Introduction
Endoscopic submucosal dissection (ESD) is an excellent instrument for en bloc resection
of precancerous and early cancerous lesions in the gastrointestinal tract but with
a steep learning curve [1]
[2]. For successful en bloc ESD, a thorough diagnostic evaluation of the lesion should
be done prior to endoscopic resection [3]. However, this can be challenging even with use of endoscopic classification methods
such as Kudo’s pit-pattern or the Japanese NBI expert team (JNET) classification [4].
Feeding information back to endoscopists after histopathological assessment could
improve their diagnostic ability and, in total, improve the quality of ESD. This method
of “mapping and virtual reconstruction” involves close collaboration between an endoscopist
and a pathologist. In countries like Korea and Japan, it is implemented routinely
for pathological work-up of ESD specimens, and correlates all conspicuous features
observed in the endoscopic image with its corresponding histopathology [5].
In this report, we demonstrate the principle of mapping and virtual reconstruction
after ESD resection of a large colonic adenoma, which had a small region of shallow
submucosal cancer.
Methods
A laterally spreading tumor (LST) of about 40 mm diameter in the proximal rectum was
examined extensively prior to ESD resection. The surface and vascular pattern of the
lesion was described using Sano’s classification as well as the JNET classification
[6]
[7]. Endoscopic examination of the lesion was done with an Olympus GIF-190 series endoscope
using white light, narrow-band imaging (NBI), indigocarmine-chromoendoscopy (IC) and
near-focus magnification.
After ESD resection, the specimen was stretched and pinned at its outer borders onto
a cork plate using standard office pins. A macroscopic close-up image of the specimen
was taken using a standard digital camera. The specimen was fixed in formalin for
about 24 hours after which a second macroscopic image was taken. A third photographic
image was taken after cutting and sectioning of the specimen. Each section was numbered
to make mapping and reconstruction possible. Finally, the histopathology of each section
was projected onto the macroscopic image. Most importantly, suspicious regions or
areas with interesting features seen on endoscopy were correlated with the histology.
Lines of different colors were used to designate the extent of normal tissue, tissue
with low-grade dysplasia, high-grade dysplasia and cancer.
Results
The endoscopic images of the laterally spreading tumor with a diameter of about 40 mm
in the rectum are shown in [Fig. 1a] and [Fig. 1b]. A greater part of the tumor has a granular surface type but a small area at the
right side with a non-granular surface is also present. This non-granular area remained
reddish after IC staining.
Fig. 1 a White light. b IC-Staining. c NBI magnified view (Sano IIIa/JNET 2B). d ESD resection.
A small region (yellow arrow) with an irregular surface was evaluated more closely
using the magnified NBI mode ([Fig. 1c]). Here an irregular vascular pattern and diminished vessel density can be seen ([Fig. 1c]). This area can be classified as Sano Type IIIA and JNET Type 2B. [Fig. 1d] shows the appearance after ESD.
The macroscopic images of the specimen before and after formalin fixation as well
as sectioning are shown in [Fig. 2a] and [Fig. 2b]. Again, the yellow arrow points at the suspicious region described above. In [Fig. 3a] the image of the sectioned specimen is shown. Sections are numbered from top to
bottom to make identification possible. Each section is examined histologically after
which the histologic results are transferred onto the macroscopic images. This process,
called mapping, results in an image with lines of different color depicting the various
histologic patterns ([Fig. 3b]).
Fig. 2 a Macroscopic image before fixation. b Macroscopic image after fixation. The upper part of the image is the oral side while
the lower part is the anal side of the lesion.
Fig. 3 a Macroscopic image after sectioning. b Macroscopic image after mapping. The upper part of the image is the oral side while
the lower part is the anal side of the lesion. The red lines show the area of pT1
adenocarcinoma.
In [Fig. 4a], [Fig. 4b], [Fig. 4c], and [Fig. 4d], the histopathology of section number 4 (counting from the top of the image) is
demonstrated. It now becomes clear that the area with a granular type surface showed
high-grade intraepithelial neoplasia while the non-granular type region had low-grade
dysplasia. Histology of the small suspicious area with Sano class IIIA and JNET Type
2B pattern showed a shallow submucosal invasive cancer (pT1 G2 600 µm).
Fig. 4 a Transition from normal mucosa to tubulo-villous adenoma – depicted as the yellow
line (normal mucosa) and the dark blue line (tubulo-villous adenoma) in [Fig. 3b]. b Tubular adenoma with high grade intraepithelial neoplasia – dark blue line in [Fig. 3b]. c Adenocarcinoma pT1, G2, sm2 (600 µm) – depicted as the red line in Fig. 3b. d Tubular adenoma with low grade intraepithelial neoplasia – depicted as the neon blue
line in [Fig. 3b].
Discussion
In this histopathological work-up of an ESD specimen from a rectal LST with a small
region of superficial submucosal cancer, we demonstrated that the process of mapping
leads to excellent topographical reconstruction of the lesion and the specimen. An
exact correlation of the histopathology with all subtle features observed during the
endoscopic examination was possible. The lesion shown in this report was chosen deliberately
to demonstrate the principles and benefits of reconstruction and mapping. Ordinarily,
the pathologic report in this case would not have provided an exact correlation with
the various features observed during endoscopy.
Sometimes, it may be difficult to compare the endoscopic view to the corresponding
region of interest on the specimen. This problem can be easily solved by pressing
two surface marks with an argon-plasma-coagulation (APC) probe at both sides of the
region of interest, making it more visible for the pathologist. Furthermore, an APC
mark can be placed at the proximal end of the lesion, which helps orientate the specimen
after en bloc resection.
Mapping of ESD specimens is not yet common practice in the West [8]. Detailed mapping of specimens demonstrated in this paper goes further than the
pathologic workup described by Kumarasinghe et al [9] in which a standardized protocol for handling, grossing, and assessing specimens
was proposed. However, it does not include the detailed mapping demonstrated in our
paper. Although the advantages of this mapping process are obvious, one limitation
may be time constraints for endoscopists and pathologists, which could hinder its
implementation in daily routine practice. Also, it involves a close discussion between
the pathologist and the endoscopist, which may not always be possible or practical
in a daily routine. A compromise for the European setting could be to reduce the mapping
and reconstruction only to those regions of interest that were particularly conspicuous
during endoscopic evaluation or at least to those regions that had a cancerous histopathology.
Conclusion
If performed routinely, mapping could improve the diagnostic ability of endoscopists
because all macroscopic details can be correlated to their histology [8]. Large ESD specimens may have only a small region of cancer or even multifocal regions
of cancer, such as in Barrett’s specimens, which may not be identified during endoscopy.
After mapping, these regions can be correlated not only onto the image of the specimen
but also onto the original endoscopic image. This correlation will educate and train
endoscopists, and could lead to better pretherapeutic evaluation of lesions prior
to endoscopic resection.
