Intraoperative ultrasound in minimally invasive surgery for deep endometriosis: time for new approaches

Endometriosis is a complicated and growing health concern, affecting approximately 2–3% of individuals assigned female at birth worldwide1, 2. Advances in ultrasound and magnetic resonance imaging (MRI) allow assessment of disease, to some extent obviating the ‘(g)old standard’ of diagnostic laparoscopy3. Use of standardized examination criteria for deep endometriosis (DE) on transvaginal ultrasound assessment, as well as terms and definitions of DE, as suggested by the International Deep Endometriosis Analysis (IDEA) group, allow for prediction of the involvement of most pelvic structures with satisfactory diagnostic accuracy4, 5. Despite these significant advances, transvaginal ultrasound for DE is operator dependent and prone to limitations, for example, in cases of severely distorted pelvic anatomy or cul-de-sac obliteration5. This necessitates the employment of innovative tools and techniques, allowing better intraoperative planning and precise intraoperative decisions. One such tool is intraoperative ultrasound (IO-US), which has been employed since the 1960s for the detection of common bile duct and renal stones in cases of profound inflammation6, 7. The application of IO-US in gynecology, specifically for DE, has been more limited, despite the progress of sonography in gynecological diagnostics7. In this Opinion, we undertook a comprehensive search of the literature using terms for intraoperative ultrasound/sonography and deep endometriosis in PubMed (MEDLINE), ScienceDirect and Google Scholar. All study types written in the English language were considered for inclusion. Herein, we summarize the evidence on the use of IO-US in surgery for DE, providing perspectives on the use of IO-US and its applicability in disease management. IO-US describes a wide variety of sonographic techniques that can be employed in all types of gynecological surgery. It is considered a low-risk tool with a ‘real-time’ scanning nature. Different ultrasound probes can be used depending on the surgical procedure, route and purpose of use. In a recent systematic review, Galazis et al.8 described five different types of IO-US applied in fertility-sparing gynecological surgery: transabdominal, transvaginal, transrectal, laparoscopic and contact ultrasound. For transabdominal ultrasound, the ultrasound probes used have a frequency range of 1–5 MHz. For transvaginal and transrectal ultrasound, a variety of probe types can be used, including convex-array, two-dimensional, three-dimensional and transverse endocavitary transducers, with frequencies ranging from 2–20 MHz. Additionally, laparoscopic ultrasound (using probes with a frequency range of 2–15 MHz) and contact ultrasound, which includes the use of linear, convex or T-shaped finger-grip probes during open surgery (using probes with a frequency range of 1–15 MHz), have also been described8-10. Each type of ultrasound technique can find an application in a different setting. For example, contact ultrasound has been used in open surgery, primarily for the detection of metastasis in oncological cases. Transabdominal ultrasound has been used to guide dilation and curettage and hysteroscopic procedures or the placement of intrauterine devices to reduce the risk of perforation7. Owing to the proximity of the transvaginal probe to ovarian tissue compared with that used in transabdominal ultrasound, transvaginal ultrasound has been applied in the context of guidance for fertility/ovary-sparing surgery of adnexal masses, paving the way for future studies in endometrioma surgery11. With respect to endometriosis, researchers have attempted excision of rectal endometriosis under transvaginal ultrasound guidance, but procedures can also be performed under laparoscopic ultrasound guidance. As far as laparoscopic ultrasound is concerned, it can be used for detection of lesions up to 3–5 mm in diameter during conventional laparoscopy and robotic surgery12. The ultrasound probes used are typically linear or convex-array transducers, which are attached to a flexible or rigid long shaft, or the probe is placed on the specimen (i.e. tissue to be examined) with the help of a laparoscopic or robotic grasper. In terms of size, all laparoscopic ultrasound probes fit the standard trocars used during surgery, thus additional incisions or further actions for procedural set-up are not required. Scanning can be achieved by contact, with or without slight pressure on the specimen, or stand-off with the help of saline solution, in which the pathological specimen is immersed and so contrast is increased9, 10, 12. Despite the scarcity of high-quality evidence supporting the use of IO-US and the lack of standardized techniques, IO-US may bolster perioperative outcomes and completeness of endometriosis excision. Given the flexibility of the probes, almost all locations of the abdomen can be scanned. A detailed list of proposed applications of IO-US, along with evidence of its application in the existing literature (if any), can be found in Table 1. Here we present the available evidence and opinions from these studies on the use of IO-US in specific locations. Puppo (2021)23, Hardman (2023)28, Alec (2024)29 Puppo (2021)23, Hardman (2023)28, Alec (2024)29 At present there is no evidence in the literature to support the use of IO-US for anterior compartment, lateral compartment and extrapelvic endometriosis. In terms of the urinary bladder, preoperative diagnostic modalities (transvaginal ultrasound and MRI) may safely exclude the disease with up to 100% specificity, while sensitivity ranges from 55–64%, indicating a need for improvement5. Intraoperative localization of the endometriotic lesion (especially in relation to the ureters), estimation of the depth of invasion and evaluation of excision completeness may strengthen the additive value of IO-US. Contrary to the relatively easy accessibility of the urinary bladder, the parametrium may be more complicated to describe and to assess preoperatively, as explained by the most recent addendum to the IDEA criteria13. Of note, recent advances in transvaginal ultrasound also allow for visualization of the sacral nerve roots and plexus, which will likely reveal an increase in the true prevalence of such lesions, a parameter that remains to be elucidated14. If the anatomical complexity of the parametrium is considered in combination with the severe functional impairments described after parametrectomy (e.g. bladder voiding difficulties), it becomes clear that this type of surgery requires maximum precision, ideally using sonographic guidance15, 16. We believe that laparoscopic ultrasound could assess the dimensions of the nodule, the depth of invasion and, importantly, the anatomical relationship of the lesions with vascular structures, nerves, ureters and muscle tissue more accurately than other IO-US modalities. Assessment could be undertaken before and after surgical dissection by direct placement of the ultrasound probe on the tissue. Nonetheless, it should be noted that there is no described standard for IO-US assessment of the parametrium using laparoscopic probes, which should be addressed in future studies. Furthermore, IO-US may play a role in differential diagnosis; for example, between endometriosis of nerves vs nerve tumors, as described previously with transvaginal ultrasound17. In addition, the use of elastography may be tested as a means for resection guidance, given the differences in the elastic properties of different types of tissue, which already has an application in the setting of peripheral nerve evaluation18. A significant advantage of IO-US regarding the posterior compartment relates to improved estimation of the number of lesions, lesion size, percentage of involved bowel circumference and depth of infiltration, which are key factors in the choice of surgical treatment technique, namely partial-thickness (including shaving) or full-thickness discoid excision or bowel resection and reanastomosis19-21. In turn, improving the accuracy of the appropriate choice of surgical technique will most likely result in decreased morbidity, with lower immediate and delayed complication rates22. Of note, IO-US can be particularly helpful in cases of severe rectum retraction caused by adhesions or large endometriomas, leading to an underestimation of disease extent preoperatively. Figures 1 and 2 demonstrate the use of IO-US in conventional and robot-assisted laparoscopy for rectal DE. Based on the above premises, Puppo et al.23 studied nine consecutive patients undergoing laparoscopic surgery for rectal DE. IO-US was conducted systematically, after bowel mobilization and bowel filling with normal saline, using a laparoscopic convex transducer. The operating surgeon was blinded to the IO-US measurements during the operation. In two (22%) cases, the researchers demonstrated more accurate estimation of the craniocaudal diameter compared with preoperative transvaginal ultrasound, attributed to better visualization of the smooth muscle hyperplasia and fibrosis, depth of invasion and the presence of stenosis. These observations led to a change in planned procedures from disc excision to segmental resection, demonstrating a possible additive value of IO-US to preoperative ultrasound in DE surgery23. Another important benefit of IO-US is detection of lesions of the cecum and ileum, high-sigmoid lesions (not detectable on transvaginal ultrasound), multifocal/multicentric disease of the rectosigmoid as well as non-visualized, palpable satellite lesions24, 25. The latter theory relating to palpable satellite lesions was introduced originally by Roman et al.25 and seems to affect at least 20% of patients with rectosigmoid DE. This study stressed the need for palpation of the whole rectosigmoid during DE surgery. When palpation is not feasible, such as in natural orifice specimen extraction procedures or in robot-assisted surgery, for which haptic feedback is currently limited, IO-US can offer real-time, objective visualization of the bowel, extending beyond the boundaries of optical vision26. For example, Mikhail et al.27 reported a case of isolated sigmoid DE resected under IO-US guidance. Of note, the lesion could not be identified macroscopically with the camera, as the serosa of the sigmoid colon was not affected. Apart from localization and determination of lesion dimensions, numerous case reports have illustrated the application of IO-US and functional IO-US in the guidance of partial- or full-thickness discoid resection of the rectal wall. From a clinical standpoint, this could potentially influence residual disease and recurrence rates. Hardman et al.28 described the use of laparoscopic ultrasound pre- and post-shaving to assess shaving completeness, evaluated as the absence of hypoechogenic lesions on the rectal surface, in a case of rectal DE. A similar concept using transvaginal ultrasound was presented in the study of Alec et al.29, which evaluated the presence of residual disease on the rectum after CO2 laser shaving in two patients. Finally, Yanai et al.30 published a case report on elastography using a laparoscopic probe (L51K; Hitachi, Tokyo, Japan). The authors identified endometriosis-associated fibrosis and its relationship with the rectal muscularis layer prior to rectal shaving. Following the conventional shaving technique with cold scissors, elastography was used again to confirm the absence of ‘stiff’ endometriotic tissue30. While these studies provide new insights into assessment of the completeness of endometriotic tissue removal, issues of technique feasibility and standardization remain, along with a lack of prospective evidence evaluating their efficacy in a larger population. IO-US has been evaluated in the uterine-sparing surgical treatment of adenomyosis, which aims to achieve cytoreduction of the affected myometrium31-33. Ota et al.32 examined 11 patients undergoing laparoscopic adenomyomectomy. The authors conducted real-time elastography using an intra-abdominal laparoscopic ultrasound probe to resect pathological tissue while preserving as much of the unaffected myometrium as possible. Recurrence of adenomyosis during study follow-up (14 months) was not observed. Long-term outcomes, including fertility and the risk of uterine rupture in pregnancy, were not examined. Compared to resection without IO-US guidance, the use of elastography seems to be a feasible alternative32. Diaphragmatic endometriosis affects about 1–2% of all cases of endometriosis and constitutes a perplexing diagnostic and surgical challenge, even in the presence of symptoms (i.e. upper right quadrant pain or shoulder pain), given the suboptimal diagnostic performance of preoperative tests34. Ceccaroni et al.35 attempted to classify lesions of the diaphragm as superficial foci, nodules (infiltrating lesions) and plaques (fibrotic thickening of diaphragm causing adhesions). Based on this, a variety of operative techniques have been introduced, namely removal of the parietal peritoneum, partial- or full-thickness diaphragmatic resection with suture of the defect and/or mesh placement per laparoscopy (conventional or robot assisted), along with thoracoscopic techniques36. Despite these advances, there is no standardized approach to diaphragmatic surgery; most importantly, there is no precise tool to predict the necessity of performing a full-thickness diaphragm resection apart from visual inspection of the lesions. Although a technique to scan the diaphragm intraoperatively has not been described, we believe that IO-US could be used to provide information on pathologies of the thoracic part of the diaphragm, depth of resection and probable multifocality, similar to cases of bowel DE. Based on the evidence we have discussed, it can be hypothesized that IO-US may emerge as a useful tool for enhancing surgical precision and outcome. Of course, the question that arises is who should perform IO-US in everyday practice: radiologists, sonologists or gynecologists? Leonardi37 coined the term ‘gynecologic surgeon sonologist’ to describe expert gynecologists practicing minimally invasive surgery with the ability to perform and interpret gynecological ultrasound examinations. As with all new techniques, increased exposure and training lead to better test performance, which is something that has been stressed in the newest consensus statement on non-invasive imaging for endometriosis5. We support the idea of the gynecological surgeon performing preoperative ultrasonography but also, in case of necessity, IO-US. This dual role could allow for real-time visualization of anatomical structures and pathological conditions, facilitating surgical efficacy and precision. IO-US could work as a complementary tool in the hands of the surgeon, never undermining the importance of preoperative imaging and the multidisciplinary surgical treatment in complex cases, as recommended by guidelines3, 38. There are some limitations of IO-US that need to be considered. IO-US remains operator dependent which, in the absence of a standardized scanning technique and reporting of findings, hinders application in clinical practice and reproducibility. Furthermore, the surgical anatomy clearly differs from the presurgical approach and sonomorphology defined by the IDEA criteria4, 13. To date, there have been no high-quality studies on the true accuracy of IO-US when used in a female pelvis that has undergone surgical dissection of spaces and tissue layers previously infiltrated by DE or affected by secondary adhesions. Another important issue in the implementation of IO-US is the need for additional equipment and specialized training. In terms of equipment, this is paired with practical problems, such as the need for sterilization after each use which is complicated by the potential number of cases per day, along with additional costs for the institution. Furthermore, for extrapelvic structures, IO-US may require the use of an additional trocar close to the area to be scanned. Regarding training, IO-US is not incorporated in ultrasound training curricula. This issue is particularly relevant for surgeons with limited experience in ultrasound12. Despite the novelty of most published reports on IO-US, robust and replicable evidence (ideally of Level IA) is still required for the technique to be used as part of routine clinical practice. Before embarking on prospective (randomized) trials assessing the advantages of IO-US compared with standard operative techniques and its additive effect to preoperative sonography, we need to focus on standardization of this technique in all pelvic compartments and extrapelvic lesions. Table 1 details potential applications of IO-US with no evidence in the literature yet available. Studies on the learning curves of IO-US are also required to guide inexperienced surgeons who want to begin using this technique. As highlighted by Hardman et al.28, future high-quality research should address the efficacy of IO-US in ensuring completeness of disease resection in different types of rectal endometriosis resection. Furthermore, researchers need to investigate if IO-US can reduce overtreatment, for example increasing partial- or full-thickness discoid excisions instead of segmental resections, but also decrease recurrence rates28. To summarize, IO-US constitutes a low-risk tool with the potential for more precise detection of endometriotic foci in all pelvic compartments intraoperatively. This could be particularly helpful for cases of large endometriomas obstructing visualization of the rectum preoperatively. Apart from diagnostic applications (e.g. relationship of the lesion with adjacent structures), it seems that the assessment of upper rectal or sigmoidal lesions along with sonographic guidance of endometriosis excision and evaluation of excision completeness may be the most appropriate potential clinical applications of IO-US. In our opinion, this technique could be performed by gynecologic surgeon sonologists, given their immediate involvement in the surgical procedures. To what extent IO-US may influence surgical efficacy remains to be elucidated in future studies, before being introduced to routine care. Data sharing is not applicable to this article as no new data were created or analyzed in this study.