De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Open Access
RESEARCH ARTICLE
© 2010 De Chia a e al; licensee BioMed Cen al L d. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons
A ibu ion License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in
any medium, p o ided he o iginal wo k is p ope ly ci ed.
Resea ch a icle
Se um CD26 is ela ed o his opa hological polyp
ai s and beha es as a ma ke o colo ec al cance
and ad anced adenomas
Lo e a De Chia a
†1
, Ana M Rod íguez-Piñei o
†1
, F ancisco J Rod íguez-Be ocal
1
, Osca J Co de o
2
, Da id Ma ínez-
A es
3
and Ma ía Páez de la Cadena*
1
Abs ac
Backg ound: Se um CD26 (sCD26) le els we e p e iously ound diminished in colo ec al cance (CRC) pa ien s
compa ed o heal hy dono s, sugges ing i s po en ial u ili y o ea ly diagnosis. The e o e we aimed o es ima e he
u ili y o he sCD26 as a bioma ke o CRC and ad anced adenomas in a high- isk g oup o pa ien s. The ela ionship o
his molecule wi h polyp cha ac e is ics was also add essed.
Me hods: sCD26 le els we e measu ed by ELISA in 299 symp oma ic and asymp oma ic pa ien s who had unde gone
a colonoscopy. Pa ien s we e diagnosed as ha ing no colo ec al pa hology, non-in lamma o y o in lamma o y bowel
disease, polyps (hype plas ic, non-ad anced and ad anced adenomas) o CRC.
Resul s: A a 460 ng/mL cu -o , he sCD26 has a sensi i i y and speci ici y o 81.8% (95% CI, 64.5-93.0%) and 72.3%
(95% CI, 65.0-77.2%) o CRC ega ding no o benign colo ec al pa hology. Clinicopa hological analysis o polyps
showed a ela ionship be ween he sCD26 and he g ade o dysplasia and he p esence o ad anced adenomas.
Hence, a 58.0% (95% CI, 46.5-68.9%) sensi i i y de ec ing CRC and ad anced adenomas was ob ained, wi h a speci ici y
o 75.5% (95% CI, 68.5-81.0%).
Conclusions: Ou p elimina y esul s show ha measu emen o he sCD26 is a non-in asi e and easonably sensi i e
assay, which could be combined wi h o he s such as he aecal occul blood es o he ea ly diagnosis and sc eening
o CRC and ad anced adenomas. Addi ional compa a i e s udies in a e age- isk popula ions a e necessa y.
Backg ound
Colo ec al cance (CRC) is one o he ou mos p e alen
cance s in Wes e n coun ies due o a low eco e y a e
in ad anced s ages, when mic ome as ases may be
al eady p esen . I de elops om p ecance ous lesions
(adenomas) ha a e easily emo ed by polypec omy, a
p ocedu e ha educes CRC incidence by 75-90% [1];
mo eo e , ea men o ea ly diagnosed umou s has a
good p ognosis [2]. Fu he mo e, acco ding o he MIS-
CAN-COLON simula ion model, a 23% educ ion in
CRC mo ali y would be achie ed wi h he sc eening o a
leas 70% o he a ge popula ion [3]. The e o e sc een-
ing o CRC aims o educe mo ali y a es by de ec ion
and emo al o ea ly-s age umou s, and incidence a es
by iden i ica ion and esec ion o polyps [4].
The e is a g ea a ie y o me hods o he de ec ion o
CRC. A e iew o he ongoing sc eening ini ia i es wo ld-
wide was ecen ly published [5]. The la es Join Colo ec-
al Cance Sc eening Guidelines [6] di ide he a ailable
es s in o wo ca ego ies: hose p ima ily aimed o de ec
cance (blood and DNA s ool es s), and hose di ec ed o
de ec also ad anced lesions (endoscopic and adio-
g aphic me hods).
Wi hin he a ailable in asi e es s a e he lexible sig-
moidoscopy, he double-con as ba ium enema and he
colonoscopy. The la e is ecommended nowadays as he
gold s anda d, hough i s miss a es ha e been es ima ed
as 22% o adenomas and 2-6% o CRC [7]. Howe e ,
bowel p epa a ion cons i u es he mos objec ionable
aspec o he p ocedu e, undamen al o a p ope
* Co espondence: mpae[email p o ec ed]
1 Uni e sidad de Vigo. Facul ad de Biología. Depa amen o de Bioquímica,
Gené ica e Inmunología. As Lagoas-Ma cosende s/n. 36310 Vigo, Spain
† Con ibu ed equally
Full lis o au ho in o ma ion is a ailable a he end o he a icle
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 2 o 9
sc eening colonoscopy [8]. O he limi a ions o he
colonoscopy a e he isk o complica ions, he cos s, and
access. Rega ding non-in asi e es s, he mos common
me hod, nowadays ecommended o CRC sc eening, is
he guaiac aecal occul blood es (gFOBT), wi h highly
a iable and b and-dependen sensi i i ies and speci ici-
ies [6,9] and equi ing die a y es ic ions.
Thus he e is an impe a i e need o de eloping non-
in asi e sc eening es s o he de ec ion o CRC and ade-
nomas, and hence many cu en s udies a e e alua ing
se um ma ke s. Examples o hese a e he α-de ensins
[10], he nico inamide N-me hyl ans e ase [11], he α-L-
ucosidase [12], o he colon cance -speci ic an igens
(CCSA) -2, -3 and -4 [13,14]. A majo d awback in hese
s udies is hey a e limi ed o disc imina ing be ween CRC
pa ien s and heal hy indi iduals, lea ing aside p ecu so
lesions and no including in he con ol coho indi idu-
als wi h benign pa hologies.
P e ious s udies o ou g oup de ec ed ha soluble
se um CD26 (sCD26) le els we e diminished in CRC
pa ien s as compa ed o heal hy dono s [15,16]. In hose
s udies, a sCD26 cu -o o 410 ng/mL demons a ed a
90% diagnos ic e iciency, wi h a speci ici y and sensi i -
i y o 90% as well [15]. This high diagnos ic e iciency,
e en in ea ly umou s ages, sugges ed i s po en ial u ili y
o diagnosis. Thus we conside ed o in e es o ex end
he alida ion o he sCD26 as an ea ly bioma ke o
CRC, including also p ecance ous lesions (adenomas).
One o he no el ies o he s udy is he use o a colono-
scopically heal hy coho ins ead o a blood dono coho
as he con ol popula ion. Mo eo e , his is he i s ime
he sCD26 is analyzed ega ding he clinicopa hological
cha ac e is ics o colo ec al polyps. Thus, we aim o anal-
yse he ela ionship o he se um CD26 le els wi h he
colonoscopic indings, in o de o alida e he u ili y o
his p o ein as a ma ke o CRC diagnosis.
Me hods
Popula ion
The popula ion s udied consis ed o pa ien s om he
CHUVI hospi al (Complejo Hospi ala io Uni e si a io de
Vigo; Spain) who had unde gone colonoscopy a he Gas-
oen e ology Depa men . All he p ocedu es desc ibed
we e pe o med acco ding o he clinical e hical p ac ices
o he Spanish Go e nmen . The s udy was app o ed by
he Galician E hical Commi ee o Clinical Resea ch
(2002/059) and complied wi h he ene s o he Helsinki
Decla a ion, he O iedo Ag eemen , he O ganic Law o
Da a P o ec ion 15/1999 and he Royal Dec ee 1720/
2007. Anonymi y was wa an ed h ough he use o clini-
cal his o y numbe s. Fo pa ien s willing o be included in
he s udy, in o med consen was ob ained and blood was
ex ac ed. The only exclusion c i e ion was a non-com-
ple ed colonoscopy (unde s ood as comple ed when he
caecum was eached). The 299 pa ien s included we e
bo h males and emales, wi h ages anging om 18 o 93
yea s. Clinical his o ies we e eco ded, including symp-
oms o bowel disease, pe sonal his o y o polyps, CRC
and o he cance s, amily his o y o cance , and colonos-
copy indica ion. This was mos ly due o symp oms as ec-
al bleeding (35.5%), abdominal pain (10%), dia hoea
(9%), cons ipa ion (5.7%), anaemia (5%), bu also o col-
o ec al polyp (12.7%) o cance (4.3%) su eillance, o
CRC sc eening (12.4%). All p ocedu es we e blinded.
Based on he his ological epo , pa ien s we e classi-
ied in o: no colo ec al pa hology, non-in lamma o y
bowel disease (non-IBD), in lamma o y bowel disease
(IBD), polyps and CRC; polyps we e sub-classi ied as
hype plas ic o adenomas. Ad anced adenomas we e
de ined as hose la ge han 10 mm, wi h ubulo illous o
illous his ology, o wi h high-g ade dysplasia. Pa ien s
wi h mo e han one polyp we e classi ied acco ding o he
mos ad anced lesion. CRC pa ien s we e classi ied
acco ding o Dukes' s aging. No ca cinomas in si u we e
de ec ed in he pa ien s included in he s udy.
Collec ion o blood samples and de e mina ion o he
sCD26 le els
Blood samples we e collec ed, coagula ed a oom em-
pe a u e o 20 min, and cen i uged a 2,000 g 15 min.
Se a we e s o ed a -80°C. The sCD26 concen a ion was
measu ed wi h he sCD26 ELISA ki (Bende Medsys-
ems; Vienna, Aus ia) acco ding o he manu ac u e 's
ins uc ions. Colo ime ic quan i ica ion was pe o med
wi h a mic opla e eade (model 550; Bio-Rad, USA) a
450/570 nm.
S a is ical analysis
S a is ical analyses we e pe o med wi h he SPSS pack-
age ( 16.0); es s we e wo-sided; p- alues < 0.05 we e
conside ed signi ican . No mal dis ibu ions and homo-
genei y o a iances we e e i ied by Kolmogo o -
Smi no and Le ene';s es s, espec i ely. Analysis o wo
independen samples was done by Mann-Whi ney's U,
whe eas o mo e samples we employed he K uskal-Wal-
lis es . Chi-squa e o Fische 's exac es s we e done wi h
con ingency ables, and applied o analyses ega ding
posi i i y/nega i i y o he sCD26. Bon e oni, alse dis-
co e y a e (FDR) and SGoF es s we e subsequen ly pe -
o med wi h he SGoF me a es so wa e o co ec o
mul iple compa ison [17]. The sCD26 abili y o sepa a e
heal hy om diseased pa ien s was s udied by ecei e
ope a ing cha ac e is ic (ROC) cu es. Sensi i i y and
speci ici y we e calcula ed using MedCalc ( .10.0.2).
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 3 o 9
Resul s
sCD26 le els acco ding o he colonoscopic esul s
Acco ding o he colonoscopic diagnosis, he 299 pa ien s
we e classi ied as 68 pa ien s wi h no colo ec al pa hol-
ogy (symp oma ic wi h ec al bleeding, abdominal pain,
dia hoea, anaemia, cons ipa ion, o asymp oma ic wi h
pe sonal his o y o polyps o CRC, and amily his o y o
polyps); 64 pa ien s wi h non-IBD (haemo hoids and
di e icula); 26 pa ien s wi h IBD (coli is o C ohn's dis-
ease); 108 pa ien s wi h colo ec al polyps (hype plas ic
polyps, non-ad anced adenomas and ad anced ade-
nomas); and 33 pa ien s wi h CRC.
As shown in able 1, he a e age sCD26 le el o he 6
g oups wi h no colo ec al pa hology was 641.2 ± 241.2
ng/mL; indi iduals wi h anaemia showed a conside ably
low sCD26 mean (370.8 ± 144.7 ng/mL), s a is ically di -
e en om o he non-colo ec al pa hology pa ien s (U
es p = 0.001). Rega ding pa ien s wi h colo ec al pa hol-
ogy, he non-IBD g oup exhibi ed mean sCD26 le els
simila o he non-colo ec al pa hology g oup (612.2 ±
231.0 ng/mL). In con as , pa ien s wi h IBD showed
lowe alues wi h ela i ely mo e a ia ion (434.4 ± 239.7
ng/mL). Pa ien s diagnosed as ha ing polyps (bo h hype -
plas ic and adenomas) esul ed in an in e media e mean
sCD26 concen a ion (588.0 ± 246.5 ng/mL), while
pa ien s wi h CRC egis e ed he lowes le els (403.7 ±
278.2 ng/mL).
The Kolmogo o -Smi no es demons a ed, excep
o he CRC popula ion (p = 0.035), a no mal dis ibu ion
o he sCD26 concen a ion o all he g oups. When he
le els o sCD26 we e compa ed among all he i e g oups
o pa ien s, we ound signi ican di e ences (K uskal-
Wallis es p < 0.001). Mo eo e , his di e ences we e no
ela ed o gende (U es p = 0.219) o age (≤ 50 yea s o >
50 yea s; U es p = 0.109), as epo ed in a la ge coho
s udy [18].
ROC cu e and sCD26 posi i i y
To e alua e he u ili y o he sCD26 as a umou ma ke ,
we es ima ed a cu -o o di e en ia e he CRC g oup (33
pa ien s) om he con ol popula ion (68 pa ien s wi h
no colo ec al pa hology). Fi s we cons uc ed he ROC
cu e, which esul ed in an a ea o 0.811 (95% CI, 0.721-
0.882; p < 0.0001). On he basis o his plo , able 2 shows
ou di e en po en ial cu -o alues wi h hei espec-
i e sensi i i y and speci ici y. Sensi i i y o CRC almos
eached 85% a he cu -o o 500 ng/mL, bu wi h a spec-
i ici y below 75%; when speci ici y was aised up o 83.8%
a 390 ng/mL, sensi i i y esul ed in a 57.6%. The cu -o
alue wi h he highes a e age diagnos ic pe o mance
was 460 ng/mL, showing a sensi i i y o 81.8% (95% CI,
64.5-93.0%) wi h a speci ici y o 79.4% (95% CI, 67.9-
88.3%).
A posi i e sCD26 alue was he e o e ≤ 460 ng/mL ( ig-
u e 1). The posi i i y a e inc eased om 20.6% (14/68) in
he non-colo ec al pa hology g oup o 81.8% (27/33) in
he CRC g oup. Pa ien s wi h non-IBD, IBD, hype plas ic
polyps, non-ad anced adenomas and ad anced ade-
nomas showed in e media e a es, co esponding o
28.1% (18/64), 69.2% (18/26), 22.2% (4/18), 22.5% (9/40)
and 41.7% (20/48), espec i ely.
Rela ionship be ween he sCD26 and he
clinicopa hological ea u es o polyps
We analysed he ela ionship o he sCD26 wi h he his o-
pa hological cha ac e is ics o he colo ec al polyps
ound by colonoscopy ( able 3), including: numbe o pol-
yps (1-2, 3 o mo e); size (≤ 0.5 cm, 0.6-1 cm, > 1 cm);
loca ion ( ec um, sigma, le colon, ans e se colon, igh
colon); mo phology (sessile, pedicula ed, la ); his ology
(hype plasic, adenoma) and de ailed his ology (hype -
plas ic, ubula , illous, ubulo illous); g ade o dysplasia
(absence, low-g ade, high-g ade); and p esence o
ad anced adenomas. The his ology o wo polyps was
missing. The a iables gende and age (≤ 50 and >50
yea s) we e also s udied, con i ming no s a is ical di e -
ences in ela ion o he sCD26 posi i i y (p = 0.212 and p
= 0.132, espec i ely).
The Chi-squa e es s e ealed no s a is ical di e ences
ega ding he numbe o polyps, hei size, loca ion, mo -
phology o his ology. Di e ences close o signi ican we e
obse ed be ween he sCD26 posi i i y and he g ade o
dysplasia (p = 0.056). The posi i i y a e inc eased g adu-
ally while he deg ee o dysplasia became mo e se e e:
22.2% o non-dysplas ic polyps, 32.5% o low-g ade dys-
plas ic adenomas and almos double (60.0%) o high-
g ade dysplas ic adenomas. When compa ing he sCD26
posi i i y a e o he ad anced and non-ad anced ade-
nomas popula ions, s a is ically signi ican di e ences
we e de ec ed (p = 0.048). Bon e oni, FDR and SGoF
pos -hoc es s we e no signi ican o any o hese analy-
ses.
Finally, diagnos ic pa ame e s we e calcula ed o he
de ec ion o ad anced adenomas and CRC, wi h he con-
ol g oup including all he emaining coho s ( able 4).
A he 460 ng/mL cu -o , sensi i i y esul ed in 58.0%
(95% CI, 46.5-68.9%), wi h a speci ici y o 75.5% (95% CI,
68.5-81.0%). When hese pa ame e s we e calcula ed o
he de ec ion o only CRC pa ien s, he sensi i i y
inc eased o 81.8% (95% CI, 64.5-93.0%) whe eas he
speci ici y was jus sligh ly diminished o 72.3% (95% CI,
65.0-77.2%). This alue inc eased o 90% (95% CI, 73.4-
97.8%) i he speci ici y was calcula ed o he non-symp-
oma ic g oup e sus he CRC g oup.
Discussion
The glycop o ein CD26 o dipep idyl pep idase IV
(DPPIV, E.C. 3.4.14.5) is an exopep idase o he plasma
memb ane able o elease dipep ides om he N- e minal
end o pep ides/p o eins bea ing p oline o alanine in he
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 4 o 9
penul ima e posi ion [19]. Biological luids con ain ela-
i ely high le els o sCD26, which is p esumably shed by
p o eoly ic clea age om any cell exp essing ansmem-
b ane CD26 [20]. Al hough i s o igin is s ill unclea , he
li e and he T cells a e ci ed as he mos likely sou ces
[21].
The measu emen o he sCD26 le els was pe o med
in se um om indi iduals whom, due o di e en medical
indica ions, had unde gone colonoscopy; mos o hem
e e ed abdominal, colon o ec al symp oms, o amil-
ial/pe sonal his o y o polyps o CCR. The indi iduals
wi hou colo ec al indings a e he colonoscopy we e
conside ed as he con ol coho ; he emaining we e
classi ied as: non-IBD, IBD, colo ec al polyps o CRC
pa ien s. The mean sCD26 concen a ion dec eased,
al hough non-signi ican ly, as he pa hology diagnosed
Table 1: A e age sCD26 concen a ion o he g oups s udied acco ding o he colonoscopy esul .
No colo ec al
pa hology
Clinical condi ion nMean ±
SD sCD26 (ng/mL)
Rec al bleeding 8 651.5 ± 190.2
Abdominal pain 10 586.1 ± 194.4
Dia hoea 9 649.5 ± 306.8
641.2 ± 241.2
Anaemia 7 370.8 ± 144.7
Cons ipa ion 4 782.7 ± 194.6
No symp oms* 30 698.5 ± 234.2
Colo ec al
pa hology
Diagnosis nMean ± SD
sCD26 (ng/mL)
Non-in lamma o y Haemo hoids 36 655.3 ± 240.4
bowel disease 612.2 ± 231.0
(non-IBD) Di e icula 28 556.7 ± 209.6
In lamma o y Coli is 20 413.2 ± 195.5
bowel disease 434.4 ± 239.7
(IBD) C ohn's disease 6 505.1 ± 355.8
Hype plas ic polyps 18 560.4 ± 201.3
Polyps Non-ad anced adenomas 40 611.1 ± 263.8 588.0 ± 246.5
Ad anced adenomas 48 566.7 ± 225.8
Duke's A 2 513.9 ± 240.4
Duke's B 12 369.8 ± 147.4
CRC 403.7 ± 278.2
Duke's C 15 402.0 ± 360.3
Duke's D 4 457.0 ± 323.0
SD: s anda d de ia ion o he mean; *: pe sonal his o y o polyps o CRC, o amilial his o y o CRC.
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 5 o 9
was mo e se e e (as seen in able 1), his is, om no col-
o ec al pa hology o CRC, wi h a no iceable dec ease in
he g oup wi h IBD.
In ou p e ious s udy, he con ol coho was o med
exclusi ely by heal hy dono s [15], whe eas in he p esen
s udy i was o med by indi iduals wi h con i med no col-
o ec al pa hology bu bea ing symp oms, o wi h his o y
o polyps o cance . Thus we calcula ed a new cu -o o
460 ng/mL, highe han he alue o 410 ng/mL epo ed
o heal hy dono s [15].
Acco ding o his cu -o , wi hin he no colo ec al
pa hology g oup, indi iduals wi h anaemia showed a sub-
s an ially ele a ed posi i i y a e (71.4%) as expec ed
om hei mean le els. Non-IBD exhibi ed a low posi i -
i y a e (28.1%), whe eas he IBD g oup eached 73.1%.
This pa hology is associa ed wi h a leas a 5- old
inc eased isk o CRC, ep esen ing one o he highes
isk g oups based on he in lamma ion-dysplasia-ca ci-
noma sequence [22]. In compliance wi h his sequence,
he sCD26 posi i i y a e inc eased om no colo ec al
pa hology o hype plas ic polyps and non-ad anced ade-
nomas, wi h a u he inc ease in ad anced adenomas
and CRC.
The capabili y o colo ec al polyps o de elop in o can-
ce is ela ed o he size o he lesion, he p opo ion o
illous componen and he g ade o dysplasia. In ela ion
o dysplasia, a mo phological ma ke o neoplas ic
lesions, we obse ed a di ec (posi i e) end be ween he
g ade o dysplasia and he posi i i y o he bioma ke ,
hough he e was no signi ican co ela ion be ween bo h
pa ame e s. Conce ning ad anced adenomas, a e m
commonly used o g oup adenomas ha ha e an
inc eased likelihood o malignan ans o ma ion, he
sCD26 posi i i y esul ed s a is ically signi ican .
Recen wo ks also s udied o he po en ial ma ke s in
ela ion o polyp cha ac e is ics: o he se um sul a ase
ac i i y, di e ences ega ding he numbe o adenomas
(single o mul iple) we e signi ican [23]; se um lep in,
adiponec in and esis in also di e ed be ween con ols
and pa ien s wi h adenomas o CRC, hough he e was no
ela ionship wi h dysplasia, his opa hology o polyp
localiza ion [24].
Table 2: Sensi i i y and speci ici y o he sCD26 a di e en cu -o alues o sepa a ing indi iduals wi h no colo ec al
pa hology om hose wi h CRC.
Cu -o (ng/mL) Sensi i i y (95% CI) Speci ici y (95% CI) +LR -LR
390 57.6% (38.2-74.5) 83.8% (72.9-91.6) 3.6 0.5
410 72.7% (54.5-86.7) 80.9% (69.5-89.4) 3.8 0.3
460 81.8% (64.5-93.0) 79.4% (67.9-88.3) 4.0 0.2
500 84.9% (68.1-94.8) 72.1% (59.9-82.3) 3.0 0.2
+LR: Posi i e likelihood a io; -LR: Nega i e likelihood a io.
Figu e 1 Do plo ep esen ing he sCD26 concen a ion acco ding o he colonoscopic diagnosis. Ho izon al line: 460 ng/mL cu -o . 1: No col-
o ec al indings (n = 68); 2: non-IBD (n = 64); 3: IBD (n = 26); 4: hype plas ic polyps (n = 18); 5: non-ad anced adenomas (n = 40); 6: ad anced adenomas
(n = 48); 7: CRC (n = 33).

De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 6 o 9
Table 3: Chi-squa e and Fische 's exac analyses o he clinicopa hological cha ac e is ics o polyps in ela ion o he
sCD26 posi i i y
Va iable Mean ± SD nsCD26+/n sCD26+ a e p
sCD26 (ng/mL)
Numbe
1-2 577.6 ± 244.3 28/84 33.3 0.807
3 o mo e 624.6 ± 256.1 7/24 29.2
Size
≤ 0.5 cm 525.2 ± 208.1 11/32 34.4
0.6 - 1 cm 673.9 ± 298.7 6/28 21.4 0.339
> 1 cm 579.9 ± 226.7 18/48 37.5
Loca ion
ec um 549.0 ± 190.0 7/24 29.2
sigma 617.1 ± 273.7 13/44 29.5
le colon 571.7 ± 194.5 7/21 33.3 0.475
ans e se colon 427.0 ± 157.6 4/6 66.7
igh colon 662.4 ± 324.5 4/13 30.8
Mo phology
sessile 592.8 ± 264.4 18/64 28.1
pedicula ed 585.3 ± 230.9 15/39 38.5 0.517
la 548.6 ± 126.5 2/5 40.0
His ology
hype plas ic 560.4 ± 201.3 4/18 22.2 0.413
adenomas 597.2 ± 256.1 30/88 34.1
De ailed his ology
hype plas ic 560.4 ± 201.3 4/18 22.2
ubula 605.3 ± 259.9 24/72 33.3 0.358
illous 612.6 ± 305.8 2/6 33.3
ubulo illous 529.7 ± 207.5 4/10 40.0
Dysplasia
no dysplasia 560.4 ± 201.3 4/18 22.2
low-g ade
dysplasia
598.7 ± 250.5 27/83 32.5 0.056
high-g ade
dysplasia
562.9 ± 293.1 3/5 60.0
Adenomas
non-ad anced 633.9 ± 286.9 9/40 22.5 0.048*
ad anced 566.7 ± 225.8 20/48 41.7
SD: s anda d de ia ion; nsCD26+: numbe o indi iduals wi h sCD26 le els ≤ 460 ng/mL; n : o al numbe o indi iduals pe g oup; sCD26+
a e: sCD26 posi i i y a e.
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 7 o 9
In ou expe imen al se ing we ha e also e alua ed he
diagnos ic pa ame e s o he sCD26. A he 460 ng/mL
cu -o , he sensi i i y and speci ici y o CRC e sus non-
cance g oups we e 81.8% and 72.3%, espec i ely. This
speci ici y is measu ed in he amewo k o symp oma ic
and asymp oma ic pa ien s bea ing in e media e benign
pa hologies including non-IBD and IBD as well as polyps.
When conside ing only asymp oma ic indi iduals speci-
ici y inc eases o 90%, which ag ees wi h he esul s p e-
iously published by ou g oup [15]; ne e heless a
dec ease in sensi i i y will be expec ed in his con ex .
In an asymp oma ic high- isk coho o indi iduals
wi h amilial his o y o CRC o pe sonal his o y o CRC
o adenoma, Hazazi e al. [25] sugges ed he use o a
quan i a i e immunochemical FOBT (iFOBT) o sc een-
ing o high- isk indi iduals. Excluding indi iduals wi h
anaemia, ec al bleeding and IBD, hey epo ed a sensi-
i i y and speci ici y o CRC o 100% and 85.3%, espec-
i ely.
In he classical CRC sc eening s udies in a e age- isk
indi iduals, he gFOBT (guaiac-based) is ex ensi ely
used, despi e i s wide ange o sensi i i y and speci ici y.
The mos common es s a e he Hemoccul II® ( ehy-
d a ed o un ehyd a ed) and he Hemoccul SENSA®,
hough he un ehyd a ed gFOBT is he one ecom-
mended o sc eening [26]. Fo he un ehyd a ed Hemoc-
cul II®, he mos accu a e diagnos ic pa ame e s ha e
been es ima ed wi h one- ime es ing on a Chinese [27]
and an Ame ican popula ions [28]; in bo h, colonoscopy
was pe o med o all indi iduals wi h posi i e o nega i e
gFOBT esul s, epo ing a sensi i i y o CRC o 25 and
12.9%, and a speci ici y o 80 and 95.2%, espec i ely
[27,28]. These s udies e lec a poo sensi i i y, al hough
i is sligh ly imp o ed wi h epea ed annual o biennial
es ing (54-80%), eaching up o 97.7% speci ici y [29,30].
On he o he hand, a 50% sensi i i y was ob ained o a
one- ime ehyd a ed es ing combined wi h sigmoidos-
copy, hough no speci ici y was epo ed, pe haps owing
o an inc ease in he numbe o alse posi i es due o
ehyd a ion [31].
When a highly sensi i e es like he Hemoccul
SENSA® is used, a 71-79% sensi i i y is eached wi h sin-
gle es ing, and abou 85% wi h mul iple es ing, wi h co -
esponding speci ici ies o 86% and 95% [32,33].
Howe e , hese pa ame e s a e p obably o e es ima ed as
hese s udies lacked colonoscopic examina ion o he
nega i e cases.
Besides guaic-based FOBT, iFOBT has been ecen ly
o e ed as an al e na i e o a e age- isk sc eening. The
s udies epo ed up o da e ha e shown ha he iFOBT is
mo e adequa e o sc eening because o i s high speci ic-
i y since i de ec s human globin [26].
Th oughou ou s udy, e idence was ga he ed ega d-
ing he u ili y o he sCD26 in he de ec ion o ad anced
Table 4: Sensi i i y and speci ici y o he sCD26 a he cu -o alue o 460 ng/mL
Coho nsCD26+/n Sensi i i y (95% CI)
CRC 27/33 81.8% (64.5-93.0)
Ad anced adenomas 20/48 41.7% (27.6-56.8)
CRC and ad anced adenomas 47/81 58.0% (46.5-68.9)
Non-ad anced adenomas and ad anced adenomas 30/88 34.1% (24.3-45.0)
Coho nsCD26-/n Speci ici y (95% CI)
No colo ec al indings-no symp oms 27/30 90.0% (73.4-97.8)
No colo ec al indings 54/68 79.4% (67.9-88.3)
No colo ec al indings, non-IBD, IBD, hype plas ic polyps, non-ad anced and ad anced
adenomas
191/264 72.3% (65.0-77.2)
No colo ec al indings, non-IBD, IBD, hype plas ic polyps and non-ad anced adenomas 163/216 75.5% (68.5-81.0)
nsCD26+: numbe o indi iduals wi h sCD26 le els ≤ 460 ng/mL; nsCD26-: numbe o indi iduals wi h sCD26 le els > 460 ng/mL; n : o al
numbe o indi iduals pe g oup.
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 8 o 9
adenomas. In sepa a ing CRC and ad anced adenomas
om all o he g oups, he sCD26 exhibi ed a 58.0% sensi-
i i y and a 75.5% speci ici y. Fo he same pa hologies
(CRC and ad anced adenomas) in an asymp oma ic high-
isk coho , Hazazi e al. [25] epo ed o a quan i a i e
iFOBT a sensi i i y and speci ici y o 65.3% and 87.5%,
espec i ely.
Rega ding s udies pe o med in a e age- isk indi idu-
als o he de ec ion o CRC and ad anced adenomas,
gFOBT has shown a sensi i i y o 10.8-14.3% [27,28] and
a speci ici y o 79.2% [27], and consequen ly is no ec-
ommended o he de ec ion o ad anced lesions [4,6].
On he o he hand, iFOBT showed a sensi i i y o 33.1%
and a speci ici y o 97.5%, hough hese pa ame e s we e
gi en o only dis al ad anced neoplasm as compa ed o
lexible sigmoidoscopy [25].
In ela ion o o he expe imen al se um bioma ke s o
ad anced adenomas, he CCSA-2 has shown a 97.3% sen-
si i i y wi h a 78.4% speci ici y conside ing no mal
colonoscopy, hype plas ic polyps and non-ad anced ade-
nomas [13], whe eas o CCSA-3 and -4 a combined sen-
si i i y o 91.3% and a speci ici y o 78.7% was epo ed
[14]. While sCD26 shows lowe sensi i i y, he speci ici y
is equi alen e en including pa ien s wi h con ounding
pa hologies such as IBD, which we e no p esen in o he
s udies.
Al hough sCD26 seems o pe o m adequa ely as a
blood bioma ke o CRC and ad anced adenomas, inde-
penden o he equen bu in e mi en bleeding unlike
gFOBT o iFOBT, ou s udy p esen s some limi a ions
ha should be conside ed: i) he symp oma ic popula ion
included is a high- isk o CRC, wi h an ele a ed p e a-
lence o colo ec al pa hology; ii) al hough no di e ences
ega ding age we e de ec ed, he age ange o he pa ien s
di e s om ha ecommended o sc eening; iii) he
classi ica ion o he pa ien s in o he ca ego ies p oposed
esul ed in se e al sub-g oups wi h a small numbe o
pa ien s. The e o e, u he esea ch in a la ge popula ion
and unde a sc eening con ex is desi able, along wi h he
compa ison o a sensi i e gFOBT o iFOBT, and o he
expe imen al non-in asi e me hods.
Conclusions
Ou esul s show ha measu emen o he sCD26 is a
non-in asi e and easonably sensi i e assay, which could
be combined wi h o he s such as he aecal occul blood
es , o he ea ly diagnosis and sc eening o CRC and
ad anced adenomas. Wi h his aim, we a e cu en ly ini-
ia ing a mul icen ic, p ospec i e, double-blinded s udy
in an a e age- isk popula ion, whe e he pe o mance o
he quan i i e iFOBT and he sCD26 assay will be
assessed and compa ed ega ding he gold s anda d
colonoscopy.
Compe ing in e es s
The au ho s decla e ha hey ha e no compe ing in e es s.
Au ho s' con ibu ions
LD: measu emen o he sCD26 le els, da a analysis, s a is ical e alua ion, man-
usc ip p epa a ion. AMRP: da a analysis, s a is ical e alua ion, manusc ip
p epa a ion. FJRB: S udy design, coo dina ion o he s udy. OJC: da a analysis,
manusc ip p epa a ion. DMA: Pa ien ec ui men , collec ion o samples and
clinical da a. MPD: S udy design, coo dina ion o he s udy, inal e ision o he
manusc ip . All au ho s ead and app o ed he inal manusc ip .
Acknowledgemen s
We hank he CHUVI Gas oen e ology Uni , especially D . Palla és-Pe al, D .
Ma ín-G anizo and Nini.
Au ho De ails
1Uni e sidad de Vigo. Facul ad de Biología. Depa amen o de Bioquímica,
Gené ica e Inmunología. As Lagoas-Ma cosende s/n. 36310 Vigo, Spain,
2Uni e sidad de San iago de Compos ela. Depa amen o de Bioquímica y
Biología Molecula . Edi icio CIBUS, Campus Su . 15782 San iago de Compos ela,
Spain and 3Complejo Hospi ala io Uni e si a io de Vigo. Se icio de Diges i o.
Calle Piza o 22. 36204 Vigo, Spain
Re e ences
1. Winawe SJ, Zaube AG, Ho MN, O´B ien MJ, Go lieb LS, S e nbe g SS,
Wa e JD, Schapi o M, Bond JH, Panish JF, Ack oyd F, Shike M, Ku z RC,
Ho nsby-Lewis L, Ge des H, S ewa ET, The Na ional Polyp S udy
Wo kg oup: P e en ion o colo ec al cance by colonoscopic
polypec omy. The Na ional Polyp S udy Wo kg oup. N Engl J Med 1993,
329:1977-81.
2. O'Connell JB, Magga d MA, Ko CY: Colon cance su i al a es wi h he
new Ame ican Join Commi ee on Cance six h edi ion. J Na l Cance
Ins 2004, 96:1420-5.
3. Vogelaa I, an Ballegooijen M, Sch ag D, Boe R, Winawe SJ, Habbema JD,
Zaube AG: How much can cu en in e en ions educe colo ec al
cance mo ali y in he U.S.? Mo ali y p ojec ions o scena ios o isk-
ac o modi ica ion, sc eening, and ea men . Cance 2006,
107:1624-33.
4. Smi h RA, Cokkinides V, B awley OW: Cance sc eening in he Uni ed
S a es, 2008: a e iew o cu en ame ican cance socie y guidelines
and cance sc eening issues. CA Cance J Clin 2008, 58:161-179.
5. Benson VS, Pa nick J, Da ies AK, Nadel MR, Smi h RA, A kin WS: Colo ec al
cance sc eening: a compa ison o 35 ini ia i es in 17 coun ies. In J
Cance 2008, 122:1357-1367.
6. Le in B, Liebe man DA, McFa land B, And ews KS, B ooks D, Bond J, Dash
C, Gia diello FM, Glick S, Johnson D, Johnson CD, Le in TR, Pickha d PJ,
Rex DK, Smi h RA, Tho son A, Winawe SJ, Ame ican Cance Socie y
Colo ec al Cance Ad iso y G oup, US Mul i-Socie y Task Fo ce, Ame ican
College o Radiology Colon Cance Commi ee: Sc eening and
su eillance o he ea ly de ec ion o colo ec al cance and
adenoma ous polyps, 2008: a join guideline om he Ame ican
Cance Socie y, he US Mul i-Socie y Task Fo ce on Colo ec al Cance ,
and he Ame ican College o Radiology. Gas oen e ology 2008,
134:1570-1595.
7. an Rijn JC, Rei sma JB, S oke J, Bossuy PM, an De en e SJ, Dekke E:
Polyp miss a e de e mined by andem colonoscopy: a sys ema ic
e iew. Am J Gas oen e ol 2006, 101:343-50.
8. Ha ewood GC, Wie sema MJ, Mel on LJ III: A p ospec i e con olled
assessmen o ac o s in luencing accep ance o sc eening
colonoscopy. Am J Gas oen e ol 2002, 97:3186-94.
9. Hewi son P, Glasziou P, Wa son E, Towle B, I wig L: Coch ane sys ema ic
e iew o colo ec al cance sc eening using he ecal occul blood es
(hemoccul ): an upda e. Am J Gas oen e ol 2008, 103:1541-9.
10. Melle C, E ns G, Schimmel B, Bleul A, Thieme H, Kau mann R, Mo hes H,
Se mache U, Claussen U, Halbhube KJ, Von Eggeling F: Disco e y and
iden i ica ion o alpha-de ensins as low abundan , umo -de i ed
se um ma ke s in colo ec al cance . Gas oen e ology 2005, 129:66-73.
Recei ed: 22 July 2009 Accep ed: 28 June 2010
Published: 28 June 2010
This a icle is a ailable om: h p://www.biomedcen al.com/1471-2407/10/333© 2010 De Chia a e al; licensee BioMed Cen al L d. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.BMC Cance 2010, 10:333
De Chia a e al. BMC Cance 2010, 10:333
h p://www.biomedcen al.com/1471-2407/10/333
Page 9 o 9
11. Roessle M, Rollinge W, Palme S, Hagmann ML, Be nd P, Engel AM,
Schneidinge B, P e e M, And es H, Ka l J, Bodenmulle H, Ruscho J,
Henkel T, Roh G, Rossol S, Rosch W, Langen H, Zolg W, Tacke M:
Iden i ica ion o nico inamide N-me hyl ans e ase as a no el se um
umo ma ke o colo ec al cance . Clin Cance Res 2005, 11:6550-6557.
12. Ayude D, Fe nandez-Rod iguez J, Rod iguez-Be ocal FJ, Ma inez-Zo zano
VS, de Ca los A, Gil E, de la Cadena Paez M: Value o se um alpha-L-
ucosidase ac i i y in he diagnosis o colo ec al cance . Oncology
2000, 59:310-316.
13. Leman ES, Schoen RE, Magheli A, Sokoll LJ, Chan DW, Ge zenbe g RH:
E alua ion o colon cance -speci ic an igen 2 as a po en ial se um
ma ke o colo ec al cance . Clin Cance Res 2008, 14:1349-1354.
14. Leman ES, Schoen RE, Weiss eld JL, Cannon GW, Sokoll LJ, Chan DW,
Ge zenbe g RH: Ini ial analyses o colon cance -speci ic an igen (CCSA)-
3 and CCSA-4 as colo ec al cance -associa ed se um ma ke s. Cance
Res 2007, 67:5600-5605.
15. Co de o OJ, Ayude D, Noguei a M, Rod íguez-Be ocal FJ, de la Cadena
MP: P eope a i e se um CD26 le els: diagnos ic e iciency and
p edic i e alue o colo ec al cance . B J Cance 2000, 83:1139-1146.
16. Ayude D, Paez de la Cadena M, Co de o OJ, Noguei a M, Ayude J,
Fe nandez-B ie a A, Rod iguez-Be ocal FJ: Clinical in e es o he
combined use o se um CD26 and alpha-L- ucosidase in he ea ly
de ec ion diagnosis o colo ec al cance . Dis Ma ke s 2004, 19:267-272.
17. Ca ajal-Rod íguez A, de Uña-Al a ez J, Rolán-Al a ez E: A new mul i es
co ec ion (SGoF) ha inc eases i s s a is ical powe when inc easing
he numbe o es s. BMC Bioin o ma ics 2009, 10:209.
18. De Chia a L, Rod íguez-Piñei o AM, Co de o OJ, Rod íguez-Be ocal FJ,
Ayude D, Ri as-He ada FJ, Páez de la Cadena M: Soluble CD26 le els and
i s associa ion o epidemiologic pa ame e s in a sample popula ion.
Dis Ma ke s 2009, 27:311-316.
19. Chen W-T, Kelly T: Sep ase complexes in cellula in asi eness. Cance
Me as asis Re 2003, 22:259-269.
20. Boonacke E, Van Noo den CJF: The mul i unc ional o moonligh ing
p o ein CD26/DPPIV. Eu J Cell Biol 2003, 82:53-73.
21. Go ell MD, Gysbe s V, McCaughan GW: CD26: a mul i unc ional in eg al
memb ane and sec e ed p o ein o ac i a ed lymphocy es. Scand J
Immunol 2001, 54:249-264.
22. Zisman TL, Rubin DT: Colo ec al cance and dysplasia in in lamma o y
bowel disease. Wo ld J Gas oen e ol 2008, 14:2662-2669.
23. Ma usiewicz M, K zys ek-Ko packa M, Diakowska D, G abowski K, Augo K,
Blachu K, Pa adowski L, Kus zeba-Wojcicka I, Pias M, Banas T: Se um
sul a ase ac i i y is mo e ele a ed in colonic adenomas han cance s.
In J Colo ec al Dis 2008, 23:383-387.
24. Kumo A, Daniel P, Pie uczuk M, Malecka-Panas E: Se um lep in
adiponec in, and esis in concen a ion in colo ec al adenoma and
ca cinoma (CC) pa ien s. In J Colo ec al Dis 2009, 24:275-281.
25. Hazazi R, Rozen P, Leshno M, Le i Z, Samuel Z, Waked A, Vilkin A, Maoz E,
Bi ken eld S, Ni Y: Can pa ien s a high isk o signi ican colo ec al
neoplasms and ha ing no mal quan i a i e aecal occul blood es
pos pone elec i e colonoscopy? Alimen Pha macol The 2010,
31:523-33.
26. Allison JE, Sakoda LC, Le in TR, Tucke JP, Tekawa IS, Cu T, Pauly MP,
Shlage L, Pali z AM, Zhao WK, Schwa z JS, Ransoho DF, Selby JV:
Sc eening o colo ec al neoplasms wi h new ecal occul blood es s:
upda e on pe o mance cha ac e is ics. J Na l Cance Ins 2007,
99:1462-1470.
27. Sung JJ, Chan FK, Leung WK, Wu JC, Lau JY, Ching J, To KF, Lee YT, Luk YW,
Kung NN, Kwok SP, Li MK, Chung SC: Sc eening o colo ec al cance in
Chinese: compa ison o ecal occul blood es lexible sigmoidoscopy
and colonoscopy. Gas oen e ology 2003, 124:608-614.
28. Impe iale TF, Ransoho DF, I zkowi z SH, Tu nbull BA, Ross ME: Fecal DNA
e sus ecal occul blood o colo ec al-cance sc eening in an a e age-
isk popula ion. N Engl J Med 2004, 351:2704-2714.
29. Ha dcas le JD, Chambe lain JO, Robinson MHE, Moss SM, Ama SS, Bal ou
TW, James PD, Mangham CM: Randomised con olled ial o aecal-
occul -blood sc eening o colo ec al cance . Lance 1996,
348:1472-1477.
30. Mandel JS, Bond JH, Chu ch TR, Sno e DC, B adley GM, Schuman LM,
Ede e F: Reducing mo ali y om colo ec al cance by sc eening o
ecal occul blood. Minneso a Cance Con ol G oup. N Engl J Med 1993,
328:1365-1371.
31. Liebe man DA, Weiss DG: One- ime sc eening o colo ec al cance wi h
combined ecal-occul blood es ing and examina ion o he dis al
colon. N Engl J Med 2001, 345:555-560.
32. Allison JE, Tekawa IS, Ransom LJ, Ad ain AL: A compa ison o ecal occul -
blood es s o colo ec al-cance sc eening. N Engl J Med 1996,
334:155-159.
33. Renne G, Renne HS, Mi on E, Pe e bu g Y: Popula ion colo ec al
cance sc eening wi h ecal occul blood es . Cance Epidemiol
Bioma ke s P e 2001, 10:1165-1168.
P e-publica ion his o y
The p e-publica ion his o y o his pape can be accessed he e:
h p://www.biomedcen al.com/1471-2407/10/333/p epub
doi: 10.1186/1471-2407-10-333
Ci e his a icle as: De Chia a e al., Se um CD26 is ela ed o his opa hologi-
cal polyp ai s and beha es as a ma ke o colo ec al cance and ad anced
adenomas BMC Cance 2010, 10:333