Human Reproduction Vol.19, No.4 pp. 996±1002, 2004
Advance Access publication March 11, 2004
DOI: 10.1093/humrep/deh150
Preferential in®ltration of large bowel endometriosis along
the nerves of the colon
V.Anaf1,4, I.El Nakadi2, Ph.Simon1, J.Van de Stadt2, I.Fayt3, Th.Simonart3 and J.-C.Noel3
Departments of 1Gynaecology, 2Digestive Surgery and 3Pathology, Academic Hospital Erasme, Free University of Brussels (ULB),
808 Route de Lennik, 1070 Brussels, Belgium
4
To whom correspondence should be addressed. E-mail: [email protected]
BACKGROUND: Little is known about the mode and the extent of in®ltration of endometriotic lesions in the large
bowel. METHODS: In 31 patients undergoing large bowel resection for severe deep-in®ltrating endometriosis of the
sigmoid and rectum with severe digestive symptoms, we performed a prospective morphological, histological and
immunohistological study (using the monoclonal antibodies S100 for the detection of the nerves and CD10 for the
detection of the endometriotic stromal cells) on the large bowel resection specimen. The evaluation of invasion of the
large bowel by endometriosis was performed by studying the presence, localization and mean number of lesions in
the different layers of the colon, the relationship between endometriosis and the nerves of the colon, the nerve
density in the respective layers of the large bowel and the presence of endometriosis on the resection margins.
RESULTS: The most richly innervated layers of the large bowel are the most intensely involved by endometriosis.
We found that 53 6 15% of endometriotic lesions were in direct contact the nerves of the colon by means of
perineurial or endoneurial invasion. The mean largest diameter of the lesion does not seem to be correlated with the
depth of in®ltration. The margins were positive in 9.7% of cases. In cases of positive margins, the endometriotic
lesions were in close histological relationship with the nerves. CONCLUSIONS: There is a close histological
relationship between endometriotic lesions of the large bowel and the nerves of the large bowel wall. Endometriotic
lesions seem to in®ltrate the large bowel wall preferentially along the nerves, even at distance from the palpated
lesion, while the mucosa is rarely and only focally involved.
Key words: deep-in®ltrating endometriosis/endometriosis/large bowel endometriosis/perineurial invasion
Introduction
The digestive tract represents the third anatomic localization of
endometriosis after the peritoneum and the ovaries, with an
incidence of at least 5% of all cases of women with
endometriosis (Prystowsky et al., 1987; Weed and Ray,
1987). A distinction must be drawn between the presence of
endometriotic foci on the large bowel serosa, which represents
peritoneal endometriosis, and true large bowel endometriosis,
which represents `deeply in®ltrating endometriosis', de®ned as
the presence of endometrial-like glands and stroma more than
5 mm under the peritoneum (Cornillie et al., 1990). One of the
characteristics of deep-in®ltrating endometriosis is its strong
association with pain. Symptoms of large bowel endometriosis
can result from the endometriotic lesion itself or from the
repercussions of the lesion on the physiology and anatomy of
the large bowel. However, they also depend on the site of
involvement on the digestive tract. Symptoms often include
dysmenorrhea and dyspareunia, possibly as a result of the high
incidence of concomitant pelvic lesions or due to the presence
of endometriosis in the large bowel itself. The exact mechanism by which endometriosis causes pain is still debated.
996
Numerous theories have been proposed to explain pain
mediation by endometriotic tissue, including the production
and release of prostaglandins, in¯ammatory mediators such as
kinins, histamine, interleukins, etc., and ®brosis and cyclical
haemorrhages. Digestive symptoms usually occur when the
intestinal lumen is distorted or when peristalsis is affected.
Patients can present with an indolent progressive course with
intermittently relapsing cramping mid-abdominal pain and
abdominal distension, or with acute small or large bowel
obstruction. Intestinal dysfunction can cause constipation if the
lesion is located in the distal colon, and diarrhea or loose stools
if it is in the small intestine. Symptoms are initially cyclical but
tend to become permanent when the lesions progress (Weed
et al., 1987; Zwas and Lyons, 1991; Berqvist, 1993). Of course,
not all patients with large bowel endometriosis require surgery.
Treatment should be individualized, depending on the presence
of symptoms, the location of lesions and the patient's desire to
preserve fertility. In the absence of obstructive symptoms and
with well-tolerated pain, treatment can be initiated with
hormonal treatments, and close radiological and clinical
supervision. With obstructive symptoms and severe pain,
Human Reproduction vol. 19 no. 4 ã European Society of Human Reproduction and Embryology 2004; all rights reserved
Large bowel endometriosis in®ltration along colon nerves
surgery is indicated. Until now there has been no consensus
about how extensive or aggressive large bowel surgery should
be performed. Does anterior rectal resection or sigmoid
segmental resection represent overtreatment? Is discoid resection preferable? Should the resection be limited to the edges of
the palpated induration, or more widely performed? Until now
all these important questions remain unanswered. Not all small
asymptomatic lesions will inexorably progress into large
lesions and cause bowel obstruction. However, there are
numerous reports in the literature supporting that digestive
endometriosis is a progressive disease, and that an `in®ltration
or invasion phenomenon' occurs at a certain moment in the
evolution of the lesions (Clayton et al., 1999; Anaf et al.,
2000a). However, very little is known about how endometriotic
lesions in®ltrate or progress into the bowel wall itself. We
recently demonstrated that there is a close histological
relationship between some deep retroperitoneal endometriotic
lesions and the subperitoneal nerves, and that this relationship
is correlated with pain scores (Anaf et al., 2000b). Such a
relationship has never before been demonstrated in an
intraabdominal organ such as the large bowel. Therefore, we
performed a prospective morphological, histological and
immunohistological study on 31 specimen of large bowel
resection for severe endometriotic involvement. The aim of
this study is not to try to de®ne which type of surgery should be
performed for large bowel endometriosis, which would require
a large prospective randomized or comparative study.
Nevertheless, we believe that a better comprehension of the
`in®ltration phenomenon' of endometriosis in the large bowel
could help us in the future to better de®ne the most adequate
surgical treatment for patients, and to use the most selective
tools that will help us to determine which part of an organ is
involved and should be treated.
Materials and methods
Patients
We performed large bowel resection for symptomatic endometriosis
in 31 patients presenting with pain (n = 31, 100%) and infertility (n =
17, 55%).
The mean follow-up of patients was 26 6 5.7 months. Patients'
characteristics, symptoms and surgical procedures are summarized in
Table I.
Surgical procedures
In 18 patients (58%), we performed segmental sigmoid resection and
simultaneous resection of concomitant endometriotic lesions by a
laparoscopically assisted technique, as described previously (Anaf
et al., 2000c).
In cases of low lesions (rectosigmoid junction, rectum, rectovaginal
septum), we performed an anterior rectal resection by the same
laparoscopically assisted technique as for sigmoid lesions, but with a
lower dissection of the rectum and a partial posterior colpectomy in 10
patients (32%). In three patients (10%) anterior rectal resection was
performed by laparotomy.
Three conditions were required for the choice of the place of
proximal and distal large bowel division. First, the large bowel was
divided in an area free of any induration at manual palpation; second,
in an area free of any serosal or muscular endometriotic implant (when
it is under the Douglas pouch); and third, at a distance of at least three
centimeters from the edges of the palpated lesion.
Evaluation of the resection margins
For the study of the resection margins, two sections of 4 mm were
performed each at the proximal and distal edges of each large bowel
resection specimen. The ®rst was stained with haematoxylin and eosin
and the second was used for S100 immunohistochemistry for the
detection of nerve structures.
Immunohistochemistry
After large bowel resection, the specimens were immediately ®xed in
4% formaldehyde for 12 h and then embedded in paraf®n.
The large bowel specimen was examined and several measurements
were performed. The length of the large bowel resection, the largest
diameter of the lesion and the largest diameter of the serosal implants
were calculated using a graduated scale. For each specimen of
resection, ®ve serial macrosections of 4 mm were performed within the
endometriotic lesion, cut in the direction of the largest diameter of the
lesion. The ®rst section was stained with haematoxylin and eosin, the
other sections were used for immunohistochemistry using S100 and
CD10 monoclonal antibodies and controls.
S100 protein is a highly sensitive marker for myelinated nerves that
are normally present in the rectovaginal septum. S100 monoclonal
antibody is directed against an acidic, dimeric calcium binding protein
(molecular weight 21 000 kDa) composed of different combinations of
alpha and beta subunits, and it is present in the nucleus and cytoplasm
of Schwann cells. S100 protein is structurally similar in the calciumbinding domains to calmodulin, an important transducer of calciummediated signals.
CD10 is a very sensitive and diagnostically useful immunohistochemical marker of normal endometrial stroma and of endometrial
stromal neoplasms. It is also very useful in demonstrating endometrial
Table I. Patients' characteristics
n
Age (years)
Localization of lesion [n (%)]
Symptoms [n (%)]
Surgical technique [n (%)]
31
34.1 6 7.3
Sigmoid, 13 (42)
Rectum and R-sigmoid junction, 18 (58)
Dysmenorrhoea, 31 (100)
Dyspareunia, 31 (100)
Chr. pelv. pain, 31 (100)
Const/diarrhoea, 31 (100)
Cramps, 25 (81)
Subocclusion, 25 (81)
Rectalgia, 17 (55)
Rectorrhagia, 6 (19)
Segmental resection, 18 (58)
Anter. rectal resection, 13 (42)
Laparoscopy, 28 (90)
Laparotomy, 3 (10)
Chr. pelv. pain = chronic pelvic pain; const/diarrhoea = alternation of constipation and diarrhoea; segmental resection = segmental sigmoid resection; anter.
rectal resection = anterior resection of the rectum.
997
V.Anaf et al.
stroma at ectopic sites and in con®rming a diagnosis of endometriosis
(Sumathi and McCluggage, 2002; Ondo et al., 2003), and in particular
of stromal endometriosis (Clement et al., 2000).
For immunohistochemistry, we used the antigen retrieval method
for CD10 but not for S100 protein, as described previously (Anaf et al.,
2000b). The characteristics and dilution of the primary antibodies
were: monoclonal antibody S100 protein (clone 15E2E2; dilution 1/
100; Biogenex, San Ramon, CA, USA) and the monoclonal antibody
CD10 (clone 56C6; dilution 1/50; Novocastra Laboratories,
Newcastle, UK).
To control for non-speci®c binding of the primary antibodies, nonimmune mouse serum at the same concentration as the primary
antibodies preparation were substituted as the ®rst layer of the serial
sections for S100 and CD10 immunohistochemistry. Positive controls
consisted of Schawnnoma and neuro®broma cells previously proven
to express S100 protein (Anaf et al., 2000b). Positive controls for
CD10 were normal endometrium.
Evaluation of the endometriotic invasion in the large bowel
The evaluation of the invasion of the large bowel by endometriosis
was performed by studying the presence, localization and mean
number 6 SD of lesions in the different layers of the colon, the
relationship between endometriosis and the nerves of the colon, the
nerve density in the respective layers of the large bowel and the
presence of endometriosis on the resection margins.
A histological layer of the large bowel was considered as positive
when it contained endometriotic foci or CD10-positive cells (stromal
cells). This was studied in serial sections stained with haematoxylin
and eosin and CD10 monoclonal antibody in at least 10 high power
®elds. The results are expressed as percentages of cases showing
positive staining in each of the respective layers of the large bowel
(pericolic fat and serosa, muscularis, submucosa and mucosa).
The precise involvement of the different constitutive layers of the
large bowel by endometriotic lesions was assessed by calculating the
mean number 6 SD of endometriotic foci in the different layers of the
large bowel, in at least 10 randomly selected serial sections, entailing
the whole thickness of the large bowel at low power ®eld (32).
The results are expressed as mean numbers of endometriotic lesions
6 SD in each of the respective layers of the colon.
The relationship between endometriotic lesions and the nerves of
the large bowel was assessed by counting the number of typical
endometriotic lesions and stromal endometriotic lesions in 2 cm2 of
each serial section stained with haematoxylin and eosin, S100 protein
and CD10 monoclonal antibody in each case, at low power ®eld (32).
The results are expressed as mean percentages 6 SD of
endometriotic lesions being in direct contact with nerves, or invading
nerves.
The nerve content of the different layers of the large bowel was
semiquantitatively evaluated in four large bowel serial sections
stained with the S100 protein and entailing the full thickness of the
large bowel, at low power ®eld (32) in ®ve cases.
The semiquanti®cation is expressed as follows: +++, strong
presence of nerves; ++, moderate presence of nerves; +, weak
presence of nerves; 0, absence of nerves.
Statistical analysis
Student's t-test (two-tailed) was used for the comparison of the
importance of the endometriotic involvement between the different
layers of the large bowel. The same test was used for the comparison
of the mean largest diameter of lesions with mucosal, submucosal or
only muscular involvement.
Results
Macroscopy
The mean length of the large bowel resected specimen was
16.8 6 6.9 cm.
The mean largest diameter of the lesions was 4.1 6 1.32 cm.
It is de®ned as the mean largest diameter of the induration on
the large bowel.
The mean length of the security margins on both sides of the
lesion (mean length between the lesion and the proximal or
distal part of the lesion and the site of large bowel division) was
3.9 6 3.2 cm.
An endometriotic serosal implant was present at the surface
of the lesion in all cases (100%). The mean largest diameter of
the serosal implant was 0.7 6 0.2 cm.
Evaluation of the endometriotic invasion in the large bowel
The study of the presence of endometriotic lesions and stromal
endometriosis in the different layers of the large bowel showed
the presence of endometriosis in the serosa and pericolic fat in
31 cases (100%), in the muscularis in 31 cases (100%), in the
submucosa in 21 cases (68%) and in the mucosa in eight cases
(26%) (Table II).
In all patients with an involvement of the mucosa (n = 8), the
submucosa and the muscularis were also involved. Similarly,
in all patients with an involvement of the submucosa (n = 13),
the muscularis was also involved by endometriosis.
Ten patients (32%) showed an involvement of the
muscularis only, 13 patients (42%) had an involvement of
the muscularis and the submucosa, and eight patients (26%)
had an involvement of the three layers of the large bowel.
Using Student's t-test (two-tailed), the comparison between
the mean largest diameter of the lesions showing an involve-
Table II. Evaluation of the endometriotic invasion in the large bowel
Presence of endometriosis (%) (n = 31)
Mean number of endometriotic foci 6 SD (n = 31)
Largest diameter of lesion with mucosal, submucosal or
muscular involvement (mean 6 SD)
Nerve density
Serosa and
pericolic fat
Muscularis
Submucosa
Mucosa
100
7.7 6 5.4*
100
11.8 6 2.4*
4.03 6 1.51**
68
2.8 6 3.4*
4.22 6 1.16**
26
0.5 6 1.1*
4.16 6 0.99**
++
+++
++
0/+
*P < 0.001 between all these values using the Student's t-test (two-tailed).
**P = 0.05 between all these values using the Student's t-test (two-tailed).
998
Large bowel endometriosis in®ltration along colon nerves
Figure 1. Endometriotic stromal cells (white arrow) in®ltrating along the nerves (black arrow) of
the Auerbach plexus between the smooth muscle layers of the large bowel muscularis (S-100
immunohistochemistry, the nerves appear in black in the ®gure).
ment of the mucosa, submucosa or muscularis only was not
statistically signi®cant (Table II).
The mean number of endometriotic foci within the different
layers of the large bowel per section at low power ®eld (32) 6
SD are as follows: serosa and pericolic fatty tissue, 7.7 6 5.4;
muscularis, 11.8 6 2.4; submucosa, 2.8 6 3.4; mucosa, 0.5 6
1.1 (Table II).
The mean number of endometriotic foci was higher in the
muscularis than in the serosa and pericolic fat (P < 0.001),
higher in the muscularis than in the submucosa (P < 0.001), and
higher in the submucosa than in the mucosa (P < 0.001).
S100 immunohistochemistry in the large bowel sections
showed that 53 6 15% of endometriotic lesions were in direct
contact with nerves and along the nerve pathway (Figure 1), or
invaded the nerves (Figure 2). In all lesions there was a
super®cial serosal implant that was histologically in direct
continuity with the underlying deep lesion and in close
relationship with the nerves. Endometriotic lesions located
within the internal or external muscularis are located around
the small nerve ®bres that cross the muscularis.
The nerve content of the large bowel wall was semiquantitatively expressed as follows: pericolic fat and serosa, ++;
muscularis, +++; submucosa, ++; mucosa, 0/+ (Table II).
Resection margins
Histology and immunohistochemistry performed with the
monoclonal antibody against CD10 and S100 protein showed
that the resection margins of the bowel resection were positive
for endometriosis in six cases (19%).
This means that although the resection site had been chosen
according the following three criteria: an area free of any
induration at manual palpation, free of any serosal or muscular
visible implant, and at a distance of at least 3 cm from the edges
of the palpated lesion, 9.7% of the margins were positive for
endometriosis.
All positive margins cases were located at the distal edge of
the resected specimen.
All cases with a positive bowel margin concerned patients
who underwent anterior rectal resection.
Discussion
All large bowel resected specimens, even those of the
rectosigmoid with an involvement of the rectovaginal septum,
showed the presence of a peritoneal lesion that was
histologically in direct continuity with the underlying deep
endometriotic lesion. Such histological continuity between the
super®cial and the underlying deep lesion strongly suggests
that the in®ltrating large bowel endometriotic lesion originates
from the progression or invasion of a lesion primarily located
on the serosa of the large bowel.
The large bowel displays a very well structured innervation.
Two main nerve plexi are de®ned in its wall. The ®rst one,
called the `Auerbach plexus', runs between the the external
outer muscular layer and the circular inner layer. The second
one is called the `Meissner plexus' and is located in the
submucosa, between the muscularis and the muscularis
mucosae. The study of nerve density in the different layers of
the large bowel shows that the most richly innervated layer of
the bowel is the muscularis, followed by the serosa and the
submucosa. The muscularis also contains signi®cantly more
endometriotic lesions than the other layers of the large bowel.
999
V.Anaf et al.
Figure 2. Endometriotic foci in the large bowel muscularis (S100-immunohistochemistry). The
nerve (black arrow) is almost totally dissected by stromal cells (white arrow). The dissected nerve
fascicles appear as black points in the ®gure.
In contrast, the mucosa, which is poorly innervated and only
contains scanty and small nerve ®bres originating from the
deep submucosal plexus (or `Henle's plexus', located under the
muscularis mucosae) (Levine and Haggitt, 1992), is rarely and
only focally involved by endometriosis.
Of course such observation does not automatically suggest
that there is a close relationship between the nerves of the colon
and the endometriotic lesions. Indeed, in any extrinsic process,
malignant or not, it is expected that the more external layers are
the more frequently and intensely involved by the process.
Nevertheless, we found that 53 6 15% of the endometriotic
lesions were located around the nerves (perineurial invasion) or
invaded the nerves (endoneurial invasion or intrafascicular
invasion) (Sternberg, 1992) of the large bowel. In rectovaginal
septum endometriotic nodules removed without large bowel
resection, we previously demonstrated that in patients with a
pain score >7, 27±36% of the endometriotic lesions showed
perineurial invasion or endoneurial invasion (Anaf et al.,
2000b). Moreover, if the invasion phenomenon of endometriosis was like any other extrinsic process, it would be
expected that large lesions would more frequently involve the
mucosa than small lesions. However, the mean largest diameter
of lesions with mucosal involvement is not signi®cantly
different from the largest diameter of lesions with submucosal
or only muscular involvement. In the present study, the
presence of 9.7% of positive resection margins, at a distance of
>3 cm from the palpated lesion, in a area free of any serosal
implant, free of any induration and showing endometriotic
lesions around nerves, suggests that nerves represent a
preferential route of propagation of the disease. Perineurial
or endoneurial invasion does not represent a common route of
1000
in®ltration for benign or malignant conditions. In large bowel
adenocarcinoma, perineurial invasion is a rare event that
occurs in 14% of cases and is a sign of advanced disease,
usually accompanied by other ominous pathological ®ndings
(Krasna et al., 1988). In contrast, perineurial and endoneurial
invasion are frequently observed in pancreatic carcinoma (Zhu
et al., 1999) and in adenoid cystic carcinoma of the Bartholin's
gland (Rosai, 1996; Anaf et al., 1999). This might explain why
this latter condition is often painful before it becomes palpable.
Until now, there is no clear explanation for the relationship
between nerves and endometriotic lesions. Possibly, endometriotic lesions follow the nerve pathways and extend longitudinally because nerves represent a zone of least resistance and
the muscularis a `mechanical obstacle'. Even when endometriotic lesions are located within the internal or external
muscularis they are located around the small nerve ®bres that
cross the muscularis. From the descriptive point of view, most
large bowel endometriotic lesions represent a plaque lesion
rather than a nodular lesion, as they can be encountered in the
bladder or in the rectovaginal septum (Koninckx and Martin,
1992). In addition to the possible mechanical barrier caused by
the muscularis, deep endometriotic lesions are accompanied by
an important smooth muscle hyperplasia. The observation of
epithelial invasion of the perineurial and endoneurial space was
at one time considered evidence of malignant disease. The
theoretical basis for the viewpoint that perineurial invasion was
evidence for malignancy was weakened when Rodin concluded
that the perineurial space is not a lymphatic space (Rodin et al.,
1967). We recently demonstrated that deep pelvic endometriotic lesions express the nerve growth factor (NGF) and that
the nerves from the inferior hypogastric plexus express Trk-A,
Large bowel endometriosis in®ltration along colon nerves
the speci®c receptor for NGF (Anaf et al., 2002). This raises
the question of a possible tissular chemotactism between
endometriotic lesions and the underlying nerves. It has indeed
been demonstrated that NGF is a positive chemotaxin
(Gundersen and Barrett, 1980; Yamamoto and Iseki, 1992).
NGF is a neurotrophin that plays a key role in the occurrence of
hyperalgesia.
Interestingly, lesions expressing NGF like pancreatic
carcinoma, chronic pancreatitis or deep-in®ltrating endometriosis not only invade around or into nerves, but can also be
responsible for hyperalgesia (Friess et al., 1999; Zhu et al.,
1999). Indeed, patients with a rectal or rectovaginal lesion can
present with an important exacerbation of pain when pressure
is exerted on the lesion at physical examination. It appears to be
a pain-triggering zone, suggesting a close relationship between
the endometriotic lesions and the nerves.
From the therapeutic point of view, in this series, all patients
underwent large bowel resection. However, only eight patients
had involvement of the mucosa. We must admit that in some
cases, a more conservative treatment with the preservation of
the deepest layers of the large bowel could have represented a
less aggressive option. Pioneers in the laparoscopic surgical
®eld have demonstrated that the resection of a part of the bowel
wall is feasible and safe (Koninckx et al., 1996). Our indication
for large bowel resection was based on the size of the lesion
and the degree of stenosis. Adapting the surgical treatment
according to the depth of in®ltration will probably help to
personalize the surgical option. In this study, the mean largest
diameter of the lesion does not seem to be correlated with the
degree of in®ltration, but it has recently been reported that
transvaginal and transrectal ultrasonography could be useful in
the detection of the involvement of the different layers of the
rectum (Chapron et al., 1998; Roseau et al., 2000; Koga et al.,
2003). However, such techniques are only suitable for lesions
that are accessible to the ultrasonographic probe and not to
higher lesions on the colon. Additionally, microscopic
involvement of the submucosa or mucosa will probably escape
detection.
Endometriosis is a benign disease and therefore a sparing
resection is preferred.
In this series, the resection margins were positive at ®nal
histology in 9.7% of the margins. In all cases the margins were
found to be positive in the muscularis with the presence of
endometriotic lesions located along or into nerves of the
Auerbach plexus. The exact consequences of positive margins
on the resected bowel are still not clear, but positive margins
might theoretically be responsible for potential complications
such as local recurrence or anastomotic ®stula. However such
complications were not observed in this series after mean
follow-up of 26 6 5.7 months. In order to avoid positive
margins, pre-operative biopsies of the rectum with frozen
sections might be performed. Another surgical option should
be to perform larger resections. However, in the case of a low
located lesion (involvement of the rectum well beyond the
Douglas pouch), a larger resection should require a lower
anastomosis with subsequent functional complications (diarrhoea, faecal spotting, etc.) of rectal amputation with coloanal
anastomosis.
What we can conclude from this study is that endometriotic
lesions are found in close histological relationship with the
nerves of the colon, even at distance from the palpated area,
and seem to in®ltrate the large bowel with predilection around
the nerves. Such a particular mode of in®ltration might
in¯uence in the future the choice of the most adapted surgical
option.
References
Anaf V, Buxant F, Rodesch F, Simon P, Van de Stadt J, Noel JC and Van
Geertruyden J (1999) Adenoid cystic carcinoma of the Bartholin's gland:
what is the optimal approach? Eur J Surg Oncol 26,406±409.
Anaf V, El Nakadi I, Simon Ph, Englert Y, Peny M-O, Fayt I and NoeÈl J-C
(2000a) Sigmoid endometriosis and ovarian endometriosis. Hum Reprod
15,790±794.
Anaf V, Simon Ph, El Nakadi I, Fayt I, Buxant F, Simonart T, Peny M-O and
NoeÈl J-C (2000b) Relationship between endometriotic foci and nerves in
rectovaginal endometriotic nodules. Hum Reprod 15,1744±1750.
Anaf V, Sperduto N, Simon Ph, Noel J-C and El Nakadi I (2000c)
Laparoscopically assisted segmental sigmoid resection for sigmoid
endometriosis (LASSR). Gynaecol Endosc 9,95±101.
Anaf V, Simon Ph, El Nakadi I, Fayt I, Simonart Th, Buxant F and Noel J-C
(2002) Hyperalgesia, nerve in®ltration and nerve growth factor (NGF)
expression in deep adenomyotic nodules, peritoneal and ovarian
endometriosis. Hum Reprod 17,1895±1900.
Berqvist A (1993) Different types of extragenital endometriosis: a review.
Gynecol Endocrinol 207,207±221.
Chapron C, Dumontier I, Dousset B, Fritel X, Tardif D, Roseau G, Chaussade
S, Couturier D and Dubuisson JB (1998) Results and role of rectal
endoscopic ultrasonography in patients with deep pelvic endometriosis.
Hum Reprod 13,2266±2270.
Clayton RD, Hawe JA, Lowe JC Wilkinson N and Garry R (1999) Recurrent
pain after hysterectomy and bilateral salpingo-oophorectomy for
endometriosis: evaluation of laparoscopic excision of residual excision.
Br J Obstet Gynaecol 106,740±744.
Clement PB and Young RH (2000) Two previously underemphasized features
of endometriosis: micronodular stromal endometriosis and endometriosis
with stromal elastosis. Int J Surg Pathol 8,223±227.
Cornillie FJ, Oosterlynck D, Lauweryns JM and Koninckx PR (1990) Deeply
in®ltrating pelvic endometriosis: histology and clinical signi®cance. Fertil
Steril 53,978±983.
Friess H, Zhu Z, diMola FF, Kulli C, Graber HU, Andren-Sandberg A,
Zimmermann A, Korc M, Rheinshager M and Buchler MW (1999) Nerve
growth factor and its high-af®nity receptor in chronic pancreatitis. Ann Surg
230,615±624.
Gundersen RW and Barrett JN (1980) Characterization of the turning response
of dorsal-root neurites towards nerve growth factor. J Cell Biol 87,546±554.
Koga K, Osuga Y, Yano T, Momoeda M, Yoshino O, Hirota Y, Kugu K,
Nishii O, Tsutsumi O and Taketani Y (2003) Characteristic images of
deeply in®ltrating rectosigmoid endometriosis on transvaginal and
transrectal ultrasonography. Hum Reprod 18,1328±1333.
Koninckx PR and Martin DC (1992) Deep endometriosis: a consequence of
in®ltration or retraction or possibly adenomyosis externa? Fertil Steril
58,924±928.
Koninckx PR, Timmermans B, Meuleman C and Penninckx F (1996)
Complications of CO2 laser endoscopic excision of deep endometriosis.
Hum Reprod 10,2263±2268.
Krasna MJ, Flanchbaum L, Cody RP, Shneibaum S and Ari GB (1988)
Vascular and neural invasion in colorectal adenocarcinoma. Incidence and
prognostic signi®cance. Cancer 61,1018±1023.
Levine DS and Haggitt RC (1992) Colon. In Sternberg S (ed.) Histology for
Pathologists. Raven Press, New York, NY, USA, pp. 573±582.
Ondo T, Ban S and Shimizu M (2003) CD10 is useful in demonstrating
endometrial stroma at ectopic sites and in con®rming a diagnosis of
endometriosis. J Clin Pathol 56,78±80.
Prystowsky JB, Stryker SJ, Ujiki GT and Poticha SM (1987) Gastrointestinal
endometriosis. Incidence and indications for resection. Arch Surg 123,855±
858.
Rodin AE, Larson DR and Roberts DK (1967) Nature of the perineural space
invaded by prostatic carcinoma. Cancer 20,1772±1779.
1001
V.Anaf et al.
Rosai J (ed.) (1996) Ackerman's Surgical Pathology, Vol 1, 8th edn. Mosby,
St Louis, MO, USA, pp. 837±838.
Roseau G, Dumontier I, Palazzo L, Chapron C, Dousset B, Chaussade S,
Dubuisson JB and Couturier D (2000) Rectosigmoid endometriosis:
endoscopic ultrasound features and clinical implications. Endoscopy
32,525±530.
Sternberg S (ed.) (1992) Histology for Pathologists. Raven Press, New York,
NY, USA.
Sumathi VP and McCluggage WG (2002) CD10 is useful in demonstrating
endometrial stroma at ectopic sites and in con®rming a diagnosis of
endometriosis. J Clin Pathol 55,391±392.
Weed JC and Ray JE (1987) Endometriosis of the bowel. Obstet Gynecol
69,727±730.
1002
Yamamoto M and Iseki S(1992) Nerve growth factor receptor (NGFR)-like
immunoreactivity in the perineural cells of the adult rat peripheral cells:
normal appearance and response to nerve section. Kaibogaku Zasshi
67,642±649.
Zhu Z, Friess H, diMola FF, Zimmermann A, Graber HU, Korc M and Buchler
MW (1999) Nerve growth factor expression correlates with perineural
invasion and pain in human pancreatic cancer. J Clin Oncol 17,2419±2428.
Zwas FR and Lyons DT (1991) Endometriosis: an important condition in
gastroenterology. Dig Dis Sci 36,353±364.
Submitted on September 30, 2003; accepted on November 21, 2003