Introduction
The esophagus is the only organ where morphological changes in the superficial microvasculature,
from normal squamous epithelium to invasive cancer, can be observed in vivo using
magnifying endoscopy [1]
[2]
[3]
[4]. After a report describing the intra-papillary capillary loop (IPCL), the superficial
fine vascular network of intact esophageal mucosa, was published by Inoue et al. [1], many studies have described in detail the microvascular architecture of esophageal
superficial carcinoma [2]
[3]
[5]
[6]. Kumagai et al. [3] measured the caliber of superficial blood vessels injected with Microfil. Using
surgically resected specimens, they found a linear relationship between caliber size
and depth of invasion. Inoue et al. [1] and Arima et al. [7] described a classification system based on the IPCL, and these classifications are
widely used by Japanese endoscopists [8].
Presence of a hypovascular or avascular area (AVA) is one of the criteria for determining
the depth of invasion in esophageal squamous cell carcinoma listed in the Japan Esophageal
Society’s classification based on magnifying endoscopy [9]. Arima et al. [7] reported that the presence of an AVA reflects tumor depth, and occurs when the cancer
tissue demonstrates a bulky growth pattern. They reported that the size of the AVA
is closely related to the depth of tumor invasion.
However, accurate measurement of AVA size in vivo is difficult, and thus, its estimation
remains limited. Furthermore, accurate comparison of histopathological findings with
the area of interest observed with magnifying endoscopy is difficult. The current
study was undertaken to address these issues by observing formalin-fixed specimens
obtained from endoscopic submucosal dissection (ESD) using magnifying endoscopy and
determining if accurate evaluation of the microvascular patterns of superficial esophageal
cancer is possible with this method.
Patients and methods
Patient selection
From October 2013 to September 2015, patients with esophageal squamous cell carcinoma
with an AVA detected via magnifying endoscopy, seen at the Esophageal Surgery Department,
Tokyo Medical and Dental University Hospital, and who underwent ESD were included
in this study. Those who received chemotherapy and/or radiotherapy were excluded from
this analysis. For the purpose of this study, AVA was defined as an avascular or hypovascular
area surrounded by irregular microvessels. The width of the AVA was calculated as
the distance of the widest interval between the irregular microvessels.
During the study period, we performed magnifying endoscopy in 122 cases, wherein we
detected 139 superficial esophageal squamous cell carcinoma lesions. Among these cases,
23 (16.5 %) AVA lesions were detected. Three cases of AVA were not included in this
study owing to inadequate preoperative examinations.
Twenty patients with AVA lesions (type 0-IIc, n = 14; type 0-IIc + 0-IIa, n = 5; type
0-IIc + 0-Is, n = 1) were included in this study. The study population comprised 17
men and 3 women with a mean age of 65.6 ± 8.2 (standard deviation [SD]) years (range,
41 – 78 years).
All 20 patients provided informed consent for the ESD and for the use of their resected
tumor samples for research purposes through a notice board placed in our hospital’s
outpatient area. The study was performed under a protocol approved by our hospital
ethics committee (registration number: M2015-555).
Materials and procedure
Magnifying endoscopy with blue laser imaging (BLI) (EG-L590ZW; Fujifilm, Tokyo, Japan)
was used for endoscopic diagnosis and identification of the AVAs ([Fig. 1 a]). A soft black hood was attached to the tip of the scope to obtain the appropriate
distance from the lesions and to focus on the surface of the lesions accurately during
magnification.
Fig. 1 Observation and measurement of avascular areas (AVAs). a Magnifying endoscopic image with blue laser imaging under a low-power field. b Formalin-fixed specimens, each with a 2-mm width. c Identification of the concerned areas that had AVAs under a low-power field. d Measurement of the width of the AVAs using fine electronic Vernier calipers.
ESD was performed in the Department of Endoscopic Diagnostics and Therapeutics under
sedation or in the operating room under general anesthesia. After iodine staining
and marking the borders of the lesion, glycerol solution (10 % glycerol 300 mL) along
with indigo carmine (0.6 mL) and 0.1 % adrenaline (0.6 mL) was injected into the submucosal
layer to lift the lesion from the muscularis propria. The incision of the mucosa started
at the distal margin of the lesion followed by proximal extension with a flush knife
(Fujifilm, Tokyo, Japan). Then, submucosal dissection was performed using the flush
knife and Mucosectom (Pentax, Tokyo, Japan). The resected lesion was extended, stuck
on a board, and fixed with formalin.
Magnifying observation after formalin fixation
Two to 3 days after the specimens were resected and fixed with formalin, they were
sliced, each with a 2-mm width, without cutting through the specimen completely, at
the Division of Pathology ([Fig. 1 b]). Then, the microvascular structure of the lesions was observed with magnifying
endoscopy. The surface of the lesions was examined and AVAs were determined under
a low-power field ([Fig. 1 c]); then the area was observed under a high-power field.
If the cut line passed through the AVA, a cross-sectional view of the area was obtained
via magnifying endoscopy with BLI. The AVA width with the widest interval across the
slice was measured using fine electronic Vernier calipers ([Fig.1 d]). To avoid formalin contamination of the sample, we made sure to wipe off the formalin
from the specimens before observation, conduct the observation in a short period under
well-ventilated conditions, and clean the endoscope immediately after observation
in compliance with the in-hospital rule.
Histopathological analysis
Histopathological specimens were stained with hematoxylin and eosin to assess the
depth of tumor invasion. We informed the pathologists which specimen slides had AVAs,
who then confirmed the depth of the concerned area. The thickness of the AVAs that
had the widest interval was measured using a microscope at × 100 magnification before
pathological examination.
Tumor invasion was categorized as M1 (cancer limited to the epithelium), M2 (cancer
invading the lamina propria mucosa), M3 (cancer reaching or invading the muscularis
mucosa), SM1 (cancer invading submucosa < 200 μm), or SM2 (cancer invading submucosa
> 200 μm).
Statistical analysis
Statistical data are expressed as the mean and 95 % confidence interval. The statistical
significance of the differences between groups was analyzed with the Mann-Whitney
U test. The correlation between AVA width and thickness was analyzed using Spearman’s
rank correlation coefficient. P values < 0.05 were considered statistically significant. Data analysis was performed
using the statistical package Ekuseru-Toukei 2012 (Social Survey Research Information
Co., Ltd., Tokyo, Japan).
Results
Eighteen lesions from 16 tissue samples obtained from patients who underwent ESD for
superficial esophageal cancer with AVAs were examined and included in this study.
Four cases were excluded because in 3 cases, the AVA lesions were lost after formalin
fixation owing to traumatic damage during the ESD, and in 1 case, the pathologist’s
final diagnosis was basaloid-squamous cell carcinoma ([Fig. 2]). Eighteen AVA lesions that were clearly visible on BLI after formalin fixation
were examined histologically from these groups of patients. The number of lesions
in the analyzed area based on depth of invasion was as follows: M1, 5 lesions; M2,
9 lesions; M3, 3 lesions; SM1, 1 lesion.
Fig. 2 Flowchart of all cases analyzed.
The tumors’ endoscopic and histopathological appearance is shown in [Fig. 3]. The M1 lesions formed small AVAs surrounded by crushed spot-like microvessels with
irregular calibers. Tumor thickness was very small at this area. The M2 lesions showed
expansive downward or upward growth patterns and formed slightly bigger AVAs than
the M1 lesions, which did not reach statistical significance. The M3 or SM1 lesions
usually had stretched and irregularly branched vessels at the surface of the lesion
and appeared thickened ([Table 1]).
Fig. 3 Endoscopic and histopathological appearance of tumors. a, b, c Magnifying endoscopic images with blue laser imaging (BLI) in vivo. d, e, f Side view of the concerned areas under magnifying endoscopy with BLI after formalin
fixation. g, h, i Histopathological features after staining with hematoxylin and eosin (× 100 magnification).
a, d, g The M1 lesion was flat and formed a small avascular area. b, e, h The M2 lesion was slightly thickened and showed an expansive growth pattern. c, f, i The SM1 lesion had stretched and irregularly branched vessels at the surface and
inside of the lesion. The widest interval between vessels at the surface across the
slice was measured.
Table 1
Comparison of AVA width and thickness among M1, M2, and M3/SM1 lesions.
|
n
|
Width (mm)
|
Thickness (mm)
|
|
M1
|
5
|
0.434 (0.390 – 0.478)
|
0.176 (0.119 – 0.233)
|
|
M2
|
9
|
0.578 (0.451 – 0.705)
|
0.518 (0.390 – 0.646)
|
|
M3/SM1
|
4
|
0.835 (0.601 – 1.069)
|
0.800 (0.652 – 0.948)
|
Values are presented as the mean (95 % confidence interval).
[Fig. 4 a] shows the relationship between the depth of tumor invasion and AVA width as measured
using Vernier calipers with magnifying endoscopy. The differences among the three
groups were not significant. [Fig. 4 b] shows the relationship between the depth of tumor invasion and AVA thickness measured
on the microscope at × 100 magnification. The differences between M1 and M2 and between
M2 and M3/SM1 lesions were significant. [Fig. 5] shows the significant relationship between AVA width and thickness.
Fig. 4 a Relationship between avascular area (AVA) width measured using Vernier calipers with
magnifying endoscopy and depth of tumor invasion. b Relationship between AVA thickness measured using a microscope and depth of tumor
invasion.
Fig. 5 Relationship between the width and thickness of the avascular areas.
Discussion
Recently, several superficial esophageal cancers have been detected owing to improvements
in digestive endoscopy. Specifically, image-enhanced endoscopy with magnification
including narrow-band imaging, BLI, and flexible spectral imaging color enhancement
has enabled clear visualization of the microsvasculature of the tumor surface [5]
[10]
[11]
[12], which led to the discovery that changes in microvasculature reflect the depth of
superficial esophageal cancer [2]
[3]
[4].
A previous study on AVA confirmed that tumor invasion becomes deeper as the size of
the AVA increases [7]. In the Japan Esophageal Society classification of magnifying endoscopy findings,
AVAs are categorized as AVA-small (< 0.5 mm, corresponds to M1-M2), AVA-middle (0.5 – 3.0 mm,
corresponds to M3-SM1), and AVA-large (> 3.0 mm, corresponds to SM2 or deeper). In
our study, all 5 M1 lesions were classified as AVA-small, 4 M2 lesions (total = 9
lesions) were AVA-small, and the other 5 lesions were AVA-middle. All M3 and SM1 lesions
were classified as AVA-middle.
The results of this study corroborate the findings of previous reports comparing the
endoscopic observation and estimation of AVA size with histopathological findings
by directly measuring the size of the AVAs after formalin fixation. This is the first
attempt to evaluate the size of the AVAs using endoscopically resected specimens,
which allows for more intensive inspection than does an in vivo study. In addition,
our results reveal that the thickness of the AVAs reflect the depth of the tumor.
However, the relationship between AVA width and depth of tumor invasion could not
be confirmed, presumably owing to the small number of patients. Meanwhile, this report
demonstrates the detailed correspondence between AVA width and thickness. Although
this study has the limitations of small number of cases, traumatizing the specimen
during the procedure, and ex vivo study after formalin fixation, these results are
important when we consider AVA formation and development.
AVAs form as the tumor develops expansively and pushes the IPCL sideways. Then, the
tumor breaks through the basal membrane and invades a deeper layer. When the tumor
reaches a certain size, the center of the tumor becomes ischemic and needs neovascularization
[13]. Experimental studies have shown that tumors rarely grow to greater than 2 or 3 mm3 without neovascularization [14]. Thus, there is a limit to the interval between vessels. Consequently, when the
tumor’s mass reaches a certain size and the tumor develops neovascularization, it
can be presumed that the tumor thickness reflects the depth of invasion more accurately
than the interval between vessels.
Thus, it is reasonable to define AVA as an avascular or hypovascular area in the Japan
Esophageal Society’s classification of magnifying endoscopy findings. The concept
of this definition indicates that the interval between vessels, or AVA width, cannot
reflect the tumor depth when the tumor becomes large, invades a deeper layer, and
develops neovascularization. However, determining the range of the hypovascular area
is difficult. Thus, there is a possibility that the AVA thickness will help to assess
the depth of invasion of the tumor, although it is impossible to accurately measure
the thickness of the AVA in vivo.
It has long been known before magnifying endoscopy became popular that the thickness
of superficial esophageal squamous cell carcinomas reflects invasion depth. Ohashi
et al. [15] have reported that the tumor thickness and diameter of invasion were correlated
with submucosal invasion. They concluded that a classification based on gross type,
thickness, depth of depressed lesions, shape of elevated lesions, and invasion patterns
should be evaluated to differentiate between M3 and SM tumors. In the case of AVAs,
an expansive growth type, our results support these findings in terms of the microvascular
structure and invasion type.
Conclusion
Magnifying endoscopy can be used to image the microvascular structure of the concerned
area and to assess the histopathological features accurately. Further investigations
are needed to elucidate the changes in AVA vascular patterns occurring during tumor
progression from M2 to M3 or SM cancers.
