Incidence of pelvic lymph node metastasis using modern FIGO staging and sentinel lymph node mapping with ultrastaging in surgically staged patients with endometrioid and serous endometrial carcinoma

Abstract

Objective:

We report the incidence of occult nodal metastasis in patients who underwent primary surgical staging for apparent early endometrioid or serous endometrial cancer with bilateral SLN mapping and enhanced pathology. Occult ovarian metastasis rates were also reported.

Methods:

Patients with clinical stage I serous or endometrioid endometrial cancer who underwent primary staging surgery with successful bilateral SLN mapping from 1/2005–12/2018 were retrospectively evaluated. Rates of isolated tumor cells (ITCs), micro- and macrometastatic nodal disease, and occult ovarian involvement were reported.

Results:

Of 1044 patients, 959 had endometrioid and 85 serous carcinoma. There were no positive SLNs among 510 patients with noninvasive FIGO grade 1/2 endometrioid carcinoma and <1%ITCs. Grade 1: 4.5%(9/202) with inner-half and 10%(6/62) with outer-half myoinvasion had positive SLNs. Grade 2: rates were 4%(3/76) and 20%(8/41), respectively. Grade 3: 5%(1/20) with noninvasive, 3%(1/31) with inner-half, and 24%(4/17) with outer-half myoinvasion had positive SLNs. ITC incidence increased with depth of myoinvasion—25% of deeply invasive grade 1/2 and 18% of deeply invasive grade 3 tumors. Four(10%) of 41 patients with noninvasive serous endometrial carcinoma had ITCs or positive SLNs. There were no occult ovarian metastases with grades 1/2 disease, 2/68 (3%) with grade 3 disease, and 2/85(2%) with serous endometrial carcinoma.

Conclusion:

Ultrastaging SLNs may be unwarranted in low-grade noninvasive endometrioid cancer but valuable in noninvasive serous carcinoma. Occult ovarian metastasis is uncommon in early endometrial carcinoma and occurs in 2–3% of high-risk histologies. Further research is needed to determine ITC significance, particularly with regard to adjuvant treatment.

Introduction

Endometrial cancer is the most common gynecologic malignancy, with more than 65,000 newly diagnosed cases expected in the US in 2020 [1]. Although the majority of patients present with uterine-confined disease, a significant subset of tumors will harbor occult metastases [2]. Surgical staging is recommended in medically fit women with clinical stage I endometrial cancer, along with the removal of the uterus, cervix, and bilateral adnexa. Peritoneal evaluation and washings, as well as pelvic lymph node dissection, with or without para-aortic evaluation, should also be performed. Staging is prognostic and aids in adjuvant treatment planning. An emerging body of literature is showing the potential benefit of molecular analysis to guide treatment and refine prognosis, even in early-stage endometrial cancer; however, our current practice guidelines mainly rely upon factors reported on final pathology, including uterine factors such as depth of myometrial invasion, lymphovascular invasion (LVI), spread to adjacent organs and nodal involvement, the last of which is one of the most significant predictors of outcome [3–7].

Sentinel lymph node (SLN) mapping for staging of apparent uterine-confined endometrial cancer is now considered a standard of care in many practices in the US, and it is considered an acceptable method for surgical staging by the National Comprehensive Cancer Network (NCCN) [3]. The majority of published studies demonstrating SLN mapping as a feasible alternative to lymphadenectomy are retrospective [8–13]. The FIRES trial, published in 2016 [14], was the first prospective cohort study to examine the use of SLN mapping in early-stage endometrial cancer and included all histologic subtypes across multiple institutions. This trial further reinforced the accuracy and benefit of SLN mapping, leading to broader use of SLN mapping in patients with endometrial cancer [15, 16].

SLN mapping is beneficial in that it entails the removal of the primary draining node from each hemi-pelvis and is three times more likely than traditional pelvic nodal sampling to identify a positive lymph node, even without ultrastaging [17,18]. In addition, the use of pathologic ultrastaging enhances the detection of low-volume metastases, which contributes to the detection of nodal disease [19].

Surgical staging of pelvic nodes remains the gold standard for the detection of non-bulky nodal metastasis. Although clinicians may use preoperative imaging to assess lymph nodes and routinely evaluate lymph nodes intraoperatively using gross visual inspection, these methods are insufficient for detecting microscopic nodal involvement. Published retrospective series have helped guide clinicians in this setting. For example, the often-cited landmark study by Creasman et al., published in 1987, outlined the risk of upstaging due to nodal positivity based on depth of myoinvasion and grade of tumor in more than 1,100 patients. This influential publication is limited by its use of an older International Federation of Gynecology and Obstetrics (FIGO) staging system and no minimum lymph node requirements [20]. An updated series published by Chi et al. in 2008 included 349 patients with at least 8 lymph nodes sampled and used the modern FIGO staging system [21]. Using Gynecologic Oncology Group (GOG) 210 trial data, Creasman and colleagues published an updated dataset of more than 5,800 patients across multiple histologies and a minimum of 10 lymph nodes sampled and used modern FIGO staging [22]. These studies did not incorporate the modern SLN approach, which uses increased intraoperative precision and enhanced pathology to harvest and evaluate SLNs.

With SLN mapping incorporated into everyday surgical practice, we sought to report the incidence of occult nodal metastasis in women who underwent surgical staging of apparent uterine-confined endometrial cancer with bilateral SLN mapping alone (i.e., ideal use of the Memorial Sloan Kettering Cancer Center [MSK] SLN mapping algorithm) at our institution during a 14-year period. We report the incidence of metastasis and include rates of isolated tumor cells (ITCs) according to tumor histology, grade, and depth of myoinvasion. Our aim was to provide a practical dataset for clinicians to reference when discussing with patients the risk of nodal metastasis using modern surgical staging and enhanced pathology. We also report on rates of occult ovarian metastases in the absence of any other extra-uterine disease, also aiming to provide a practical dataset for clinicians when discussing oophorectomy with premenopausal patients who may wish to conserve their ovaries for fertility purposes or bone and cardiovascular health.

Methods

Upon obtaining Institutional Review Board (IRB) approval, we retrospectively reviewed the records of 4,041 consecutive patients with newly diagnosed clinical stage I endometrial cancer treated with primary surgery at our institution between January 1, 2005 and December 31, 2018. We reviewed all cases starting in 2005 and abstracted nodal and occult ovarian metastasis data relevant for this project. We began our data abstraction starting in 2005. A prior publication from our institution had already examined a similar question using a time period that spanned 1993–2005 [21]. We had opened an IRB endometrial SLN protocol with back-up lymphadenectomy, which began accrual in 2005. Beginning in 2009 the rate of backup lymphadenectomy decreased. We published our institutional SLN algorithm in 2012 and the rate of back-up lymphadenectomy decreased further. Patients were included in the current study if they had serous or endometrioid (all grades) endometrial carcinoma and had undergone successful staging with bilateral SLN mapping and no additional back-up lymphadenectomy, no failed mapping in a hemi-pelvis, and no empty nodal packets on final pathology. Successful staging was defined as the removal of the uterus, cervix, and bilateral SLNs, with or without bilateral adnexa. SLN mapping to a lymph node was confirmed on final pathology in all cases. Clinical and pathologic data, including age, body mass index (BMI), surgical procedures performed, tumor type, FIGO grade, depth of myoinvasion, and presence of extra-ovarian disease for staging purposes, were abstracted from the medical record. Appropriate descriptive statistics were used to report rates of ITC and micro- and macrometastatic nodal disease, as well as occult ovarian involvement, found on final pathology. A previously published MSK SLN mapping algorithm [3, 23], which includes a para-aortic lymph node dissection at the discretion of the surgeon, was used in all cases (Supplemental Figure 1). All patients in whom an SLN was not mapped to a hemi-pelvis were excluded from analysis. Pathologic ultrastaging was performed in accordance with an institutional pathology protocol. In brief, all lymph nodes removed had a single section stained with hematoxylin and eosin (H&E) and were evaluated for metastases. If metastatic tumor was present, no further evaluation was performed. If the SLN was negative, ultrastaging was performed. Ultrastaging consists of a corresponding cytokeratin (AE1:AE3) immunohistochemical stain of the original H&E sample and then H&E and adjacent cytokeratin immunohistochemical staining in additional sections at 50 microns into the block to detect possible metastases missed by the initial sections. Macrometastatic disease was defined as any SLN with a metastatic tumor focus larger than 2.0 mm. Micrometastatic disease was defined as a focus of tumor measuring between >0.2 mm and ≤2.0 mm or 200 or more tumor cells seen on H&E staining. ITCs were defined as a tumor cluster of ≤0.2 mm or <200 isolated, single cells seen on H&E and cytokeratin staining. Cytokeratin-positive cells alone not seen on H&E were not defined as ITCs, and the SLN was reported as negative. ITC-positive SLNs were classified as pN0(i+) and did not alter surgical stage. ITC-positive SLNs with concurrent micro- or macrometastatic SLNs reported on final pathology were only included as micro- or macrometastatic for the purposes of data analysis.

Results

Of 1,044 patients who met the defined inclusion criteria, 959 (92%) had endometrioid carcinoma. Patient characteristics, including median age, BMI, grade, depth of myoinvasion, number of SLNs removed and stage, are detailed in Table 1. Of these 959 patients, all underwent hysterectomy with bilateral SLN mapping and 926 (97%) underwent bilateral oophorectomy at the time of surgical staging. Thirty-three patients (3%) underwent ovarian conservation at the time of their staging surgery. Inclusive of all FIGO grades, 530 tumors (55%) had no myoinvasion, 309 (32%) had <50% and 120 (13%) had ≥50% myoinvasion reported on final pathology. All patients had clinical stage I disease at initial diagnosis, with final pathology demonstrating disease stage as follows: stage IA, 805 (84%); stage IB, 82 (10%); stage II, 19 (2%); stage IIIA, 10 (1%); stage IIIC, 40 (4%); and stage IVB, 3 (<1%).

Table 1.

Clinical, surgical and pathologic characteristics for patients with endometrioid and serous endometrial carcinomas staged with bilateral sentinel lymph node mapping from 2005–2018 (N=1,044)

	Endometrioid (n=959) n (%)	Serous (n=85) n (%)
Median age, years (range)	60 (24–92)	68 (49–89)
Median BMI, kg/m2 (range)	30 (15–70)	30 (18–54)
One or both ovaries retained	33 (3)	0
Median lymph nodes removed (range)	3 (2–16)	3 (2–11)
FIGO grade
1	713 (74)	NA
2	178 (19)
3	68 (7)
Myoinvasion
None	530 (55)	41 (48)
Inner half (<50%)	309 (32)	23 (27)
Outer half (≥50%)	120 (13)	21 (25)
FIGO Stage
IA	805 (84)	50 (59)
IB	82 (9)	10 (12)
II	19 (2)	6 (7)
IIIA	10 (1)	2 (2)
IIIC	40 (4)	11 (13)
IVB	3 (<1)	6 (7)

BMI, body mass index; SLN, sentinel lymph node; FIGO, International Federation of Gynecology and Obstetrics

Of 1,044 patients who met the inclusion criteria, 85 (8%) had serous carcinoma. Patient characteristics, including median age, BMI, grade, depth of myoinvasion, number of SLNs removed and stage, are detailed in Table 1. All 85 patients underwent hysterectomy with bilateral SLN mapping, and all patients with ovaries underwent bilateral oophorectomy at the time of surgical staging. Forty-one tumors (48%) had no myoinvasion, 23 (27%) had <50%, and 21 (25%) had ≥50% myoinvasion reported on final pathology. All patients had clinical stage I disease at initial diagnosis, with final pathology demonstrating disease stage as follows: stage IA, 50 (59%); stage IB, 10 (12%); stage II, 6 (7%); stage IIIA, 2 (2%); stage IIIC, 11 (13%); and stage IVB, 6 (7%).

Pathology reports were reviewed, and SLN status was reported according to the aforementioned definitions (Methods section) of ITC as well as micro- and macrometastatic disease in SLNs. The incidence of SLN metastases reported by grade and depth of myometrial invasion in endometrioid endometrial carcinoma (n=959) is detailed in Table 2. Notably, there were no positive SLNs among the 449 patients with FIGO grade 1 noninvasive endometrioid carcinoma or the 61 patients with FIGO grade 2 noninvasive endometrioid carcinoma. Of patients with noninvasive FIGO grade 3 disease, 1 (5%) of 20 patients had a positive SLN. A positive SLN was reported for FIGO grades 1, 2, and 3 outer-half myoinvasion in 10% (6/62), 20% (8/41) and 24% (4/17) of patients, respectively.

Table 2.

Incidence of sentinel lymph node metastases reported by grade and depth of myometrial invasion in endometrioid endometrial carcinoma (n=959)

Depth of invasion	Grade 1 n (%)		Grade 2 n (%)		Grade 3 n (%)
	ITC	Micro/Macro	ITC	Micro/Macro	ITC	Micro/Macro
None	2/449 (<1)	0/449 (0)	0/61 (0)	0/61 (0)	0/20 (0)	1/20 (5)
Inner half	20/202 (10)	9/202 (4.5)	4/76 (5)	3/76 (4)	2/31 (6)	1/31 (3)
Outer half	19/62 (31)	6/62 (10)	7/41 (17)	8/41 (20)	3/17 (18)	4/17 (24)
Total	41/713 (6)	15/713 (2)	11/178 (6)	11/178 (6)	5/68 (7)	6/68 (9)

ITC, isolated tumor cells

Depth of invasion: inner half defined as <50% myoinvasion; outer half defined as ≥50% myoinvasion

Grade: refers to the International Federation of Gynecology and Obstetrics (FIGO) grading system for endometrial carcinoma

Among the 449 patients with FIGO grade 1 noninvasive endometrioid carcinoma, 2 (<1%) had ITCs; of those with outer-half myoinvasion, 19 (31%) of 62 had an ITC-positive SLN. There were no ITC-positive SLNs among the 61 patients with FIGO grade 2 noninvasive endometrioid carcinoma; of those with outer-half myoinvasion, 7 (17%) of 41 had an ITC-positive SLN. There were no ITC-positive SLNs among the 20 patients with FIGO grade 3 noninvasive endometrioid carcinoma; of those with outer-half myoinvasion, 3 (18%) of 17 had an ITC-positive SLN (Table 2).

Among 41 patients with noninvasive serous carcinoma, 2 (5%) had a positive SLN and 2 (5%) had an ITC- positive SLN. Of the 23 patients with inner-half myoinvasion, 4 (17%) had a positive SLN and no patient had an ITC-positive SLN. Among the 21 patients with outer-half myoinvasion, 7 (33%) had a positive SLN and 2 (10%) had an ITC-positive SLN (Table 3).

Table 3.

Incidence of sentinel lymph node metastases reported by depth of myometrial invasion in serous endometrial carcinoma (n=85)

Depth of invasion	ITC n (%)	Micro/Macro n (%)
None	2/41 (5)	2/41 (5)
Inner half	0/23 (0)	4/23 (17)
Outer half	2/21 (10)	7/21 (33)
Total	4/85 (5)	13/85 (15)

ITC, isolated tumor cells

Depth of invasion: inner half defined as <50% myoinvasion; outer half defined as ≥50% myoinvasion

Of the 959 patients with apparent early endometrioid carcinoma, 926 (97%) underwent bilateral oophorectomy at the time of staging surgery or had a prior oophorectomy for a benign indication. Of these 926 patients, 10 had a prior bilateral oophorectomy for a benign indication, 21 had a prior unilateral salpingo-oophorectomy (USO) for a benign indication and underwent contralateral USO at the time of staging surgery, and 2 patients underwent a subsequent oophorectomy after staging surgery given myoinvasion and LVI seen on final pathology in one patient and change from low- to high-grade histology in the other. The remaining 33 patients (3%), of whom 32 had FIGO grade 1/2 disease and 1 had FIGO grade 3 disease on final pathology, retained at least one ovary at the time of surgical staging. The patient with FIGO grade 3 disease, who retained both ovaries, is without evidence of disease at 34 months of disease surveillance. Of the 33 patients with endometrioid histology who retained at least one ovary, all are without evidence of disease at a median follow-up of 24 months (range, 0–124 months).

Table 4 presents the incidence of occult, isolated ovarian metastases reported on final pathology by grade for the endometrioid and serous endometrial carcinomas. There were no occult ovarian metastases in any patient with FIGO grade 1 or 2 endometrioid histology. Two (3%) of 68 total patients with FIGO grade 3 disease developed occult ovarian metastases. Among 85 patients with apparent early serous carcinoma who underwent bilateral oophorectomy at the time of surgical staging, 2 (2%) developed occult ovarian metastases in the absence of any other extra-uterine disease on final pathology. This included one case that was reviewed as a synchronous/metachronous carcinoma and a second case that was reviewed as metastatic from a primary serous endometrial carcinoma.

Table 4.

Incidence of isolated ovarian metastases reported by grade and histologic subtype in endometrioid (n=926^*) and serous (n=85) endometrial carcinoma

Endometrioid tumors n (%)			Serous tumors n (%)
Grade 1	Grade 2	Grade 3	NA
0/713 (0)	0/178 (0)	2/68 (3)	2/85 (2)

^*33 patients with endometrioid endometrial carcinoma retained at least one ovary at time of surgery and were not included in this table.

Grade: refers to the International Federation of Gynecology and Obstetrics (FIGO) grading system for endometrial carcinoma

Discussion

To our knowledge, this is the largest patient series to report on the incidence of nodal metastases in endometrioid and serous endometrial carcinoma using a uniform and modern staging approach. The strengths of our study include a population of more than 1,000 patients spanning a 14-year experience at a single institution, the use of a standardized surgical approach, and the incorporation of expert gynecologic pathologists’ evaluation using the MSK SLN mapping algorithm and pathologic ultrastaging.

A key finding is the detection of ITC-positive disease in 5% of patients with noninvasive serous carcinoma and in an additional 5% of patients with noninvasive endometrioid carcinoma in whom occult nodal metastases were detected. Taken together, based on the data of our series 10% of patients with serous endometrial carcinoma will have tumor cells in an SLN with no appreciable myoinvasion on final pathology. This finding supports the practice of SLN mapping with ultrastaging, even in noninvasive serous endometrial cancer. This information may be important in guiding adjuvant chemotherapy strategy. For example, in select cases of serous cancer with no myoinvasion, chemotherapy may be held, but if low-volume metastasis or ITCs are found in the SLN, further treatment discussion may be warranted.

The incidence of ITC-positive disease increased with increasing depth of myoinvasion across all grades and histologic subtypes in our series, with ITCs found in 25% of deeply invasive FIGO grade 1 and 2 tumors and in 18% of deeply invasive FIGO grade 3 tumors. Although currently staged as pN0(i+), the finding of ITCs in SLNs often impacts adjuvant treatment discussion and may influence prognosis and decisions regarding the frequency and methods of clinical surveillance.

The optimal management of ITC-positive SLNs in otherwise stage I endometrial carcinoma should not be based on a “node centric” approach and should take into consideration other uterine risk factors and the characteristics of the ITCs (e.g., cell count, distribution, patterns) to decide on adjuvant therapy. Uterine risk factors include grade, histology, myoinvasion, and LVI. Cases with ITCs can be discussed in a multidisciplinary setting of surgical oncologists, medical oncologists, pathologists, etc., in an effort to decide adjuvant treatment strategy. St. Clair et al. reviewed our institutional experience in the clinical management of ITC-positive endometrial cancer from 2005–2013. Of 844 patients included in the study, 23 had ITC-positive disease. Patients with ITC-positive disease were compared with node-positive and node-negative patients and were shown to have a more favorable 3-year recurrence-free survival than patients with macroscopic node-positive disease. ITC-positive patients were most similar to node-negative patients, with the caveat that the majority (87%) of ITC-positive patients received adjuvant chemotherapy [24]. Similar work by Plante et al., published in 2017, analyzed a group of 519 patients with endometrial cancer, of whom 31 had ITCs. The authors found that ITC-positive patients had similar 3-year progression-free survival outcomes as node-negative patients, which were statistically better than those of macroscopic-positive patients. This study shares a similar caveat in that 35% of the ITC-positive cohort received adjuvant chemotherapy ± pelvic radiotherapy, and an additional 32% received pelvic radiotherapy, with only one third of this group receiving no adjuvant treatment other than vaginal brachytherapy [25]. Both studies included multiple histologic subtypes with no uniform adjuvant treatment approach, which limits interpretation.

More recent reports on patients with ITCs in SLNs noted that these patients may have an increased risk of recurrence compared with node-negative patients with endometrial carcinoma. In 2018, Aloisi et al. reported on patterns of first recurrence in 207 patients with stage IIIC1 endometrial cancer who had undergone SLN-based surgical staging without a para-aortic lymphadenectomy. The authors reported a 17% recurrence rate among the 71 ITC-positive patients compared with a 37% recurrence rate among 135 patients with micro- or macrometastatic disease [26]. These findings suggest the presence of ITCs may be clinically relevant, impacting adjuvant treatment decisions and subsequent oncologic outcome. Unfortunately, we have insufficient evidence to support clear adjuvant treatment recommendations in ITC-positive patients, and this remains an active area of investigation. More research is needed to clarify the outcomes impact of uterine risk factors, histologic subtype, molecular subtyping, and adjuvant treatment choice in patients with ITC-positive SLNs. In addition, longer follow-up of at least 5 years is needed for ITC cases, as patterns of recurrence in low-volume disease may be delayed.

It is worth noting that in our series 57 patients with endometrioid endometrial carcinoma had ITCs in an SLN in the absence of other extra-uterine disease, and 95% of these patients had low-grade disease. This patient population is the most controversial in terms of defining adjuvant treatment recommendations, and we feel this should be a particular focus of study to determine the relevance of ITCs and the impact of systemic treatment. Current clinical practice focuses on established prognostic factors such as depth of invasion and the presence of LVI, with consideration of less-established prognostic factors such as the presence of ITCs. Future treatment decisions will also likely consider molecular classification of risk, as we await the results of a much-anticipated prospective trial from the PORTEC group. The trial is designed to examine high-intermediate risk endometrial cancer patients with regard to molecular profiling and adjuvant treatment decision making. The trial will examine how factors such as microsatellite instability and p53, CTNNB1 and POLE tumor status may help risk stratify patients to escalate or de-escalate adjuvant treatment [27].

Of note, we found only 2 (<1%) ITC-positive SLNs and no micrometastic SLNs among the 512 patients with low-grade, noninvasive endometrioid endometrial carcinoma. We feel this represents a negligible incidence of metastatic tumor detection in this population, and our current practice is to forego ultrastaging in patients with non-myoinvasive low-grade endometrial tumors. This is both time and cost effective as we consider use of valuable health care resources and how ultrastaging can benefit patient care.

In our series, we found that occult ovarian metastasis, in the absence of any other extra-ovarian disease, is uncommon in apparent early endometrial carcinoma. We report high rates of oophorectomy, as 97% of patients with endometrioid carcinoma and 100% of those with serous carcinoma underwent or completed BSO at the time of surgery. There were no occult ovarian metastases reported in low-grade endometrioid carcinoma, while 3% of patients with high-grade endometrioid and 2% of patients with serous carcinoma were found to have occult ovarian involvement. These data are consistent with population-based published literature with possible benefit of ovarian retention reaching beyond fertility preservation to include cardiovascular disease and overall mortality impact [28, 29]. Our data suggest that in premenopausal women with clinical stage I, mismatch repair intact, low-grade endometrial carcinoma, ovarian retention may be reasonable and should be discussed with patients. Our institution has developed selection criteria using the available literature on ovarian retention, as there are no uniformly applied criteria agreed upon across institutions. Much of the clinical decision making regarding ovarian retention in this population remains grounded in careful counseling and shared decision making.

We have amassed a large dataset to provide practical information to guide clinical decision making in apparent early-stage endometrial carcinoma. Our results reinforce the value of SLN mapping to detect occult tumor cells in nodes using an SLN mapping algorithm with ultrastaging. This dataset also updates our understanding of the incidence of nodal disease based upon FIGO grade, histology, and depth of myoinvasion in the era of finding ITCs in nodes. Further work is needed to determine the significance of ITCs and in which patient populations, particularly those with serous endometrial carcinoma, does the presence of ITCs merit consideration of adjuvant treatment.

Highlights.

• Low-volume metastases and isolated tumor cells are found in 10% of patients with noninvasive serous endometrial carcinoma

• Ultrastaging sentinel lymph nodes in patients with low-grade, noninvasive endometrioid endometrial carcinoma is low yield

• There were no isolated ovarian metastases in any patient with low-grade endometrioid endometrial carcinoma

• Isolated ovarian metastases in patients with high-grade endometrioid and serous endometrial carcinoma ranges from 2–3%