Abstract
Objectives:
To examine the real-life utilization of sentinel lymph node mapping for surgical staging of patients with endometrial cancer. We evaluated patterns of surgical staging in unmapped patients and studied how this practice affected adjuvant therapies and survival.
Methods:
We conducted a retrospective chart review of patients with newly diagnosed endometrial cancer who underwent minimally invasive surgical staging, including sentinel lymph node mapping with cervical injection of indocyanine green from January 2019 to December 2021. Patient demographics, surgical findings, sentinel lymph node mapping, adjuvant therapy, and recurrence rates were collected. Sentinel lymph node detection rates were calculated, and reasons for omitting lymphadenectomy in unmapped patients were evaluated.
Results:
Among 121 patients, 80 (66%) had successful sentinel lymph node mapping and 41 (34%) failed mapping. Our yearly detection rate was 63%, 68%, and 70% for 2019, 2020, and 2021, respectively. In patients with successful sentinel lymph node mapping, 73.8% were low-grade and 26.2% were high-grade histology. For patients with failed mapping, 75.6% were low-grade and 24.4% were high-grade histology. For the failed mapping cohort, 23 patients (56.1%) had a complete lymphadenectomy performed, of which 3 (13.0%) had positive lymph nodes. Reasons for omitting lymphadenectomy were documented as: (1) intraoperative pathologic evaluation; (2) inability to tolerate Trendelenburg; (3) difficulty with anatomical dissection/visualization; and (4) evidence of locally advanced disease. There were 18 incompletely staged patients, including 8 (44.4%) with low-risk disease, 2 (11.1%) with locally advanced disease, and 1 (5.6%) with serous histology. The remaining 7 (38.9%) unstaged patients were offered and/or received adjuvant radiation based on final pathology. During a short-term follow-up period, no patients in the unmapped or incompletely staged cohorts had a recurrence of the disease.
Conclusions:
The rate of sentinel lymph node detection is improving. Low-risk disease identified on intraoperative pathology was the most common reason for omitting lymphadenectomy in unmapped patients, and that practice did not seem to affect adjuvant therapy or recurrence of disease.
Introduction
Endometrial cancer (EC) is the most frequently diagnosed gynecological cancer in the United States.1,2 The number of cases of EC has been increasing, with approximately 67,880 new cases diagnosed in 2024. While the survival rates for the majority of cancers have steadily improved in the past few decades, EC remains an exception, with mortality rates continuing to increase by 1.9% from 2009 to 2018.1,2 At the time of initial diagnosis, the majority of cases are early stage and confined to the uterine cavity.2,3 Diagnosis is accomplished by histopathological analysis of endometrial biopsy, followed by surgical staging including a hysterectomy, bilateral salpingo-oophorectomy, with pelvic and/or paraaortic lymph node dissection by either sentinel lymph node (SLN) biopsy or a complete lymphadenectomy.
Lymph node (LN) evaluation is a prognostic tool that is crucial for risk stratification to determine the need for adjuvant therapy. A complete pelvic and paraaortic lymphadenectomy has associated morbidity including lymphedema, lymphocyst, cellulitis, and nerve damage. 4 Strategies, such as the Mayo criteria and SLN mapping, have been employed to avoid the morbidity of lymphadenectomy without compromising prognostic information to help with therapy decisions. The Mayo criteria was created to help determine if lymphadenectomy is indicated during the time of surgery by examining the tumor in real time in the operating room. Patients are deemed low-risk by the Mayo criteria if they have (1) tumor diameter ⩽2 cm, (2) grade 1 or 2 histology, (3) myometrial invasion (MI) ⩽ 50%. 5 This information allows to determine if the prognostic yield of LNs was worth the perioperative morbidity of a lymphadenectomy.5,6 In 2014, the National Comprehensive Cancer Network revised its guidelines to recommend SLN dissection over complete lymphadenectomy. While there are no randomized controlled trials demonstrating the exclusion of retroperitoneal staging, SLN biopsy was been adopted for surgical staging. SLN biopsy allows for the identification of nodal metastasis while decreasing the morbidity associated with complete pelvic and paraaortic lymphadenectomy.4,7 First described in 1996, the SLN mapping technique requires the injection of a dye into the cervix. The dye uptake is traced to identify the lymphatic channels by which tumor invasion may occur. Lymphadenectomy is then conducted for LNs that are identified to have taken up the dye, thereby sparing LNs outside the channel for possible metastatic spread. 4
SLN mapping has become the standard of care for EC surgical staging, avoiding the morbidity of complete lymphadenectomy, while providing prognostic value and therapeutic guidance. We examined the real-life utilization of SLN mapping at our institution and determined the rates of LN metastasis in cases of mapping failure. We further evaluated patterns of surgical staging in unmapped patients and examined how this practice may affect adjuvant therapies and survival outcomes.
Methods
Study design
We conducted a retrospective chart review of patients with confirmed diagnoses of EC who underwent SLN mapping followed by robotic hysterectomy and completion pelvic and paraaortic lymphadenectomy at Rutgers Cancer Institute from January 2019 to December 2021. This was a high-volume New Jersey only National Cancer Institue (NCI)-designated comprehensive cancer center, with five fellowship-trained Gynecologic Oncologists. The study was reviewed and approved by the Scientific Review Board and Institutional Review Board of Rutgers Cancer Institute.
Eligibility criteria
Inclusion criteria were patients with a diagnosis of EC on pre-operative biopsy who underwent minimally invasive hysterectomy and surgical staging with the injection of indocyanine green (ICG) into the cervix. Successful SLN mapping was defined as having visualization of at least one SLN; a distinction was made if a patient had unilateral or bilateral SLN mapping. Patients were excluded if they did not have a diagnosis of EC prior to surgery, if ICG cervical injection was not performed, or at the time of final pathology LN tissue was not noted in the SLN specimen.
Data collection
Patient demographics and clinical information were collected from the electronic medical record, including age, BMI, race, adjuvant treatments (if applicable), and clinical outcomes. Surgical information and tumor characteristics were collected, including histology, stage (by 2009 International Federation of Gynecology and Obstetrics surgical staging), type of surgery, SLN injection, number of SLN removed, and total LNs removed. Reasons for omitting or limiting lymphadenectomy in unmapped patients were noted in operative reports. Patients were followed for up to 3 years, and surveillance included physical exams and imaging, if indicated.
Statistical analysis
The overall SLN detection rate was calculated per year and descriptive statistics were used to analyze our data. Reasons for mapping failure and the rate of full lymphadenectomy performed in unmapped patients were collected.
Results
We identified 121 patients with newly diagnosed EC who underwent minimally invasive surgical staging, including ICG cervical injection for SLN mapping. The median age was 66, and most patients were non-Hispanic white (76.0%) followed by Asian (9.1%), then Hispanic (6.6%) (Table 1). For tumor histology, 74.4% were low-grade endometrioid, 7.4% high-grade endometroid, 8.3% serous, and 9.9% other (carcinosarcoma, clear cell, dedifferentiated, mixed) (Table 2), and most patients were stage I (80.2%) (Table 1).
Clinical characteristics by sentinel lymph node mapping outcome.
Pathologic characteristics by sentinyl lymph node mapping outcome.
In the cohort, 80 patients (66.1%) had successful SLN mapping, while 41 patients (33.9%) failed SLN mapping. Our detection rate by year is as follows—2019: 63%, 2020: 68%, 2021: 70%. The majority of patients in both the SLN mapped and unmapped groups were non-Hispanic White (78.8% vs 70.7%, respectively), followed by Asian (7.5% vs 12.2%), Hispanic (7.5% vs 4.9%), non-Hispanic Black (3.8% vs 7.3%) (Table 1). In regard to BMI, 46.3% of patients who had successful LN mapping had a BMI < 30. Of the patients who did not map, 65.9% had a BMI ⩾ 30 (Table 1). For tumor grade, in both the SLN mapped and unmapped groups, most tumors were low-grade (grade I/II endometroid) histology (73.8% vs 75.6%, respectively) (Table 2). LN metastasis was identified in 6 (7.5%) patients with mapped SLN and 3 (13%) in patients who failed to be mapped.
Out of 41 patients with failed mapping, 23 (56.1%) patients had complete lymphadenectomy performed and three patients (13.0%) had positive LNs after full LN dissection. In the failed mapping cohort, 18 patients (43.9%) did not have a lymphadenectomy performed. Documented reasons for omitting lymphadenectomy included: (1) intraoperative pathologic evaluation yielding a low probability of positive LNs based on Mayo criteria (n = 8, 44.4%), (2) inability of the patient to tolerate Trendelenburg to complete lymphadenectomy (n = 1, 16.7%), (3) difficulty with anatomical dissection and limited visualization (n = 5, 2.8%), (4) evidence of locally advanced disease, where lymphadenectomy was no longer indicated (n = 3, 16.7%). In the unstaged cohort, eight patients (44.4%) had low-risk disease, 2 (11.1%) had locally advanced disease, 1 (5.6%) had serous histology with large uterine disease, and 7 (38.9%) were offered and/or received adjuvant radiation based on final pathology. Tumor characteristics of patients in the unmapped, unstaged cohort are listed in Table 3. This cohort was followed for up to 30 months (range: 11 days–30 months), 2 are lost to follow-up and 16 have no evidence of disease recurrence.
Patient demographics and tumor characteristics of unmapped, unstaged patients.
AWOD: alive without disease; LTFU: lost to follow-up; N: no; Y: yes.
Discussion
This study was a real-life examination demonstrating the implementation and outcomes of SLN mapping for EC at a high-volume comprehensive cancer center. It demonstrated that we are improving SLN mapping with a detection rate of 70% in 2021. Our detection rate is similar to others reported in the literature, such as the FIRES trial, with a detection rate of 80%. 8 The FIRES trial was a multicenter, prospective cohort study examining SLN biopsy versus complete lymphadenectomy in detecting metastatic EC. 8 While our detection rate was lower than the FIRES trial, it was done in a real-life setting, off-trial, and is overall encouraging to see it increase over the years. With an increase in surgical volume and surgeon experience, we expect our detection to continue to rise.
We further examined real-life clinical scenarios of what occurs when SLN mapping is unsuccessful. In 56% of cases of unsuccessful SLN mapping, a complete lymphadenectomy was performed, in keeping with the standard of care. In the 18 cases where a lymphadenectomy was not performed, we investigated the underlying factors behind this decision. The most common reason for omitting lymphadenectomy was based on real-time surgical pathologic information. In eight patients (44%), frozen pathology demonstrated low-risk disease, which did not meet GOG-99 or PORTEC-1 criteria for adjuvant therapy.9,10 GOG-99 and PORTEC-1 use tumor-related factors to determine patients who are at high-intermediate risk of disease recurrence and should be offered adjuvant therapy. In PORTEC-1, a lymphadenectomy was not part of the inclusion criteria, further supporting omitting a complete lymphadenectomy in the eight low-risk, unmapped, unstaged patients. 10 LN assessment did not change treatment management for four patients given concerns of locally advanced disease and/or serous histology that necessitated adjuvant therapy irrespective of nodal status.
In the unmapped, unstaged cohort, there were eight patients who were potentially undertreated by omitting lymphadenectomy. We followed them to see what their clinical outcomes were. Over a 3 year follow-up period, all eight patients remain alive without evidence of disease recurrence. Although this group may have been undertreated by not knowing their nodal status, the real-time decision to defer SLN based on low-risk intraoperative findings has not impacted their disease-free interval to date. Other groups have examined the decision to forgo SLN mapping and its impact on oncologic outcomes. Bogani et al. 11 conducted a propensity-matched analysis comparing patients with EC who underwent hysterectomy alone versus hysterectomy with SLN mapping. They found that adjuvant therapy rates were similar between the groups. In the SLN mapping cohort, only 4% of patients received adjuvant therapy based on LN status and the remaining 96% received adjuvant therapy based on uterine factors. There was no difference in 5-year disease-free survival (p = 0.720) or overall survival (p = 0.632) between the two groups. The ASTEC trial was a prospective, randomized controlled trial examining 1408 patients who were randomized to hysterectomy and bilateral salpingo-oophorectomy with or without lymphadenectomy. ASTEC found no difference in 5-year recurrence-free survival (HR: 1.25, 95% CI: 0.93–1.66, p = 0.14) or overall survival (HR: 1.04, 95% CI: 0.74–1.45, p = 0.83). 12 The lack of impact of LN evaluation on survival in both Bogani et al. 11 and ASTEC remained true in sub-analysis when examining low, intermediate, and high-risk early disease. 12 This data is in line with our findings of no impact on survival by omitting lymphadenectomy. However, the ASTEC trial was performed prior to the utilization of SLN in lieu of lymphadenectomy. The decision to omit SLN staging is currently being evaluated in ENDO-3 (NCT04073706), which is a prospective, phase III, randomized non-inferiority trial examining hysterectomy with and without LN evaluation. 13 ENDO-3 is anticipated to enroll 760 patients who will be randomized to hysterectomy and bilateral salpingo-oophorectomy with or without sentinel node biopsy. The primary endpoint of disease-free survival at 4.5 years, with results expected in 2031. 13
Other studies have examined reasons for mapping failure, including patient factors, tumor characteristics, and technical difficulties. Patient factors include age, anatomic issues, and the presence of dense adhesions. Tumor characteristics include lymph-vascular space invasion and MI. Technical features include surgical expertise with ICG injection and the type of dye utilized.14–18 Tortorella et al. 14 similarly performed a retrospective study at a single institution to identify predictors of unsuccessful SLN mapping in patients with early stage EC. Among 327 patients, there were 78% successful mappings and 22% mapping failure. The failure rate was attributed to surgical expertise with ICG injection as well as anatomical difficulties including lysis of adhesion. All of these factors were also noted as challenges in our study. Martinez Bravo et al. 16 examined 92 patients with early stage EC who underwent SLN mapping using technetium-99-m-nanocolloid (Tc-99m). A failure rate of 23% was noted with the most common reasons for failure including age, depth of MI ⩾ 50%, and high risk of lymphatic infiltration. In our unmapped, unstaged cohort, 39% had MI ⩾ 50%, and 33% had lymphatic infiltration. Both factors could also have contributed to the failure of SLN mapping. While in our study we used ICG, the concept of different types of dyes and tracers is an important factor to consider. Retrospective data from Martinelli et al. 18 demonstrated that in 221 patients with early stage EC, ICG dye had a higher detection rate for bilateral SLN mapping as compared to Tc-99m. Over time, our detection rate improved, which can be attributed to the increased skill set of the surgeons. Future directions could be to examine the rates now, given it has been over 3 years since being evaluated, and there has been a change in surgeons at the institution.
The strength of this study is that it was conducted at a high-volume NCI-designated comprehensive cancer center with five gynecologic oncology surgeons. The study included a diverse patient population and had up to a 3-year follow-up period. Some weaknesses of this study include its retrospective design and small sample size. Given that this was a retrospective descriptive study, a power analysis for sample size was not calculated. The study was also conducted during the COVID-19 pandemic, which could have impacted the surgical volume and sample size.
This study has implications for the utilization of SLN for EC. Because this study occurred outside of a controlled trial setting, it provided an environment for real-life clinical scenarios to occur. This allowed us to better study both our success and our response to inadequate SLN mapping. In addition, from this study, we can identify the real-time driving factors in surgical scenarios to understand why certain decisions were made to adapt to specific clinical situations.
Conclusions
Overall, our study demonstrates that our rate of SLN detection is improving over time. The rate of positive LNs associated with SLN mapping failure was higher than in the SLN mapped group (13% vs 7.5%, respectively). Over half of the patients who failed SLN mapping underwent full lymphadenectomy to complete staging. Low-risk disease during intraoperative pathology assessment was the most common reason for omitting lymphadenectomy in unmapped patients. This practice did not seem to affect adjuvant therapy or the recurrence of disease. This study demonstrates the real-life implementation and application of SLN mapping. Future research should focus on optimizing SLN mapping to identify best practices for mapping and managing unmapped patients.
Footnotes
Ethical considerations
The study was reviewed and approved by the Scientific Review Board and Institutional Review Board (IRB) of Rutgers Cancer Institute of New Jersey as study Pro2022000339.
Consent to participate
Written informed consent was waived by the IRB as the research involves existing electronic medical record data.
Consent for publication
Not applicable.
Author contributions
D Glassman: Writing original draft, writing review and editing, methodology, investigation, data curation, conceptualization, project administration. R Kher: writing original draft, writing review and editing, data curation. C Ananth: methodology, investigation, conceptualization. E Girda: writing original draft, writing review and editing, methodology, investigation, data curation, conceptualization, funding, supervision, project administration.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by the Cancer Center Support Grant Award # P30 CA072720.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
In accordance with the journal’s guidelines, we will provide our data for intended analysis by a selected team by the Editorial Team for the purposes of additional data analysis or for the reproducibility of this study in other centers if such is requested. This project was presented as a poster presentation at the Society of Gynecologic Oncology 2023 Annual Meeting in Tampa, Florida. The abstract was published in the Journal of Gynecologic Oncology in September 2023.