INTRAOPERATIVE USE OF HANDHELD SPECTRAL DOMAIN …people.duke.edu/~sf59/dayani.pdf ·...

INTRAOPERATIVE USE OF HANDHELDSPECTRAL DOMAIN OPTICALCOHERENCE TOMOGRAPHY IMAGINGIN MACULAR SURGERYPOUYA N. DAYANI, MD, RAMIRO MALDONADO, MD, SINA FARSIU, PHD,CYNTHIA A. TOTH, MD

Purpose: To describe the intraoperative use of handheld spectral domain opticalcoherence tomography (SDOCT) imaging in macular surgery.

Design: Prospective, observational case series.Methods: A handheld SDOCT device was used to obtain preincision optical coherence

tomography imaging in patients undergoing vitrectomy for macular diseases. Afterremoval of the internal limiting membrane or the epiretinal membrane, repeat intraop-erative imaging was obtained. Spectral domain optical coherence tomography findingswere characterized.

Results: An efficient technique was established for obtaining intraoperative SDOCTimaging. A total of eight patients were included in the study. Four patients underwentsurgery for macular hole, three patients for epiretinal membrane, and one for vitreomaculartraction. Comparison of the preincision and intraoperative SDOCT images demonstrateddistinct changes in retinal contour and macular hole configuration. Intraoperative SDOCTimaging identified additional membranes in two patients.

Conclusion: The intraoperative use of handheld SDOCT imaging provides an efficientmethod for visualizing macular pathology. This technology may, in certain cases, helpconfirm or identify diseases that may be difficult to visualize during surgery.

RETINA 29:1457–1468, 2009

Optical coherence tomography (OCT) is a rapid,noncontact, and noninvasive method of imaging

intraocular tissue that has significantly contributed tothe understanding, diagnosis, treatment, and follow-upof retinal diseases.1–5 Spectral domain optical coher-ence tomography (SDOCT) provides a number ofadvantages over conventional time-domain OCT sys-

tems. By eliminating the need for a scanning mirror,SDOCT offers higher acquisition speed and, thus, theability to obtain significantly more images in a shorterperiod of time (approximately 20,000 A-scans persecond compared with 400 A-scans per second ontime-domain OCT). This increased speed minimizespatient movement artifact and improves image reso-lution. Some studies suggest that the higher resolutionobtained with SDOCT imaging may improve the de-tection of abnormalities.6–10 The ability to obtain vol-ume maps of an entire region and the improved reg-istration with SDOCT allow one to accurately followchanges in the same anatomic area on subsequentimaging. Some units also have added features such asfluorescein angiography, autofluorescence, indocya-nine green angiography, and scanning laser ophthal-moscope capability.

An important limitation of typical OCT units is thatthey are stationary and require a compliant patientwho can sit upright. Thus, uncooperative pediatricpatients and those with limited mobility (i.e., because

From the Department of Ophthalmology, Duke University EyeCenter, Durham, North Carolina.

Supported by North Carolina Biotechnology Center CollaborativeFunding Grant no. 2007-CFG-8005 with Bioptigen Inc., NationalInstitutes of Health R21 EY017393, and the Duke Translational Med-icine Institute Subcontract No. 12 of NIH Grant No. 5ULT-RR024128-03. Supported, in part, by a grant from the Heed Ophthal-mic Foundation (to P.N.D.).

Dr. Toth receives other support from the following commercialentities: Alcon Laboratories (honorarium, research support, androyalties), Bioptigen (honorarium, research support, and licenseagreement), Genentech (honoraria and research grant), and SirionTherapeutics (research grant). Dr. Dayani received an honorariumas a clinical consultant from OSI Therapeutics.

Reprint requests: Cynthia A. Toth, MD, Duke University EyeCenter, PO Box 3802, Durham, NC 27710; e-mail: [email protected]

1457

of musculoskeletal disorders) or access (i.e., hospitalfloor patients or those in isolation) may not be prop-erly imaged. A portable SDOCT unit (Bioptigen Inc.,Research Triangle Park, NC) is now available thatallows imaging of such patients. We have recentlydescribed methods of imaging children with shakenbaby syndrome and those with albinism using thishandheld SDOCT unit and have characterized novelretinal SDOCT findings in these two entities.10,11

An additional value of the handheld SDOCT systemis the capability to obtain noncontact intraoperativeimages in the supine position. In the current study, wereport our experience in using a handheld SDOCTimaging device to obtain high-resolution, cross-sec-tional images in patients undergoing surgery for full-thickness macular hole (FTMH), epiretinal membrane(ERM), or vitreomacular traction. Our objectives areto describe the technique of obtaining reliable intra-operative SDOCT images, to characterize the intraop-erative findings, and to identify how these imagescompare with the preoperative scan and with the sur-geon’s assessment of the macular status.

Materials and Methods

This study was approved by the Institutional Re-view Board of Duke University Medical Center. Con-secutive patients undergoing vitrectomy surgery formacular diseases were enrolled in the study. All pa-tients provided informed consent.

Patients were imaged with a portable SDOCTworkstation consisting of a handheld SDOCT scan-ning head connected, through a flexible fiberopticcable, to a mobile cart housing the computer–spec-trometer unit and viewing screen (Bioptigen Inc.).The handheld scanning head can operate while po-sitioned at any angle, allowing its use in the supinepatient.

Retinal SDOCT imaging was performed in the op-erating room at Duke University Eye Center immedi-ately before and during vitrectomy surgery. All im-ages were obtained with the attending surgeon(C.A.T.) at the head of the bed (Figure 1). Duringimage acquisition, the probe was typically stabilizedby resting the hand on the wrist rest, which wastemporarily raised to the level of the forehead. Alter-natively, the surgeon could stabilize the hand-piece byresting several fingers on the patient’s forehead. Anassistant operated the computer software while imageswere obtained by the surgeon.

During each imaging session, three types of scanswere obtained. Horizontal and vertical scans wereobtained for each setting. The settings includedsummed 10-mm linear B-scans (40 repeated B-scans

of 1,000 A-scans each), 10 � 10-mm volumetric scans(100 B-scans of 1,000 A-scans each), and 5 � 5-mmvolumetric scans (200 B-scans of 500 A-scans). Thesecustomized settings obtained radial B-scans 100 �mapart on the 10 � 10-mm volumetric scans comparedwith 25-�m spacing on the 5 � 5-mm scans. Thevolumetric scans produced a three-dimensional repre-sentation of the retinal layers. By axially projectingthe volumetric scans, a two-dimensional image anal-ogous to a fundus photo, called a summed voxelprojection (SVP), was created (Figures 2 and 3).

For each imaging session, the reference arm lengthwas adjusted according to the axial length of thepatient’s eye, which in most cases was estimated bythe patient’s age using a table.12 If patients had doc-umented axial lengths from previous surgeries, thosemeasurements were used. We also adjusted for thepatient’s refractive error (spherical equivalent). Thehandheld SDOCT unit has a focusing adjustment of 21diopters (D) (�10 to �11 D). Additional fine focusingwas performed at the time of imaging. Focusingchanges of �2 D did not seem to significantly affectthe clarity of the images.

The postprocessing software used was InVivoVue1.2 (Bioptigen Inc.). The SVP images obtained withthe handheld probe held at the head of the bed appearflipped and inverted on the display monitor because

Fig. 1. Portable SDOCT workstation.

1458 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2009 ● VOLUME 29 ● NUMBER 10

the InvivoVue postprocessing software did not allowus to correct for the orientation of the handheld probeat the time of imaging. A more detailed description ofthe image-processing algorithm used to enhance im-age quality is described in our previous publication.10

Preincision SDOCT images were obtained after ad-ministration of retrobulbar anesthesia with the patientsin the supine position without a lid speculum. Bal-anced salt solution or artificial tears was used to main-tain corneal surface integrity and to improve the clar-ity of the images. Three-port pars plana vitrectomysurgery was performed in all patients. The Alcon25-gauge vitrectomy system (Alcon Laboratories, Inc.,Fort Worth, TX) was used in five patients and the 20-gauge system in the remainder. The posterior hyaloidwas separated in eyes without a complete posterior vit-reous detachment. Indocyanine green angiography wasadded over the posterior pole in the fluid-filled eye in

patients undergoing macular hole (MH) surgery. Preser-vative-free triamcinolone was used to visualize the hya-loid in some patients with ERM. The Tano diamond-dusted membrane scraper and forceps were used to peelthe internal limiting membrane (ILM) along with anymembranes off the fovea. Cannulas were then plugged,and repeat intraoperative SDOCT images were obtainedin all patients. An additional sterile drape was placedfrom the patient’s upper lid, over the forehead, and offthe field to protect from inadvertent contact with thehandheld probe cord. The handheld probe was placedinto a sterile plastic bag, adjusted in such a way that thelens surface of the probe was aligned with a hole in thebag (created earlier by the surgeon), and suspended overthe eye for noncontact imaging. Methylcellulose wasplaced over the ocular surface. After acquisition of theimages, the additional drape was removed and thesurgeon resumed the surgery.

Fig. 2. Preoperative SDOCT of Case 1 captured immediately before surgery. A, SVP from a 10 � 10-mm volumetric scan outlining the areas scannedin B–D. B, Vertical 10 � 10-mm scan showing a FTMH, cystic thickening of the retina at the hole margin (white arrow), and partial posterior vitreousseparation. C, Vertical scan demonstrating partial vitreous separation (red arrow), cystic thickening of the parafoveal region, and normal hyperre-flectivity of the ILM before removal. D, Vertical scan demonstrating normal juxtafoveal retinal contour and ILM hyperreflectivity before surgery.

1459INTRAOPERATIVE SDOCT IMAGING ● DAYANI ET AL

Macular hole measurements were obtained withMatLab software (Version 7.0, Mathworks Inc.,Natick, MA) and verified with ImageJ (freeware, Na-tional Institutes of Health, Bethesda, MD). The scanselected for measurements was the one with the larg-est hole diameter. Four measurements were calculatedfrom each SDOCT image corresponding to the centerof the hole. These included the base diameter, mea-sured at the level of the retinal pigment epithelium; theminimum diameter, defined as the minimal extent ofthe hole; the maximum diameter, defined as the max-imum distance between hole edges; and the holeheight, measured as the distance between the retinalpigment epithelium and the inner most aspect of thehole (greatest height).

Results

A total of eight adult patients (eight eyes) wereenrolled and successfully imaged. There were fourwomen and four men with a mean age of 63.5 years(range, 45–81 years; Table 1). Four patients werephakic and four were pseudophakic. Four patientsunderwent vitrectomy for MH repair, three for ERM,and one for VMT. The MH was idiopathic in allpatients.

There was a learning curve in localizing the area ofinterest and obtaining reliable, high-quality scans.

Manual alignment was challenging because of severalfactors. The most important of these was motion arti-fact associated with stabilization of the handheldprobe because small horizontal or angular movementsof the imager’s hand would shift alignment off thepupil. Other challenges included the lack of real-timefundus image correlation and occasional poor SVPimage quality. Orientation during scanning was morechallenging because the SVP image and scans wereinverted on the viewing screen because of the hand-held scanner location over the patient’s forehead.

Despite such challenges, our ability to image im-proved significantly over the course of the study. TheSVPs were particularly helpful in identifying land-marks such as retinal vessels and the optic nerve,which could be used for orientation during imaging.We, thus, found that the wider field of view on theSVP obtained with the 10 � 10-mm scans, comparedwith the 5 � 5-mm scans, was advantageous in aidingorientation. The vessel pattern on the SVP was alsouseful in assessing image quality. Motion artifact dur-ing volumetric scans typically resulted in a brokenvascular pattern on the SVP (Figure 4). We also foundthat operating the Bioptigen system in “free run”mode allowed the selection of an optimal SDOCTvolume set to capture and save. With experience, eachimaging session lasted 4 minutes to 5 minutes, andindividual scans were acquired in 5.8 seconds. We

Fig. 3. Intraoperative SDOCT of Case 1 obtained immediately after ILM removal. A, SVP of the areas scanned. The dark spot in the lower rightcorner is a preretinal hemorrhage. Asterisks show corresponding location on B-scans of B–D. B, Retinal distortion, residual adherent ILM (whitearrow), and decreased reflectivity of the outer retina and shadowing (blue arrow) corresponding to a focal hemorrhage at the margin of ILM peeling.This site had a normal retinal contour before surgery. C and D, Note the hyperreflective areas along the inner retinal surface that correspond to theremaining ILM (yellow star). In contrast, there is a less intense hyperreflectivity in the areas in which the ILM was peeled corresponding to the nervefiber layer.


instilled Systane artificial tears (Alcon Laboratories)every two or three scans as needed to maintain ocularsurface integrity during preoperative imaging. Wefound that images obtained with artificial tears lubri-cating the cornea produced higher quality imagescompared with balanced salt solution. During surgery,Methocel 2% (Omnivision GmbH, Puchheim, Ger-many) gel over the cornea did not seem to impairimage quality.

Preincision SDOCT images confirmed the diseaseof interest in all cases. Comparison of the preoperativeand intraoperative SDOCT images yielded a numberof observations (Tables 2 and 3). These included fo-cal, iatrogenic traumatic changes (Figure 3), ILM de-fects, and distinction between areas with remainingand removed ILM (Figures 3 and 5). Residual ERMsor ILM (Figure 6) and the absence of a previouslyvisualized operculum or posterior hyaloid face werealso observed (Figures 2 and 3). In addition, alter-ations in the retinal contour and MH configurationwere observed in all cases.

Among the patients undergoing MH surgery, thereseemed to be a reduction in the mean size of the holediameter (Table 4). This change was observed withthe base diameter and with the minimum and maxi-mum diameter (Figures 5 and 6). The MH heightremained relatively stable, and there was only moder-ate change in most cases. In 3 of the 4 patients withFTMH (Patients 1–3), the retina at the edges of thehole seemed to move centrally (toward the center ofthe MH) after ILM removal, suggesting decreasedtangential traction (Figures 5 and 6). In 2 of thesepatients (Patients 1 and 3), there was an increasedvertical retinal elevation at the edges of the hole(Figure 5).

Among those undergoing surgery for ERM andVMT, definite alterations in the retinal contour wereseen after ERM/ILM peeling (Figures 4 and 7). In onepatient with ERM (Patient 6, Figure 7) and anotherwith VMT (Patient 8, Figure 8), an immediate nor-malization of the retinal contour was seen. In Patient7, there were multiple thin attachments between theretinal surface and the ERM (Figure 4A). Aftermembrane peeling, there was persistence of an ir-regular and corrugated retinal contour (Figure 4B).This appearance improved significantly on subsequentSDOCT imaging over the ensuing 4 weeks. In anotherpatient, an increased distortion of the retinal layerswas observed. Interestingly, the intraoperative imagesidentified residual membranes not visualized by theassistant surgeon during surgery in 2 patients, oneundergoing surgery for MH (Patient 2; Case 2, de-

Tab

le1.

Pat

ient

Dem

ogra

phi

csan

dS

urgi

calF

ind

ings

Pat

ient

Dem

ogra

phi

csS

urgi

calF

ind

ings

Cas

eG

end

erA

geE

tiolo

gyfo

rS

urge

ryP

ast

Ocu

lar

His

tory

Vis

ual

Acu

ityP

haki

cS

tatu

sP

VD

Gau

geof

Vitr

ecto

my

Com

ple

teP

VD

Vis

ualiz

atio

nA

id

Suc

cess

ful

Clo

sure

ofM

H

1F

58S

tage

3FT

MH

Non

e20

/50

Pha

kic

No

20N

oIC

GY

es2

F81

Sta

ge4

FTM

HA

MD

,C

E/I

OL

20/2

50–1

Pse

udop

haki

cY

es25

Yes

ICG

Yes

3M

68S

tage

3FT

MH

Fuch

sco

rnea

ld

ystr

ophy

20/3

00P

seud

opha

kic

No

20N

oIC

GY

es4

F69

Sta

ge3

FTM

HE

RM

,C

E/I

OL

HM

Pse

udop

haki

cN

o25

No

ICG

Yes

5M

70E

RM

BR

VO

,C

ME

20/8

0P

haki

cY

es25

Yes

PF

Ken

alog

N/A

6M

45E

RM

PD

R,

CM

E20

0E

Pha

kic

No

20N

oN

/AN

/A7

F55

ER

MN

one

20/6

4�P

haki

cY

es25

Yes

N/A

N/A

8M

62V

MT

PD

R,

CM

E,

CE

/IO

L20

/400

Pse

udop

haki

cY

es25

Yes

PF

Ken

alog

N/A

AM

D,

age-

rela

ted

mac

ular

deg

ener

atio

n;B

RV

O,

bra

nch

retin

alve

inoc

clus

ion;

CE

/IO

L,ca

tara

ctex

trac

tion

with

intr

aocu

lar

lens

;C

ME

,cy

stoi

dm

acul

ared

ema;

ER

M,

epire

tinal

mem

bra

ne;

F,fe

mal

e;FT

MH

,fu

llth

ickn

ess

mac

ular

hole

;M

,M

ale;

PD

R,

pro

lifer

ativ

ed

iab

etic

retin

opat

hy;

PF,

pre

serv

ativ

efr

ee;

PV

D,

pos

terio

rvi

treo

usd

etac

hmen

t;S

F6,

sulfu

rhe

xaflu

orid

e.


scribed later) and another for ERM (Patient 7). Noimmediate postoperative complications were observedin any of the study patients. MH closure was attainedwith a single procedure in all patients undergoing MHrepair.

Case 1

A 58-year-old woman (Patient 1) with no ocularhistory presented with decreased vision and best-cor-rected visual acuity of 20/50. She was found to have aStage 3 FTMH, which was confirmed by OCT, andunderwent surgical repair.

Preincision, intraoperative SDOCT imaging con-firmed a FTMH with cystic thickening at the edges(Figure 2). There is a partial vitreous separation withpersistent attachments primarily inferiorly. Normalcontour and hyperreflectivity of the inner retinal lay-ers, including the ILM, can be appreciated.

Intraoperative images obtained after ILM removalshowed an area of irregular retinal contour inferonasalto the fovea with persistent hyaloidal attachment, de-creased reflectivity of the outer retinal layers, andshadowing (Figure 3). This area corresponds to themargin of ILM peeling and a focal intraretinal hem-orrhage that was appreciated intraoperatively. Sharpcontrast is observed between the hyperreflective areaswith persistent ILM and the adjacent retina where theILM was peeled. The intraoperative SDOCT imagesalso show that both edges of the hole are more verti-cally elevated compared with the preoperative scansand seem to be approaching one another (Figure 5).

Case 2

A pseudophakic 81-year-old woman (Patient 2)with a history of age-related macular degenerationpresented with vision loss in her better-seeing eye. On

Fig. 4. Spectral domain optical coherence tomography of Patient 7. A, Diffuse epiretinal membrane with focal area of discontinuity. Note multiplethin strands between the retinal surface and the membrane (white arrow). B, Irregular retinal contour with release of attachments after membranepeeling. C, SVP of preincision scan showing normal vessel appearance. D, SVP of preincision scan showing motion artifact and a broken vascularpattern.


Tab

le2.

Pre

-an

dIn

trao

per

ativ

eO

CT

Find

ings

inP

atie

nts

Und

ergo

ing

MH

Sur

gery

Pre

oper

ativ

eO

CT

Find

ings

Intr

aop

erat

ive

OC

TFi

ndin

gs

Cas

eN

o.E

RM

Vis

ualiz

eIL

MC

ME

SR

FO

per

culu

m/P

reho

leO

pac

ityA

dd

ition

alP

reop

erat

ive

Find

ings

Res

idua

lFo

veal

ILM

Res

idua

lE

RM

Cha

nge

inS

RF

Cha

nge

inC

ME

Cha

nge

inR

eflec

tivity

Iatr

ogen

icC

hang

es

1N

YY

NN

N/A

NN

/AN

YIn

crea

sed

ILM

RR

thic

keni

ngw

ithad

here

ntst

rand

(ILM

)at

edge

2N

YY

NN

N/A

YN

/AN

NN

N3

NY

YN

Y—

pos

terio

rhy

aloi

dfa

ceN

/AN

N/A

NN

NFo

cals

had

owin

gfr

omhe

me

atIL

Med

ge4

NY

YY

Y—

pos

terio

rhy

aloi

dfa

ceFo

calh

yper

Rat

RP

Ele

vel,

opac

ities

inm

id-l

evel

ofho

le

NN

NY

NN

CM

E,

cyst

oid

mac

ular

edem

a;N

,no

;N

/A,

not

app

licab

le;

R,

refle

ctiv

ity;

RP

E,

retin

alp

igm

ent

epith

eliu

m;

SR

F,su

bre

tinal

fluid

;Y

,ye

s.

Tab

le3.

Pre

inci

sion

and

Intr

aop

erat

ive

OC

TFi

ndin

gsin

Pat

ient

sU

nder

goin

gE

RM

Sur

gery

Cas

eN

o.E

RM

Cha

ract

eris

tics

Ret

inal

Con

tour

Cha

ract

eris

tics

ofH

yalo

idR

esid

ualC

entr

alE

RM

Ret

inal

Con

tour

Iatr

ogen

icFi

ndin

gs

5D

iffus

ew

ithou

tfo

cal

trac

tion

Irre

gula

rw

ithm

oder

ate

thic

keni

ngN

/AN

oIr

regu

lar

thic

keni

ngw

ithin

crea

sein

ragg

edap

pea

ranc

e;in

crea

sed

hyp

erre

flect

ivity

ofin

ner

retin

alla

yers

;fa

int

mem

bra

ne/

stra

nds

atfo

veal

edge

s

Foca

lexc

resc

ence

sw

ithsh

adow

ing;

edge

ofm

emb

rane

/ILM

dire

cted

ante

riorly

6D

iffus

ew

ithm

ultip

lefir

mat

tach

men

ts;

dou

ble

laye

rof

ER

M/s

chis

is

Rag

ged

cont

our;

diff

use

CM

Ew

ithla

rge

cyst

sH

yper

refle

ctiv

ityan

terio

rto

hyal

oid

face

No

Nor

mal

izat

ion

ofre

tinal

cont

our

with

red

uced

ragg

edap

pea

ranc

eTw

ofo

cale

xcre

scen

ces

with

shad

owin

g

7D

iffus

eE

RM

with

mul

tiple

,fo

cal,

firm

atta

chm

ents

;ar

eaof

dis

cont

inui

ty

Irre

gula

rth

icke

ning

;m

ultip

leat

tach

men

tto

ER

M

N/A

Res

idua

lper

ifove

alE

RM

Irre

gula

rw

ithp

ersi

sten

ttr

actio

nfr

omE

RM

and

per

sist

ence

ofra

gged

app

eara

nce

Foca

lret

inal

elev

atio

nw

ithsh

adow

ing

8D

iffus

ew

ithou

tfo

cal

trac

tion

Cys

ticed

ema

with

fove

altr

actio

nel

evat

ion;

hyp

erre

flect

ivity

ofou

ter

laye

rs

Foca

l,fir

mat

tach

men

tto

fove

alce

nter

No

Ab

senc

eof

foca

lele

vatio

n;d

iffus

eth

icke

ning

with

hyp

erre

flect

ivity

ofou

ter

laye

rs;

foca

lhyp

erre

flect

ive

dot

sal

ong

retin

alsu

rfac

e(p

ossi

bly

tria

mci

nolo

nep

artic

les)

Foca

lret

inal

elev

atio

n

CM

E,

cyst

oid

mac

ular

edem

a;N

/A,

not

app

licab

le.


examination, visual acuity was 20/250 in the affectedeye, and there was nonfoveal geographic atrophy anda Stage 4 FTMH, which was confirmed by SDOCTimaging.

She subsequently underwent surgical repair. Afterinitial ILM removal, the surgeons disagreed as towhether further ILM remained at the upper margin ofthe hole. The intraoperative SDOCT images obtainedafter initial ILM removal demonstrated remainingmembrane that was partially separated and scrolled atthe temporal edge of the hole (Figure 6). On the basisof the indocyanine green angiography staining of thetissue and the scrolled appearance, the residual mem-brane was believed to be ILM. One feature of thisimage that is most striking is the contrast between thecontour of the hole margin where there is persistenttraction from remaining ILM and the hole edge that isfree of ILM. Surgery was resumed and the residualILM was removed. Repeat SDOCT intraoperative im-aging verified absence of the remaining ILM and ahole contour that appeared more elevated and lesstaut, confirming reduced tangential traction.

Discussion

In this study, we report the use of a handheldSDOCT unit to obtain preincision and intraoperative

images in patients undergoing macular surgery. To thebest of our knowledge, this is the first report of theintraoperative use of such technology. Imaging ob-tained during surgery provided unique informationregarding the location of persisting ILM or epiretinalmembrane. This served as a valuable teaching toolduring surgery to demonstrate findings to the surgicalassistant that were difficult to identify visually. Inaddition, interesting observations were made that con-firm and may expand our understanding of the patho-physiology of MHs. There was a learning curve inefficiently acquiring reliable images, and our abilityto obtain high-quality images improved consider-ably with time. Artificial tears, improved ability tohold the probe, and the understanding of how SVPswere displayed were factors that enhanced our im-aging capability.

Idiopathic FTMH affects approximately 0.3% ofpeople and results in reduced central vision.13 Opticalcoherence tomography is currently considered thegold standard for the diagnosis of MHs and has greatlycontributed to our understanding of its pathogenesisand response to therapy. Advances in MH surgeryhave led to improved surgical outcome and visualrecovery. Surgical repair typically involves pars planavitrectomy, removal of posterior cortical vitreous, and

Fig. 5. Comparison of the preoperative and intraoperative SDOCT of Patient 1 demonstrating change in the MH configuration. A, Preoperative imagedemonstrating a FTMH and cystic retinal thickening. The red lines depict the MH height and the base diameter. B, Postoperative scan showing a morepeaked appearance of the hole edge and a smaller base diameter. Also, note the distinction of the hyperreflectivity along the retinal surface in the areawith remaining ILM (white arrow) compared with the adjacent retina in which the ILM was removed.


infusion of expansile gas with postoperative pronepositioning. A meta-analysis on 1,654 eyes treatedusing different techniques reported an 87.5% ana-tomic success rate.14 Data suggest that removal of theILM may improve the anatomic success of MH repairby relieving tangential traction and by preventing pro-liferation of epiretinal tissue.15–20 Accordingly, amongeyes in which the ILM was not removed at the time ofsurgery, recurring ERMs seem to be the main cause ofMH reopening.21 In our series, residual ILM not vi-sualized by the assistant surgeon at the time of surgerywas identified on the intraoperative SDOCT imagingin one of four patients undergoing MH repair. Afterremoval of the residual ILM segment, there was adefinite change in the MH morphology (Figure 6).Given the apparent role of ILM in MH pathogenesis,and the difficulty in visualizing this structure duringsurgery, intraoperative SDOCT may prove to be auseful adjunct in select cases. This may provide analternative to the current technique of indocyaninegreen angiography staining, which has been associ-ated with ocular toxicity and vision loss.17,22–25 Intra-

operative SDOCT imaging also identified persistenttraction and distortion from residual membranes in apatient undergoing vitrectomy for ERM (Patient 7).Surgery was resumed and additional peeling was per-formed with relief of traction. In a previous publica-tion, we reported on a 6-month infant undergoingvitrectomy for a double layer of ERM.10 One monthafter the initial procedure, a deeper layer of remainingmembrane was identified on SDOCT, and a secondvitrectomy was performed. Intraoperative SDOCTmay, therefore, also be valuable in the managementof ERMs, in particular in cases with limited intra-operative visualization or double-layer (schisis)membranes.

In addition to clinical factors affecting surgical out-come, several studies have assessed the correlationbetween preoperative OCT findings and the anatomicand visual success of MH surgery. For example, OCTstudies suggest that the base and minimum diametersof MHs are predictive of postoperative success andthat visual results are adversely affected by increasinghole size.26,27 The significance of MH height and base

Fig. 6. Spectral domain optical coherence tomography of Patient 2 captured preoperatively and after initial and complete ILM removal. A, SVP froma 5 � 5-mm volumetric scan. B, Preoperative image showing a FTMH in an area of perifoveal geographic atrophy. C, Intraoperative images obtainedafter initial ILM removal show a “scrolled,” partially separated ILM at the margin of the MH. Note the difference in retinal contour at the right sideof the hole where the ILM is now absent compared with the left margin where the ILM persists. The red line depicts the maximum hole diameter.D, OCT scan after complete ILM removal shows a more elevated and less taut retinal morphology compared with previous scans with the MHapproaching closure.


diameter was addressed by Kusuhara et al,28 usingOCT imaging, to calculate the MH index. The MHindex, defined as the greatest height of the hole (fromthe retinal pigment epithelium to the vitreoretinal in-terface) divided by its base diameter, was the onlyfactor identified in the study that significantly corre-lated with postoperative vision. This study found thata larger hole height and a smaller base diameter wereassociated with better visual outcome.28 These find-ings are supported by a study that showed a positivecorrelation between both the MH index and the trac-tional hole index (ratio of the maximal height to theminimum diameter) and visual outcome.29 On thecontrary, a study by Haritoglou et al30 found a nega-tive correlation between visual outcome and MHheight.

The success of repeat vitrectomy surgery for pa-tients with persistent FTMH has also been correlatedwith OCT configuration. Hillenkamp et al31 found thatthose with a cuff of subretinal fluid at the hole marginon OCT imaging had a higher success rate for ana-tomic closure and visual outcome. The size, type oftamponade, or duration of the hole before the initialsurgery did not seem to correlate with surgical out-come. The authors explain that the elevated configu-ration may facilitate the centripetal movement of ret-inal tissue over the fovea that is presumably necessaryfor hole closure. As opposed to the “stuck down”appearance, the increased height may be an indicator

Fig. 7. Spectral domain optical coherence tomography of Patient 6. A,Epiretinal membrane with multiple attachments (white arrows) andtraction. Note the corrugated retinal appearance between the attach-ments and the hyperreflective opacities anterior to the membrane. B,Decreased retinal traction after membrane peel (white arrows) withnormalization of the retinal contour.

Tab

le4.

Mea

sure

men

tof

Pre

inci

sion

and

Intr

aop

erat

ive

Hol

eD

imen

sion

s(in

mic

rons

)

Bas

eD

iam

eter

Max

imum

Dia

met

erM

inim

umD

iam

eter

Mac

ular

Hol

eH

eigh

tM

acul

arH

ole

Ind

ex

Cas

eP

rein

cisi

onIn

trao

pC

hang

eP

rein

cisi

onIn

trao

pC

hang

eP

rein

cisi

onIn

trao

pC

hang

eP

rein

cisi

onIn

trao

pC

hang

eP

rein

cisi

onIn

trao

p

184

067

0�

170

960

800

�16

070

051

1�

189

485

584

990.

580.

872

920

960

4093

087

1�

5991

879

0�

128

451

413

�38

0.49

0.43

311

1010

80�

3012

7111

57�

114

590

650

6075

884

183

0.68

0.78

410

9010

50�

4095

210

6010

833

048

015

069

971

415

0.64

0.68

Mea

n99

094

0�

5010

2897

2�

5663

560

8�

2759

863

840

Med

ian

1005

1005

095

696

610

645

581

�64

592

649

57S

.D.

131

187

162

165

245

142

153

183

Intr

aop

,In

trao

per

ativ

e.


of decreased traction on retinal tissue or evidence oftraction elevation as a result of the upward peeling.

In the current study, we compared the SDOCTimages taken immediately before surgery and imme-diately after the removal of ILM of four eyes under-going MH repair. In three of these patients, there wasa reduction in the base diameter of the hole and anincrease in MH height after vitrectomy and ILM re-moval. Interestingly, the only patient who did notshow such changes had a Stage 4 FTMH with anexisting posterior vitreous separation (Patient 2). Thisfinding highlights the role of persistent posterior cor-tical vitreous attachments in the pathogenesis of MHs.The earlier findings suggest an immediate alteration inthe tangential and anteroposterior forces on the holeand provide additional insight into the role of vitrec-tomy and ILM removal in MH surgery. Given that thecortical vitreous and the ILM were both removedbefore repeat imaging, the individual role of thesesteps on MH morphology remains uncertain. Futurestudies may answer this question by obtaining intra-operative OCT imaging after both the removal ofposterior cortical vitreous and the ILM.

The limitations of the current report include a smallsample size and potential for investigator bias. Despitesuch limitations, we report a safe and efficient methodof obtaining intraoperative SDOCT imaging in thesupine patient using a handheld SDOCT unit. We alsocharacterize, for the first time, intraoperative OCTimages obtained immediately after macular surgery.Our study found a decrease in the base diameter of theMH among patients without complete posterior vitre-ous separation. We predict that SDOCT integrationinto surgical viewing, potentially into the microscope,may provide useful information during fine macularsurgical maneuvers in the near future.

Key words: epiretinal membrane, Fourier domainOCT, handheld, intraoperative, macular hole, OCT,

optical coherence tomography, spectral domain op-tical coherence tomography, vitreomacular traction.

References

1. Huang D, Swanson EA, Lin CP, et al. Optical coherencetomography. Science 1991;254:1178–1181.

2. Hee MR, Puliafito CA, Wong C, et al. Optical coherencetomography of macular holes. Ophthalmology 1995;102:748–756.

3. Gaudric A, Haouchine B, Massin P, Paques M, Blain P,Erginay A. Macular hole formation: new data provided byoptical coherence tomography. Arch Ophthalmol 1999;117:744–751.

4. Puliafito CA, Hee MR, Lin CP, et al. Imaging of maculardiseases with optical coherence tomography. Ophthalmology1995;102:217–229.

5. Toth CA, Narayan DG, Boppart SA, et al. A comparison ofretinal morphology viewed by optical coherence tomographyand by light microscopy. Arch Ophthalmol 1997;115:1425–1428.

6. Gupta V, Gupta P, Singh R, Dogra MR, Gupta A. Spectral-domain cirrus high-definition optical coherence tomography isbetter than time-domain stratus optical coherence tomographyfor evaluation of macular pathologic features in uveitis. Am JOphthalmol 2008;145:1018–1022.

7. Koizumi H, Spaide RF, Fisher YL, Freund KB, Klancnik JMJr, Yannuzzi LA. Three-dimensional evaluation of vitreo-macular traction and epiretinal membrane using spectral-do-main optical coherence tomography. Am J Ophthalmol 2008;145:509–517.

8. Stopa M, Bower BA, Davies E, Izatt JA, Toth CA. Correlationof pathologic features in spectral domain optical coherencetomography with conventional retinal studies. Retina 2008;28:298–308.

9. Khanifar AA, Koreishi AF, Izatt JA, Toth CA. Drusen ultra-structure imaging with spectral domain optical coherence to-mography in age-related macular degeneration. Ophthalmol-ogy 2008;115:1883–1890.

10. Scott AW, Farsiu S, Enyedi LB, Wallace DK, Toth CA.Imaging the infant retina with a hand-held spectral-domainoptical coherence tomography device. Am J Ophthalmol 2009;147:364.e2–373.e2.

11. Chong G, Farsiu S, Freedman SF, et al. Abnormal fovealmorphology in ocular albinism imaged with spectral domain

Fig. 8. A, Preincision 5 � 5-mm SDOCT scan of Patient 8 demonstrating vitreomacular traction with cystic thickening of the retina. The white arrowdemonstrates the attachment of the hyaloid to the fovea. B, Intraoperative 5 � 5-mm SDOCT scan showing the edge of the peeled hyaloid (whitearrow). The hyperreflective spots along the retinal surface likely represent residual triamcinolone particles.


optical coherence tomography. Arch Ophthalmol 2009;127:37–44.

12. Gordon RA, Donzis PB. Refractive development of the humaneye. Arch Ophthalmol 1985;103:785–789.

13. la Cour M, Friis J. Macular holes: classification, epidemiology,natural history and treatment. Acta Ophthalmol Scand 2002;80:579–587.

14. Mester V, Kuhn F. Internal limiting membrane removal in themanagement of full-thickness macular holes. Am J Ophthal-mol 2000;129:769–777.

15. Kwok AK, Lai TY, Yuen KS, Tam BS, Wong VW. Macularhole surgery with or without indocyanine green stained inter-nal limiting membrane peeling. Clin Experiment Ophthalmol2003;31:470–475.

16. Smiddy WE, Feuer W, Cordahi G. Internal limiting membranepeeling in macular hole surgery. Ophthalmology 2001;108:1471–1476; discussion 1477–1478.

17. Tognetto D, Grandin R, Sanguinetti G, et al. Internal limitingmembrane removal during macular hole surgery: results of amulticenter retrospective study. Ophthalmology 2006;113:1401–1410.

18. Uemoto R, Yamamoto S, Aoki T, Tsukahara I, Yamamoto T,Takeuchi S. Macular configuration determined by optical co-herence tomography after idiopathic macular hole surgery withor without internal limiting membrane peeling. Br J Ophthal-mol 2002;86:1240–1242.

19. Ando F, Sasano K, Ohba N, Hirose H, Yasui O. Anatomic andvisual outcomes after indocyanine green-assisted peeling ofthe retinal internal limiting membrane in idiopathic macularhole surgery. Am J Ophthalmol 2004;137:609–614.

20. Brooks HL Jr. Macular hole surgery with and without internallimiting membrane peeling. Ophthalmology 2000;107:1939–1948; discussion 1948–1949.

21. Yoshida M, Kishi S. Pathogenesis of macular hole recurrenceand its prevention by internal limiting membrane peeling.Retina 2007;27:169–173.

22. Engelbrecht NE, Freeman J, Sternberg P Jr, et al. Retinalpigment epithelial changes after macular hole surgery withindocyanine green-assisted internal limiting membrane peel-ing. Am J Ophthalmol 2002;133:89–94.

23. Gass CA, Haritoglou C, Schaumberger M, Kampik A. Func-tional outcome of macular hole surgery with and withoutindocyanine green-assisted peeling of the internal limitingmembrane. Graefes Arch Clin Exp Ophthalmol 2003;241:716–720.

24. Kanda S, Uemura A, Yamashita T, Kita H, Yamakiri K,Sakamoto T. Visual field defects after intravitreous adminis-tration of indocyanine green in macular hole surgery. ArchOphthalmol 2004;122:1447–1451.

25. Miura M, Elsner AE, Osako M, Iwasaki T, Okano T, Usui M.Dissociated optic nerve fiber layer appearance after internallimiting membrane peeling for idiopathic macular hole. Retina2003;23:561–563.

26. Ip MS, Baker BJ, Duker JS, et al. Anatomical outcomes ofsurgery for idiopathic macular hole as determined by opticalcoherence tomography. Arch Ophthalmol 2002;120:29–35.

27. Ullrich S, Haritoglou C, Gass C, Schaumberger M, Ulbig MW,Kampik A. Macular hole size as a prognostic factor in macularhole surgery. Br J Ophthalmol 2002;86:390–393.

28. Kusuhara S, Teraoka Escano MF, Fujii S, et al. Prediction ofpostoperative visual outcome based on hole configuration byoptical coherence tomography in eyes with idiopathic macularholes. Am J Ophthalmol 2004;138:709–716.

29. Ruiz-Moreno JM, Staicu C, Pinero DP, Montero J, Lugo F,Amat P. Optical coherence tomography predictive factors for mac-ular hole surgery outcome. Br J Ophthalmol 2008;92:640–644.

30. Haritoglou C, Neubauer AS, Reiniger IW, Priglinger SG, GassCA, Kampik A. Long-term functional outcome of macularhole surgery correlated to optical coherence tomography mea-surements. Clin Experiment Ophthalmol 2007;35:208–213.

31. Hillenkamp J, Kraus J, Framme C, et al. Retreatment offull-thickness macular hole: predictive value of optical coher-ence tomography. Br J Ophthalmol 2007;91:1445–1449.


INTRAOPERATIVE USE OF HANDHELD SPECTRAL DOMAIN …people.duke.edu/~sf59/dayani.pdf ·...

Documents

Transcript of INTRAOPERATIVE USE OF HANDHELD SPECTRAL DOMAIN …people.duke.edu/~sf59/dayani.pdf ·...