Recei ed: 4 No embe 2022
-
Accep ed: 8 Feb ua y 2023
DOI: 10.1002/ueg2.12375
ORIGINAL ARTICLE
Colo ec al polyps: Ta ge s o luo escence‐guided
endoscopy o de ec high‐g ade dysplasia and T1 colo ec al
cance
Nik Dekke s
1
|Elham Zonoobi
2
|Hao Dang
1
|Ma s I. Wa me dam
2
|
S ijn C obach
3
|Alexand a M. J. Lange s
1
|Jolein an de K aan
1
|
Denise E. Hilling
2,4
|Koen C. M. J. Pee e s
2
|Fabian A. Holman
2
|
Alexande L. Vah meije
2
|Co nelis F. M. Sie
2,5
|James C. H. Ha dwick
1
|
Ju jen J. Boons a
1
1
Depa men o Gas oen e ology and
Hepa ology, Leiden Uni e si y Medical Cen e ,
Leiden, The Ne he lands
2
Depa men o Su ge y, Leiden Uni e si y
Medical Cen e , Leiden, The Ne he lands
3
Depa men o Pa hology, Leiden Uni e si y
Medical Cen e , Leiden, The Ne he lands
4
Depa men o Su gical Oncology and
Gas oin es inal Su ge y, Uni e si y Medical
Cen e Ro e dam, Ro e dam, The
Ne he lands
5
Pe cu os BV, Leiden, The Ne he lands
Co espondence
Nik Dekke s, Albinusd ee 2, 2333 ZA Leiden,
The Ne he lands.
Email: [email p o ec ed]
Funding in o ma ion
H2020 Ma ie Skłodowska‐Cu ie Ac ions,
G an /Awa d Numbe : 857894‐CAST;
Nede landse Ve eniging oo
Gas oen e ologie, G an /Awa d Numbe :
Gas os a onds, 04–2020
Abs ac
Backg ound: Di e en ia ing high‐g ade dysplasia (HGD) and T1 colo ec al cance
(T1CRC) om low‐g ade dysplasia (LGD) in colo ec al polyps can be challenging.
Inco ec ecogni ion o HGD o T1CRC oci can lead o a need o addi ional
ea men a e local esec ion, which migh no ha e been necessa y i i was
ecognized co ec ly. Tumo ‐ a ge ed luo escence‐guided endoscopy migh help o
imp o e ecogni ion.
Objec i e: Selec ing he mos sui able HGD and T1CRC‐speci ic imaging a ge
om a panel o well‐es ablished bioma ke s: ca cinoemb yonic an igen (CEA), c‐
mesenchymal‐epi helial ansi ion ac o (c‐MET), epi helial cell adhesion molecule
(EpCAM), ola e ecep o alpha (FRα), and in eg in alpha‐ be a‐6 (α β6).
Me hods: En bloc esec ion specimens o colo ec al polyps ha bo ing HGD o
T1CRC we e selec ed. Immunohis ochemis y on pa a in sec ions was used o
de e mine he bioma ke exp ession in no mal epi helium, LGD, HGD, and T1CRC
(sco es o 0–12). The di e en ial exp ession in HGD‐T1CRC componen s compa ed
o su ounding LGD and no mal componen s was assessed, jus as he sensi i i y
and speci ici y o each ma ke .
Resul s: 60 specimens we e included (21 HGD, 39 T1CRC). Posi i e exp ession
(sco e >1) o HGD‐T1CRC componen s was ound in 73.3%, 78.3%, and 100% o
cases o CEA, c‐MET, and EpCAM, espec i ely, and in <40% o FRαand α β6.
Nega i e exp ession (sco e 0–1) o he LGD componen occu ed mo e equen ly
o CEA (66.1%) han c‐MET (31.6%) and EpCAM (0%). The di e en ial exp ession
in he HGD‐T1CRC componen compa ed o he su ounding LGD componen was
Nik Dekke s and Elham Zonoobi sha ed he i s au ho ship.
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion‐NonComme cial‐NoDe i s License, which pe mi s use and dis ibu ion in any
medium, p o ided he o iginal wo k is p ope ly ci ed, he use is non‐comme cial and no modi ica ions o adap a ions a e made.
© 2023 The Au ho s. Uni ed Eu opean Gas oen e ology Jou nal published by Wiley Pe iodicals LLC on behal o Uni ed Eu opean Gas oen e ology.
282
-
Uni ed Eu opean Gas oen e ol J. 2023;11:282–292. wileyonlinelib a y.com/jou nal/ueg2
ound o CEA in 66.7%, o c‐MET in 43.1%, o EpCAM in 17.2%, o FRαin 22.4%,
and o α β6 in 15.5% o he cases. Mo eo e , CEA showed he highes combined
sensi i i y (65.0%) and speci ici y (75.0%) o he de ec ion o an HGD‐T1CRC
componen in colo ec al polyps.
Conclusion: O he es ed a ge s, CEA appea s he mos sui able o speci ically
de ec HGD and T1 cance oci in colo ec al polyps. An in i o s udy using umo ‐
a ge ed luo escence‐guided endoscopy should con i m hese indings.
KEYWORDS
ad ance, bioma ke s, colo ec al cance , colo ec al polyps, CRC, dysplasia, endoscopy,
luo escence, esec ion, a ge
INTRODUCTION
Since he in oduc ion o popula ion‐based sc eening p og ams, a
g owing numbe o la ge colo ec al polyps ha e been de ec ed.
1
These
la ge polyps can o en be emo ed by local esec ions.
2,3
To de e mine
he p e e ed local esec ion echnique, i is c ucial o es ima e he isk
o high‐g ade dysplasia (HGD) o ea ly s age T1 colo ec al cance
(T1CRC). P e e ably, polyps suspec ed o ha bo a ocus o HGD o
T1CRC a e emo ed en bloc o acili a e comple e his ological
assessmen , a e which he need o addi ional ea men is de e -
mined.
4
In con as , inco ec ecogni ion o a ocus o HGD o T1CRC
may lead o inapp op ia e ea men (i.e. piecemeal esec ion) and he
need o oncological su ge y, he eby unnecessa ily exposing pa ien s
o he isk o su gical mo bidi y and mo ali y. Al hough he epo ed
pe cen ages a y g ea ly, i is clea ha he e is oom o imp o emen
in he op ical diagnosis o T1CRCs. Among expe s, he a e o un-
ecognized T1CRCs is s ill 13%–22%,
5,6
whe eas among endoscopis s
a he communi y le el, he a e o misclassi ied T1CRCs can inc ease
up o 81%.
7
Un o una ely, he addi ional alue o imaging modali ies
such as endoscopic ul asound o magne ic esonance imaging seems
limi ed. Tumo ‐ a ge ed luo escence op ical imaging (FOI) migh help
o imp o e he ecogni ion o a ocus o HGD o T1CRC in colo ec al
polyps du ing endoscopic assessmen , possibly aiding he p ocess o
decision‐making o he p e e ed local esec ion echnique.
Nea ‐in a ed FOI is a p omising echnique ha combines he
adminis a ion o a a ge ed luo escen con as agen wi h he use
o Nea ‐in a ed ligh . I allows o eal‐ ime op ical imaging by
selec i ely highligh ing cells ha exp ess ce ain molecula a ge s.
8
In he su gical ield, umo ‐ a ge ed FOI has been applied o
di e en aspec s o CRC,
8
including in aope a i e de ec ion and
dema ca ion,
9
and in aope a i e imaging o me as ases.
10
In he
endoscopic ield, FOI has been applied o aid polyp de ec ion
11
and o
e alua e neoadju an ea men esponse in locally ad anced ec al
cance .
12
Fluo escence‐guided endoscopy enables in aluminal isu-
aliza ion o polyps based on speci ic biomolecula ea u es by using
luo escen ly labeled molecula p obes ha bind o speci ic molecula
a ge s o which a ace is adminis e ed p io o imaging.
13
These
luo escen ‐ a ge ing ace s can be adminis e ed in a enously,
o ally, o as sp ay dyes. By adding a laye o in o ma ion o he
con en ional endoscopic assessmen o polyps, his s a egy can
po en ially imp o e he accu acy o op ical diagnosis and he eby
imp o e eal‐ ime clinical decision‐making o he p e e ed local
esec ion echniques o la ge polyps. To he bes o ou knowledge,
no s udy has ocused on he abili y o umo ‐ a ge ed FOI o de ec
oci o HGD o T1CRC in colo ec al polyps. Be o e emba king on a
clinical s udy, examining which bioma ke is mos sui able as a FOI
umo a ge is necessa y.
Key summa y
Summa ize he es ablished knowledge on his subjec
�To de e mine he p e e ed local esec ion echnique o
colo ec al polyps, i is c ucial o es ima e he isk o high‐
g ade dysplasia (HGD) o ea ly s age colo ec al cance
(T1CRC).
�The accu acy o op ical diagnosis is no op imal, espe-
cially in la ge polyps.
�Tumo ‐ a ge ed luo escence‐guided endoscopy migh
help o imp o e he ecogni ion o a ocus o HGD o
T1CRC in colo ec al polyps.
�The mos sui able imaging a ge o speci ically de ec -
ing a ocus o HGD o T1CRC in colo ec al polyps is
cu en ly unknown.
Wha a e he signi ican and/o new indings o his s udy?
�I is easible o de ec HGD and T1CRC oci in colo ec al
polyps in i o by s aining o umo ‐speci ic a ge s.
�O he es ed a ge s, ca cinoemb yonic an igen (CEA)
mos equen ly showed di e en ial exp ession in he
HGD and T1CRC componen s compa ed o su ounding
polyp issue wi h low‐g ade dysplasia.
�An in i o s udy is needed o con i m CEA as sui able
a ge o speci ically de ec HGD o T1CRC oci in
colo ec al polyps by luo escence‐guided endoscopy.
DEKKERS ET AL.
-
283
The a ge selec ion o imaging pu poses depends on di e en
cha ac e is ics, including he di e en ial exp ession in he a ge
issue compa ed o no mal issue.
14
Enhanced p o ein exp ession in
he a ge issue and low o e en absen exp ession in no mal issue
a e p e equisi es. T1CRCs o en eside in polyps ha consis o
se e al s ages o dysplasia. A sui able a ge should be able o
dis inguish a ocus o HGD o T1CRC om he su ounding LGD
componen o a polyp. P omising a ge s in CRC de ec ion, wi h
a ailable luo escence a ge ing p obes, include ca cinoemb yonic
an igen‐ ela ed adhesion molecule 5 (CEACAM5, om he e on o be
e e ed o as CEA), c‐mesenchymal‐epi helial ansi ion ac o (c‐
MET), epi helial cell adhesion molecule (EpCAM), ola e ecep o
alpha (FRα) and in eg in α β6. Ca cinoemb yonic an igen is a
memb ane‐bound glycop o ein wi h known exp ession in he ma-
jo i y o CRCs and li le exp ession in no mal mucosa.
15
c‐MET is he
memb ane‐bound hepa ocy e g ow h ac o ecep o in ol ed in
p oli e a ion and in asion. C‐MET o e exp ession has been demon-
s a ed in he sequence o colo ec al adenoma‐ca cinoma sequence
as an ea ly e en .
16,17
EpCAM is a ansmemb ane glycop o ein
in ol ed in cell‐cell in e ac ions and cell‐s oma adhesions ha is
gene ally o e exp essed in epi helial malignancies such as colo ec al
cance .
18
FRαis a memb ane‐bound olic acid‐binding and ans-
po ing p o ein, wi h highe exp ession in CRCs han in no mal mu-
cosa o adenoma.
19
α β6 is an in eg in sub ype ha is exp essed only
in epi helial cells, wi h signi ican ly inc eased exp ession in epi helial
umo s.
20
Al hough se e al s udies ha e epo ed he enhanced
exp ession o hese ma ke s in CRC, nei he ha e s udied he di -
e en ial exp ession be ween a componen o HGD o T1CRC and he
su ounding componen o LGD in colo ec al polyps.
We aimed o selec he mos HGD‐T1CRC speci ic luo escence‐
guided endoscopy a ge o an in i o pilo s udy.
MATERIALS AND METHODS
Popula ion
Fo malin‐ ixed pa a in‐embedded issue blocks om pa ien s who
unde wen en bloc endoscopic submucosal dissec ion (ESD) (be ween
Feb ua y 2013 and No embe 2019) in he Leiden Uni e si y Med-
ical Cen e o la e al sp eading polyps ha bo ing a ocus o T1CRCs
o HGD loca ed in he ec um o sigmoid we e e ie ed om he
pa hology depa men . To inc ease he sample size, we also included
a andom sample o 10 FFPE blocks om pa ien s who unde wen en
bloc endoscopic mucosal esec ion (EMR) o non‐g anula T1CRCs in
he ec um o sigmoid (be ween Feb ua y 2013 and No embe
2019). P io o inclusion, slides we e eexamined by a pa hologis
specialized in gas o‐in es inal pa hology (S.C.) and e‐s aged
acco dingly. Pa ien s wi h an insu icien amoun o issue we e
excluded. E hical app o al was ob ained om he Medical E hical
Commi ee o he Leiden Uni e si y Medical Cen e , and he
equi emen o ob aining in o med consen was wai ed ( e e ence:
B20.016, 11‐06‐2020). The s udy p o ocol con o ms o he e hical
guidelines o he 1975 Decla a ion o Helsinki as e lec ed in a p io i
app o al by he ins i u ion's human esea ch commi ee.
Clinical a iables
Demog aphic pa ien cha ac e is ics (sex, age) and clinical da a (polyp
mo phology, p ocedu e‐ ela ed pa ame e s, his ology pa ame e s)
we e collec ed om elec onic medical eco ds. En bloc esec ion was
de ined as mac oscopic emo al o he lesion in a single piece. High‐
g ade dysplasia was de ined as a chi ec u al abno mali y and se e e
cy ologic a ypia wi hou in asion h ough he muscula is mucosae.
T1 colo ec al cance was de ined as umo s wi h umo in asion
h ough he muscula is mucosae and in o, bu no beyond, he
submucosa.
An ibodies and immunohis ochemis y
Based on hema oxylin‐eosin (HE)‐s ained slides, a pa hologis
specialized in gas o‐in es inal pa hology (S.C.) selec ed a ep esen-
a i e o malin‐ ixed pa a in‐embedded issue block o each pa ien ,
con aining as many s ages o dysplasia (no mal, LGD, HGD, and
T1CRC) as possible. Selec ed issue blocks we e sec ioned (4 μm) and
moun ed on adhesi e slides. Sec ions we e depa a inized wi h
xylene o 15 min, ehyd a ed in dec easing e hanol concen a ions
and hen insed in demine alized wa e . Subsequen ly, endogenous
pe oxidase was blocked using 0.3% hyd ogen pe oxide (Me ck Mil-
lipo e, Ne he lands) in demine alized wa e o 20 min. Speci ica ions
ega ding an igen e ie al and an ibodies a e p o ided in supple-
men a y Table S1. A e wa d, he slides we e insed in phospha e‐
bu e ed saline (PBS, pH7.4), and s ained wi h he app op ia e sec-
onda y an ibody (EnVision an i‐mouse o an i‐ abbi ho se adish
pe oxidase) (Dako) o 30 min, ollowed by ano he washing s ep.
Immuno eac ions we e isualized wi h diaminobenzidine subs a e
bu e (Dako) a e 10 min, coun e s ained using Maye 's hema ox-
ylin solu ion (Sigma‐Ald ich, USA), and dehyd a ed a 37°C be o e
being moun ed wi h Pe ex (Leica Mic osys ems, Ge many). Nega i e
(PBS) and conjuga e con ol (only seconda y an ibody) we e included
o ule ou nonspeci ic s aining.
To ensu e ha he di e en s ages o dysplasia we e he same
h oughou all sec ioned slides o one block, he i s and las slides
om each block we e s ained wi h HE and examined by a pa hologis
specialized in gas oin es inal pa hology (S.C.).
Sco ing me hod
All s ained slides we e digi ally scanned (Ins elliSi e Ul a‐Fas
Scanne , Philips). HE‐slides o each case we e u ilized o de e -
mine one clea egion o each p esen s age o dysplasia (S.C. and N.
D.). These egions we e hen ma ked on he digi ally scanned HE
slide. Subsequen ly, he same egions we e ma ked in he emaining
284
-
UNITED EUROPEAN GASTROENTEROLOGY JOURNAL
slides o ha case ha we e s ained wi h s udy ma ke s (N.D.). The
ma king p ocess o one case is shown in supplemen a y Figu e S1.
The a ge exp ession in all s ages o dysplasia was quan i ied using
he immuno eac i e sco e (IRS). The IRS was calcula ed by mul i-
plying he posi i e cell p opo ions (PS) and s aining in ensi y sco e
as p e iously desc ibed.
21
PS ep esen ed he pe cen age o posi-
i ely s ained cells and anged be ween 0 and 4 (0 =no posi i e
cells; 1 = <10% posi i e cells; 2 =10–50%, 3 =51–80%, 4 =
>80%). In ensi y sco e ep esen ed s aining in ensi y and anged
be ween 0 and 3 (0 =no colo eac ion; 1 =mild eac ion;
2=mode a e eac ion; 3 =in ense eac ion). The o al IRS was a
ange be ween 0 and 12 and was u he subdi ided in o subg oups
(0–1 =nega i e, 2–3 =mild, 4–8 =mode a e, 9–12 =s ongly
posi i e). Th ee obse e s independen ly e alua ed he ma ke
exp ession (N.D., J.B., and J.H.). All cases wi h disag eemen
ega ding he IRS subg oup we e discussed un il a consensus was
eached. The a e age o he indi idual sco es wi hin he same
subg oup esul ed in he de ini i e IRS.
S a is ical analyses
S a is ical analyses we e pe o med using IBM SPSS e sion 24.0
(Chicago, USA) and G aphPad P ism 6 (La Jolla, CA, USA). The di -
e en ial exp ession o each bioma ke was s udied by sub ac ing
he exp ession sco es o he LGD o no mal colon componen om
he exp ession sco es o he adjacen HGD‐T1CRC componen om
he same slide. Sensi i i ies and speci ici ies o HGD‐T1CRC de ec-
ion we e calcula ed om he mean s aining sco es by ecei e
ope a ing cha ac e is ic (ROC) cu es. In he wo ma ke s ha
showed he g ea es di e en ial exp ession, he in luence o
mo phological polyp cha ac e is ics on he occu ence o nega i e
exp ession in he HGD‐T1CRC componen was explo ed using he
chi‐squa ed es . Mean IRS o di e en s ages o dysplasia we e
compa ed using he Wilcoxon ank es . A p‐ alue ≤0.05 was
conside ed s a is ically signi ican .
RESULTS
In o al, issue blocks o 39 T1CRC pa ien s and 21 pa ien s wi h
HGD we e included (Figu e 1). A componen o no mal colon issue
was p esen in all cases, bu due o he small size o his componen ,
i was deemed insu icien o sco ing in 3/60 cases. In wo o he
cases, no dis inc LGD componen could be iden i ied. Fo hese 5
cases, componen s o he o he s ages o dysplasia we e included in
he esul s.
Pa ien and polyp cha ac e is ics a e shown in Table 1. The
o e all median polyp size was 40 mm ( ange 8–100). The median
polyp size was 15 mm ( ange 8–20) in he en bloc EMR subg oup and
40 mm ( ange 14–100) in he ESD subg oup. Mac oscopic polyp
mo phology was la ele a ed in 24 (40%) and sessile in 36 (60%).
Polyps we e mainly loca ed in he ec um o ec osigmoid.
Exp ession o ma ke s in di e en s ages o dysplasia
Posi i e cell p opo ions and in ensi y sco es a ied widely o all
ma ke s h oughou he coho in no mal, LGD, HGD and T1CRC
issues. Figu e 2shows all indi idual s aining sco es (IRS) in no mal,
LGD and HGD o T1CRC o each a ge ; hese sco es a e inde-
penden ly a anged in ascending o de pe a ge o illus a e he
dis ibu ions ac oss he coho . The mean IRS o each a ge in he
di e en s ages o dysplasia is shown in Table 2.
FIGURE 1 Flowcha o pa ien selec ion. CRC, colo ec al cance ; EMR, endoscopic mucosal esec ion; ESD, endoscopic submucosal
dissec ion; HGD, high‐g ade dysplasia.
DEKKERS ET AL.
-
285
I he HGD o T1CRC componen showed posi i e exp ession
(i.e. s aining sco e >1), CEA and FRαwe e p edominan ly exp essed
on he apical memb ane o he HGD‐T1CRC componen , while c‐
MET, EpCAM, and α β6 showed a mo e memb anous, ci cum e en-
ial s aining pa e n in he HGD‐T1CRC componen s. Figu e 3shows
he posi i e exp ession pa e n o he di e en a ge s; o each
a ge , he mos ep esen a i e case o posi i e exp ession in he
HGD‐T1CRC componen was selec ed. An example o he s aining
pa e n in he en i e polyp o all a ge s in he same case o T1CRC is
shown in Figu e 4.
Exp ession o ma ke s in no mal issues
Ca cinoemb yonic an igen exp ession was nega i e in 56/57 (98.2%).
C‐MET exp ession was nega i e in all cases. Epi helial cell adhesion
molecule exp ession was posi i e in all cases, showing mode a e (41/
57, 71.9%) o s ong (16/57, 28.1%) exp ession. Fola e ecep o alpha
exp ession was nega i e in 56/57 (98.2%). α β6 exp ession was
nega i e in 49/56 (87.5%).
Exp ession o ma ke s in low‐g ade dysplasia
Ca cinoemb yonic an igen exp ession was nega i e in 37/57 (64.9%).
C‐MET exp ession was nega i e in 19/58 (32.8%), mos o cases
showed a mode a e exp ession (22/58, 37.9%). Epi helial cell adhe-
sion molecule exp ession was posi i e in all cases, showing mode a e
(10/58, 17.2%) o s ong (48/58, 82.8%) exp ession. Fola e ecep o
alpha exp ession was nega i e in 47/58 (81.0%). α β6 exp ession was
nega i e in 48/58 (82.8%).
Exp ession o ma ke s in high‐g ade dysplasia o T1
colo ec al cance
Ca cinoemb yonic an igen exp ession was posi i e in 44/60
(73.3%), showing a s ong exp ession in 16/60 (26.7%). C‐MET
exp ession was posi i e in 47/60 (78.3%), showing a s ong
exp ession in 3/60 (5%). Epi helial cell adhesion molecule exp es-
sion was posi i e in all cases, showing a s ong exp ession in 57/
60 (95%). Fola e ecep o alpha exp ession was posi i e in 18/60
(30.0%), s ong exp ession was obse ed in 1/60 (1.7%). α β6
exp ession was posi i e in 12/60 (20.0%), bu s ong exp ession
was no obse ed.
Di e en ial HGD‐T1CRC exp ession compa ed o
no mal
The s aining in ensi y was highe in he HGD‐T1CRC componen
han he adjacen no mal componen o CEA in 46/57 (80.7%), c‐
MET in 46/57 (80.7%), EpCAM in 40/57 (70.2%), FRαin 20/57
(35.1%), and α β6 in 14/56 (25.0%) (Figu e 5). Ca cinoemb yonic
an igen showed he g ea es inc ease in IRS. I he e was an inc ease,
he HGD‐T1CRC componen sco ed on a e age 6.0 poin s highe
(95% CI 5.0–7.0) han he adjacen no mal componen . Fo c‐MET
his was 4.4 (95% CI 3.8–5.1) (Table 3).
TABLE 1Clinical‐pa hological cha ac e is ics o he s udy
coho .
Numbe o cases n=60 (%)
Pa ien cha ac e is ics
Sex, male 44 (73.3)
Age, yea s, median ( ange) 65 (35–84)
T ea men
ESD 51 (85.0)
En bloc EMR 9 (15.0)
Polyp cha ac e is ics
Loca ion
Sigmoid 11 (18.3)
Rec osigmoid 10 (16.7)
Rec um 39 (65.0)
Diame e polyp, mm, median ( ange) 40 (8–100)
G oss mo phology
Fla ele a ed 24 (40.0)
Sessile 36 (60.0)
Pa is classi ica ion
Is 29 (48.3)
0‐IIa 5 (8.3)
0‐IIa +Is 18 (30.0)
0‐IIa +c 8 (13.3)
G anula i y
G anula 20 (33.3)
Non‐g anula 40 (66.7)
Maximal deg ee o dysplasia
HGD 21 (35.0)
T1CRC 39 (65.0)
Adenoma componen (n=42)
Tubula 12 (20.0)
Villous 4 (6.7)
Tubulo illous 25 (41.7)
Se a ed 1 (1.7)
No e: Values a e n(%) unless o he wise de ined.
Abb e ia ions: CRC, colo ec al cance ; EMR, endoscopic mucosal
esec ion; ESD, endoscopic submucosal dissec ion; HGD, high‐g ade
dysplasia; LST, la e al sp eading umo .
286
-
UNITED EUROPEAN GASTROENTEROLOGY JOURNAL
FIGURE 2 S aining sco es o CEA, c‐MET, EpCAM, FRα, and α β6 in no mal colo ec al issue, low‐g ade dysplasia (LGD) and high‐g ade
dysplasia (HGD) o T1 colo ec al cance (T1CRC) we e exp essed as immuno eac i e sco es. The o al immuno eac i e sco es we e
independen ly a anged in ascending o de o demons a e he dis ibu ions ac oss ou coho . CEA, ca cinoemb yonic an igen; c‐MET, c‐
mesenchymal‐epi helial ansi ion ac o ; CRC, colo ec al cance ; EpCAM, epi helial cell adhesion molecule; FRα, ola e ecep o alpha; IRS,
immuno eac i e sco e.
TABLE 2The mean immuno eac i e sco e (IRS) o he componen o no mal issue, low‐g ade dysplasia (LGD), high‐g ade dysplasia
(HGD), and T1 colo ec al cance (T1CRC) (minimum 0, maximum 12).
Ta ge
No mal mean
IRS (n=57)
LGD mean IRS
(n=58)
HGD‐T1CRC mean
IRS (n=60)
p‐ alue no mal e sus
HGD‐T1CRC
p‐ alue LGD e sus
HGD‐T1CRC
CEA 0.08 1.83
a
4.78 <0.001 <0.001
c‐MET 0.02 2.64 3.58 <0.001 0.003
EpCAM 8.79 10.95 11.75 <0.001 0.011
FRα0.04 0.70 1.23 0.001 0.051
α β6 0.42
a
1.01 0.76 0.075 0.930
Abb e ia ions: CEA, ca cinoemb yonic an igen; c‐MET, c‐mesenchymal‐epi helial ansi ion ac o ; EpCAM, epi helial cell adhesion molecule; FRα, ola e
ecep o alpha; IRS, immuno eac i e sco e.
a
Da a o one case was missing due o a b oken slide on which he IRS o ha componen could no be assessed p ope ly.
Di e en ial HGD‐T1CRC exp ession compa ed o low‐
g ade dysplasia
The s aining in ensi y was highe in he HGD‐T1CRC componen
han he su ounding LGD componen o CEA in 38/57 (66.7%), c‐
MET in 25/58 (43.1%), EpCAM in 10/58 (17.2%), FRαin 13/58
(22.4%), and α β6 in 9/58 (15.5%) (Figu e 5). Fo CEA, i he e was an
inc ease in IRS, HGD‐T1CRC componen s sco ed on a e age 4.5
poin s highe (95% CI 3.4–5.5) han adjacen LGD componen s. Fo c‐
MET, his was 3.0 (95% CI 2.2–3.9) (Table 3).
Sepa a e esul s o he HGD and T1CRC subg oups can be
ound in he supplemen a y esul s.
Speci ici y and sensi i i y o HGD‐T1CRC de ec ion
Va ying he limi o de ec ion, we examined wha p opo ion o HGD‐
T1CRC componen s would be isualized (sensi i i y) and wha p o-
po ion o LGD componen s would be isualized (speci ici y). A
ecei ing ope a ing cha ac e is ic (ROC) was plo ed o each ma ke
o selec he op imal cu ‐o sco es (i.e. sco es wi h he g ea es
combined sensi i i y and speci ici y). Using hese op imal cu ‐o
sco es, sensi i i y and speci ici y o de ec ion o he HGD‐T1CRC
componen e sus su ounding LGD we e 65.0% and 75.0% o
CEA (cu ‐o >2.5), 55.0% and 60.3% o c‐MET (cu ‐o >3.5), 93.3%
and 22.4% o EpCAM (cu ‐o >11), 16.7% and 93.1% o FRα(cu ‐
o >3.5), and 28.3% and 75.9% o α β6 (cu ‐o <0.5). Supplemen-
a y Figu e S2 shows he ROC cu es o de ec ion o he HGD‐T1
componen s compa ed o he no mal and LGD componen s.
Co ela ion be ween ca cinoemb yonic an igen and c‐
mesenchymal‐epi helial ansi ion ac o exp ession
and mo phological cha ac e is ics
Fo CEA and c‐MET, nega i e s aining in he HGD o T1CRC
componen did no s a is ically di e be ween la ele a ed and
sessile polyps, g anula and non‐g anula polyps, and smalle o
DEKKERS ET AL.
-
287
FIGURE 3 Posi i e s aining pa e n o all a ge s in he high‐g ade dysplasia (HGD) o T1 colo ec al cance (T1CRC) componen . Fo each
a ge , an illus a i e case was selec ed wi h posi i e exp ession (i.e. s aining sco e >1) in he HGD‐T1CRC componen . The egion enclosed by
he ec angle wi h dashed line consis s o HGD o T1CRC. An o e iew image (le ) and enla gemen o he HGD‐T1CRC egion ( igh ) a e
p o ided o each a ge . (a) CEA exp ession and (b) c‐MET exp ession. (c) EpCAM exp ession. (d) FRαexp ession. (e) α β6 exp ession. CEA,
ca cinoemb yonic an igen; c‐MET, c‐mesenchymal‐epi helial ansi ion ac o ; EpCAM, epi helial cell adhesion molecule; FRα, ola e ecep o .
288
-
UNITED EUROPEAN GASTROENTEROLOGY JOURNAL
la ge polyps (dicho omized, using 40 mm as cu ‐o ), all p>0.05
(supplemen a y Table S2). Addi ional in o ma ion ega ding he cases
wi h nega i e CEA s aining in he HGD o T1CRC componen is
p o ided in supplemen a y esul s.
DISCUSSION
This s udy is he i s o e alua e he sui abili y o CEA, c‐MET,
EpCAM, FRα, and α β6 as possible a ge s o de ec a ocus o HGD
o T1CRC in la ge colo ec al polyps using umo ‐ a ge ed
luo escence‐guided endoscopy. Ou esul s indica e ha CEA
shows he mos di e en ial exp ession o he HGD‐T1CRC
componen o he es ed ma ke s. The e o e, CEA appea s o be
he mos p omising a ge o in i o es ing.
Ca cinoemb yonic an igen ou pe o med he o he ma ke s by
showing he g ea es di e en ial HGD‐T1CRC exp ession, espe-
cially compa ed o he LGD componen . In compa ison o CEA,
posi i e exp ession (i.e. Immuno eac i e sco e >1) in he HGD‐
T1CRC componen was ound mo e equen ly o c‐MET (47/60,
78.3%). Howe e , c‐MET lacked he deg ee o di e en ial exp es-
sion wi h LGD because he exp ession in he LGD componen was
FIGURE 4 O e all s aining pa e n o all a ge s in he same case o T1 colo ec al cance (T1CRC). The egion enclosed by he ec angle
wi h dashed line consis s o HGD‐T1CRC. (a) Endoscopic images and esec ion specimen a e endoscopic submucosal dissec ion. (b) HE slide.
(c) CEA exp ession. (d) c‐MET exp ession. (e) EpCAM exp ession. ( ) FRαexp ession. (g) α β6 exp ession. HE hema oxylin‐eosin. CEA,
ca cinoemb yonic an igen; c‐MET, c‐mesenchymal‐epi helial ansi ion ac o ; EpCAM, epi helial cell adhesion molecule; FRα, ola e ecep o .
DEKKERS ET AL.
-
289
also posi i e in a conside able amoun o cases, which was in line
wi h p e ious s udies. C‐MET was e en success ully used as an in
i o FOI a ge o polyps.
11
EpCAM showed a posi i e exp ession
in he HGD‐T1CRC componen mos equen ly o all es ed bio-
ma ke s (60/60, 100%) bu ha dly showed any di e en ial exp es-
sion wi h he LGD componen s. The numbe o cases wi h a posi i e
exp ession in he HGD‐T1CRC componen o FRα(18/60, 30.0%)
and α β6 (12/60, 20.0%) we e oo low o be conside ed as sui able
a ge s.
Fo CEA, posi i e exp ession in he HGD‐T1CRC componen was
seen in 44/60 (73.3%) cases. This was sligh ly lowe han he p e i-
ously epo ed 87%–99% in s udies ha mainly included mo e
ad anced CRC s ages.
15,22
Since no all HGD‐T1CRC componen s
show posi i e exp ession o CEA, no all pa ien s will bene i om
umo ‐ a ge ed FOI a ge ing CEA. I would be p e e able o be able
o selec hose pa ien s who would. This s udy could no iden i y
mo phological polyp cha ac e is ics ha we e associa ed wi h nega-
i e umo exp ession. Howe e , i should be kep in mind ha he
cu en s udy may be unde powe ed o iden i y ele an ac o s.
Mo eo e , se um CEA le els do no appea in o ma i e o p edic ing
exp ession le els.
23
Addi ionally, sc eening CEA exp ession on p e‐
ope a i e biopsies does no appea o be a easible selec ion s a -
egy because, in acco dance wi h he mo i e o his s udy, ecognizing
and hus being able o ake biopsies om he HGD‐T1CRC compo-
nen in la ge polyps can be challenging. CEA's impe ec umo
exp ession a e may hampe i s clinical implemen a ion as HGD‐and
T1CRC‐speci ic FOI a ge s. Howe e , he pe ec a ge has ye o
be disco e ed and CEA appea s o be he mos p omising. Al e na-
i ely, a combina ion o wo complemen a y a ge s could be
conside ed. Based on ou esul s, c‐MET and EpCAM could enhance
he de ec ion o HGD‐T1CRC e sus no mal issue, bu his does no
con ibu e o be e dis inc ion be ween HGD‐T1CRC oci e sus
LGD componen s compa ed o single a ge CEA, which was he aim
o his s udy.
Based on he esul s o his s udy, we a e conduc ing a clinical
pilo s udy o assess whe he i is possible o speci ically de ec an
HGD o T1CRC componen in non‐peduncula ed ec al polyps using
SGM‐101, a luo och ome‐labeled an i‐CEA monoclonal an ibody.
A e in a enous adminis a ion o his luo escen CEA‐ a ge ing
ace , imaging will be pe o med using a luo escence‐endoscope.
Fo his clinical s udy, i should be aken in o accoun ha immu-
nohis ochemical s udies can only pa ly mimic he in i o si ua ion
whe e se e al o he ac o s can po en ially in luence he pe o -
mance. These ac o s include issue pene a ion, backg ound s aining,
immunological esponse and sensi i i y o he NIR‐came a sys em.
Howe e , despi e hese challenges, he easibili y and sa e y o
luo escence‐labeled con as agen s a ge ing CEA o in i o umo
imaging ha e al eady been shown. Fo example, SGM‐101 showed
enhanced di e en ia ion be ween no mal and cance ous issues in
panc ea ic cance and CRC.
24
Addi ionally, i s applica ion du ing CRC
su ge y in luenced clinical decision‐making.
9
A p omising no el clin-
ical applica ion o CEA‐ a ge ed luo escen agen s migh be du ing
endoscopic assessmen o colo ec al polyps whe e i could help o
imp o e he ecogni ion o HGD‐T1CRC oci and he e o e aid he
p ocess o decision‐making o he p e e ed local esec ion
echnique.
Al hough he esul s a e p omising, he p esen s udy has some
limi a ions. The main d awbacks a e he ela i ely small numbe o
cases and he use o semiquan i a i e immunohis ochemis y o
measu e p o ein exp ession. E en hough immunohis ochemis y is
ou inely used, i equen ly lacks s anda diza ion and he e o e
in e p e a ion o s aining pa e ns migh be he e ogeneous. Ou
FIGURE 5 Di e en ial exp ession plo s. Di e en ial exp ession
sco es we e calcula ed by sub ac ing he IRS o he no mal o low‐
g ade dysplasia (LGD) componen om he IRS o he HGD‐T1CRC
componen . Di e en ial exp ession sco es we e independen ly
a anged and connec ed in ascending o de o demons a e he
dis ibu ions ac oss he coho . (a) shows he di e en ial
exp ession plo o HGD‐T1CRC componen s compa ed o
su ounding no mal colo ec al issue. (b) shows he di e en ial
exp ession plo o HGD‐T1CRC componen s compa ed o
su ounding componen s o LGD. CEA, ca cinoemb yonic an igen;
c‐MET, c‐mesenchymal‐epi helial ansi ion ac o ; CRC, colo ec al
cance ; EpCAM, epi helial cell adhesion molecule; FRα, ola e
ecep o alpha; IRS, immuno eac i e sco e; LGD, low‐g ade
dysplasia.
290
-
UNITED EUROPEAN GASTROENTEROLOGY JOURNAL
