6. Surgical therapy of endometrial carcinoma – Endometrial Cancer

Background

Traditionally, radical hysterectomy (resection of the parametria) was recommended for stage pT2 endometrial carcinoma (involvement of the cervical stroma) [503]. This recommendation was based on small case series, such as Tamussino et al. [504], who found continuous progression of carcinoma from the cervical stroma into the parametria in 2 of 16 patients with cervical involvement. The Japanese GOTIC study group [315] retrospectively analyzed data from 300 EC patients with suspected macroscopic cervical involvement. Seventy-four women had received radical, 112 modified radical, and 114 simple hysterectomy. The type of hysterectomy did not affect the rate of local recurrence, progression-free survival or overall survival, even when clear cervical involvement was demonstrated on the hysterectomy specimen. Intraoperative complications and postoperative voiding dysfunction were found significantly more often in the groups with radical or modified radical hysterectomy.

A recent meta-analysis of 10 retrospective cohort studies involving 2,866 patients showed no significant advantage of radical hysterectomy for overall survival (HR 0.92; 95%-CI 0.72-1.16; P = 0.484) or progression-free survival (HR 0.75; 95%-CI 0.39-1.42; P = 0.378). Even after adjuvant radiotherapy was considered, there was no advantage of radical hysterectomy [502]. In case of parametrial involvement, stage pT3b is present, which should be treated with radical hysterectomy under the aspect of R0 resection.

Background

There are no published studies explicitly addressing the role of lymphadenectomy in advanced endometrial cancer. Multiple retrospective studies address the benefits of “optimal” cytoreductive surgery in patients with stage III and IV endometrial cancer [527], [528]. Each study was able to show a statistically significant benefit in terms of PFS and OS when optimal cytoreduction could be achieved.

No direct evidence exists in the  literature for the recommendation of sentinel-assisted systematic lymphadenectomy in patients with EC type I, pT2 to pT4, M0, G1-3. The recommendation is based on the estimated unfavorable prognosis and the assumption of a higher detection rate of lymph node metastases and an associated higher probability of benefit from adjuvant therapy.

Background

Type II carcinomas are more poorly differentiated and have a worse prognosis than type I tumors. Relative to the proportion of all endometrial cancers (10–20% of cases), they have a disproportionate share of endometrial cancer-related deaths (40%) [532].

No study separately reports the effect of systematic LNE versus no lymphadenectomy in patients with type II endometrial carcinoma, as they are mostly included in the so-called “high-risk” group of advanced or metastatic endometrioid carcinomas. In the SEPAL study alone, type II carcinomas (n = 55) were classified into an intermediate-risk (FIGO I and II) or high-risk (FIGO III u. IV) group based on stage. As will be shown below, pelvic and para-aortic LNE in intermediate and high-risk compared with pelvic LNE alone showed a reduction in the risk of death [533]. In addition, patients with type II carcinomas and proven pelvic metastasis (FIGO IIIC1) had more frequent occult metastases on ultrastaging of the para-aortic lymph nodes than patients with type I carcinomas [533].

No evidence from randomized trials exists in the retrieved literature for the recommendation of sentinel-assisted systematic lymphadenectomy in patients with EC type II. The recommendation is based on the prognosis being considered unfavorable and the assumption of a higher detection rate of lymph node metastases and an associated higher probability of benefit from adjuvant therapy.

Background

The effect of performing para-aortic lymphadenectomy on survival in endometrial cancer was investigated in a retrospective cohort analysis, the so-called SEPAL study [533].

Retrospectively, 671 patients with endometrial cancer treated at two tertiary centers with complete systematic pelvic (n = 325 patients) or combined pelvic and para-aortic lymphadenectomy were studied (n = 346) (January 1986 to June 2004). Patients at intermediate or high risk of recurrence were offered adjuvant radio- or chemotherapy. The primary outcome was OS. Here, OS was significantly better in the pelvic and para-aortic than in the pelvic LNE group (HR = 0.53, 95%-CI 0.38–0.76; p = 0.0005). Combined pelvic and para-aortic LNE had a positive effect in 407 patients at intermediate or high risk of recurrence (p = 0.0009), but not in low-risk patients. In multivariate analysis, pelvic and para-aortic LNE in intermediate and high-risk compared with pelvic LNE alone reduced the risk of death (0.44, 0.30–0.64; p < 0.0001). Analysis of 328 intermediate- or high-risk patients treated with adjuvant radio- or chemotherapy showed that both pelvic and para-aortic LNE (0.48, 0.29–0.83; p = 0.0049) and adjuvant chemotherapy each independently resulted in longer survival (0.59, 0.37–1.00; p = 0.0465).

It should be critically noted that the difference between pelvic only versus pelvic + para-aortic LNE in terms of HRs was substantially greater in the SEPAL trial than in the studies that examined LNE versus no LNE. The study is characterized by the risk of substantial bias [533] .

Odagiri et al [536] performed a systematic LNE from the femoral annulus to the renal vein in 266 patients with EC. A mean of 62.5 lymph nodes were removed (range 40–119). Forty-two women (15.8%) had lymph node metastases, of which 16 (38%) were exclusively pelvic, 7 (16.7%) were exclusively para-aortic and 19 (45.2%) were pelvic plus para-aortic. That is, para-aortic lymph nodes were involved in approximately 60% of women with lymph node metastases. 11% of positive lymph nodes were found above the inferior mesenteric artery [536].

Alay et al [537] performed systematic pelvic plus para-aortic lymphadenectomy to the renal pedicle in 204 EC patients. A mean of 69.1 lymph nodes were removed (range 33–122). 44 patients (21.6%) had lymph node metastases, 27 of which were para-aortic. 11 of these women had exclusively para-aortic metastases above the inferior mesenteric artery, 4 exclusively below this vessel and 12 in both locations [537].

The above and other retrospective analyses [536] show that in EC with lymph node metastases, the para-aortic lymph nodes are affected in at least 50%. In accordance with the lymphatic drainage of the corpus uteri, which is not insignificantly along the ovarian vessels, the lymph nodes between the inferior mesenteric artery and the renal pedicle are affected in relevant frequency in cases of lymph node involvement.

A recent systematic review with meta-analysis including eight studies and 2,793 patients showed that para-aortic plus pelvic LNE resulted in prolonged overall survival compared with pelvic LNE alone in patients with intermediate and high risk EC (HR= 0.52¸95%-CI = 0.39-0.69, P < 0.001) [535]. In patients with low risk EC, additional para-aortic LNE to the renal pedicle did not significantly improve overall survival. However, given the inherent limitations of retrospective studies, the authors of the meta-analysis call for validation of the results with sufficiently large RCTs [535].

A recent analysis of 3,650 women with lymph node-positive EC from the SEER database showed that compared with patients with pelvic LNE alone, those with additional para-aortic LNE had lower overall (HR= 0.74;95%-CI= 0.63-0.88) and EC-specific mortality (HR= 0.79; 95%-CI = 0.66-0.95) [538].

Background

Surgical therapy for uterine carcinosarcoma has traditionally been similar to other endometrial cancer therapy and should include hysterectomy and bilateral adnexal extirpation [539].

In 2008, Nemani et al. [529] analyzed the courses of 1,855 operated patients with stage I (n = 1,099), II (n = 245), and III (n = 353) uterine carcinosarcoma from the SEER database; 965 women (57%) had received lymphadenectomy. A median of 12 lymph nodes had been removed. 119 (14%) of lymphadenectomized patients had positive lymph nodes. The 5-year overall survival was significantly better in patients with LNE than in those who did not have lymphadenectomy (49% versus 35%). Median survival was 54 months (CI 44–72) versus 25 months (CI 22–29). Adjuvant radiotherapy did not improve overall survival [529].

The study, like many retrospective analyses in the SEER database (see above), is at considerable risk of bias. For example, the finding that the number of lymph nodes removed (< 12 versus > 12) had no effect on improving survival is not plausible. The patients with positive lymph nodes were scored as stage IIIC, resulting in stage migration (improvement in both stage I by removal of the pN1 patients and stage III by inclusion of the patients who were putatively stage I but had pN1 microscopically). However, since lymph node metastases are found in more than 10% of uterine carcinosarcoma, systematic lymphadenectomy seems a reasonable measure.

A retrospective analysis of 1,140 women with uterine carcinosarcoma from Dutch databases showed an improvement in overall survival when more than 10 lymph nodes had been removed (HR = 0.67; 95%-CI = 0.50-0.89; P= 0.006 in multivariate analysis). Additional adjuvant radiotherapy and/or chemotherapy improved overall survival in nodal-positive patients or those without LNE, but not in nodal-negative women with carcinosarcoma [530].

No evidence from randomized trials exists in the retrieved literature to recommend sentinel-assisted systematic lymphadenectomy in patients with carcinosarcoma of the uterus. The recommendation is based on the estimated unfavorable prognosis and the assumption of a higher detection rate of lymph node metastases and an associated higher probability of benefit from adjuvant therapy.

Background

Sentinel node biopsy (SNB) is an established concept in many tumors for the safe detection of tumor-involved lymph nodes and has advantages over conventional systematic lymphonodectomy due to less trauma. Studies have investigated SNB as an alternative and also additive to systematic LND.

Various labeling procedures and injection techniques for intraoperative identification of sentinel lymph nodes have been evaluated so far in endometrial cancer mainly in single-center studies; see for example [540], [541]. The Sentiendo study [542], [543], [544], [545]is the most important multicenter study. Detection rates achieved with intrauterine or cervical injection varied between 35 and 100%. A meta-analysis including 1,385 patients showed a sensitivity of the method of 81% and a false-negative rate of 19% [546]. Algorithms involving complete pelvic lymphadenectomy of the respective hemi-pelvis in the absence of unilateral detection increased the sensitivity to 95% and decreased the false-negative rate to 5%.

The FIRES trial is a multicenter prospective cohort study of the use of the sentinel lymph node procedure in early endometrial cancer [439]. At the 10 participating U.S. hospitals, 385 clinical stage 1 patients were labeled by cervical injection with indocyanine green. In subsequent surgery, accumulating SNL-lymph nodes were detected, removed, and pelvic +/- para-aortic lymphonodectomy was performed. SNL mapping with complete pelvic lymphonodectomy was performed in 340 and additional para-aortic lymphonodectomy in 196 (58%) patients. At least one SNL-lymph node was visualized in 293 (86%) of the patients. Forty-one (12%) patients had affected LK, of which 36 had at least one labeled SNL lymph node. Metastases in SNL-lymph nodes were detected in 35 (97%) of the 36 labeled patients resulting in a sensitivity of 97.2% (95%-CI 85.0-100), and a negative predictive value of 99.6% (97·9-100) for the detection of LK metastasis. The authors concluded that because of its diagnostic accuracy in detecting invaded lymph node, SNL-LND is a safe alternative to systematic LND in endometrial cancer.

Approximately 4.1 to 5.6% of all unselected patients with low-risk endometrial cancer have lymph node metastases [507].

In the retrospective analysis by J. Mueller and N. R. Abu-Rustum et al. [547], 959 patients with clinical stage I endometrial carcinoma were evaluated: no positive (macro- or micrometastases) SLNs were found among the 510 patients with non-invasive FIGO grade 1/2 endometrial carcinoma.

In grade 1 and pT1a, 4.5% (9/202) and in grade 1 and pT1b, 10% (6/62) had positive SLNs.

For grade 2 and pT1a, it was 4% (3/76) and for grade 2 and pT1b, 20% (8/41).

For grade 3: 5% (1/20) with non-invasive endometrial cancer, 3% (1/31) with invasion of the inner myometrial layer and 24% (4/17) with invasion of the outer myometrial layer had positive SLNs.

The PORTEC-II study [407] investigated the impact of lymphatic vessel invasion (LVSI) risk factors, p53 mutation, and L1CAM expression on the likelihood of recurrence in women with high-intermediate risk EC. Extensive LVSI (see Chapter 4.5) was found to be a strong and independent risk factor for pelvic and distant recurrences (hazard ratio 8.73 (p = 0.005) and 5.36 (p = 0.001), respectively) and also for disease-related survival (HR 7.16, p  < 0.001).

Sentinel lymphadenectomy is associated with low morbidity and mortality, but also allows better staging in the low-risk endometrial cancer group [439], [548].

A retrospective study recently demonstrated a potential benefit of sentinel lymphadenectomy in these endometrial carcinomas. 279 patients with endometrial carcinoma at low risk of recurrence (FIGO stage 1, endometrioid histology, grades 1 and 2) were divided into three groups: 103 (36.9%) had no lymphadenectomy, 118 (42.3%) had SLN removal, and 58 (20.8%) patients had pelvic and/or para-aortic lymphadenectomy. After a median follow-up of 33 months, the SLN group, compared with the group in which no lymphadenectomy was performed, showed a trend toward better recurrence-free and overall survival. However, when comparing peri-operative morbidities, no significant differences were found between the group without lymphadenectomy and that with SLN removal [548]. Thus, in endometrial cancer with low risk of recurrence, SLN removal allows identification of the low percentage of patients who may benefit from adjuvant therapy.

A recent analysis of the SEER database showed in 11,603 patients with stage IA G1 endometrial cancer G1– G3 that neither systematic LNE nor sentinel node biopsy improved survival [549].

Ultrastaging allows intensive pathologic workup of identified sentinel lymph nodes. As a result, lymph node metastases are detected in up to 5% in the low-risk population that would not be identified by conventional pathologic evaluation. However, the significance of micrometastases and single-cell metastases identified in this manner is unclear. Performing SNL lymphodectomy appears safe even in early stages (FIGO I, G1/G2), where the likelihood of lymphogenic metastasis is low, does not lead to higher surgical morbidity and may influence adjuvant therapy. There may be an improvement in oncologic prognosis for cases with LVSI [548].

Especially in low and intermediate risk endometrial carcinomas, ultrastaging detects more positive sentinel nodes [437], [439], [440]. In a good half of the cases of positive sentinel nodes, these are based on isolated tumor cells (ITC) or micrometastases [437], [439], [440] whose relevance for prognosis is at least unclear, possibly even irrelevant [437], [440]. Therefore, the generous performance of sentinel node biopsy with ultrastaging may also lead to unnecessary upstaging and initiation of adjuvant therapies that only increase morbidity without improving prognosis [437], [440]. Micrometastases (> 0.2 mm to ≤2 mm, pN1 (mi)) are considered by most experts to be metastatic disease indicating adjuvant therapy [436], [237].

In a retrospective study, patients with micrometastases were shown to have the same DFS as nodal-negative women when the micrometastases considered nodal-positive were treated adjuvantly [436]. Whether this is due to a therapeutic effect or lack of relevance of micrometastases cannot be clarified by this study.

The prognostic value of ITCs (pN0 (i+)) is unclear [237]. Adjuvant therapy is recommended only if additional risk factors (e.g. LVSI, p53 mutation, type II carcinoma) are present [526], [237], [437], [440], [531].

The definitive classification of an endometrial carcinoma into the “low-risk group” is made postoperatively after receipt of the definitive pathology report. A proportion of endometrial carcinomas are accordingly classified postoperatively into a “high-risk group”, with a risk of lymph node metastases of up to 40% [550]. The lack of information about lymph node status often leads to generously indicated adjuvant radiotherapy in these patients, which would have been omitted if appropriate knowledge about lymph node status had been available [551], [552]; Sharma et al. 2011; [553]. Therefore, among others, Sinno et al. have suggested that if sentinel lymph nodes are identified bilaterally, a frozen section of the uterus to determine the depth of infiltration should be omitted [554], [548].

A multi-institutional retrospective study compared long-term outcomes after LND, LND plus SNL, and SNL alone [531]. Using a propensity score matching algorithm, 180 patients with SNL-LND (90 SNL alone, 90 SNL plus LND) were identified and compared with 180 patients with lymphonodectomy. 10% of all patients had metastatically affected lymph nodes. Comparison of DFS and OS showed no significant difference between the three groups, even when the cohort was divided into low-, intermediate-, and high-risk cancers. The addition of SNL lymphonodectomy allowed better detection of nodal-positive patients than systematic LND alone. In the cohort with SNL plus lymphonodectomy (17% lymph node-positive patients), 16% were diagnosed in an SNL lymph node and only 1% in lymph nodes removed by systematic lymphonodectomy.

A similar result was found in a prospective multicenter cohort study (SENTOR study) of patients with intermediate- to high-grade carcinomas [525]. 156 patients, including 126 with high-grade endometrial cancer, were included. All underwent SNLB and pelvic LND, and 101 patients (80%) with high-grade endometrial carcinomas also underwent para-aortic LND. SNL detection rates were 97.4% per patient (95%-CI, 93.6%-99.3%), 87.5% per hemipelvis (95%-CI, 83.3%-91.0%) and 77.6% bilaterally (95%-CI, 70.2%-83.8%).

Of the 27 patients (17%) with lymph node metastases, 26 were correctly identified by the SNLB algorithm, resulting in a sensitivity of 96% (95%-CI, 81%-100%), a false-negative rate of 4% (95%-CI, 0%-19%), and a negative predictive value of 99% (95%-CI, 96%-100%). Only one patient (0.6%) was misclassified by the SNL algorithm. Seven of the 27 nodal-positive patients (26%) were found outside the usual lymphonodectomy areas or only by immunohistochemistry, so as a conclusion, the use of the SNL method may improve the detection of lymph node metastases in high-grade carcinomas.

However, it should be considered that up to more than 50% of positive sentinel nodes are due to micrometastases or isolated tumor cells (ITC), the clinical relevance of which is unclear or questionable [437], [438], [440], [524], [525], [Bogani, G et al. 2020]). Moreover, in the studies showing higher sensitivity of sentinel node biopsy, systematic LNE according to US standards was performed in the comparison groups (11–20 lymph nodes, no presacral LNE, no LNE above the inferior mesenteric artery). It is therefore not surprising that positive sentinel nodes were found in these regions, which would have been missed in the “systematic” LNE [439], [524] [440].

A retrospective cohort study in patients with type 2 EC sought to answer the question of whether SNL LND may also be an alternative to systematic pelvic and para-aortic LND in clear cell or serous EC (type II) at high risk for lymphogenic metastasis [523]. Lymph node staging was performed by SNL-LND in patents operated on at Memorial Sloan Kettering between 2006 and 2013 (n = 118) and by systematic lymphonodectomy in patients at the Mayo Clinic between 2004 and 2008 (n = 96). Although overall survival was not significantly different, the nodal-negative SNL group had worse DFS than the systematic lymphonodectomized nodal-negative group.

In conclusion, therefore, there is currently no direct evidence of an advantage of sentinel lymphadenectomy or sentinel-assisted lymphadenectomy over systematic lymphadenectomy in terms of recurrence-free or overall survival in patients with EC. Study data are also lacking for the therapeutic consequences of diagnosing micrometastases and isolated tumor cells in sentinel lymph nodes. Prospective randomized trials have not established whether adjuvant chemotherapy, radiotherapy or combined chemoradiation therapy is beneficial in the presence of micrometastases or isolated tumor cells in sentinel lymph nodes with respect to recurrence-free or overall survival.

Background

The procedure should start with laparoscopy. Only after any adhesions or other pathologies have been removed should ICG be injected.

If the LSK reveals suspicion for “bulky nodes”, the injection of ICG is no longer useful.

The ICG powder (25 mg/ampoule) is dissolved in 10 ml of aqua for injection, this gives a solution of 2.5 mg/ml. Cervical injection is superior to hysteroscopic peritumoral injection [555]. Various injection techniques have been described in the literature. Injection of 1ml each (0.5 ml submucosal and 0.5 ml stromal) at 2 (at 3 and 9 o'clock) or 4 (2, 4, 8, 10 o'clock) sites in the cervical region has proven effective. Pelvic spread of ICG along the lymphatic vessels is clearly identifiable after 6 to 10 minutes. If bilateral spread of ICG does not occur and the SLN cannot be identified on one side of the pelvis, ICG may be re-injected. However, the total daily dose should be less than 5 mg/kg bw.

If there is only unilateral visualization of the sentinel even with repeated injections, a systematic pelvic LNE is recommended on the ICG-negative side (except for low-risk Ia, G1-2).

Various labeling procedures and injection techniques for intraoperative identification of sentinel lymph nodes have been evaluated so far in endometrial carcinoma, mainly in single-center studies; see for example [540], [541]. For the most important multicenter study on this, the Sentiendo study (ClinicalTrials.gov, number NCT00987051), see [542], [543], [544], [545]. Detection rates obtained with intrauterine or cervical injection varied from 35 to 100%. A meta-analysis including 1,385 patients showed a sensitivity of the method of 81% and a false-negative rate of 19% [546]. Algorithms involving complete pelvic lymphadenectomy of the respective hemi-pelvis in the absence of unilateral detection increased the sensitivity to 95% and decreased the false-negative rate to 5%. In the Swedish prospective SHREC study of 257 patients with high-risk EC, intracervical injection of ICG achieved a sensitivity of 98% and a negative predictive value of 99.5%. The bilateral staining rate was 95% [556].-The Canadian prospective SENTOR study found a sensitivity of 96% and a negative predictive value of 99% in 156 patients with intermediate and high-risk EC after intracervical injection of ICG [525].