Several studies have evaluated the accuracy of transvaginal ultrasound (TVS) in the prediction of deep infiltrating endometriosis (DIE)[1-6]; however, few studies have reported on the ability of TVS to predict pouch of Douglas (POD) obliteration. POD obliteration is diagnosed when the peritoneum of the POD is no longer visible during surgery, and it occurs as a result of adhesion formation between the posterior aspect of the uterus/cervix and surrounding structures such as the anterior rectum, rectosigmoid bowel, rectovaginal septum, vagina or uterosacral ligaments. The primary causes of POD obliteration include pelvic inflammatory disease and DIE. POD obliteration may also occur as a result of extensive superficial endometriosis infiltrating the peritoneum of the POD and adjacent structures.
The surgical finding of POD obliteration is known to be associated significantly with symptoms such as dyspareunia and rectal bleeding[7, 8]. The ability to diagnose POD obliteration at initial pelvic ultrasound examination not only may be helpful in explaining the possible cause of pelvic pain symptoms, but also may shed light on the potential severity of underlying endometriotic disease.
When POD obliteration is encountered at laparoscopy, advanced laparoscopic skills are required to separate the adhesions in the POD and excise any associated posterior-compartment DIE lesions. The risk of bowel endometriosis and the need for bowel surgery is known to be up to three times higher in women with suspected endometriosis if the POD is obliterated at laparoscopy[7, 9]. Given the increased complexity of the surgery and prolonged operating times associated with POD obliteration, it is very useful for the laparoscopic surgeon to know the status of the POD preoperatively. Inability to predict POD obliteration with posterior compartment DIE on preoperative TVS has two important implications. One is the potential for incomplete surgical cytoreduction of posterior compartment endometriotic disease when a general gynecological surgeon without the necessary operative skills is unexpectedly confronted with POD obliteration intraoperatively. Affected women then have to be referred to a tertiary laparoscopic unit for a second laparoscopic procedure. The second is the need for unplanned intraoperative consultation and surgical input by a colorectal surgeon, when an advanced laparoscopic surgeon is confronted with POD obliteration and bowel disease during the surgery.
Knowing whether or not the POD is obliterated ensures that the woman is referred to the most appropriately skilled laparoscopic surgeon from the outset and streamlines patient counseling and consent preoperatively, in terms of expectations for surgery and the possible need for bowel intervention with its associated morbidities.
Both TVS and magnetic resonance imaging (MRI) have been used to predict DIE involving the posterior compartment. In a recent study by Savelli et al., 69 women were evaluated with TVS and with double-contrast barium enema (DCBE) to predict posterior compartment DIE preoperatively. With regard to the prediction of bowel DIE, TVS vs DCBE gave accuracy of 91% vs 45%, sensitivity of 91% vs 43%, specificity of 100% vs 100%, positive predictive value (PPV) of 100% vs 100% and negative predictive value (NPV) of 29% vs 6%. The TVS assessment for bowel endometriosis included placing gentle pressure with the transvaginal probe in the area of interest, to determine the fixation of the endometriotic nodule to adjacent structures. Another recent study by Vimercati et al. examined 90 women with suspected DIE using preoperatively both contrast-enhanced magnetic resonance-colonography (CE-MR-C) and TVS to predict DIE. In the diagnosis of adhesions of the rectosigmoid colon, TVS and MRI had an accuracy of 82% and 98%, sensitivity of 68% and 100%, specificity of 96% and 96%, PPV of 94% and 96%, NPV of 76% and 100%, positive likelihood ratio of 16 and 23 and negative likelihood ratio of 0.3 and 0. The conclusion from these two studies was that TVS is an accurate and cost-effective imaging method in the prediction of DIE involving the posterior compartment. However, neither of these studies reported on the ability of TVS to predict specifically POD obliteration.
The few ultrasound studies that have used TVS to predict POD obliteration preoperatively have shown promising results, with sensitivities and specificities ranging from 72–83% and 97–100%, respectively[1, 7, 12]. In a study published in this issue of the Journal, we report on the use of the TVS ‘sliding sign’ technique in the preoperative prediction of POD obliteration. The sliding sign uses gentle pressure by the TVS probe to assess whether the anterior rectum glides freely across the posterior aspect of the cervix and posterior vaginal wall and whether the rectosigmoid glides freely over the posterior aspect of the posterior upper uterus. When sliding occurs (i.e. the sliding sign is positive) in both of these anatomical regions the POD is considered to be unobliterated. When sliding does not occur in either anatomical region the sliding sign is considered negative and the POD is considered obliterated. In women with suspected endometriosis undergoing laparoscopy we found that the sliding sign had an accuracy of 93% in the preoperative prediction of POD obliteration. While this sign appears to be a simple way to determine whether or not the POD is obliterated, we acknowledge that our study involved a limited number of patients (n = 100) and that the laparoscopic surgeons were not blinded to the preoperative TVS findings. Furthermore, we cannot rule out the possibility that the interpretation of POD obliteration at laparoscopy may have varied amongst each of the seven laparoscopic surgeons involved in the study. In another study in this issue, Hudelist et al. show that the TVS sliding sign can be used to predict the likelihood of DIE lesions involving the bowel based on adhesions to the posterior uterus, with an accuracy of 93%. However, it is important to recognize that although the risk of bowel endometriosis is increased in the presence of POD obliteration at laparoscopy, not all women with POD obliteration will have a bowel nodule. Similarly, not all women with a bowel nodule will have POD obliteration[7, 9]. It would be incorrect to assume that a negative TVS sliding sign is an absolute predictor of bowel DIE, when it is possible to directly visualize a bowel nodule whether or not the POD is obliterated.
As with any new diagnostic technique, validation is required through reproducibility studies prior to implementation into standard clinical practice. We recently published a reproducibility study which involved offline viewing of pre-recorded TVS video sets of 30 women presenting with chronic pelvic pain, in order to determine POD obliteration using the sliding sign technique. The four gynecological sonologists demonstrated near-perfect inter- and intraobserver correlation in evaluating the TVS sliding sign for prediction of POD obliteration and their diagnostic accuracy using this method was high, with an accuracy of 93–100%, sensitivity of 93–100%, specificity of 91–100%, PPV of 78–100% and NPV of 98–100%. However, an important limitation of this reproducibility study was that the TVS sliding sign videos were performed and pre-recorded by a sonologist who was experienced in the assessment of POD obliteration using this technique. A further limitation is that cases classified as ‘unsure’ for POD obliteration by the four study observers were excluded from the diagnostic accuracy analysis. Therefore, a repeatability study involving live scanning would be required to assess the true ability of different observers to assess the sliding sign for the presence/absence of POD obliteration.
In modern practice, most ultrasound units do not assess routinely for features such as POD obliteration or DIE during a standard pelvic ultrasound scan. Therefore, despite the report of a ‘normal’ pelvis on TVS, a significant proportion of women with chronic pelvic pain will have endometriosis present at subsequent laparoscopy. Women with suspected endometriosis are known to have an increased risk of POD obliteration at laparoscopy, ranging from 16 to 30%[7, 9, 12]. An ultrasound examination which has not assessed the posterior compartment and is reported as ‘normal’ is not helpful for surgical planning in women with chronic pelvic pain, as POD obliteration and/or DIE lesions may have been overlooked.
The ability to diagnose DIE lesions preoperatively using TVS requires skill and experience in visualization of the posterior compartment. The studies in this issue, however, suggest that the TVS sliding sign test to predict POD obliteration/bowel DIE is a simple and easily learned first-line ultrasound technique that could be considered as part of the standard preoperative work-up for all women in whom endometriosis is suspected. If a negative TVS sliding sign, indicating POD obliteration, is present at the initial scan, this should result in referral of these women to an advanced gynecological ultrasound unit for further preoperative imaging to determine the presence, location and extension of DIE lesions. In areas in which advanced gynecological ultrasound is not available, the TVS sliding sign could still be used as a first-line test to predict POD obliteration and the increased risk of bowel endometriosis, allowing for women to be referred to an advanced laparoscopic surgeon with consultation with a colorectal surgeon.
Improved preoperative diagnosis of POD obliteration and posterior compartment DIE should ensure appropriate referral, surgical planning and counseling of women with suspected endometriosis, and potentially improve the overall management of these women. It is our hope that, in the future, a report that concludes ‘normal pelvis’ should have documented a positive sliding sign and excluded coexisting posterior compartment DIE.