Preferential infiltration 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 1 Gynaecology, 2 Digestive Surgery and 3 Pathology, Academic Hospital Erasme, Free University of Brussels (ULB), 808 Route de Lennik, 1070 Brussels, Belgium

4 To whom correspondence should be addressed. e-mail: vincent.anaf{at}ulb.ac.be


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: Little is known about the mode and the extent of infiltration of endometriotic lesions in the large bowel. METHODS: In 31 patients undergoing large bowel resection for severe deep-infiltrating 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 ± 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 infiltration. 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 infiltrate 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-infiltrating endometriosis/endometriosis/large bowel endometriosis/perineurial invasion


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
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., 1987Go; Weed and Ray, 1987Go). 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 infiltrating endometriosis’, defined as the presence of endometrial-like glands and stroma more than 5 mm under the peritoneum (Cornillie et al., 1990Go). One of the characteristics of deep-infiltrating 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. Numerous theories have been proposed to explain pain mediation by endometriotic tissue, including the production and release of prostaglandins, inflammatory mediators such as kinins, histamine, interleukins, etc., and fibrosis 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., 1987Go; Zwas and Lyons, 1991Go; Berqvist, 1993Go). 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, 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 ‘infiltration or invasion phenomenon’ occurs at a certain moment in the evolution of the lesions (Clayton et al., 1999Go; Anaf et al., 2000aGo). However, very little is known about how endometriotic lesions infiltrate 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., 2000bGo). 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 define 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 ‘infiltration phenomenon’ of endometriosis in the large bowel could help us in the future to better define 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
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
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 ± 5.7 months. Patients’ characteristics, symptoms and surgical procedures are summarized in Table I.


View this table:
[in this window]
[in a new window]
 
Table I. Patients’ characteristics
 
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., 2000cGo).

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 µm were performed each at the proximal and distal edges of each large bowel resection specimen. The first 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 fixed in 4% formaldehyde for 12 h and then embedded in paraffin.

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, five serial macrosections of 4 µm were performed within the endometriotic lesion, cut in the direction of the largest diameter of the lesion. The first 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 calcium-binding domains to calmodulin, an important transducer of calcium-mediated 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 stroma at ectopic sites and in confirming a diagnosis of endometriosis (Sumathi and McCluggage, 2002Go; Ondo et al., 2003Go), and in particular of stromal endometriosis (Clement et al., 2000Go).

For immunohistochemistry, we used the antigen retrieval method for CD10 but not for S100 protein, as described previously (Anaf et al., 2000bGo). 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-specific binding of the primary antibodies, non-immune mouse serum at the same concentration as the primary antibodies preparation were substituted as the first layer of the serial sections for S100 and CD10 immunohistochemistry. Positive controls consisted of Schawnnoma and neurofibroma cells previously proven to express S100 protein (Anaf et al., 2000bGo). 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 ± 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 fields. 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 ± 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 field (x2).

The results are expressed as mean numbers of endometriotic lesions ± 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 field (x2).

The results are expressed as mean percentages ± 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 field (x2) in five cases.

The semiquantification 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
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Macroscopy
The mean length of the large bowel resected specimen was 16.8 ± 6.9 cm.

The mean largest diameter of the lesions was 4.1 ± 1.32 cm. It is defined 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 ± 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 ± 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).


View this table:
[in this window]
[in a new window]
 
Table II. Evaluation of the endometriotic invasion in the large bowel
 
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 involvement of the mucosa, submucosa or muscularis only was not statistically significant (Table II).

The mean number of endometriotic foci within the different layers of the large bowel per section at low power field (x2) ± SD are as follows: serosa and pericolic fatty tissue, 7.7 ± 5.4; muscularis, 11.8 ± 2.4; submucosa, 2.8 ± 3.4; mucosa, 0.5 ± 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 ± 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 superficial 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 fibres that cross the muscularis.



View larger version (121K):
[in this window]
[in a new window]
 
Figure 1. Endometriotic stromal cells (white arrow) infiltrating 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 figure).

 


View larger version (133K):
[in this window]
[in a new window]
 
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 figure.

 
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
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
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 superficial and the underlying deep lesion strongly suggests that the infiltrating 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 defined in its wall. The first 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 significantly more endometriotic lesions than the other layers of the large bowel. In contrast, the mucosa, which is poorly innervated and only contains scanty and small nerve fibres originating from the deep submucosal plexus (or ‘Henle’s plexus’, located under the muscularis mucosae) (Levine and Haggitt, 1992Go), 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 ± 15% of the endometriotic lesions were located around the nerves (perineurial invasion) or invaded the nerves (endoneurial invasion or intrafascicular invasion) (Sternberg, 1992Go) 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., 2000bGo). 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 significantly 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 infiltration 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 findings (Krasna et al., 1988Go). In contrast, perineurial and endoneurial invasion are frequently observed in pancreatic carcinoma (Zhu et al., 1999Go) and in adenoid cystic carcinoma of the Bartholin’s gland (Rosai, 1996Go; Anaf et al., 1999Go). 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 fibres 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, 1992Go). 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., 1967Go). 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, the specific receptor for NGF (Anaf et al., 2002Go). 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, 1980Go; Yamamoto and Iseki, 1992Go). 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-infiltrating endometriosis not only invade around or into nerves, but can also be responsible for hyperalgesia (Friess et al., 1999Go; Zhu et al., 1999Go). 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 field have demonstrated that the resection of a part of the bowel wall is feasible and safe (Koninckx et al., 1996Go). 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 infiltration 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 infiltration, 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., 1998Go; Roseau et al., 2000Go; Koga et al., 2003Go). 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 final 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 fistula. However such complications were not observed in this series after mean follow-up of 26 ± 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 infiltrate the large bowel with predilection around the nerves. Such a particular mode of infiltration might influence in the future the choice of the most adapted surgical option.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 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.[CrossRef]

Anaf V, El Nakadi I, Simon Ph, Englert Y, Peny M-O, Fayt I and Noël J-C (2000a) Sigmoid endometriosis and ovarian endometriosis. Hum Reprod 15,790–794.[Abstract/Free Full Text]

Anaf V, Simon Ph, El Nakadi I, Fayt I, Buxant F, Simonart T, Peny M-O and Noël J-C (2000b) Relationship between endometriotic foci and nerves in rectovaginal endometriotic nodules. Hum Reprod 15,1744–1750.[Abstract/Free Full Text]

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.[CrossRef]

Anaf V, Simon Ph, El Nakadi I, Fayt I, Simonart Th, Buxant F and Noel J-C (2002) Hyperalgesia, nerve infiltration and nerve growth factor (NGF) expression in deep adenomyotic nodules, peritoneal and ovarian endometriosis. Hum Reprod 17,1895–1900.[Abstract/Free Full Text]

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.[Abstract]

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.[Medline]

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.[Medline]

Cornillie FJ, Oosterlynck D, Lauweryns JM and Koninckx PR (1990) Deeply infiltrating pelvic endometriosis: histology and clinical significance. Fertil Steril 53,978–983.[Medline]

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-affinity receptor in chronic pancreatitis. Ann Surg 230,615–624.[CrossRef][Medline]

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.[Abstract]

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 infiltrating rectosigmoid endometriosis on transvaginal and transrectal ultrasonography. Hum Reprod 18,1328–1333.[Abstract/Free Full Text]

Koninckx PR and Martin DC (1992) Deep endometriosis: a consequence of infiltration or retraction or possibly adenomyosis externa? Fertil Steril 58,924–928.[Medline]

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 significance. Cancer 61,1018–1023.[Medline]

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 confirming a diagnosis of endometriosis. J Clin Pathol 56,78–80.[Free Full Text]

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.[Medline]

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.[CrossRef][Medline]

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 confirming a diagnosis of endometriosis. J Clin Pathol 55,391–392.[Abstract/Free Full Text]

Weed JC and Ray JE (1987) Endometriosis of the bowel. Obstet Gynecol 69,727–730.[Abstract]

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.[Medline]

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.[Abstract/Free Full Text]

Zwas FR and Lyons DT (1991) Endometriosis: an important condition in gastroenterology. Dig Dis Sci 36,353–364.[Medline]

Submitted on September 30, 2003; accepted on November 21, 2003.