122 Original article
Utility of intrapapillary capillary loops seen on
magnifying narrow-band imaging in estimating
invasive depth of esophageal squamous cell
carcinoma
Authors
Institutions
submitted 29. April 2014
accepted after revision
22. September 2014
Bibliography
DOI http://dx.doi.org/
10.1055/s-0034-1390858
Published online: 15.1.2015
Endoscopy 2015; 47: 122–128
© Georg Thieme Verlag KG
Stuttgart · New York
ISSN 0013-726X
Corresponding author
Haruhiro Inoue, MD, PhD
Digestive Disease Center
Showa University
Northern Yokohama Hospital
35-1 Chigasakichuo
Tsuzuki-ku
Yokohama 224-8503
Japan
Fax: +81-45-9497000
haruinoue777@gmail.com
Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
Hiroki Sato1, Haruhiro Inoue1, Haruo Ikeda1, Chiaki Sato1, Manabu Onimaru1, BuHussain Hayee2, Chainarong Phlanusi1,
Esperanza Grace R. Santi3, Yasutoshi Kobayashi4, Shin-ei Kudo1
Institutions are listed at the end of article.
Background and study aims: Intrapapillary capil-
lary loops (IPCLs) have been used to estimate his-
topathological atypia and the invasion depth of
squamous cell carcinoma (SCC). The aim of this
study was to evaluate the clinical significance of
IPCLs.
Patients and methods: A total of 358 consecutive
patients with esophageal neoplasia on magnify-
ing narrow-band imaging (M-NBI) were studied.
The lesions were categorized according to the
IPCL classification of Inoue et al. and were subse-
quently resected. Resected specimens were histo-
pathologically analyzed to determine the invasion
depth. The inter- and intraobserver agreements in
the interpretation of IPCL images were also inves-
tigated.
Results: A total of 446 lesions were diagnosed on
M-NBI as IPCL type V lesions, which were further
classified as 185 IPCL type V1, 109 type V2, 104
type V3, and 48 type Vn. Sensitivity and specifici-
Introduction
! helps to determine the depth of invasion [7, 8].
Advances in magnifying endoscopy have allowed
the identification of intrapapillary capillary loops
(IPCLs) on the surface of the esophagus [1, 2].
IPCLs are capillary loops arising perpendicularly
from smooth branching vessels in the subepithe-
lium (●" Fig. 1). In squamous cell carcinoma (SCC)
of the esophagus, Inoue et al. observed IPCL pat-
tern changes such as dilatation, weaving, change
in caliber, and variety in shape, the so-called
“four characteristic markers of cancer.” These fea-
tures are also useful in estimating atypia and the
depth of invasion [3]. Furthermore, the use of nar-
row-band imaging (NBI) has facilitated the detec-
tion of SCC by enhancing the superficial microvas-
cular patterns, which appear as brownish areas
with a demarcation line [4 – 6] (● " Fig. 2). The
combination of magnifying endoscopy and NBI
(M-NBI) has proven to be valuable in clearly iden-
tifying variations in the IPCL pattern. This enables
Downloaded by: Collections and Technical Services Department. Copyrighted material.
ty of IPCL type V1-2 for invasion confined to the
epithelium or lamina propria mucosa (m1-2)
were 89.5 % (95 % confidence interval [CI] 85.4 % –
92.7 %) and 79.6 % (95 %CI 72.3 % – 85.7 %), respec-
tively. Sensitivity and specificity of IPCL type V3
for invasion confined to the muscularis mucosa
or slight submucosal invasion (m3-sm1) were
58.7 % and 83.8 %, respectively. Sensitivity and
specificity of IPCL type Vn for deeper invasion
(sm2-3) were 55.8 % and 98.6 %, respectively. In-
terobserver agreement was substantial (κ = 0.609,
0.641, and 0.705), as was intraobserver agree-
ment (κ = 0.705 and κ = 0.819).
Conclusion: Changes in the morphology of IPCLs
on M-NBI correlated with the depth of SCC inva-
sion, and results were reproducible and reliable
among observers. Identification of IPCL type V1-
2 proved useful for the intraprocedural identifica-
tion of m1-2 lesions, which are considered an ab-
solute indication for endoscopic resection.
pretreatment endoscopic diagnosis of atypia and
Close association between IPCL caliber changes
and depth of tumor invasion has been demon-
strated [9]. Kumagai et al. reported that the cali-
ber of tumor vessels gradually increases as tu-
mors invade more deeply [10 – 12], and Santi et
al. reported that neovascularization (Vn) in inva-
sive carcinoma comprises vessels of nearly three
times the diameter of vessels in carcinoma con-
fined to the mucosa [13].
However, no detailed investigation has examined
the association between changing patterns of
IPCLs on M-NBI and the depth of invasion in his-
topathology. The usefulness of IPCLs observed
with M-NBI for estimating the depth of invasion
is still unclear in the clinical setting. The present
study was conducted to address this research gap.
 Original article 123

Branching vessel Intra-epithelial papillary

capillary loops (IPCL)

Obliquely running
vessel

Downloaded by: Collections and Technical Services Department. Copyrighted material.

Submucosal vein

Fig. 2 Narrow-band imaging of intrapapillary capillary loops type V,

Fig. 1 Schematic representation of the vascular network of esophageal showing brownish area with clear demarcation line.
mucosa.

Patients and methods IPCL Type I

!
Patients
The study was performed between 2002 and June 2013 at Showa
IPCL Type II
University Northern Yokohama Hospital, a tertiary referral center
in Japan. A total of 358 consecutive patients with lesions endo-
scopically diagnosed as IPCL type V, corresponding to carcinoma
in situ and SCC, were analyzed. Patients with residual relapse or IPCL Type III
those who had undergone chemotherapy or radiation therapy
were excluded from the study, as IPCLs were considered to have
been altered by treatment. IPCL Type IV
The present study was carried out in accordance with the Decla-
ration of Helsinki. Written informed consent was obtained from
all patients.
IPCL Type V-1 m1
Classification of IPCLs and depth of invasion Dilation, meandering, irregular
The Japanese classification of esophageal carcinoma categorizes caliber, and form variation
neoplastic lesions confined within the epithelial layer as “carci-
noma” [14]. However, these are defined as “high grade dysplasia
IPCL Type V-2 m2
(HGD)/carcinoma in situ” in Western classification systems [15, Extension of IPCL Type V-1
16]. Deeper invasion is not required for the diagnosis of carcino-
ma in the Japanese classification. Carcinoma in situ in Western
classification and carcinoma confined within mucosal epithelium IPCL Type V-3 m3, sm1
Advanced or deeper
in the Japanese classification were defined as “m1” in this study.
destruction of IPCL
Similarly, carcinomas with invasion into the lamina propria mu-
cosa, the muscularis mucosa, superficial invasion into the sub- IPCL Type Vn sm2
mucosa (SM1), and massive invasion into the submucosa (SM2- Generation of or deeper
3) were defined as “m2,” “m3,” “sm1,” and “sm2-3,” respectively. new tumor vessel

According to the IPCL classification by Inoue [2], histopathologi-

cal atypia evolves as alterations in the IPCL pattern become more
Fig. 3 Classification of intrapapillary capillary loops (IPCLs) pattern. Type I:
pronounced (i. e. IPCL type I to type V). IPCL type V is classified
normal pattern. Type II: IPCLs have one or two out of four characteristic
into the following subtypes [8] (● " Fig. 3).
changes, and elongation and/or dilatation is commonly seen. Type III: IPCLs
▶ IPCL type V1: all four characteristic findings (dilatation, weav- have minimal changes. Type IV: IPCLs have three out of four characteristic
ing, change in caliber, and variety in shape) are identified. IPCL changes described in Type V. Type V: IPCLs have all four characteristic
type V1 is considered to correspond to carcinoma in situ (m1) changes indicating carcinoma in situ: dilatation, tortuous running, caliber
(●
" Fig. 4). changes, and different shapes in each IPCL.
▶ IPCL type V2: elongation of deformed IPCLs is seen, and the
loop structure in IPCLs is conserved in its original condition.

Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
124 Original article
Fig. 4 Intrapapillary capillary loops (IPCL) type V1. All four characteristic
findings (i. e. dilatation, weaving, variation in caliber, and variation in shape)
are identified. IPCL type V1 corresponds to carcinoma in situ (m1).
Fig. 5 Intrapapillary capillary loops (IPCL) type V2. Elongation of deformed
IPCLs is seen, and the loop structure is conserved. IPCL type V2 corresponds
to carcinoma confined within the lamina propria mucosae (m2).
IPCL-type V2 is considered to correspond to carcinoma con-
fined within the lamina propria mucosa (m2) (● " Fig. 5).
▶ IPCL type V3: irregular course of IPCLs is identified with a par-
tially destroyed loop structure. IPCL V3 is considered to corre-
spond to carcinoma within the muscularis mucosa with slight
invasion into the submucosa (m3-sm1) (● " Fig. 6).
▶ IPCL type Vn: newly formed abnormal vessels (Vn) are identi-
fied with apparent caliber change, and the original loop struc-
ture of IPCLs is completely destroyed. IPCL Vn is considered to
correspond to carcinoma with deeper invasion into the sub-
mucosa (sm2-3) (● " Fig. 7).
Endoscopic procedure and treatment approach
All endoscopic procedures were carried out under intravenous
anesthesia. The magnification endoscope Q240Z was used until
Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
Downloaded by: Collections and Technical Services Department. Copyrighted material.
Fig. 6 Intrapapillary capillary loops (IPCL) type V3. Irregular course of
IPCLs is identified with a partially destroyed loop structure. IPCL type V3
corresponds to carcinoma within the muscularis mucosa with slight inva-
sion into the submucosa (m3-sm1).
Fig. 7 Intrapapillary capillary loops (IPCL) type Vn. Newly formed abnor-
mal vessels (Vn) are identified with apparent caliber change. The original
loop structure is completely destroyed. IPCL type Vn corresponds to carci-
noma invading into the submucosa (sm2-3).
2007, and the H260Z (Olympus Co., Tokyo, Japan) was used
thereafter. A soft black plastic cap (2 mm depth, MB-46; Olym-
pus) was attached to the distal end of the scope to maintain an
appropriate distance from the target area and to provide a clear
view during magnification.
Macroscopic morphology of the tumor was examined under
white-light endoscopy and classified according to the Paris classi-
fication [15, 17]. M-NBI was employed subsequently to assess the
IPCL pattern, and tumors were categorized according to the clas-
sification of Inoue et al. (●
" Fig. 3) [18, 19]. IPCLs were categorized
into IPCL type I to type V, and if categorized into IPCL type V, they
were then subcategorized into type V1, V2, V3, or Vn. For larger
lesions with multiple IPCL patterns, the highest quality M-NBI
image from the macroscopically worst area (e. g. high protrusion,

deep depression, and large nodule) was evaluated. Tumor inva-
sion depth was determined in a comprehensive manner by con-
ventional white-light endoscopy, NBI magnification, and in some
cases, endoscopic ultrasonography (EUS). Only characteristic
findings were included for the purposes of this study, and equi-
vocal findings were not accepted. Chromoendoscopy using Lu-
gol’s solution was performed to confirm the presence of non-
staining lesions and to more clearly delineate the tumor margins.
The study was conducted in a Japanese tertiary referral center
that specializes in gastroenterology. All procedures were per-
formed by trained gastroenterologists who had extensive experi-
ence and who had undergone specific training by Professor In-
oue. Therefore, maintenance of the quality of NBI magnification
and accuracy of diagnosis was assured during the study period.
The basic structure and maximum magnification rate of the two
endoscopes used during the study period were identical, and no
difference in the diagnostic capabilities of these two endoscopes
is likely.
All patients underwent computed tomography (CT) scanning be-
fore any treatment. Tumor morphology on conventional endos-
copy, IPCL pattern on M-NBI, and presence or absence of lymph
node metastasis (LNM) on CT were considered when choosing
the therapeutic approach.
Patients with a preoperative diagnosis of carcinoma confined to
the mucosal membrane (cT1a) with an absence of LNM on CT un-
derwent endoscopic mucosal resection (EMR) or endoscopic sub-
mucosal dissection (ESD) [20]. For patients with more invasive le-
sions extending into the submucosa (cT1b), or with LNM on CT,
radical surgery was recommended.
The primary end point was to determine the diagnostic correla-
tion between IPCL type V1-2 and pathological m1-m2 invasion.
IPCL type V1-2 corresponds to lesions of the epithelium and the
lamina propria mucosa (m1-2). Such lesions have almost no risk
of LNM [21, 22], and are an absolute indication for endoscopic re-
section.
A secondary end point was to investigate the diagnostic correla-
tion between IPCL type V3 and m3-sm1, and between IPCL type
Vn and sm2-3. IPCL type V3, which corresponds to muscularis
mucosa and SM1 invasion (m3-sm1), has about a 10 % risk of
LNM following the pathological findings such as lymphatic inva-
sion, and is therefore considered to be a relative indication for
endoscopic resection [20, 23, 24]. IPCL type Vn lesions, which cor-
respond to SM2-3 (sm2-3), are best treated by esophagectomy.
Patients who were not suitable for esophagectomy because of un-
derlying medical conditions underwent chemoradiotherapy or
ESD as alternative treatments.
Intra- and interobserver agreement using IPCL
classification
Inter- and intraobserver agreement values for the assessment of
IPCLs were calculated. A total of 93 consecutive lesions, repre-
senting the final cases collected between 2012 and 2013, were
selected for this purpose. In each case, two digital still pictures
demonstrating the most characteristic pattern changes of IPCLs
under M-NBI were chosen. The 93 pairs of M-NBI images were ar-
ranged at random for the assessment of interobserver agree-
ment. The assessment was performed by two independent, ex-
ternal reviewers (B.H. [reviewer B] and E.S. [reviewer C]) of differ-
ent nationalities, and with varied experience in the use of M-NBI.
The definition of each IPCL finding and typical images were made
available before the assessment. The 93 images were classified
into one of the three subcategories of IPCL type V1-2, V3, and
Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
Original article 125
Vn. Kappa values were calculated between our consensus (repre-
senting reviewer A) and the opinions of reviewer B and reviewer
C.
For the assessment of intraobserver agreement, the same set of
93 images was re-arranged at random and sent to reviewers B
and C a week after their initial exposure to the images. All prepa-
ration for this substudy was performed by C.S. and H.S., and all re-
viewers were blinded.
Histopathological assessment
Endoscopically and surgically resected specimens were fixed in
10 % formalin and were split vertically into 2-mm-thick slices;
whole mount sections were then prepared. Tissue samples were
embedded in paraffin wax and sliced into 3-μm sections. Speci-
mens were stained with hematoxylin and eosin, and if invasion
into the muscularis mucosa or a deeper layer was suspected, ad-
ditional immunohistochemical staining was performed using
Downloaded by: Collections and Technical Services Department. Copyrighted material.
monoclonal antibody QBEND/10 (CD34) and desmin for endothe-
lial cells and muscle cells, respectively.
A positive CD34 staining was defined as any vessel demonstrat-
ing a brown lining on CD34 immunostaining, which indicated
the presence of “vascular invasion.” Desmin stains muscle and
helps to verify the depth of tumor invasion; interruption of the
stained muscle layer by neoplastic cells indicated a breach in the
muscularis mucosa and therefore invasion of the submucosa.
In addition, SM1 (superficial invasion into the submucosa) was
defined as invasion of ≤ 200 μm below the muscularis mucosa
into the submucosa in the EMR or ESD specimen, or up to one-
third in the surgical specimen. Deeper invasion was defined as
SM2-3.
Statistical analysis
All analyses were performed using STATA version 11.2 (STATA
Corp., College Station, Texas, USA). Continuous variables are
expressed as means ± SD, and categorical variables as number
(%). The 95 % confidence intervals (CIs) were also calculated.
Inter- and intraobserver agreements were assessed using kappa
statistics and the quadratic weight method [25]. Kappa values
were interpreted according to Landis and Koch [26], where
kappa = 0 demonstrated absence of agreement; < 0.20, slight
agreement; 0.21 – 0.40, fair agreement; 0.41 – 0.60, moderate
agreement; 0.61 – 0.80, substantial agreement; and > 0.81, almost
perfect agreement.
Results
!
Clinicopathological features of SCC in the esophagus
A total of 446 lesions in 358 consecutive patients were analyzed.
Clinicopathological characteristics are shown in ● " Table 1.
According to the macroscopic types, there were 22 0-I + lesions
(4.9 %), 48 0-IIa + lesions (10.8 %), 71 0-IIb + lesions (15.9 %), 291
0-IIc + lesions (65.2 %), and 14 0-III + lesions (3.1 %). The mean le-
sion size was 20.0 ± 20.4 mm. EMR, with or without a cap fitted
to the endoscope, was performed in 86 patients (19.3 %). ESD
was performed in 298 patients (66.8 %), and surgery was per-
formed in 62 patients (13.9 %). Histopathological studies revealed
132 m1 lesions (29.6 %), 162 m2 (36.3 %), 52 m3 (11.7 %), 23 sm1
(5.2 %), and 77 sm2-3 (17.3 %) lesions. There were no adverse
events during any endoscopic procedure.
126 Original article

Table 1 Clinicopathological features. Table 2 Relationship between intrapapillary capillary loop patterns and his-
No. of patients/lesions 358/446 topathological depth of invasion.
Sex (male/female), n 299/58 IPCL category Histopathological depth of invasion1
Age, mean ± SD (range), years 67 ± 9.3 (45 – 92)
Macroscopic types, n (%) m1-m2 m3-sm1 sm2-3 Total
0-I+ 22 (4.9) V1 169 12 4 185
0-IIa+ 48 (10.8) V2 94 14 1 109
0-IIb+ 71 (15.9) V3 31 44 29 104
0-IIc+ 291 (65.2) Vn 0 5 43 48
0-III+ 14 (3.1) Total 294 75 77
Size of lesions, mean ± SD (range), mm 20 ± 20.4 (1 – 120)
IPCL, intrapapillary capillary loop.
Treatment, n (%) 1
m1: tumor limited to the epithelium; m2: tumor invading the lamina propria muco-
Endoscopic mucosal resection 1 86 (19.3) sa; m3: tumor invading the muscularis mucosa; sm1: tumor with superficial invasion
Endoscopic submucosal dissection 298 (66.8) (≤ 200 μm or one-third below the muscularis mucosa into the submucosa); sm2-3:
tumor with massive invasion (> 200 μm or one-third into the submucosa).
Surgery with lymph node resection 62 (13.9)
Histopathology, n (%)
Carcinoma in situ (m1) 2 132 (29.6)
Carcinoma confined to lamina propria mucosa (m2) 162 (36.3) Table 3 Interobserver and intraobserver agreement (kappa value).

Downloaded by: Collections and Technical Services Department. Copyrighted material.

Carcinoma confined to muscularis mucos (m3) 52 (11.7)
Carcinoma invading into SM1 (sm1)3 23 (5.2)
Carcinoma invading into SM2 (sm2-3)4 77 (17.3)
1
With or without a cap fitted to the endoscope.
2
m1: tumor limited to the epithelium.
3
sm1: tumor with superficial invasion (≤ 200 μm or one-third below the muscularis
mucosa into the submucosa).
4
sm2-3: tumor with massive invasion ( > 200 μm or one-third into the submucosa).
Association between IPCL pattern changes and
histological diagnosis
All lesions with IPCL type V were subclassified into IPCL type V1
(n = 185), V2 (n = 109), V3 (n = 104), and Vn (n = 48). The validity
measures of the different IPCL stages on M-NBI and invasion
depth as determined by histopathological examination are pres-
ented in ●" Table 2.
IPCL type V1-2 showed excellent correspondence with histopa-
thologically proven cancer limited to m1-m2, with a sensitivity
and specificity of 89.5 % (95 %CI 85.4 % – 92.7 %) and 79.6 % (95 %CI
72.3 % – 85.7 %), respectively.
IPCL type V3, corresponding to m3-sm1 lesions, had a sensitivity
of 58.7 % (95 %CI 46.7 % – 69.9 %) and specificity of 83.8 % (95 %CI
79.7 % – 87.4 %), whereas IPCL type Vn, which corresponds to
sm2-3 lesions, had a sensitivity and specificity of 55.8 % (95 %CI
44.1 % – 67.2 %) and 98.6 % (95 %CI 96.9 % – 99.6 %), respectively.
Inter- and intraobserver agreement
Interobserver agreement of A – B, B – C, and C – A was κ = 0.609,
0.641, and 0.705, respectively. Intraobserver agreement of B and
C were κ = 0.705 and 0.819, respectively (●
" Table 3).
Discussion
! Vn implies that these vascular changes are not high incidence
Esophageal cancer is difficult to recognize by routine white-light
endoscopy without dye staining. Lugol’s chromoendoscopy has
been used to aid detection; however, this staining is time con-
suming, and may cause retrosternal burning and allergic reac-
tions. Although Lugol’s staining and NBI have comparable results
for SCC detection, Lugol’s solution is also associated with an in-
creased incidence of false-positive lesions and cannot predict
the depth of invasion [5, 27].
One cannot overemphasize the role of tumor invasion depth in
the determination of patient treatment and prognosis. However,
tumor invasion progresses gradually, and the accuracy of endo-
A–B B–C C–A
[95 %CI] [95 %CI] [95 %CI]
Interobserver 0.609 0.641 0.705
agreement [0.552 – 0.676] [0.587 – 0.697] [0.608 – 0.797]
B C
Intraobserver 0.705 0.819
agreement [0.629 – 0.829] [0.790 – 0.906]
CI, confidence interval.
scopic diagnosis is considered limited because of the absence of
apparent anatomical landmarks between the epithelium, lamina
propria mucosa, SM1, and SM2-3 lesions. On the other hand, it
has been postulated that invasion into the muscularis mucosa
would cause changes in the appearance of the minute vascular
structure of papillary capillary loops. Accordingly, determining
the correspondence between IPCL types V1-2 and m1-m2 was
defined as a primary goal in the present study. Indeed, the diag-
nostic correlation of IPCL types V1 and V2 for m1-m2 was excel-
lent, with a sensitivity and specificity of 89.5 % and 79.6 %, respec-
tively. Excessive inflammation or keratinization, which obscured
the vascular patterns, and tumors that showed “minute invasion
into the muscularis mucosa,” which is considered difficult to
identify endoscopically, prevented these results from being even
more remarkable.
IPCL type V3 and IPCL type Vn had a specificity of 83.8 % and 98.6 %,
for predicting m3-sm1 and sm2-3 lesions, respectively. If IPCL
type V3 or type Vn is identified within the brownish areas de-
monstrating IPCL types V1 and V2, the possibility of LNM needs
to be considered when considering patient management. Al-
though considered specific markers for m3-sm1 and sm2-3 le-
sions, respectively, the low sensitivity of IPCL type-V3 and type-
rate markers. Therefore, if the muscularis mucosa or submucosal
invasion is suspected on the basis of tumor morphology, but IPCL
type V3 or type Vn is not seen, EUS should be used to predict the
depth of invasion. The wide confidence intervals may be due to
the low number of cases.
The inter- and intraobserver agreement study showed substan-
tial agreement, confirming that IPCL classification has a strong
reproducibility and reliability among observers with different
backgrounds and with different levels of experience. Endoscopic
diagnosis using IPCL classification relies on experience, and it is
important to understand that there is a learning curve in the in-

Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128

Fig. 8 Endocytoscopy
images using GIF-Y0002
(prototypes from Olym-
pus Co., Tokyo, Japan),
at × 380 magnification
(tissue field of view,
700 × 600 μm). a Intra-
papillary capillary loops
(IPCL) type V1. b IPCL
type V2. The caliber
change of IPCLs is clear-
ly identified.
terpretation of IPCL images. Recently, an automated endoscopic
diagnostic system has been developed for the diagnosis of colo-
rectal polyps by quantifying the changes in pit pattern [28]. In
the future, an automated system to evaluate the pattern of IPCLs
may allow even greater predictive accuracy (● " Fig. 8).
The IPCL classification has some limitations. In the present study,
histological results of all lesions with IPCL type V were carcinoma
in situ or carcinoma. However, IPCL type IV (“severe dysplasia”),
incompletely unstained with iodine, is also thought to corre-
spond to HGD and is sometimes removed by endoscopic resec-
tion in our daily practice. According to Western histopathological
criteria, both HGD corresponding to IPCL type IV and the so-
called carcinoma in situ corresponding to IPCL type V1 are classi-
fied together. Histopathological atypia is a gradual change, and
HGD includes all noninvasive neoplastic epithelial lesions, some
of which could be labeled as carcinoma in situ. Further studies fo-
cusing on IPCL types IV and V1 are needed to clarify their rela-
tionship to the histological atypia present in these lesions.
Another point worth mentioning is that the use of vascular pat-
terns to diagnose tumor depth should only serve as an adjunct to
other validated measures (i. e. morphological characteristics ob-
served on conventional endoscopy or EUS). When microvascula-
ture cannot be observed, modalities such as EUS may be em-
ployed to predict invasion depth. At present, it is still risky to as-
sess the depth of invasion by vascular pattern alone. This was a
single-center, observational study, and future multicenter, con-
trolled studies are needed. The accuracy of postprocedural diag-
nosis cannot be evaluated from the present study, and this also
represents an area of future study.
Arima et al. reported the concept of an “avascular area” [29],
which may complement the IPCL pattern in predicting the depth
of tumor invasion. Moreover, the Japan Esophageal Society has
developed a new classification of magnifying endoscopy using
both IPCLs and the avascular area, and studies using this classifi-
cation are ongoing [30].
Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
Original article 127
In conclusion, changes in the morphology of IPCLs on M-NBI are
well correlated with the depth of SCC invasion. Identification of
IPCL type V1-2 is promising for the intraprocedural identification
of m1-2 lesions, which are considered an absolute indication for
endoscopic resection.
Competing interests: None
Institutions
1
Digestive Disease Center, Showa University, Northern Yokohama Hospital,
Yokohama, Japan
2
Department of Gastroenterology, King’s College Hospital NHS Foundation
Trust, London, United Kingdom
3
Department of Gastroenterology, De La Salle University Medical Center,
Manila, Philippines
4
Kobayashi Internal Medicine Clinic, Kobe, Japan
References
1 Inoue H, Honda T, Yoshida T et al. Ultra‐high magnification endoscopy
Downloaded by: Collections and Technical Services Department. Copyrighted material.
of the normal esophageal mucosa. Dig Endosc 1996; 8: 134 – 138
2 Inoue H. Magnification endoscopy in the esophagus and stomach. Dig
Endosc 2001; 13: 40 – S41
3 Inoue H, Honda T, Nagai K et al. Ultra‐high magnification endoscopic
observation of carcinoma in situ of the esophagus. Dig Endosc 1997;
9: 16 – 18
4 Gono K, Obi T, Yamaguchi M et al. Appearance of enhanced tissue fea-
tures in narrow-band endoscopic imaging. J Biomed Opt 2004; 9:
568 – 577
5 Muto M, Minashi K, Yano T et al. Early detection of superficial squa-
mous cell carcinoma in the head and neck region and esophagus by
narrow band imaging: a multicenter randomized controlled trial. J
Clin Oncol 2010; 28: 1566 – 1572
6 Yoshida T, Inoue H, Usui S et al. Narrow-band imaging system with
magnifying endoscopy for superficial esophageal lesions. Gastrointest
Endosc 2004; 59: 288 – 295
7 Inoue H. [Endoscopic diagnosis of tissue atypism (EA) in the pharyn-
geal and esophageal squamous epithelium; IPCL pattern classification
and ECA classification]. Kyobu Geka 2007; 60: 768 – 775
8 Inoue H, Cho J, Satodate H et al. High-magnification endoscopic diagno-
sis of the stratified squamous esophageal cancer. Stomach Intestine
2003; 38: 1629 – 1640
9 Kaga M, Inoue H, Kudo SE et al. Microvascular architecture of early
esophageal neoplasia. Oncol Rep 2011; 26: 1063 – 1067
10 Kumagai Y, Inoue H, Nagai K et al. Magnifying endoscopy, stereoscopic
microscopy, and the microvascular architecture of superficial esopha-
geal carcinoma. Endoscopy 2002; 34: 369 – 375
11 Kumagai Y, Toi M, Kawada K et al. Angiogenesis in superficial esopha-
geal squamous cell carcinoma: magnifying endoscopic observation
and molecular analysis. Dig Endosc 2010; 22: 259 – 267
12 Kumagai Y, Toi M, Inoue H. Dynamism of tumour vasculature in the ear-
ly phase of cancer progression: outcomes from oesophageal cancer re-
search. Lancet Oncol 2002; 3: 604 – 610
13 Santi EG, Inoue H, Ikeda H et al. Microvascular caliber changes in intra-
mucosal and submucosally invasive esophageal cancer. Endoscopy
2013; 45: 585 – 588
14 The Japan Esophageal Society. Japanese classification of esophageal
cancer. 10th: edition Tokyo: Kanehara Co; 2008
15 Bosman FT, Carneiro F, Hruban RH, Theise ND. WHO classification of tu-
mours of the digestive system. Lyon: IARC Press; 2010
16 Rice TW, Blackstone EH, Rusch VW. 7th edition of the AJCC Cancer Stag-
ing Manual: esophagus and esophagogastric junction. Ann Surg Oncol
2010; 17: 1721 – 1724
17 [Anonymous] The Paris endoscopic classification of superficial neo-
plastic lesions: esophagus, stomach, and colon: November 30 to De-
cember 1, 2002. Gastrointest Endosc 2003; 58: 3 – 43
18 Minami H, Inoue H, Ikeda H et al. Usefulness of background coloration
in detection of esophago-pharyngeal lesions using NBI magnification.
Gastroenterol Res Pract 2012; 2012: 529782
19 Chai NL, Ling-Hu EQ, Morita Y et al. Magnifying endoscopy in upper
gastroenterology for assessing lesions before completing endoscopic
removal. World J Gastroenterol 2012; 18: 1295 – 1307
128 Original article
20 The Japan Society. Guidelines for the clinical and pathologic studies on
carcinoma of the esophagus. 10th edn. Tokyo: Kanehara Shuppan;
2010
21 Eguchi T, Nakanishi Y, Shimoda T et al. Histopathological criteria for ad-
ditional treatment after endoscopic mucosal resection for esophageal
cancer: analysis of 464 surgically resected cases. Mod Pathol 2006; 19:
475 – 480
22 Araki K, Ohno S, Egashira A et al. Pathologic features of superficial
esophageal squamous cell carcinoma with lymph node and distal me-
tastasis. Cancer 2002; 94: 570 – 575
23 Oyama T, Miyata Y, Shimaya S. Lymph nodal metastasis of m3, sm1
esophageal cancer. Stomach Intestine 2002; 37: 71 – 74
24 Higuchi K, Tanabe S, Koizumi W et al. Expansion of the indications for
endoscopic mucosal resection in patients with superficial esophageal
carcinoma. Endoscopy 2007; 39: 36 – 40
25 Fleiss JL, Cohen J. The equivalence of weighted kappa and the intraclass
correlation coefficient as measures of reliability. Educational and Psy-
chological Measurement 1973; 33: 613 – 619
Sato Hiroki et al. Intrapapillary capillary loops and cancer invasion … Endoscopy 2015; 47: 122–128
26 Landis JR, Koch GG. The measurement of observer agreement for cate-
gorical data. Biometrics 1977; 33: 159 – 174
27 Ide E, Carneiro FO, Frazao MS et al. Endoscopic detection of early esoph-
ageal squamous cell carcinoma in patients with achalasia: narrow-
band Imaging versus Lugol’s Staining. J Oncol 2013; 2013: 736756
28 Takemura Y, Yoshida S, Tanaka S et al. Computer-aided system for pre-
dicting the histology of colorectal tumors by using narrow-band ima-
ging magnifying colonoscopy (with video). Gastrointest Endosc 2012;
75: 179 – 185
29 Arima M, Tada M, Arima H. Evaluation of microvascular patterns of su-
perficial esophageal cancers by magnifying endoscopy. Esophagus
2005; 2: 191 – 197
30 Oyama T, Ishihara R, Takeuchi M et al. Usefulness of Japan Esophageal
Society classification of magnified endoscopy for the diagnosis of su-
perficial esophageal squamous cell carcinoma. Gastrointest Endosc
2012; 75: AB456
Downloaded by: Collections and Technical Services Department. Copyrighted material.