Progression of Deep Infiltrating Rectosigmoid Endometriotic Nodules

Antoine Netter; Perrine d'Avout-Fourdinier; Aubert Agostini; Isabella Chanavaz-Lacheray; Marta Lampika; Marilena Farella; Clotilde Hennetier; Horace Roman


Hum Reprod. 2019;34(11):2144-2152. 

In This Article


Our results suggest that in 28% of patients who have a deep infiltrating rectosigmoid nodule, growth will occur over a 3-year period. This risk of progression is directly related to the presence of menstruation during this time, as nodule growth will occur in 39% of women free of amenorrhoea. These findings are of major importance when selecting expectant management rather than surgery or hormonal suppression for women with rectosigmoid endometriotic nodules.

Our study has several limitations. The inter-MRI interval of 3 years is of short duration, when compared to the potentially long period of development of endometriotic nodules over an average of 30 years, from menarche to menopause. We therefore cannot extrapolate our findings to nodules followed up over any period of longer duration. The sample size is small, due to exclusion criteria which eliminated half of the patients presenting for review with two consecutive MRIs. Our cut-off values defining the progression and the regression of nodules may be considered restrictive, as in the group of women with nodules demonstrating progression, mean length and thickness increased 2-fold when compared to baseline measures. Due to the absence of predefined criteria, the cut-off measurements used to confirm the progression or regression of rectosigmoid nodules were chosen according to published criteria used to define solid tumour growth (Eisenhauer et al., 2009). Our findings cannot be extrapolated to asymptomatic women with rectosigmoid nodules, because all patients included in our analysis reported either pain or infertility or both.

All MRI examinations were performed externally, prior to patient review in our centre. As there is currently no international consensus regarding MRI patient preparation, protocols or reporting criteria (Bazot et al., 2017), there is heterogeneity of MRI techniques, which were performed by radiologists from a range of facilities in France. Our inclusion criteria required the availability of three main sequences, images of high quality and without artefact hampering nodule assessment. These criteria were strict; thus, 17 patients, despite presenting with rectosigmoid nodules demonstrated on two MRIs, were ultimately ineligible for enrolment in our study. We cannot confirm the effect these criteria may have had on our results. However, these limitations did not allow us to monitor nodule growth, but only to demonstrate that their size may increase, particularly in patients having periods.

Another limitation of the study was the inclusion of patients with a prior history of pelvic surgery (48.8%). However, none of these patients had previously undergone surgery for deep infiltrating endometriosis; thus, it is unlikely that previous surgeries had an impact on the growth of any deep endometriotic nodules.

One may object that our study population is heterogeneous, due to the inclusion of women undergoing various hormonal treatments, women who were pregnant and/or breastfeeding and women with ongoing menses. The sample is also heterogeneous in terms of length of interval between MRIs and length of interruptions in medical treatment during the interval. In reality it is difficult, if not impossible, to conduct a long-term study in patients with endometriosis completely free of medical or surgical treatment. Symptomatic patients, who are likely to benefit from hormonal or surgical treatments, have a lower probability of eligibility for inclusion at the time of follow-up. Further studies are required to assess the risk of progression of rectosigmoid nodules over the whole 30-year reproductive period.

Our study has several strengths. As our cut-off values are restrictive, the definition and therefore confirmation of growth or regression of nodules is robust. We chose to enrol only patients with rectosigmoid nodules, because their measurement is both comprehensive and reproducible: the length is measured longitudinally while the thickness is measured orthogonally to the bowel wall. We were able to accurately estimate the impact of amenorrhoea on the growth, stability or regression of nodules, because all patient medical records consistently documented duration of menstruation, amenorrhoea, pregnancy, breastfeeding and medical treatment occurring in the inter-MRI interval. All MRIs were reviewed by a senior radiologist experienced in deep endometriosis.

Regarding the technique for radiological measurement of nodules, our results demonstrate good intra- and inter-observer reproducibility for measurement of both length and thickness of lesions. To our knowledge, there are no specific recommendations regarding the technique for measurement of endometriotic nodules on MRI. We deliberately chose not to measure the width, the area or the volume of deep endometriotic nodules for two reasons: (i) these measurements can be challenging and are therefore less reproducible due to the irregular shape of endometriotic nodules and (ii) the length and the thickness of nodules are clinically relevant measurements, as the length is routinely used to define the feasibility of bowel disc excision, while the thickness is employed to estimate the feasibility of rectal shaving (Abrão et al., 2015; Roman et al., 2016; Donnez and Roman, 2017).

It has already been proven in the field of radiology for malignancy that there is unavoidable intra- and inter-observer variability when solid tumour measurements are expressed in millimetres. For this reason, it has been recommended that confirmation of progression or regression of solid tumours be done via the use of thresholds: progression is defined as an increase in tumour diameter of >20%, regression is defined as a decrease in diameter of >30% and stability is defined as any growth or regression between these two thresholds (Eisenhauer et al., 2009). To increase the strength of our study, we chose large thresholds (>20% variation in length and in thickness), which defined nodules where size increased 2-fold as 'progressive'. Using these thresholds, the definition and therefore the statement of progression of nodules are more reliable for readers.

In our centre, we surgically manage more than 130 patients per year with deep rectosigmoid nodules; however, very few of these patients present with two prior pelvic MRIs of good quality and without surgical procedures during the interval between MRIs. In addition, despite our small sample size, we report robust data regarding the probability of growth or regression of endometriotic nodules, as well as statistically significant association between duration of amenorrhoea and progression of endometriotic nodules. The low number of patients did not allow us to compare amenorrhoea by cause (pregnancy, lactation, GnRH agonist or continuous combined oral contraception), although this may be interesting to study further in future studies.

The pathogenesis of deep infiltrating endometriosis is not yet fully understood (Vercellini et al., 2014; Gordts et al., 2017). Retrograde menstruation through the fallopian tubes is as yet only a theory, mostly based on indirect observations such as the common locations of lesions and the correlation between the frequency of the disease and various factors that could increase retrograde flow. Neither the mechanism of the development of deep endometriotic lesions is known nor the time at which they arise. The progression of deep endometriotic nodules has been demonstrated in three female baboons, in which nodules were induced and then followed up at 6 and 12 months, with specific analyses of gland morphology, collective cell migration and nerve fibre density. Invasion and nodule innervation increased between 6 and 12 months of follow-up. The authors suggested that nerve fibres may play a role in the development of lesions, as has previously been observed in women (Orellana et al., 2017). One cohort study on 500 women who underwent laparoscopy for endometriosis showed no correlation between patient age and the stage of endometriosis (Savaris et al., 2014). Fedele et al., 2004 conducted a prospective cohort study on 88 women presenting with asymptomatic rectal endometriotic nodules with follow-up over 68 months. Lesions were assessed every 6 months by rectal ultrasonography. Only six patients (6.8%) demonstrated progression of nodule size. The authors concluded that progression of an asymptomatic rectal nodule is unlikely to occur. However, the findings of this study cannot be compared to ours, as our patients were symptomatic, reporting either pain or infertility or both. Several biological studies have suggested that endometriotic lesions tend to regress naturally and fibrose rather than proliferate (Zhang et al., 2016a,b; González-Foruria et al., 2017), which may correlate to our group of patients with stable lesions.

There is no consensus on optimal method and frequency of follow-up in women with deep rectosigmoid nodules (Leyland et al., 2010; Dunselman et al., 2014; Collinet et al., 2018). On the basis of our results, it appears reasonable to recommend induction of continuous amenorrhoea in patients with rectosigmoid nodules in whom surgical management has not been performed. There is a very low risk of progression of deep endometriotic nodules infiltrating the rectosigmoid in women with amenorrhoea induced by medical therapy, pregnancy or lactation. In those patients who do not benefit from amenorrhoea, due to desire to conceive or refusal of medical treatment, we recommend close surveillance with symptoms and routine imaging (every 1–2 years) to allow early detection of growth and progression of rectosigmoid nodules. This recommendation also concerns young patients who wish to conceive, in whom first line IVF has been recommended. In these patients, especially if there is a long time to conception, there is often ongoing menstruation and a lack of amenorrhoea.

Finally, we did not find any previous reports in the literature describing the occurrence of a new rectosigmoid nodule, as occurred in two of the patients in our study. This observation is compatible with our demonstration of the progression of rectosigmoid endometriotic lesions.

The rationale for the assessment of progression of deep rectosigmoid lesions is based on two main concerns. First, cases of bowel occlusion in women desiring pregnancy, either spontaneously or by ART, have been reported in the literature (Roman et al., 2015). Although the incidence of this serious complication is low, its consequences may be disastrous. In our opinion, physicians should consider this outcome in women with large rectosigmoid nodules, which appear to grow (Roman et al., 2015). Second, the growth of nodules may change the surgical approach, render surgical procedures more challenging and increase the risk of unfavourable postoperative outcomes. In large nodules > 30 mm diameter, conservative rectal surgery is less feasible, which may subsequently impact on functional outcomes (Abrão et al., 2015; Donnez and Roman, 2017). Furthermore, the increase in risk of complications is not only related to nodule size but also to the direction of extension and subsequent involvement of the ureters, splanchnic nerves or sacral roots.

The efficacy of hormonal contraception or GnRH agonists to reduce pain symptoms related to endometriosis has been widely studied (Fedele et al., 1993; Vercellini et al., 1993; Leone Roberti Maggiore et al., 2014). Our results suggest that hormonal treatments, which induce amenorrhoea may prevent growth of nodules and may even result in the regression of lesions. Continuous hormonal treatment and induction of amenorrhoea are recommended in patients with rectosigmoid nodules in whom surgery is, for various reasons, not performed. In infertile women managed with ART, continuous hormonal treatment is recommended during the interval between two consecutive IVF cycles, especially if there is no chance of natural pregnancy (absence of fallopian tubes, sperm abnormalities, absence of sexual intercourse due to deep dyspareunia).

However, it must be emphasized that regression or stability of nodule size may not be associated with relief of pain. In our study, pelvic pain was still present in 100.0% of women with a lesion that regressed and in 92.3% of those with a stable nodule, compared to 91.7% of women with a nodule that progressed (P = 0.807). This means that in numerous women free of nodule progression, amenorrhoea does not allow to definitively avoid the surgery. On the other hand, even though surgery for colorectal endometriosis provides an overall improvement of pain and quality of life, complete postoperative relief of pain is not guaranteed, often due to the presence of adenomyosis. Therefore, surgery should be carefully considered especially in women with amenorrhoea presenting with troublesome symptoms such as bowel sub-occlusion, severe dyschezia and deep dyspareunia.