (Fig 4E, H).

Figure 5 demonstrates the pupil responses to red light in healthyparticipants and RP patients. Retinitis pigmentosa patients fromboth groups demonstrated lower PPC and MCV and longerLMCV compared with healthy participants, but to a smallerextent than the response to the blue light. Thus, in group B, themean PPC and MCV were less than 5 SEs from the mean ofhealthy participants in 35 (46% of the VF) and 57 (75% of theVF) test points, respectively (Fig 5C, F). Similarly, the meanLMCV was more than 5 SEs from the mean of healthyparticipants in 50 test points (66% of VF) compared with 68points (89% of the VF) in response to blue light (Fig 5I). GroupA demonstrated a milder decline in pupil responses to red lightcompared with group B, with only 4 (5% of the VF) and 9 (12%of the VF) test points in which the PPC and MCV were lessthan 5 SEs from the mean of healthy participants, respectively(Fig 5B, E). The mean LMCV was more than 5 SEs from themean of healthy participants in 35 test points (46% of the VF;Fig 5H).

Variability in Latency of Maximum ContractionVelocity as a Diagnostic Tool for RetinitisPigmentosa

As shown in Figures 2 and 4, LMCV was relatively constant inhealthy participants in response to blue and red light in most testpoint locations, ranging from 0.6 to 0.8 seconds. By contrast,this parameter was highly variable between different test pointlocations in RP patients, ranging from 0.6 to 1.7 seconds andfrom 0.6 to 1.4 seconds in response to the blue and red light,respectively (Figs 4 and 5). To evaluate the extent of thevariability in LMCV between different test point locations ofthe VF, the mean response for each subject was determined (i.e.,the mean LMCV among the subject’s 76 test points). Then, themean absolute deviation was calculated as the mean of theabsolute difference between the mean and the measurements ineach of the test points. Figure 6A demonstrates for eachparticipant a boxplot depicting the distribution of the LMCVparameter for all test points in response to the red light.The mean absolute deviation in LMCV in response to the redlight between different test points of each participant wassignificantly higher in patients with RP than in healthyparticipants (P < 10�6, ManneWhitneyeWilcoxon test). Inaddition, the ManneWhitneyeWilcoxon statistic test indicatedthat a classification method based on measurement of LMCV inresponse to red light would have an AUC of 0.97. Similarly, themean absolute deviation of LMCV in response to the blue lightbetween different test points of each participant was significantlyhigher in patients with RP than in healthy participants (P ¼10�4, WilcoxoneManneWhitney test; AUC, 0.93; data notshown). There was no significant difference in the absolute mean

1903


Figure 4. Visual field (VF) results showing the mean (AeC) percentage of change of pupil size (PPC), (DeF) maximum contraction velocity (MCV), and(GeI) latency of maximum contraction velocity (LMCV) parameters recorded in response to blue light in (A, D, G) healthy participants and in patientsfrom (B, E, H) group A and (C, F, I) group B in each of the 76 test targets of the 16.2� VF. Color coding is based on number of standard errors (SEs) fromthe mean of healthy participants and is shown on the right. White ¼ mean value equals or is less than 1 SE from the mean of healthy participants in eachlocation; dark gray ¼ mean values differs in more than 5 SEs from the mean of healthy participants in each location. deg ¼ degree.


deviation of the PPC and MCV parameters between RP patientsand healthy controls (data not shown).

The mean absolute deviation in LMCV in response to the redlight correlated negatively with the fraction of functional subjectiveVF as determined by CDA-GVF (Spearman r, �0.45; P ¼ 0.13;Fig 6B). The highest mean absolute deviation in LMCV (>0.3)was found in patients with no light detection by CDA-GVF. Bycontrast, lower mean absolute deviation in LMCV (<0.3) wasdemonstrated in patients with some functional CDA-GVF.

Clustering Analysis

Next we examined whether fewer test point locations could be usedfor RP diagnosis in an attempt to optimize the analysis ofpupillometer-based perimetry and reduce testing time. Figure 7shows the mean and standard deviation of the AUC obtainedafter a random selection of test locations. Randomly reducing thenumber of test points to a total of 15 points did not significantlyreduce the AUC in RP diagnosis based on the absolute meandeviation of LMCV. The AUC of 0.90 (highlighted by a dashed

1904


red line) suggests that the probability of LMCV to discriminatesuccessfully a (randomly selected) RP patient from a (randomlyselected) healthy subject is 0.9.

Individual Retinitis Pigmentosa Cases

To illustrate the pattern of recorded pupil response valuescompared with the results of subjective CDA-GVF testing, theindividual reports for 3 RP patients are presented in Figures 8, 9,and 10.

Patient 30: No Light Detection by Chromatic Dark-AdaptedGoldmann Visual Field. Patient 30 showed no light detection byCDA-GVF (Fig 8A, E). The PPC and MCV parameters were lessthan 5 SEs from the mean of healthy participants in 75 of the 76test points (99% of the VF) in response to blue light (Fig 8B,C). The LMCV parameter was more than 5 SEs from the meanof healthy participants in 48 of the 76 test points (63% of theVF; Fig 8D).

The pupil responses to red light also were diminished signifi-cantly throughout the VF, with 69 and 74 of the 76 test points (91%


Figure 5. Visual field (VF) results showing the mean (AeC) percentage of change of pupil size (PPC), (DeF) maximum contraction velocity (MCV), and(GeI) latency of maximum contraction velocity (LMCV) parameters in response to red light in (A, D, G) healthy participants and in patients from (B, E,H) group A and (C, F, I) group B in each of the 76 test targets of the 16.2� VF. Color coding is based on number of standard errors (SEs) from the mean ofhealthy participants and is shown on the right, as described in Figure 4. deg ¼ degree.


and 77% of the VF), respectively, presenting PPC and MCV lessthan 4 SEs from the mean of healthy participants (Fig 8F, G). TheLMCV parameter was more than 5 SEs from the mean of healthyparticipants in 38 test points (50% of the VF; Fig 8H). The meanabsolute deviation in LMCV in response to the red light forthis patient was the largest recorded in this study (0.47 second;Fig 6A).

Patient 4: Tunnel Vision by Chromatic Dark-AdaptedGoldmann Visual Field. Patient 4 demonstrated tunnel vision onCDA-GVF (Fig 9A, E). The map of the PPC parameter in responseto blue light correlated well with the CDA-GVF map. Thus, inperipheral test point locations (nonseeing by CDA-GVF), the PPCvalues were less than 5 SEs from the mean of healthy participants,whereas in central locations, PPC values were only 1 to 3 SEs lessthan the mean of healthy participants (Fig 9B). The map of MCVdemonstrated substantial reduction of MCV throughout the VF,with MCV less than 4 SEs from the mean of healthy participantsin all 76 test points (100% of the VF; Fig 9C). The LMCVparameter was 1 to 2 SEs from the mean of healthy participantsin 20 test points (26% of the VF), mostly in central locations(Fig 9D).


A similar tunnel vision pattern of pupil responses was obtainedin response to the red light. The PPC and MCV parameters wereequal to or only 1 SE from the mean of healthy participants in 14and 8 central test point locations, correlating to 18% and 11% ofthe VF, respectively. By contrast, the PPC and MCV recorded innearly all peripheral test points were more than 5 SEs less than themean of healthy participants (Fig 9F, G). The LMCV parameterwas close to normal (1e2 SEs from the mean of healthyparticipants) in 38 test points (50% of the VF). Most of thesetest points were located at the center of the VF (Fig 9H). Themean absolute deviation in LMCV in response to the red lightfor this patient was intermediate (0.26 second; Fig 6A).

Patient 3: Larger Fraction of Functional Chromatic Dark-Adapted Goldmann Visual Field. Patient 3 had a significantfunctional portion of the 16.2� VF as determined by the subjectiveCDA-GVF (Fig 10A, E). The pupil responses of this patient to boththe red and blue light were equal to or 1 to 2 SEs from the mean ofhealthy participants in most test points located in functional areasdetermined by the CDA-GVF. Most areas that demonstrated sub-stantially reduced PPC and MCV and longer LMCV comparedwith control values (more than 2 SEs from the mean of healthy

1905


Figure 6. The variability in parameter latency of maximum contraction velocity (LMCV) in response to red-light stimuli between different visualfield (VF) test point locations in retinitis pigmentosa (RP) patients correlates with their chromatic dark-adapted Goldmann VF (CDA-GVF) results. A,Box-and-whisker plot showing that RP patients demonstrate significantly higher variability in LMCV in response to red-light stimuli between different VFlocations compared with control subjects. The absolute difference between the mean LMCV (of the 76 points) and the recordings in 76 points wasdetermined for each subject (y-axis). x-axis ¼ subject number. Boxes represent the first and third quartiles and the band inside the box represents the secondquartile (the median). The whiskers extend to the most extreme data point, which is no more than 1.5 times the interquartile range from the box. B, Graphshowing negative linear correlation between the mean absolute deviation in LMCV in response to red light and CDA-GVF. y-axis ¼ the mean absolutedeviation of LMCV (of the 76 points) in response to red light. x-axis ¼ the fraction of functioning VF in response to red light that was calculated for RPpatients based on their subjective CDA-GVF results. Solid line ¼ linear regression.


participants) were located mostly in nonseeing areas by CDA-GVF(Fig 10). The mean absolute deviation in LMCV in response to thered light for this patient was small (0.16 second; Fig 6A).

Discussion

This study demonstrated the feasibility of using a chromaticmultifocal pupillometer for objective assessment of VF

Figure 7. The computational random selection analysis of the mean ab-solute deviation in latency of maximum contraction velocity parameter:the mean and standard deviation of the area under the receiver operatingcharacteristic curve (AUC) obtained after a random selection of testlocations (y-axis) versus the number of test points used (x-axis).

1906


defects in patients with RP. Test point locations inwhich PPC and MCV were less than 4 SEs from the mean ofhealthy participants correlated with areas that wereabnormal (nonseeing) by CDA-GVF. The RP patients withsevere VF loss presented more testing points that differedsubstantially from the mean of healthy participantscompared with patients with a moderate loss of VF,particularly in response to the blue-light stimuli. Patientswith some functional VF demonstrated reduced pupil lightresponse (PLR) particularly in peripheral test points and inresponse to the blue-light stimuli. The pathologic features ofRP are characterized by a loss of rod function that exceedsthe reduction of cone function, and the VF loss typicallybegins with peripheral VF constriction.3,28 Our finding thatpupil response to blue light in RP patients was more affectedthan the pupil response to red-light stimuli strongly suggeststhat the pupil response to blue light measured by ourchromatic pupillometer in response to focal light stimuli ismediated mainly by rods, whereas the pupil response to redlight is mediated mainly by cones. Hence, a chromaticmultifocal pupillometer may enable objective noninvasiveassessment of the function of rods and cones at distinctlocations of the VF for the first time. These results aresupported by our previous study that demonstrated reducedpupil responses in peripheral VF locations in response toblue light32 and by the studies from Kawasaki et al,35

Kardon et al,36,37 and Lorenz et al38 that demonstratedthat transient pupil response to full-field low-intensityblue-light stimulus reflects rod activity and that transientpupil response to red light is driven predominantly by cones.

Curcio et al39 reported that the average horizontaldiameter of the rod-free zone in the human fovea is 0.350mm (1.25�).


Figure 8. A, E, Chromatic dark-adapted Goldmann visual field (CDA-GVF) results in patient 30 demonstrated no light detection. The (B, F) percentage ofchange of pupil size (PPC), (C, G) maximum contraction velocity (MCV), and (D, H) latency of maximum contraction velocity (LMCV) parameters ofthis patient in response to (BeD) blue-light and (FeH) red-light stimuli recorded in each of the 76 points of the 16.2� visual field are presented. Colorcoding is the same as described in Figure 4. deg ¼ degree.


Because the 4 most central pupillometer test point lo-cations are located at 1.85�, the pupil responses to bluelight recorded from these 4 central test points may bemediated by foveal rods. Since the pupil responsesrecorded in this study were transient and because dim lightwas used in the current test protocol, it is less likely thatintrinsically photosensitive retinal ganglion cells areinvolved in pupil response to blue-light stimuli presentedat the 4 central test points. However, because cones arehighly concentrated in the fovea, it is possible that some ofthe pupil response to blue light in the 4 central test pointsis mediated by S cones. We plan to examine this possi-bility in a future study by testing coneerod dystrophypatients.

The intensity of blue-light stimulus used in our study was5-fold lower than that of the red-light stimulus. Neverthe-less, pupil responses to blue-light stimulus in healthy par-ticipants were stronger than the responses to red-lightstimuli in the same test point locations (Fig 2). Thesefindings correlate with those of our previous study usingthe first chromatic pupillometer prototype32 and with thefindings of Kawasaki et al,35 Kardon et al,36,37 and Lorenzet al,38 and may be explained by the lower number of conescompared with rods in the human retina39 and by the smaller


receptive fields of cones and their lower sensitivity for lightcompared with rods.

The LMCV parameter recorded in response to red-lightstimuli seems to be a useful tool for noninvasive andobjective diagnosis of RP with an AUC of 0.97. Ourcomputer-based random clustering analysis suggested thatshortening test duration may be possible with computa-tional clustering, without reducing the sensitivity and ac-curacy of RP diagnosis. Thus, testing only 15 test points inresponse to red-light stimuli, which is predicted to take 1minute, would enable diagnosis of RP with an AUC of 0.9.Importantly, our study group included RP patients atdifferent stages of the diseases, some with substantialfunctional 16.2� VF (such as patient 3) and some patientswho had no light detection (such as patient 30), furtheremphasizing the high specificity and sensitivity of theLMCV score. In coming clinical trials using a chromaticmultifocal pupillometer, we will test a larger VF (30�)using a larger cohort of patients and healthy participants.The future study also will include more thorough compu-tational optimization for clustering based on the promisingresults of random reduction of target locations. In addition,future trials are planned in which the pupillometry testingof patients at early stages and healthy participants will be

1907


Figure 9. A, E, Chromatic dark-adapted Goldmann visual field (CDA-GVF) results in patient 3 demonstrating the subjective visual field (VF) for (A) blueand (E) red light. The dashed green line marks the borders of the 16.2� VF. The (B, F) percentage of change of pupil size (PPC), (C, G) maximumcontraction velocity (MCV), and (D, H) latency of MCV (LMCV) parameters of the patient in response to (BeD) blue-light and (FeH) red-light stimulirecorded in each of the 76 points of the 16.2� VF are presented. Color coding is the same as described in Figure 4. deg ¼ degree.


performed in a masked fashion to examine the applicabilityof pupillometry testing as a diagnostic tool.

Previous studies suggested direct correlation betweenpupil response amplitude, maximum velocity, and latencyin response to full-field white-light stimulus. The authorsnoted that most of the latency in healthy participants wasthe result of delay in iris smooth muscle contraction andthat a relatively small part is the result of conductionalong the pupil reflex pathway.40 Because apoptosis inthe iris muscle increases with age, we included age-matched groups with a wide range of ages (26e77years). To the best of our knowledge, our study enablesmapping different pupil light response parameters indifferent locations of the VF in response to red and bluelight for the first time. Our findings suggest that in mostcases, maximum pupil contraction velocity correlateddirectly with change in pupil diameter. However, someexceptions were noted (for example, in patient 4). Afuture clinical trial with a larger study group may enabledetermination of the correlation between the differentpupil response parameters in this new setting of responseto focal chromatic light stimuli.

One of the major limitations of subjective perimetry istesteretest variability. A good correlation was demonstratedin PPC and MCV in testeretest analysis. The large bin size

1908


of the LMCV parameter (0.033 second) and outliers may bethe cause for the low testeretest correlation of LMCV. Afuture clinical trial will examine the test reliability inrepeated testing of a larger number of healthy participantsand patients.

The chromatic multifocal pupillometer device presentedhere was built to enable mapping of the central VF. Wechose to examine the device in RP patients because of theirpathologic features that facilitate the differentiation betweencone and rod function. Future clinical trials will includepatients with other blinding diseases, such as patients withmacular degeneration and glaucoma, using a device thatenables 30� VF testing.

The major limitation of this study is the size of thestudy cohorts. This is a preliminary study that is beingfollowed by a larger clinical trial with more types of dis-eases. Several studies demonstrated a correlation betweenstructural abnormalities on optical coherence tomography(OCT), including decreased outer nuclear layer and outersegment thickness and the loss of local VF sensitivity ofRP patients.41e44 Wen et al45 demonstrated strong positivecorrelation between cone function preservation measuredby multifocal electroretinography, VF sensitivity, and theremaining thickness of the photoreceptor layer byspectral-domain OCT in RP patients. Recent adaptive


Figure 10. Chromatic dark-adapted Goldmann visual field (CDA-GVF) results in retinitis pigmentosa patient 4 demonstrating the subjective visual field(VF) for (A) blue and (E) red light. The dashed green line marks the borders of the 16.2� VF. The (B, F) percentage of change of pupil size (PPC), (C, G)maximum contraction velocity (MCV), and (D, H) latency of MCV (LMCV) parameters of this patient in response to (BeD) blue-light and (FeH) red-light stimuli recorded in each of the 76 points of the 16.2� VF are presented. Color coding is the same as in Figure 4. deg ¼ degree.


optics scanning laser ophthalmoscopy studies demon-strated early detection of reduced cone density and irreg-ular cone mosaic spatial arrangement in the macula of RPpatients with best-corrected visual acuity of 20/20.46e48

Future trials will include follow-up pupillometry andspectral-domain OCT imaging that will allow analysis oflongitudinal changes in pupillary response with respect todisease progression and examination of the correlationbetween functional pupillometry findings and spectral-domain OCT structural findings. In addition, it would beinteresting to examine the correlation between the chro-matic multifocal pupillometry function testing and adap-tive optics scanning laser ophthalmoscopy imaging. In thecurrent study, the pupillometer test results were comparedwith CDA-GVF that enables evaluation of cone and rodresponses. In future studies, the pupillometer results willbe compared with the more commonly used Humphreyperimetry (Carl Zeiss Meditec, Inc., Jena, Germany).

Taken together, our results suggest direct correlationbetween seeing and pupil response function of the retina.Different parameters of pupil response can be used for VFdetermination (PPC and MCV) and disease diagnosis(stimulus color and LMCV). The pupillometer test is pre-dicted to be less stressful for tested subjects because patientsare unaware of the test results and the test may facilitate


objective perimetry and assessment of function of retinalphotoreceptors with minimal patient cooperation and mini-mal technician training.

References

1. Sharma P, Sample PA, Zangwill LM, et al. Diagnostic toolsfor glaucoma detection and management. Surv Ophthalmol2008;53(suppl 1):S17–32.

2. Rotenstreich Y, Fishman GA, Lindeman M, et al. The appli-cation of chromatic dark-adapted kinetic perimetry to retinaldiseases. Ophthalmology 2004;111:1222–7.

3. Grover S, Fishman GA, Brown J Jr. Patterns of visual fieldprogression in patients with retinitis pigmentosa. Ophthal-mology 1998;105:1069–75.

4. Trope GE, Britton R. A comparison of Goldmann and Hum-phrey automated perimetry in patients with glaucoma. Br JOphthalmol 1987;71:489–93.

5. Kim LS, McAnany JJ, Alexander KR, et al. Intersessionrepeatability of Humphrey perimetry measurements in patientswith retinitis pigmentosa. Invest Ophthalmol Vis Sci 2007;48:4720–4.

6. Seiple W, Clemens CJ, Greenstein VC, et al. Testeretestreliability of the multifocal electroretinogram and Humphrey

1909


http://refhub.elsevier.com/S0161-6420(16)30405-5/sref1




















visual fields in patients with retinitis pigmentosa. DocumentaOphthalmologica 2004;109:255–72.

7. Heijl A, Lindgren A, Lindgren G. Test-retest variability inglaucomatous visual fields. Am J Ophthalmol 1989;108:130–5.

8. Artes PH, Iwase A, Ohno Y, et al. Properties of perimetricthreshold estimates from full threshold, SITA standard, andSITA fast strategies. Invest Ophthalmol Visual Sci 2002;43:2654–9.

9. Turalba AV, Grosskreutz C. A review of current technologyused in evaluating visual function in glaucoma. Semin Oph-thalmol 2010;25:309–16.

10. Chauhan BC, Garway-Heath DF, Goni FJ, et al. Practicalrecommendations for measuring rates of visual field change inglaucoma. Br J Ophthalmol 2008;92:569–73.

11. Advanced Glaucoma Intervention Study Group. AdvancedGlaucoma Intervention Study 2: visual field test scoring andreliability. Ophthalmology 1994;101:1445–55.

12. Harms H. Grundlagen, methodik und bedeutung derpupillenperimetrie für die physiologie und pathologie dessehorgans. Graefes Arch Clin Exp Ophthalmol 1949;149:1–68.

13. Reuther R, Krastel H, Alexandridis E. Disturbances of thepupil reflex associated with homonymous hemianopic para-central scotoma [in German]. Arch Psychiatr Nervenkr1981;229:259–66.

14. Asakawa K, Shoji N, Ishikawa H, et al. New approach for theglaucoma detection with pupil perimetry. Clin Ophthalmol2010;4:617–23.

15. Hong S, Narkiewicz J, Kardon RH. Comparison of pupil peri-metry and visual perimetry in normal eyes: decibel sensitivity andvariability. Invest Ophthalmol Vis Sci 2001;42:957–65.

16. Aoyama T. Pupillographic perimetry. the application to clin-ical cases [in Japanese]. Nihon Ganka Gakkai Zasshi 1977;81:1527–38.

17. Chang DS, Xu L, Boland MV, et al. Accuracy of pupilassessment for the detection of glaucoma: a systematic reviewand meta-analysis. Ophthalmology 2013;120:2217–25.

18. Alexandridis E, Krastel H. Pupillographic perimetry using the“octopus”. Fortschr Ophthalmol 1989;86:692–4.

19. Yoshitomi T, Matsui T, Tanakadate A, et al. Comparison ofthreshold visual perimetry and objective pupil perimetry inclinical patients. J Neuroophthalmol 1999;19:89–99.

20. Schmid R, Luedtke H, Wilhelm BJ, et al. Pupil campimetry inpatients with visual field loss. Eur J Neurol 2005;12:602–8.

21. Skorkovska K, Wilhelm H, Ludtke H, et al. How sensitive ispupil campimetry in hemifield loss? Graefes Arch Clin ExpOphthalmol 2009;247:947–53.

22. Skorkovska K, Ludtke H, Wilhelm H, et al. Pupil campim-etry in patients with retinitis pigmentosa and functional vi-sual field loss. Graefes Arch Clin Exp Ophthalmol 2009;247:847–53.

23. Jensen W. A description of a method for objective perimetry[in German]. Graefes Arch Clin Exp Ophthalmol 1976;201:183–91.

24. Sabeti F, Maddess T, Essex RW, et al. Multifocal pupillo-graphic assessment of age-related macular degeneration.Optom Vis Sci 2011;88:1477–85.

25. Maddess T, Ho YL, Wong SS, et al. Multifocal pupillographicperimetry with white and colored stimuli. J Glaucoma2011;20:336–43.

26. Carle CF, James AC, Kolic M, et al. Blue multifocal pupil-lographic objective perimetry in GlaucomaBlue mfPOP inglaucoma. Invest Ophthalmol Vis Sci 2015;56:6394–403.

27. Sabeti F, Nolan CJ, James AC, et al. Multifocal pupillographyidentifies changes in visual sensitivity according to severity ofdiabetic retinopathy in type 2 DiabetesmfPOP to identify

1910


patients with and without DR. Invest Ophthalmol Vis Sci2015;56:4504–13.

28. Hartong DT, Berson EL, Dryja TP. Retinitis pigmentosa.Lancet 2006;368:1795–809.

29. Birch DG, Anderson JL, Fish GE. Yearly rates of rod and conefunctional loss in retinitis pigmentosa and cone-rod dystrophy.Ophthalmology 1999;106:258–68.

30. Holopigian K, Greenstein V, Seiple W, et al. Rates of changediffer among measures of visual function in patients withretinitis pigmentosa. Ophthalmology 1996;103:398–405.

31. Ross DF, Fishman GA, Gilbert LD, et al. Variability of visualfield measurements in normal subjects and patients with reti-nitis pigmentosa. Arch Ophthalmol 1984;102:1004–10.

32. Skaat A, Sher I, Kolker A, et al. Pupillometer-based objectivechromatic perimetry in normal eyes and patients withretinal photoreceptor dystrophies. Invest Ophthalmol Vis Sci2013;54:2761–70.

33. Rotenstreich Y, Chibel R, Yahia SH, et al. The first prototypeof chromatic pupillometer for objective perimetry in retinaldegeneration patients. SPIE BiOS Int Soc Optics Photonics2015. 93070Qe93070Q-7.

34. Marmor MF, Fulton AB, Holder GE, et al. ISCEV standard forfull-field clinical electroretinography (2008 update). DocOphthalmol 2009;118:69–77.

35. Kawasaki A, Crippa SV, Kardon R, et al. Characterization ofpupil responses to blue and red light stimuli in autosomaldominant retinitis pigmentosa due to NR2E3 mutation. InvestOphthalmol Vis Sci 2012;53:5562–9.

36. Kardon R, Anderson SC, Damarjian TG, et al. Chromaticpupil responses: preferential activation of the melanopsin-mediated versus outer photoreceptor-mediated pupil light re-flex. Ophthalmology 2009;116:1564–73.

37. Kardon R, Anderson SC, Damarjian TG, et al. Chromaticpupillometry in patients with retinitis pigmentosa. Ophthal-mology 2011;118:376–81.

38. Lorenz B, Strohmayr E, Zahn S, et al. Chromatic pupillometrydissects function of the three different light-sensitive retinalcell populations in RPE65 deficiency. Invest Ophthalmol VisSci 2012;53:5641–52.

39. Curcio CA, Sloan KR, Kalina RE, et al. Human photoreceptortopography. J Comp Neurol 1990;292:497–523.

40. Ellis CJ. The pupillary light reflex in normal subjects. Br JOphthalmol 1981;65:754–9.

41. Rangaswamy NV, Patel HM, Locke KG, et al. A comparison ofvisual field sensitivity to photoreceptor thickness inretinitis pigmentosa. Invest Ophthalmol Vis Sci 2010;51:4213–9.

42. Hood DC, Lazow MA, Locke KG, et al. The transition zonebetween healthy and diseased retina in patients with retinitispigmentosa. Invest Ophthalmol Vis Sci 2011;52:101–8.

43. Jacobson SG, Cideciyan AV, Aleman TS, et al. Usher syn-dromes due to MY07A, PCDH15, USH2A or GPR98 muta-tions share retinal disease mechanism. Hum Mol Genet2008;17:2405–15.

44. Jacobson SG, Aleman TS, Sumaroka A, et al. Diseaseboundaries in the retina of patients with usher syndromecaused by MY07A gene mutations. Invest Ophthalmol Vis Sci2009;50:1886–94.

45. Wen Y, Klein M, Hood DC, et al. Relationships amongmultifocal electroretinogram amplitude, visual field sensitivity,and SD-OCT receptor layer thicknesses in patients with reti-nitis pigmentosa. Invest Ophthalmol Vis Sci 2012;53:833–40.

46. Makiyama Y, Ooto S, Hangai M, et al. Macular cone abnor-malities in retinitis pigmentosa with preserved central visionusing adaptive optics scanning laser ophthalmoscopy. PLoSOne 2013;8:e79447.






































































































































47. Park SP, Lee W, Bae EJ, et al. Early structural anomaliesobserved by high-resolution imaging in two related cases ofautosomal-dominant retinitis pigmentosa. Ophthal Surg LasersImaging 2014;45:469–73.


48. Park SP, Hong IH, Tsang SH, et al. Cellular imagingdemonstrates genetic mosaicism in heterozygous carriers ofan X-linked ciliopathy gene. Eur J Hum Genet 2013;21:1240–8.

Footnotes and Financial Disclosures

Originally received: January 11, 2016.Final revision: May 21, 2016.Accepted: May 23, 2016.Available online: July 15, 2016. Manuscript no. 2016-92.1 The Goldschleger Eye Institute, Sheba Medical Center, Tel-Hashomer,Israel.2 Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.3 Department of Ophthalmology, Wolfson Medical Center, Holon, Israel.4 Biostatistics Unit, Gertner Institute for Epidemiology and Health PolicyResearch, Ramat Gan, Israel.5 Unit of Cardiovascular Epidemiology, Gertner Institute for Epidemiologyand Health Policy Research, Ramat Gan, Israel.

Financial Disclosure(s):The author(s) have made the following disclosure(s): Y.R.: Financial sup-port (to institution) e Accutome Inc., Malvern, PA; Patent e ShebaMedical Center, Tel-Hashomer, Israel

Supported by Accutome, Inc., Malvern, PA; Accutome, Inc., participated inreview of the manuscript but had no role in the design or conduct of thisresearch.

Author Contributions:

Conception and design: Rotenstreich, Sher

Analysis and interpretation: Chibel, Sher, Achiron, Skaat, Berchenko,Oberman, Kalter-Leibovici, Freedman, Rotenstreich

Data collection: Chibel, Sher, Ben Ner, Mhajna, Hajyahia

Obtained funding: Rotenstreich

Overall responsibility: Rotenstreich

Abbreviations and Acronyms:AUC ¼ area under the receiver operating characteristic curve; CDA-GVF ¼ chromatic dark-adapted Goldmann visual field; LMCV ¼ latencyof maximum contraction velocity; MCV ¼ maximum contraction velocity;OCT ¼ optical coherence tomography; PPC ¼ percentage of change ofpupil size; RP ¼ retinitis pigmentosa; SE ¼ standard error; VF ¼ visualfield.

Correspondence:Ygal Rotenstreich, MD, Goldschleger Eye Institute, Sheba MedicalCenter, Tel Hashomer 52621, Israel. E-mail: [email protected].

1911










mailto:[email protected]

mailto:[email protected]

Documents

Chromatic Multifocal Pupillometer for Objective Perimetry and … · 2017-06-12 · white and chromatic stimuli arranged in a dartboard pattern was developed (nuCoria Field Analyzer;