Catholic Health Services, Long Island, 1000 Montauk Highway West Islip, NY 11795, USAUnited States
Also available as PDF (figures and tables appear in the PDF).
Objective The study sought to evaluate the historical surgical-pathological trends in Gynecologic Oncology Group (GOG) 33 in a cohort of patients who underwent robot-assisted staging of uterine carcinomas. Materials and Methods This is a retrospective study from June 2016 through December 2018 at Catholic Health Services Hospitals in Long Island (CHSLI), NY, USA. All patients underwent robotic surgical staging with hysterectomy, bilateral salpingo-oophorectomy, and pelvic and para-aortic lymphadenectomy. Fifty patients were included, and patients were separated into cohorts with endometrioid or non-endometrioid histology for analysis. Patients were staged using the International Federation of Gynecology and Obstetrics (FIGO) 2009 classification. Results For all patients undergoing surgical staging, the rate of pelvic and para-aortic nodal metastases occurred at 10% and 8%, respectively. Sixty percent of the grade 1 endometrioid tumors had less than 50% myometrial invasion. No patients had positive para aortic lymph nodes in the absence of pelvic node involvement. High-risk histologies were associated with nodal disease thirteen percent of the time. A correlation between increasing depth of invasion and positive lymph nodes was demonstrated. Increasing grade of tumors was correlated with the frequency of nodal metastasis. Intraperitoneal spread was highly correlated to metastatic lymph nodes. Deep invasion was positively correlated with nodal disease. Conclusion The results of this study validate the trends previously known from GOG 33 in a population undergoing robot-assisted surgical staging for uterine carcinomas.
Keywords: Gynecologic Oncology Group (GOG) 33; Robotic; Surgery; Uterine cancer; Lymph nodes; Trends
This is a retrospective study from June 2016 through December 2018 at Catholic Health Services Hospitals in Long Island (CHS- LI), NY, USA. All patients underwent robotic surgical staging with hysterectomy, bilateral salpingo-oophorectomy, and pelvic and para-aortic lymphadenectomy. Fifty patients were included, and patients were separated into cohorts with endometrioid or non-endometrioid histology for analysis. Patients were staged using the International Federation of Gynecology and Obstetrics (FIGO) 2009 classification.
For all patients undergoing surgical staging, the rate of pelvic and para-aortic nodal metastases occurred at 10% and 8%, respectively. Sixty percent of the grade 1 endometrioid tumors had less than 50% myometrial invasion. No patients had positive paraaortic lymph nodes in the absence of pelvic node involvement. High-risk histologies were associated with nodal disease thirteen percent of the time. A correlation between increasing depth of invasion and positive lymph nodes was demonstrated. Increasing grade of tumors was correlated with the frequency of nodal metastasis. Intraperitoneal spread was highly correlated to metastatic lymph nodes. Deep invasion was positively correlated with nodal
The results of this study validate the trends previously known from GOG 33 in a population undergoing robot-assisted surgical staging for uterine carcinomas.
Gynecologic Oncology Group (GOG) 33; Robotic; Surgery; Uterine cancer; Lymph nodes; Trends.
ndometrial cancer is the most common gynecologic malignancy in the United States, afflicting approximately 61,380 E 1 annually. It is well-known today that most clinical stage I endometrial carcinomas are limited to the uterus, but 20% have occult metastatic disease on final pathology after surgical staging.2 Prior to 1988, endometrial adenocarcinomas were clinically staged diseases, primarily relying upon physical examination and radiological evidence. A landmark trial by the Gynecologic Oncology Group
cc Copyright 2019 by Wadhwa NA. This is an open-access article distributed under Creative Commons Attribution 4.0 International License (CC BY 4.0), which
(GOG), known as GOG 33, was conducted in 1987 and paved the way for the adoption of surgical staging for all endometrial carcinoma. The study prospectively examined patients with clinical stage 1 carcinoma of the endometrium and evaluated the surgicalpathologic behaviors of these tumors. The results were remarkable that 22% of patients assumed to have stage 1 uterine confined disease were found to have extrauterine spread at the time of surgery with microscopic pelvic and/or para-aortic metastases in 11% of women.3 After the results of this study were published, the International Federation of Gynecology and Obstetrics (FIGO) soon
thereafter reclassified endometrial carcinomas as a surgically staged
The patients in GOG 33 underwent abdominal surgical staging, including hysterectomy with pelvic and para-aortic lymphadenectomy. Since the publication of the study, minimally invasive techniques have become the standard of care for the comprehensive surgical staging of endometrial carcinomas.4,5 In gynecologic surgery, the 1990s were notable for the advancement of laparoscopic techniques, as well as for the development of more advanced surgical equipment. Multi-institutional clinical trials followed to compare laparotomy and laparoscopy. For example, in 2009, the GOG LAP 2 study was the most prominent prospective trial to compare these techniques during surgical staging procedures. Laparoscopy was found to be feasible, safe, and equivalent in nodal counts, and resulted in fewer complications and shorter hospital stays.5
Since abdominal staging has become largely replaced with laparoscopy, robotic surgery is increasingly utilized within gynecologic oncology. For example, the FIRES trial published in 2017 investigated the diagnostic accuracy of sentinel lymph nodes in uterine cancer using robotic surgery. The results of the trial showed a high sensitivity (97.2%), and a low false negative rate (3%) with the use of robotic surgery.6 Advantages of robotic surgery include improved visualization with three dimensional (3D) optics, greater dexterity with wristed instruments, and more comfortable ergonomics. The authors of this study sought to evaluate the historical surgical-pathological trends in GOG 33 in a cohort of patients with uterine carcinomas who underwent robot-assisted staging of uterine carcinomas. Subset analyses were performed based upon histology and high-risk factors.
This study is a retrospective, institutional review board (IRB)-exempt analysis from June 2016 through December 2018. Subjects included were women who underwent surgical staging of endometrial cancer at Good Samaritan Hospital in West Islip (GSHWI), NY, USA which is part of the Catholic Health Services Hospitals in Long Island (CHSLI), NY, USA. Sixty-five patients were identified during this time period. Fifteen patients were excluded for a variety of reasons, such as the inability to perform a complete surgical staging or the necessity to convert to a laparotomy.
Table 1.1. Descriptive Statistics for Scale Variables of Interest
n 50 50 50 50 48 48 45 50 No 28 56.0 Missing 0 0 0 0 2 2 Mean 63.50 76.00 81.82 31.84 28.21 172.9 2.909 96.62 No 40 80.0 Median 61.50 64.00 80.75 31.86 25.00 173.0 2.600 82.50 Yes 29 58.0 SD 11.17 32.97 19.91 7.638 14.85 60.38 1.916 99.55 No 21 42.0 Range 43.00 110.0 97.80 33.80 72.00 347.0 9.500 475.0 Yes 14 28.0 Minimum 46.00 60.00 45.50 18.80 0.000 50.00 0.000 0.000 No 36 72.0 Maximum 89.00 170.0 143.3 52.60 72.00 397.0 9.500 475.0 Fifty patients with carcinoma of the endometrium were evaluated who had undergone a robot-assisted total laparoscopic hysterectomy, bilateral salpingo-oophorectomy, bilateral pelvic lymphadenectomy, and para-aortic lymphadenectomy. No patient was excluded if lymphadenectomy was performed, irrespective of the number of lymph nodes removed.
4,5 As performed in GOG 33, the surgical pathologic findings of these patients were evaluated. Additionally, two cohorts were developed, dependent upon the uterine histology, and subgroup analyses were separately performed. Thirty-five patients with endometrioid endometrial adenocarcinoma were grouped together, as were fifteen patients with high-risk histologies. For the purposes of this study, the high-risk histologies included uterine serous, clear cell, carcinosarcoma, and undifferentiated adenocarcinomas. The FIGO stage was classified according to the 2009 guidelines.
A robot-assisted total laparoscopic hysterectomy and bilateral salpingo-oophorectomy was performed. In the pelvis, the bilateral retroperitoneal spaces were opened and the lymph bearing tissues were removed. The anatomic boundaries included the deep circumflex iliac vein distally; the midpoint of the common iliac artery proximally; the genitofemoral nerve laterally and the superior vesical artery medially. The obturator nerve served as the posterior boundary. If a suspicious node was identified in any area, it was removed. The peritoneum over the aorta and vena cava was opened. The fat pad above the inferior mesenteric artery extending to the proximity of the renal vessels was removed in toto. Peritoneal cytology was obtained at the conclusion of the lymphadenectomy.
Descriptive analyses were performed to determine the significance of each factor to the frequency of positive pelvic and paraaortic nodes. The findings of this study were compared to the results in GOG 33.
Overall, with all uterine histologies combined, our results were significant for a pelvic and para-aortic nodal metastases rate of 10% and 8%, respectively. Four percent (4%) had spread to the adnexa, and 2% had other extrauterine metastases at the time of surgery.
Table 1.2. Descriptive Statistics
Yes 22 44.0 Smoking History
5 0 Yes 10 20.0 Diabetes Mellitus
Hypertension
Hyperlipidemia
When analyzed as distinct cohorts, the results in the endometrioid and high-risk subgroups were different. Patients with endometrioid uterine histologies had a rate of metastasis of 9% to the pelvic lymph nodes, 6% to the para-aortic nodes, 3% to the adnexa, and 3% had other extrauterine metastases at the time of surgery. In the high-risk cohort, our results found that 13% of patients had pelvic
Table 2. Surgical-Pathologic Findings Surgical-Pathologic Findings of Endometrioid Adenocarcinomas Surgical-Pathologic Findings of Non-Endometrioid Adenocarcinomas
1A 22 63% 1A 8 53% 1B 8 23% 1B 3 20% II 1 3% II IIIA 1 3% IIIA 1 7% IIIB 0 0% IIIB 0 0% IIIC1 1 3% IIIC1 0 0% IIIC2 1 3% IIIC2 2 13% IVA 0 0% IVA 0 0% IVB 1 3% IVB 0 0%
Endometrioid Adenocarcinoma 35 100% MMMT/Carcinosarcoma 2 13% Grade Clear Cell 5 33% 1 Well 20 57% Serous 6 40% 2 Moderate 10 29% Undifferentiated 2 13% 3 Poor 5 14% Grade Myometrial Invasion 1 Well 0 0% Endometrium only 11 31% 2 Inner 1/3 8 23% 3 Middle ½ 8 23% Myometrial Invasion Deep 1/3 8 23% Endometrium only 5 33% Peritoneal Cytology Inner 1/3 5 33% Positive 6 17% Middle ½ 3 20% Negative 22 63% Deep 1/3 2 13% Not Collected 5 14% Peritoneal Cytology Indeterminate 2 6% Positive 7 47% Adnexa Involvement Negative 7 47% Positive 1 3% Not Collected 1 7% Negative 34 97% Indeterminate 0 0%
Positive 3 9% Positive 1 7% Negative 32 91% Negative 14 93%
Positive 2 6% Positive 2 13% Negative 33 94% Negative 13 87%
Positive 1 3% Positive 2 13% Negative 34 97% Negative 13 87%
Positive 6 17% Positive 0 0% Negative 29 83% Negative 15 100% nodal metastases, 13% had para-aortic lymphadenopathy, 7% had spread to the adnexa, and 0% had other extrauterine metastases at the time of surgery.
The distribution of the study parameters are detailed in Table 1. There were more patients with endometrioid adenocar-
1 7%
Moderate 0 0% Poor 15 100%
Positive 2 13% Negative 13 87%
cinoma. Thirty five patients (70%) had endometrioid histology, and fifteen (30%) had a high-risk histology. Well-differentiated and poorly-differentiated tumors were evenly present as 20 patients (40%) had well differentiated tumors, and 20 patients (40%) had poorly differentiated tumors. Forty-two percent (42%) of patients had tumors with middle and deep myometrial invasion (greater than 50% myometrial invasion). Twenty-six percent (26%) of patients had positive cytology. Sixteen percent (16%) of patients had tumors with lymphovascular space invasion (Table 2).
Most endometrioid tumors were grade 1 and confined to the endometrium. Sixty percent (60%) of the grade 1 endometrioid tumors were present within the endometrium or superficial myometrium. Sixty-six percent (66%) of the high-risk histology tumors were present within the endometrium or superficial myometrium (Table 3).
Table 3. Histologic Grade and Depth of Invasion Histologic Grade and Depth of Invasion in Endometrioid Adenocarcinomas
Endometrium Only 8 (40%) 2 (22%) 1 (17%) 11 (31%) Superficial 4 (20%) 1 (11%) 3 (50%) 8 (23%) Middle 5 (25%) 2 (22%) 1 (17%) 8 (23%) Deep 3 (15%) 4 (44%) 1 (17%) 8 (23%) Total 20 (100%) 9 (100%) 6 (100%) 35 (100%)
Histologic Grade and Depth of Invasion in Non-Endometrioid Adenocarcinomas
Endometrium Only 0 0 5 (33%) 5 (33%) Superficial 0 0 5 (33%) 5 (33%) Middle 0 0 3 (20%) 3 (20%) Deep 0 0 2 (13%) 2 (13%)
Total 0 0
Five patients (10%) had metastasis to the pelvic and/or to the para-aortic lymph nodes. Four patients (8%) had nodal metastasis to both pelvic and para-aortic lymph nodes. Zero patients had positive para-aortic lymph nodes in the absence of pelvic node involvement. Eighty percent of patients who had positive pelvic lymph nodes also had positive para-aortic lymph nodes (Table 4).
Table 4. Relationship of Positive Pelvic Nodes to Aortic Nodes Relationship of Positive Pelvic Nodes to Aortic Nodes in Endometrioid Adenocarcinomas Aortic
Negative 32 (91%) 0 (0%) 32 (91%) Positive 1 (3%) 2 (6%) 3 (9%) Total 33 (94%) 2 (6%) 35 (100%) Relationship of Positive Pelvic Nodes to Aortic Nodes in Non-Endometrioid Adenocarcinomas Aortic
Negative 13 (87%) 0 (0%) 13 (87%) Positive 0 (0%) 2 (13%) 2 (13%) Total 13 (87%) 2 (13%) 15 (100%)
Differing histologic cell types were related to pelvic lymph nodes metastasis. Clear cell and undifferentiated histologies were associated with positive pelvic and para-aortic lymph nodes 29% of the time. Any non-endometrioid histology was associated with nodal disease 13% of the time (n=2). Endometrioid adenocarcinoma affected pelvic and para-aortic lymph nodes 9% of the time (n=3).
Grade of the tumors was correlated with the frequency of nodal metastasis. No grade 1 tumors had any positive lymph
Table 5. Frequency of Nodal Metastasis Among Risk Factors Pelvic Aortic
MMMT/ 2 0 0% Carcinosarcoma Clear Cell 5 1 20% 1 20% Serous 6 0 0% 0 0% Undifferentiated 2 1 50% 1 50% Endometrioid 35 3 9% 2 6% Adenocarcinoma
1 Well 20 0 0% 0 0% 2 Moderate 10 1 10% 1 10% 15 15 3 Poor 20 4 20% 3 15% (100%) (100%)
Endometrium only 16 0 0% 0 0% Inner 1/3 13 0 0% 0 0% Middle 1/3 11 2 18% 1 9% Deep 1/3 10 3 30% 3 30%
Positive 13 2 15% 2 15% Negative 29 1 3% 0 0% Not Collected 6 1 17% 1 17% Indeterminate 2 1 50% 1 50%
Positive 2 0 0% 0 0% Negative 48 5 10% 4 8%
Positive 1 1 100% 1 100% Negative 49 4 8% 3 6%
Positive 8 3 38% 3 38% Negative 42 2 5% 1 2%
nodes. Four patients with grade 3 tumors had positive lymph nodes (20%). The same trend was present for para-aortic lymph nodes (Tables 5 and 6).
Table 6. Grade, Depth of Invasion, and Pelvic Node Metastasis Grade, Depth of Invasion, and Pelvic Node Metastasis of Endometrioid Adenocarcinomas
Endometrium Only 0 (0%) 0 (0%) 0 (0%) N=11 Inner 0 (0%) 0 (0%) 0 (0%) N=8 Middle 0 (0%) 0 (0%) 1 (100%) N=8 Deep 0 (0%) 1 (25%) 1 (100%) N=8 N=8
Grade, Depth of Invasion, and Pelvic Node Metastasis of Non-Endometrioid Adenocarcinomas
Endometrium Only 0 (0%) 0 (0%) 0 (0%) N=5 Inner 0 (0%) 0 (0%) 0 (0%) N=5 Middle 0 (0%) 0 (0%) 1 (33%) N=3 Deep 0 (0%) 0 (0%) 1 (50%) N=2
A correlation between increasing depth of invasion and positive lymph nodes was demonstrated. No patients with tumor confined to the endometrium or inner myometrium (less than 50% myometrial invasion) had nodal metastasis. 23% of patients with pelvic node disease and 25% of patients with para-aortic disease had tumors with at least 50% myometrial invasion.
Thirteen patients (26%) had positive cytology. Those with non-endometrioid histologies had positive cytology 47% of the time, compared to 17% with patients with endometrioid adenocarcinoma. Overall, 15% of patients with positive cytology also had positive lymph nodes. All patients in this study with positive washings were affected with pelvic and para-aortic nodal disease.
Adnexal involvement was present in 2 patients (4%). Regardless of histology, adnexal involvement was not related to nodal
Intraperitoneal spread was highly correlated to metastasis to the pelvic and para-aortic lymph nodes. In the one patient with intraperitoneal spread, both the pelvic and para-aortic nodes were affected. However, the absence of intraperitoneal spread does not exclude nodal disease. In 49 patients with negative intraperitoneal spread, 4 patients (8%) were affected by positive nodal disease. This trend was seen in endometrioid adenocarcinoma tumors and high-risk tumors.
Lymphovascular space invasion (LVSI) was present in 17% and 13% of endometrioid and high-risk tumor histologies, respectively. Non-endometrioid tumors affected by LVSI had positive pelvic and para-aortic lymph nodes 50% of the time. Endome- Grade trioid adenocarcinoma with LVSI was seen to have positive pelvic and para-aortic lymph nodes in thirty-three percent of cases.
Grade and depth of invasion were evaluated in the correlation of nodal disease. Even though both are risk factors, depth of invasion is more influential. Deep invasion of any histology of tumor was correlated with nodal disease. No patients with superficial invasion were affected by positive lymph nodes. All five patients with nodal disease were noted to have at least 50% myometrial invasion (Tables 6 and 7).
Table 7. Grade, Depth of Invasion, and Aortic Node Metastasis Grade, Depth of Invasion, and Aortic Node Metastasis of Grade Endometrioid Adenocarcinomas Grade
Endometrium Only 0 (0%) 0 (0%) 0 (0%) N=11 Inner 0 (0%) 0 (0%) 0 (0%) N=8 Middle 0 (0%) 0 (0%) 0 (0%) N=8 Deep 0 (0%) 1 (20%) 1 (100%) N=8
Grade, Depth of Invasion, and Aortic Node Metastasis of Non-Endometrioid Adenocarcinomas Grade
Endometrium Only 0 (0%) 0 (0%) 0 (0%) N=5 N=5 Inner 0 (0%) 0 (0%) 0 (0%) N=5 N=5 Middle 0 (0%) 0 (0%) 1 (33%) N=3 Deep 0 (0%) 0 (0%) 1 (50%) N=2
GOG 33 was a landmark trial to instill the reform for uterine cancers to become a surgically staged disease. Our primary objective was to compare the trends from GOG 33 to the surgicalpathologic trends noted in this retrospective study of patients who underwent robot-assisted laparoscopic staging of uterine cancers. The two cohorts of patients analyzed were those with endometrial
endometrioid adenocarcinomas and those with non-endometrioid uterine adenocarcinomas.
Overall, several trends noted in our data paralleled the findings from GOG 33 regarding the frequency of nodal involvement and depth of tumor invasion. As expected, the non-endometrioid cohort contained high-risk uterine subtypes, and this group demonstrated an increased rate for metastatic spread outside of the uterus with involvement of the adnexa, pelvic, and para-aortic lymph nodes. All uterine subtypes also showed a positive correlation between lack of tumor differentiation and an increased incidence of lymph node spread. None of the patients with welldifferentiated tumors were noted to have nodal spread. Conversely, all of the patients with pelvic or para-aortic lymph node metastasis were seen to have poorly differentiated grades of tumor, regardless of histology. There was one patient with a moderately differentiated tumor grade who had spread to the pelvic and para-aortic lymph nodes. Lastly, all of the patients with positive pelvic and para-aortic nodal metastases had greater than 50% myometrial invasion.
Our results show two important differences between GOG 33 and other contemporary trials. The results from GOG 33 show a 2% rate of isolated positive para-aortic nodes without pelvic nodal involvement. In a contemporary trial from 2008, Mariani et al describe a 16% isolated positive para-aortic rate in those patients with deeply invasive, high grade tumors. Our study, on the other hand, shows a 0% isolated positive para-aortic lymph node rate with robot-assisted surgery, in both endometrioid and non-endometrioid tumors. We show that all patients with positive para-aortic lymph nodes also had positive pelvic lymph nodes.
A second difference between our results and GOG 33 is the concomitant rate of positive aortic nodes when the pelvic nodes harbor metastatic disease. GOG 33 showed approximately one-third of aortic nodes will be positive with pelvic nodal disease. Instead, our results show a higher rate with 80% aortic involvement. It is possible that our study is underpowered to make the same observations as previous studies. Another explanation may be that the magnification present during robotic surgery allows the surgeon to better visualize enlarged lymph nodes, leading to their removal, with fewer reported instances of isolated paraaortic tumor spread. Further studies with robot-assisted surgical staging are needed to evaluate this possibility.
Our data is significant in that it parallels the results of previous landmark studies, providing further validation for robotassisted laparoscopic surgical staging of endometrial carcinoma. Previous research has evaluated conventional laparoscopic surgical staging. The GOG LAP 2 trial compared conventional laparoscopy to laparotomy and demonstrated decreased hospital inpatient stays and blood loss with low rates of recurrence. Additionally, other publications have compared the outcomes of robotic surgery to conventional laparoscopy. Corrado et al examined the surgical and oncological outcomes in robotic surgical staging in obese patients with endometrial cancer. The authors demonstrated the safety, feasibility, and reproducibility of robotic surgical staging. Nieto et al examined the rates of robotic-assisted procedures, as well as survival, in non-endometrioid endometrial cancers. The authors report that 75% of all minimally invasive surgeries were performed robotically without adversely impacting survival.10 Additional studies have evaluated other outcomes, such as blood loss, nodal counts, and post-operative complications in laparoscopic 11,12 surgical staging.
Despite the work previously performed from the above studies, no other study has evaluated the surgical-pathological trends in robot-assisted surgical staging for uterine cancers as did GOG 33. In summary, the results of this study validate the trends previously known from GOG 33, in a population undergoing robot-assisted surgical staging for uterine carcinomas.
The authors declare that they have no conflicts of interest.