Robotic surgery for endometrial cancer and sentinel lymph node mapping

Introduction

Endometrial (uterine) cancer is the most common malignancy of the female genital tract in the United States, with an estimated 66,570 new cases and 12,940 deaths in 2021. In 2019, there were an estimated 807,860 women living with uterine cancer in the United States. Approximately 3.1% of women will be diagnosed with uterine cancer at some point during their lifetime. Hysterectomy remains the mainstay in the treatment of endometrial cancer, which was traditionally accomplished via open laparotomy until the 1990s. Since then, multiple studies demonstrated the feasibility of a minimally invasive laparoscopic approach. ^, The Gynecologic Oncology Group (GOG) phase III LAP2 trial showed noninferiority of laparoscopy versus laparotomy in terms of recurrence and survival. Robotic-assisted laparoscopy has not been prospectively compared in randomized trials to conventional laparoscopy in endometrial cancer surgical staging, but several retrospective cohort comparison studies suggest equivalency in staging results. ^,

Lymph node (LN) metastasis is one of the most important prognostic factors affecting recurrence and survival in endometrial cancer (EC). The presence or absence of LN involvement is a major determinant in staging and affects decisions regarding adjuvant therapies. The incidence of LN metastasis ranges from 0 to approximately 28% in apparent clinical stage I endometrial cancers, with the actual risk dependent upon uterine tumor histology, lesion size, depth of invasion (DOI), and lymphovascular space involvement (LVSI). Traditionally, LN metastases are detected surgically with systematic pelvic and aortic lymphadenectomy. However, comprehensive lymphadenectomy is associated with adverse outcomes including additional blood loss, lymphedema, lymphocysts, and neuralgia. Moreover, a clinically significant percentage of node-negative patients suffer from lymphatic recurrence, suggesting possible deficiencies in the diagnosis and treatment of lymphatic dissemination.

The sentinel lymph node (SLN) biopsy procedure improves sensitivity for the detection of metastases by locating the most likely involved lymph node (i.e., the SLN) and evaluates the SLN with ultra-sectioning and immunohistochemical (IHC) stains. These enhanced pathology techniques increase the detection of isolated tumor cells (ITCs) and micro-metastases, further increasing sensitivity for the detection of lymph node metastases. Morbidity related to surgical staging is reduced by limiting the number of lymph nodes removed. The SLN is the most proximal lymph node in a sentinel pathway emanating from the primary tumor. In the case of uterine cancer, lymphatic pathways leave the uterus bilaterally through the lower uterine segment parametria and either proceed laterally crossing over the superior vesicle artery to the lateral pelvic sidewall nodes, or posteriorly branching to pararectal, internal iliac, presacral, common iliac, or infra-mesenteric aortic lymph nodes. Fundal routes of drainage via the ovarian blood supply are also present; however, mapping the aorta using cervical injection of dyes has not been consistent or well studied. The targeted biopsy of sentinel nodes results in less disruption of pelvic lymphatic drainage and less morbidity than comprehensive lymphadenectomy. SLN mapping is also diagnostically superior to traditional lymphadenectomy because it can detect unexpected drainage pathways that are not universally captured by traditional pelvic lymphadenectomy, such as presacral, parametrial, and internal iliac lymph node basins in approximately 5% of cases.

History

The sentinel node concept is a relatively contemporary milestone in the understanding of solid malignancy metastasis. Virchow, in the nineteenth century, theorized that lymph nodes filter particulate matter from lymph. The experimental work of Halsted and others supported the hypothesis of sequential dissemination through the lymphatic system. After incidental reports of nonguided sentinel node procedures, it was the introduction of the technique of intra-operative lymphatic mapping in 1989 that initiated the widespread use and general acceptance of this approach.

The sentinel lymph node was first hypothesized by Gould in 1960, who observed that the most likely involved LN was located in close proximity to the primary tumor during en bloc dissection of parotid cancer. The first SLN mapping procedure was described two decades later by Cabanas in penile cancer by cannulating the dorsal lymphatics of the penis and performing lymphangiography. Morton et al. published a landmark study in 1992 using blue dye to introduce a new concept concerning the dynamic physiology of lymphatic drainage, described as quite variable from patient to patient. They fundamentally changed the historical thoughts about lymphatic dissemination, emphasizing that in many patients there was an orderly spread of metastasis from the primary site to sentinel lymph nodes, and then elsewhere via the bloodstream. Pathology ultra-staging of sentinel nodes that uses multiple sections of paraffin-embedded tissue and sensitive cytokeratin stains was first reported in breast cancer by Giuliano and colleagues in 1994. Lymphatic mapping entered the domain of gynecologic cancer first in vulvar cancer by Levenback et al., followed by endometrial cancer by Burke et al. using blue dye, and finally arrived in cervical cancer 3 years later by Echt et al.

The clinical utility of sentinel node biopsies has been demonstrated in several different malignancies. The NSABP 32 trial in breast cancer illustrated that the 8-year disease-free survival, regional control rate, and overall survival were statistically equivalent between the SLN group and the axillary dissection group. Another randomized multicenter trial in breast cancer also concluded that upper extremity morbidity including swelling, sensory loss, mobility loss and quality-of-life (QOL) scores after sentinel node biopsy are better with SLN biopsy. SLN biopsy is now considered the standard of care in vulvar cancer since the publication of the GROINSS-V-I study and GOG-173.

Surgical technique and the surgical algorithm

Sentinel node mapping is feasible with laparoscopic, robotic-assisted laparoscopic, and open abdominal approaches. A meta-analysis of 26 studies on the utilization of SLN biopsy in EC reported no association of sensitivity or detection rate to the surgical approach. The cervix is a midline organ that has bilateral lymphatic drainage in the pelvis. The uterine corpus, however, drains both to bilateral pelvic node beds through the parametria and to the para-aortic basins through the infundibulopelvic (IP) lymphatics and/or via posterior presacral lymphatics. The cervix is the most commonly injected site for endometrial cancer mapping compared to other methods such as fundal and hysteroscopic tumor injections. Corpus injection using transvaginal sonography (TVS) or hysteroscopy is cumbersome, is less reproducible, and has more variability in mapping success than mapping with cervical injections. The National Comprehensive Cancer Network (NCCN) guidelines recommend cervical injection of dyes or a radioactive tracer 1 to 3 mm below the mucosa at either two cardinal points or four cardinal points in the cervix (i.e., 3 and 9 o’clock positions, or 2, 4, 8, 10, or 3, 6, 9, 12 o’clock positions). In our experience, four-point injection has a higher rate of successful mapping than two-point technique especially with less experienced surgeons, although there is no documented difference in outcomes apparent from retrospective comparisons. Aliquots of 0.5 mL of a 1 mg/mL solution of indocyanine green (ICG) dye should be used at each site in the four-quadrant technique, or 1 mL injected at 3 and 9 o’clock. Care should be taken to ensure that the injection is within the lateral cervical stroma with enough depth at 3 to 5 mm to avoid dye leaking back out into the vagina, and also to avoid deep injection into the posterior cul-de-sac, which stains peritoneum obscuring the SLN pathways. Deeper injections in the cervix up to 10 mm are proposed for uterine cancers in an attempt to map the infundibulopelvic lymphatics; however, efficacy and interpretation of these findings is lacking.

Indocyanine green and blue dyes will travel quickly into the parametrial lymphatics within minutes. Visualization through dissection of the retroperitoneum should ideally be initiated within 15 minutes of injection to avoid “washout” of too much dye in the retroperitoneum. Timing of the dye injection is an important consideration in patients with prior abdominal surgeries who may require adhesiolysis before initiating the SLN procedure. Injection of dye should be delayed until access to the pelvis is confirmed in patients with complex surgical abdomens. SLN mapping is initiated by searching for pathways emanating from cervical stroma through the parametria, usually in proximity to the uterine artery. From there, the pathways will diverge most often to the medial external iliac and proximal obturator nodes, and less commonly to common iliac nodes. However, approximately 10% to 15% of pathways will diverge to areas not usually dissected in traditional pelvic lymphadenectomy including internal iliac, pararectal, and presacral lymph nodes through the posterior lymphatic pathways. Posterior pathways have been reported more clinically relevant with node metastasis in patients with type II high-risk histologies more than endometrioid histology. Presacral pathways occasionally cross the common iliac artery and terminate in infra-mesenteric aortic lymph nodes. When opening the retroperitoneum and developing paravesical and pararectal spaces, avoidance of cutting lymphatic pathways by instead performing blunt dissection and tissue spreading will allow more time for dye transit in the event nodes are not initially identified. If one hemi-pelvis lacks propagation of dye, it is beneficial to allow more time by dissecting the opposite hemi-pelvis, isolating and dividing the IP ligaments, or developing the bladder flap. If mapping still fails, a second injection of dye is often helpful, usually if the initial dye injection was too superficial and may have leaked out of the needle track rather than diffuse into the stroma. This concept was supported by Maramai et al. who showed that cervical reinjection resulted in improvement of the bilateral detection rate from 73% to 94%, and thereby reduced the number of side-specific required lymphadenectomies. Consideration should be given to the patient’s history, as sentinel mapping can be negatively affected by prior history of pelvic radiation, chronic inflammatory conditions such as diverticulosis, or lymphoma/leukemia. Also, in regard to their endometrial cancer diagnosis, LVSI, nonendometrioid histology, and intraoperative finding of enlarged lymph nodes were identified as independent risk factors for unsuccessful mapping. The complete surgical algorithm per NCCN is summarized in Fig. 41.1 . ^,

Observational studies have suggested that unilateral pelvic SLN status does not accurately predict the presence or absence of metastasis on the contralateral side. Stephens et al. reported a 75% congruency in bilateral mapping, meaning that 25% of sentinel nodes would be inaccurately dissected if the surgeon relied solely on symmetry for SLN location. Barlin et al. retrospectively applied a surgical algorithm that incorporated side specific lymphadenectomy, as well as removal of any suspicious nodes and peritoneal lesions in a study of 498 endometrial cancer patients undergoing SLN mapping. The proposed endometrial surgical algorithm was estimated to reduce the false negative rate (FNR) from 14.9% to 1.9%. The NCCN guidelines recommend utilization of this surgical algorithm for SLN mapping. ^, Cormier and colleagues subsequently analyzed the efficacy of the NCCN surgical algorithm in endometrial cancer on 1385 patients from a historical database. They observed that 37 of 190 node positive patients had false negative SLNs. Retrospectively applying the Barlin et al. surgical algorithm reduced the FNR from 19% to 5%. Sinno et al. proposed a hypothetical algorithm called the restricted frozen section algorithm that would apparently reduce lymphadenectomy rates. Retrospective application of their algorithm to 114 patients with apparent uterine confined, grade 1/2 endometrial cancer and complex atypical hyperplasia reduced pelvic lymphadenectomy rates to 9.2% in the restricted frozen section algorithm compared to 36.8% in the NCCN surgical algorithm. We utilize this proposed algorithm, where pelvic and para-aortic lymphadenectomy is performed only if SLN mapping fails and if high-risk uterine features are identified on frozen section. If the primary tumor meets Mayo criteria (less than 2 cm, hyperplasia or grade 1 histology, and no myometrial invasion), full lymphadenectomy of a hemi-pelvis that failed mapping is not necessary and can be omitted.

Multiple studies have reported the risk of isolated aortic metastases in endometrial cancers at approximately 1% to 3% in pelvic node-negative patients. The decision to dissect aortic nodes should be determined based on the frozen section information of the primary tumor (grade, depth of invasion, LVSI) or the presence of grossly positive pelvic nodes. When patients are determined to be at risk for aortic metastasis, the target should be the infra-renal para-aortic lymph nodes in addition to infra-mesenteric nodes. Not all patients are candidates for infra-renal dissection based on body habitus or comorbid medical conditions, and the infra-renal lymph nodes can be radiographically monitored for recurrence as an alternative.

The evolution of SLN detection methods

A variety of colored dyes have been used for SLN mapping. The use of methylene blue and isosulfan blue (ISB) dyes are the least expensive technique for mapping because they are visualized in white light and do not require the cost of additional imaging systems. Other limitations include ISB’s risk of anaphylactic reactions (1 per 1000 injections) and cost, as it contains human albumin. While less costly, injection of methylene blue dye can result in paradoxical methemoglobinemia leading to a falsely low serum O ₂ saturation. Methylene blue is not FDA approved for lymphatic mapping and has occasionally been associated with skin necrosis in extremity mapping of melanoma. The overall detection rate of SLNs by blue dye alone in endometrial cancer is 93%, with bilateral detection ranging from 44% to 81%.

Technetium-99 (Tc99) is a radiocolloid that uses a gamma camera or single-photon emission computerized tomography (SPECT) scan for detection. The “short” physical half-life of the isotope and its biological half-life of 1 day allows scanning procedures that collect data rapidly but keeps patient radiation exposures low. Tc-99 can be injected the day prior to surgery or on the same day of surgery. A preoperative lymphoscintigram/SPECT scan is taken 30 minutes after the injection and provides the surgeon with the location of the sentinel node. Gamma laparoscopic or hand-held gamma probes are required to help dissect the nodes intra-operatively. Tc-99 has been widely used in breast cancer and malignant melanoma, and more recently in vulvar cancer and endometrial cancer. In a prospective multicenter SENTI-ENDO Trial, cervical dual injection of [99mTc]-labeled colloid and patent blue was used in 125 patients with early-stage endometrial cancer. Ballester et al. reported an overall detection rate of 89% with a sensitivity of 84% and a negative predictive value (NPV) of 97%. How et al. reported an overall detection rate of 92% (bilateral detection rate of 72%), sensitivity of 89%, and NPV of 99% using a pericervical combined injection of patent blue and [99mTc]-labeled colloid in 100 patients. However, the Tc-99 procedure’s issues of cost and requirements for conscious injection of the tumor sites in the nuclear medicine department are inconvenient and somewhat limiting.

ICG is a tricarbocyanine dye that fluoresces when illuminated with 806 nm near-infrared light (NIR). The dye comes in a 25 g powder vial and is diluted with 25 mL of sterile water for injection. The fluorescent light is then captured using an NIR imaging system. It was initially used in vascular procedures for imaging a vessel’s blood flow to document patency; however, its applications have expanded to many other procedures including biliary tract imaging. ICG was first utilized in breast cancer in 2005 by Kitai et al. and then in melanoma by Fujiwara et al. in 2009. In 2012, the combination of blue dye and ICG was described in endometrial cancer with high detection rates by Rossi et al. (88%) and Holloway et al. (100%). Retrospective case series suggest that ICG alone has comparable sensitivity to a combination of colorimetric and radiotracer dyes and is superior to blue dyes alone. A prospective randomized trial showed that ICG plus ISB had overall and bilateral detection rates of 96% and 84% compared to ISB only, which had rates of 76% and 40%, respectively. Fluorescent dyes are considered superior in the detection of SLNs in obese women. Figs. 41.2–41.4 display lymphatic dissemination of ICG in normal light and NIR imaging after retroperitoneal dissection. The main limitation of ICG is rare allergic reactions (1/12,000), possibly more common in patients with iodine allergies. Manufacturers are developing advanced molecules of fluorescent dyes and videoscopes to improve the precision in finding the sentinel nodes. Activatable fluorescent probes (smart probes) are being tested to delineate cancer cells in vivo. Future prospects include the detection of tumor margin in the cervix and vulva, detection of metastatic nodes in vivo, and precise imaging of metastatic disease.