Adenocarcinoma of the uterine corpus and sarcomas of the uterus

Pathology

Careful evaluation of the uterus by the pathologist is essential for proper diagnosis and treatment of corpus cancer ( Fig. 5.5 ). Gross inspection of a bivalved uterus at the time of hysterectomy can offer an impression of the size of the lesion, its location (involvement of fundus, lower uterine segment, or cervix), and the depth of tumor penetration into the myometrium (depth of invasion). A clinically enlarged uterus may be caused by increasing tumor volume, but this should not be the only gauge for significant local disease. Obviously, many patients can have enlarged uteri because of factors other than AC. Carcinoma of the endometrium may start as a focal discrete lesion, as in an endometrial polyp. It may also be diffuse in several different areas, in some situations involving the entire endometrial surface. As the tumor grows, it can become larger or spread within the endometrium or myometrium. Endometrial cancer may disseminate to regional lymph nodes, by embolization or direct extension into the pelvis or vagina, or hematogenously to distant organs ( Fig. 5.6 ). The risk of spread is related to several factors, including depth of invasion into the myometrium, tumor grade, and histologic type. Further adding to the complexity, is the observation that type 2 carcinomas may not have grossly visible tumor but can be deeply infiltrative microscopically.

Endometrioid AC, the most common histologic type, is usually preceded by a predisposing lesion, atypical endometrial hyperplasia ( Table 5.3 ). Only hyperplasia with cellular atypia is considered a true “precancerous” precursor of AC of the endometrium. For most patients with endometrioid-type tumors, particularly grade I or II lesions, and hyperplasia, hyperestrogenism is the etiologic basis. Pathologically, endometrial cancer is characterized by the presence of glands in an abnormal relationship to each other, with the hallmark of little, if any, intervening stroma between the glands. There can be variations in the size of the glands, and infolding is common. The cells are usually enlarged, as are the nuclei, along with nuclear chromatin clumping and nucleolar enlargement. Mitosis may be frequent. Differentiation of AC (grades I, II, or III) is important prognostically and is incorporated into FIGO surgical staging ( Fig. 5.7 ). Most studies suggest that 60% to 65% of all endometrial cancers are of the endometrioid subtype.

For almost a century, it has been recognized that a squamous component may be associated with an AC of the endometrium. This occurs in about 25% of patients. Historically, patients with a squamous component were further stratified according to whether the squamous component appeared benign (designated adenoacanthoma [AA]) or malignant (designated adenosquamous carcinoma [AS]). It was suggested that AA indicated a good prognosis and those with AS had a poor survival. Today, this distinction has been questioned concerning its prognostic importance. Zaino and coworkers, in reporting data from the GOG, suggest that the notation of squamous component irrespective of differentiation does not affect survival. Patients with clinical stage I and stage II cancers were evaluated, and 456 with typical AC and 175 with squamous differentiation (AC + SQ) were identified. The latter were subdivided into 99 with AA and 69 with AS. Multiple known prognostic factors were compared with differentiation of glandular and squamous component of the tumor. Age, depth of myometrial invasion, architecture, nuclear grade, and combined grade were similar for AC and AC + SQ, although patients with AA were better differentiated than those with AS and had less myometrial invasion. Both glandular and squamous differentiation correlated with frequency of pelvic and paraaortic node metastasis. Nodal metastasis, when it was stratified for grade and depth of invasion, was similar in AC and AC + SQ patients. The differentiation of squamous component is closely correlated with the differentiation of the glandular element, and the glandular element is a better predictor of outcome. It would therefore appear that the previous designation of AA and AS has no added predictive property than differentiation of glandular component and probably should be dropped as a diagnostic term. The authors suggest the term squamous differentiation instead, with differentiation of the glandular component noted as the important prognostic factor.

Secretory AC ( Fig. 5.8 ) is an uncommon type of endometrial cancer. It usually represents well-differentiated carcinoma with progestational changes. It is difficult to differentiate from secretory endometrium. Survival is good and comparable to that associated with the pure AC. Although it is an interesting histologic variant, the separation of the entity as it relates to treatment and survival is probably not warranted.

Increasing emphasis has been placed on the importance of uterine serous carcinoma (USC). This subtype represents approximately 5% to 10% of all ACs but is highly aggressive, representing up to 40% of endometrial cancer-related deaths. Unlike the more common variants, USC is not associated with hyperestrogenism and frequently develops in the setting of atrophic endometrium. It is more commonly seen in older and non-White patients. Hendrickson, in the early 1980s, noted that in more than 250 endometrial cancers, only 10% had histologic features of USC, but these accounted for 50% of all treatment failures. The histopathologic appearance resembles a high-grade serous carcinoma of the ovary that has a propensity for vascular or lymphatic vascular space involvement (LVSI), as well as extrauterine spread, which can occur in the absence of myometrial invasion. Well-formed papillae are lined by neoplastic cells with grade III cytologic features ( Fig. 5.9 ). Differentiation between papillary architecture and syncytial metaplasia with benign endometrial alterations must be made because the papillary architecture alone does not designate USC. The uterus may appear grossly normal, without a visible lesion, but can have extensive myometrial invasion. Most USC are aneuploid and have a high S-phase and can be pure or admixed with other histologic types (endometrioid, clear cell, carcinosarcoma). USC commonly arise in endometrial polyps, with varying degrees of myometrial invasions. Current models of pathogenesis hypothesize that USC is a progressive lesion, arising from endometrial intraepithelial carcinoma (EIC), a finding supported by molecular assessments of paired EIN and USC tissue samples ( ; ).

Clear cell carcinomas ( Fig. 5.10 ) are also infrequent, accounting for 1% to 6% of all endometrial cancer, with distinct histologic criteria. Clear cell tumors are characterized by large polyhedral epithelial cells that may be admixed with typical non–clear cell ACs. Some authorities accept the mesonephritic-type hobnail cells as part of this pattern, but others believe that this histologic type should be excluded from the clear cell category. Silverberg and DeGiorgi and Kurman and Scully suggested a worse prognosis for clear cell AC than for pure AC. This was confirmed in studies by Christopherson and coworkers. Even in stage I disease, only 44% of patients with clear cell carcinomas survive 5 years. Neither the FIGO classification nor nuclear grade correlates with survival. Photopulos and associates, in a review of their material, noted that patients with this entity were older and tended to have a worse prognosis. They did note that patients with stage I clear cell carcinomas had a 5-year survival rate similar to that of patients with stage I pure AC of the endometrium.

Tumor grade

In addition to histologic type, pathologists assign a measure of tumor differentiation, known as grade, to endometrial cancers. Grade I lesions are well differentiated, frequently associated with estrogen excess, closely resemble hyperplastic endometrium, and are generally associated with a favorable prognosis. Grade III lesions are poorly differentiated, do not resemble normal endometrium, and frequently have a poorer prognosis. Grade II tumors are moderately differentiated and have an intermediate prognosis. Both architectural criteria and nuclear grade are used to classify. Architectural grade is related to the proportion of solid tumor growth, with grade I having an AC with less than 5% of the tumor exhibiting solid sheets, grade II having 6% to 50% of the neoplasm arranged in solid sheets of neoplastic cells, and grade III having greater than 50% of the neoplastic cells in solid masses. Regions of squamous differentiation are excluded from this assessment. The FIGO rules for grading state that notable nuclear atypia, inappropriate for architectural grade, raises the grade of a grade I or grade II tumor by one. By convention, serous and clear cell histologies are considered grade III or high-grade tumors.

Stage of disease: Depth of invasion, cervical involvement, adnexal involvement, and nodal metastasis

Staging of patients with malignancies is designed to have prognostic value by classifying the extent of tumor. The survival rate in regard to stage of disease has been consistent, and Table 5.6 and Fig. 5.11 show the 5-year survival rate reported by FIGO (FIGO 1988 staging). The pattern of disease spread in endometrial cancer was evaluated in a prospective study performed by the GOG (GOG protocol 33) and reported by Creasman and colleagues. This study is required reading for physicians who care for patients with endometrial cancer. GOG protocol 33 was a surgical-pathologic study of 621 patients with clinical stage I endometrial cancer who were uniformly treated with total abdominal hysterectomy (TAH), bilateral salpingo-oophorectomy (BSO), peritoneal cytologic evaluation, and selective pelvic and paraaortic lymphadenectomy. Before this study, it was presumed most patients with endometrial cancer were at risk for nodal metastases, and nearly all patients required some form of pelvic radiation therapy (either preoperatively or postoperatively). The study demonstrated important relationships between pathologic factors and risk of nodal disease. For example, factors associated with nodal disease included higher tumor grade, deeper myometrial invasion, cervical involvement, (+) cytology, and LVSI. Because of this study, FIGO accepted a surgical-based staging system, including lymphadenectomy, of endometrial cancer in 1988.

For endometrial cancer, FIGO staging reflects the progression of disease within the uterus and at extrauterine sites. Approximately 75% of patients with endometrial cancer present with disease limited to the uterus ( Table 5.7 ). For stage I disease, depth of tumor invasion into the myometrium is an important prognostic factor. The degree of myometrial invasion is a consistent indicator of tumor virulence ( Fig. 5.12 ) . DiSaia and associates noted that recurrences were directly related to depth of myometrial invasion in patients with stage I cancer treated primarily with surgery ( Table 5.8 ). The Annual Report of FIGO demonstrated a decrease in the survival rate as myometrial penetration increased ( Table 5.9 ). Lutz and coworkers determined that the depth of myometrial penetration was not as important as the proximity of the invading tumor to the uterine serosa. Patients whose tumors invaded to within 5 mm of the serosa had a 65% 5-year survival rate, but patients whose tumors were more than 10 mm from the serosa had a 97% survival rate. The depth of myometrial invasion is associated with the other prognostic factors, such as tumor grade. As noted by DiSaia and associates, the survival rate of patients with poorly differentiated lesions and deep myometrial invasion is poor in contrast to that of patients who have well-differentiated lesions but no myometrial invasion. This suggests that virulence of the tumor may vary considerably, and as a result, therapy should depend on the combination of prognostic factors.

Location of the tumor within the endometrial cavity is important because tumors low in the cavity may involve the cervix earlier than fundal lesions. The prognosis for women with cervical stromal involvement (stage II) is worse than those with stage I disease. Cervical involvement is often a surrogate marker for extrauterine disease spread or for risk of local recurrence. Data from GOG 33 showed that those with disease of the lower uterine segment had a higher incidence of pelvic lymph node metastases (16%) than did those with only fundal disease (8%). There is a similar frequency of paraaortic nodal metastases: a 16% incidence from disease of the lower uterine segment and a 4% incidence when only fundal disease is present. Previously, endocervical curettage (ECC) was commonly used to determine whether the patient had cervical involvement. Many false-positive results occur with this technique, however. In addition, the prognostic significance of endocervical glandular spread (formerly stage IIa) has been challenged. The new surgical staging adopted by FIGO in 2009 includes only patients with cervical stromal invasion in the stage II category. In some cases, cervical biopsy or ECC is required in the pretreatment planning of a patient with suspected cervical involvement.

Patients with stage III or IV disease may have a heterogeneous mix of disease characteristics. It is well recognized that endometrial cancer can and frequently does metastasize to the adnexa. Patients with adnexal or serosal involvement are categorized as having stage IIIA disease and are considered to have a higher risk of peritoneal recurrence. In GOG 33, 5% of patients had adnexal involvement. Adnexal metastasis was significantly associated with involved pelvic (32%) and paraaortic (20%) lymph nodes. In a large surgical trial comparing laparoscopy with open staging (GOG Lap 2 trial), 5% of the 2616 patients enrolled were found to have stage IIIa disease (FIGO 1988 staging). In an analysis of 222 patients with clinical stage I carcinoma of the endometrium reported by DiSaia and colleagues, 16 (7%) were found to have metastasis in the adnexa. This finding correlated with many but not all of the other prognostic factors. Spread to the adnexa did not seem to be related to the size of the uterus. The grade of the disease did not appear prognostically important in regard to this in that 6% of patients with grade I tumors had adnexal disease compared with only 10% if poorly differentiated carcinoma was present. The depth of invasion did appear to be significant, however, in that 4% of patients with only the endometrium involved had adnexal spread compared with 24% who had adnexal metastases if deep myometrial invasion was present. If tumor was limited to the fundus of the uterus, only 5% of patients had disease in the adnexa; however, if the lower uterine segment or the endocervix was involved, one-third had spread to the adnexa. When metastasis was present in the adnexa, 60% of patients had malignant cells in the peritoneal cytologic fluid compared with only 11% if the adnexa were not involved. Recurrences appeared in only 14% of these individuals who did not have metastasis to the adnexa compared with recurrences in 38% of patients with adnexal metastasis.

Metastatic spread to lymph nodes (stage IIIc) has prognostic and therapeutic implications. It is clear that lymph node dissection, to date, is the most sensitive way to identify nodal disease. The GOG 33 study demonstrated that as the depth of tumor invasion or tumor grade increased, the frequency of pelvic and paraaortic nodal metastases also increased ( Tables 5.10 and 5.11 ). For all 621 patients, 58 (9%) had positive pelvic nodes, and 34 (6%) had positive paraaortic nodes. Patients with grade I, II and III tumors had pelvic nodal metastases of 3%, 9%, and 18% respectively. Similarly, patients with no myometrial invasion, inner one-third, middle one-third, and outer one-third invasion had 1%, 5%, 6%, and 25% pelvic nodal disease, respectively. The highest risk group included patients with grade III tumors and outer one-third invasion who had pelvic nodal involvement in 34% of cases ( Table 5.12 ). Without a lymph node dissection, one may grossly estimate the probability of nodal involvement based on these data to select for or against the use of adjuvant therapy. This strategy potentially results in undertreatment or overtreatment of patients, however. In two large prospective randomized trials, routine lymphadenectomy resulted in the identification of more patients with nodal disease (9% to 13%) compared with when only clinically suspicious nodes were removed (1% to 3%). It is clear that patients with nodal disease have a poorer prognosis (3- to 5-year survival rate of 50% to 75%) and different patterns of failure (nodal, distant) than patients with negative nodes (3- to 5-year survival rate of 80% to 95%, vaginal cuff failures predominate). Patients with nodal involvement include those with pelvic nodal disease (IIIc1) or paraaortic involvement (IIIc2). The sub-staging was created in 2009 given the belief of different outcomes associated with different levels of nodal involvement. The number of involved nodes and the extent of resection of grossly involved nodes also affect outcome. In addition, Mariani and McMeekin have suggested that patients with positive nodal disease plus other stage III defining features (adnexal or serosal involvement) have a much poorer prognosis than those with nodal disease only. Patients with stage IV or distant disease spread (intraperitoneal, lung, liver) have a particularly poor prognosis, with 5-year survival rate of less than 20%. The extent of surgical resection has been suggested to alter the prognosis in patients with advanced-stage disease. As with ovarian cancer, the extent of resection of disease reflects the biology of the tumor, aggressiveness of the surgery, and response to postoperative therapies.

Surgical management of endometrial cancer

Sampling of the endometrium in symptomatic postmenopausal patients is recommended as the first diagnostic technique. If histologic findings are “negative,” the patient is observed. An endometrial curettage is recommended if the patient continues to be symptomatic after the negative endometrial biopsy result. After a tissue diagnosis of endometrial cancer is established, the patient should be assessed for surgical options of therapy. Most patients require routine complete blood counts (CBC), metabolic profile, and a metastatic evaluation with chest radiography. Routine use of computed tomography (CT) or MRI scans has not shown to be useful and should be reserved for patients with high-risk tumor types, evidence of intraperitoneal disease, positive chest radiograph, or clinical evidence of metastatic disease. Most patients with endometrial cancer are candidates for definitive surgery, including surgical staging. Routes of hysterectomy include vaginal, abdominal, laparoscopic (total or assisted), and robotic and have broadened the surgical options for patients.

The surgical evaluation of most patients with endometrial cancer requires a thorough inspection of the peritoneal cavity, collection of cytologic washings, and sentinel lymph node (SLN) dissection in addition to hysterectomy. Cytologic evaluation of peritoneal fluids, or washings, has been a common surgical step in staging endometrial cancer because of associations with extrauterine disease and prognosis. After the peritoneal cavity is opened, an assessment of the amount of peritoneal fluid in the pelvis is made. If none is present, approximately 100 mL of normal saline solution is injected into the pelvis and aspirated. This can be done with the suction irrigation device at the time of minimally invasive surgical staging. Peritoneal cytologic evaluation may be performed in patients undergoing surgery for endometrial cancer, although FIGO no longer uses cytology as a stage-defining characteristic. Omental biopsy may be considered in patients with gross spread to the omentum or adnexa or in cases with high-risk histologies such as serous or clear cell tumors. In a large GOG study of 621 staged patients, 6% of patients with clinical stage I disease were noted to have intraperitoneal disease.

The definitive treatment for patients with endometrial cancer is hysterectomy. The hysterectomy should be extrafascial, and removal of the upper vagina does not appear to decrease vault recurrences ( Fig. 5.13 ). Extirpation of the uterus removes the primary tumor and can provide important information that can be used to estimate risk of spread to lymph nodes or risk of recurrence. Removal of the adnexa is thought to be important given that approximately 5% of endometrial cancers have metastatic disease to the ovaries, fallopian tubes, or both. In addition, synchronous ovarian and endometrial cancers are not infrequent, particularly in younger patients. Historically, total hysterectomy and BSO have been the hallmarks of therapy for endometrial cancer. There are no data to suggest that the route of hysterectomy influences RFS or OS outcomes for patients; as such, the approach (robotic, laparoscopic, or abdominal) that is the safest and facilitates rapid recovery is recommended. Minimally invasive surgical approaches are preferred given the reduced morbidity and rapid recovery without decrement in oncologic outcome, as supported by the LAP-2 publications.

Several studies note the role of vaginal hysterectomy in highly selected patients with endometrial cancer for whom surgical staging cannot be performed safely using alternate approaches (significant cardiopulmonary medical comorbidities, advanced age). Because factors associated with the choice of vaginal hysterectomy (morbid obesity, medical comorbidities) are often associated with favorable uterine characteristics (low-grade lesions, smaller uterus), it is not surprising that survival rates are comparable to those of the abdominal approach. Chan and coworkers reviewed 51 medically compromised patients treated with vaginal hysterectomy and reported 3- and 5-year disease-specific survival rates of 91% and 88%, respectively. Of note, approximately 50% of women could not have the ovaries removed at the time of vaginal hysterectomy . Assessment of lymph nodes is also not feasible with the vaginal approach and should be a component of patient counseling. Smith evaluated 63 patients with obesity or medical comorbidities and found that vaginal hysterectomy was safe and well tolerated in this patient population. Vaginal hysterectomy may represent a reasonable option for patients who may not tolerate other approaches or for whom surgical staging is not being considered (atypical hyperplasia, some patients with grade I cancers).

Over time, there has been a more widespread use of lymphadenectomy in the management of patients with endometrial cancer. Lymph node dissection provides the best estimate of spread of disease (vs. palpation or use of imaging studies), the lymph node status is prognostically important (as evident from the different survival rates seen with different stages of disease), and patients who are found to have positive or negative nodes receive different postoperative therapy versus patients with unknown status of lymph nodes. This may be particularly relevant in the context of recent clinical trials suggesting survival benefits with incorporation of adjuvant chemotherapy in patients with node positive endometrial cancer.

Unfortunately, complete pelvic and/or paraaortic lymphadenectomy is associated with intraoperative and postoperative complications including nerve and vascular injury, lymphocyst formation, lymphedema, and neuralgia and is difficult to accomplish in morbidly obese patients ( ; ). Furthermore, two large prospective phase III trials failed to show a survival advantage with incorporation of lymph node dissection into the surgical staging algorithm.

In a large multicentric study (A Study in the Treatment of Endometrial Cancer [ASTEC]), 1408 patients were randomly assigned to TAH or BSO with or without pelvic lymphadenectomy (PLA). With a median follow up of 37 months, there was no difference in survival between the two groups. This was an intent-to-treat study and has been criticized as having a relatively large noncompliance in regard to lymphadenectomy and subsequent therapy. In the lymphadenectomy arm, almost half had no nodes or a small number (≤9) removed. This study was further complicated by the fact that many of these patients were secondarily randomly assigned into a postoperative radiation study, not taking into consideration surgical findings. para-aortic lymphadectomy (PALA) was not the standard management, although some patients had these nodes removed.

The second study was published by Panici and colleagues. The Italian study randomly assigned 514 stage I patients to systematic PLA or no lymphadenectomy. Although 13% of patients in the lymphadenectomy arm had known metastasis versus 3% in the no-lymphadenectomy arm, there was no difference in recurrence or OS. The protocol required a minimum of 20 lymph nodes to be removed; however, the lymphadenectomy was limited to the pelvis, although PALA could be removed at the discretion of the surgeon. Postoperative radiation was left to the discretion of the treating physician. Adjuvant therapy was similar in the two groups. Whether or not this had an effect on survival is unknown because this was not standardized.

In an effort to mitigate adverse events associated with comprehensive lymphadenectomy, while recognizing the prognostic value of defining lymph node status, investigators explored the role of sentinel lymph node dissection in the management of patients with endometrial cancer. The initial feasibility of sentinel lymph node dissection in endometrial cancer was described in retrospective, single institution studies and smaller prospective trials ( ; ). The Fluorescence Imaging for Robotic Endometrial Sentinel lymph node biopsy (FIRES) trial was designed to estimate the sensitivity and negative predictive value of sentinel-lymph-node mapping using robotic assisted fluorescence imaging of the tracer indocyanine green (ICG) in detecting lymphatic metastases in patients with endometrial cancer ( Table 5.16 ) ( ). This multicenter, cohort study enrolled a total of 385 patients with endometrial cancer of any histology, and clinical stage 1 disease, without physical examination or radiographic imaging suggestive of extrauterine disease. Investigators injected ICG into the cervix at a dose of 0.5 mg/mL by diluting 1 mL of the stock solution (2.5 mg/mL) into 4 mL of sterile water. A spinal needle was used to inject 1 mL (0.5 mg) of the ICG solution into the uterine cervix at 3 and 9 o’clock of the ectocervix to a 1 cm depth, achieving a total dose of 1 mg. Successful mapping was defined by observing a channel leading from the cervix directly to at least one candidate lymph node in at least one hemi-pelvis. Identified sentinel lymph nodes were then retrieved and labelled for location ( Fig. 5.14 ). Completion bilateral lymphadenectomy (removal of all remaining non-sentinel lymph node tissue within the relevant nodal basins) was then done on all patients according to the GOG surgical handbook. PLA was required in all patients, with PALA at the discretion of the operating physician. Excellent surgical quality control was incorporated into this trial. Pathologists assessed SLNs for metastases by hematoxylin and eosin (H&E) staining, followed by pancytokeratin AE1 and AEs IHC ultra-staging if H&E assessment was negative. Importantly, 102 patients (29%) had high-grade histology (grade III endometrioid, serous and clear cell cancer). Mapping identified at least one SLN in 293 (86%) of 340 patients. Two of three patients with a mapped isolated paraaortic lymph node had metastatic disease, with otherwise negative pelvic LN. They found an impressive 97.2% sensitivity and 99.6% negative predictive value ( ). In this large study, SLN mapping was not successful in 13.8% of eligible patients.

Following publication of the FIRES trial, SLN dissection was adopted into the staging algorithm for endometrial cancer (both minimally invasive and open approaches), with the NCCN recognizing SLN dissection as an appropriate surgical option (Category 2A). It should be emphasized that the success of SLN mapping is dependent on appropriate application of the algorithm, which includes completion of side-specific comprehensive lymph node dissection in cases of failed mapping and the removal of any suspicious appearing LN irrespective of mapping. Lastly, in cases where SLN mapping is not successful, providers can use uterine factors to guide and inform need for nodal assessment.

An unanticipated consequence of incorporating SNL dissection into uterine cancer care has been the identification of isolated tumor cells (ITCs) (<0.2 mm) on pathologic nodal evaluation. In the FIRES trial, 19 patients had low-volume disease identified in the SLN, with 10 subjects having ITCs. ITCs do not upstage patients (N0(i+)), and to date, studies suggest no difference in outcome based on choice of adjuvant therapy ( ). In a large, multi-institutional, retrospective study of 175 patients with stage I/II endometrioid endometrial cancer and SLN ITCs, adjuvant therapy did not impact oncologic outcomes (RFS), suggesting observation may be a reasonable approach in this setting (Backes, 2021, #2599). Longer follow-up time and sample sizes are required to confirm these findings.

When comprehensive lymphadenectomy is required, the retroperitoneal spaces in the pelvis are opened in routine fashion. The vessels are outlined, and the lymph node–bearing tissue along the external iliacs from the bifurcation to the inguinal ligament is removed. The obturator fossa anterior to the obturator nerve is cleaned of lymphoid tissue. Lymph nodes along the common iliacs are also removed. The left and right paraaortic nodes are approached by retracting the small intestine into the upper abdomen and incising the peritoneum over the upper common iliac artery and lower aorta. The main vessels are outlined, and the ureter is retracted laterally. On the right, the tissue overlying the vena cava and the aorta are removed en bloc, beginning at the bifurcation of the aorta and extending caudad. On the left, the left common nodes are frequently quite lateral. Using this technique, one should have a total of 20 to 30 pelvic and paraaortic lymph nodes available for histologic evaluation. The upper limit of the dissection (unless enlarged nodes are noted above this area) is usually the inferior mesenteric artery (IMA), but some surgeons suggest that a comprehensive lymphadenectomy should extend to the level of the renal vessels.

In the context of prior studies suggesting two nodal drainage basins—pelvic and paraaortic—the initial adoption of sentinel LN algorithms was accompanied by a concern for paraaortic failures in patients with negative pelvic sentinel LN. Prior studies indicated that when lymph nodes are positive, approximately 50% to 60% of the time paraaortic nodes are involved. In a retrospective study from the Mayo Clinic, 137 patients at high risk for nodal involvement who underwent PALA (PAL+) were compared with those who did not undergo surgical evaluation of the paraaortic nodes (PAL−). The 5-year survival rate was 85% for PAL + patients compared with 77% for PAL − patients. In 51 patients with pelvic or paraaortic node metastasis, survival was 77% for PAL + patients compared with 42% in the PAL − group. It is generally thought that the rate of isolated paraaortic lymph node metastasis is less than 5%; metastasis isolated to the paraaortic lymph node and metastasis above the IMA are seen. Mariani and colleagues reported on a study of 482 patients with endometrial cancer at the Mayo Clinic. A total of 281 underwent lymphadenectomy to the renal vessels, and 22% had positive lymph nodes; of these, 51% had both positive pelvic and paraaortic nodes, 33% had positive pelvic nodes only, and 16% had isolated involvement of the paraaortic nodes. Of note, 46% of those with isolated paraaortic lymph nodes were only positive above the IMA, and 77% had at least one metastatic node located above the level of the IMA. However, the authors do report that there were no positive lymph nodes detected in those with grade I or II tumors with a tumor size of smaller than 2 cm with less than 50% invasion, raising the question of the benefit of lymphadenectomy in the truly low-risk group. Abu-Rustum and colleagues demonstrated a 1% risk of isolated paraaortic lymph node metastasis in both the low- and high-grade tumors.

In a meta-analysis which included 3536 patients in six studies, investigators showed higher positive pelvic nodal detection rates, no difference in paraaortic nodal detection rates, and no difference in overall recurrence or nodal recurrence rates in SLN compared with complete lymphadenectomy ( ). Specific to high-risk histology, a prospective study examined women with grade III endometrioid, serous, clear cell, and carcinosarcoma by SLN biopsy followed by full pelvic and paraaortic lymphadenectomy; they found 95% sensitivity and 98% negative predictive value and bilateral mapping rates of 58% and unilateral mapping rates of 40%, supporting the use of SLN in high-risk patients ( ).

Since 1988 when FIGO changed endometrial staging from clinical to surgical, the debate as to whether the lymphadenectomy is only diagnostic, which is an important determinate, or whether it could also be therapeutic has persisted. Kilgore and associates, in evaluating 649 patients, noted that those who underwent multiple-site lymph node removal had significantly better survival rates than those patients who had no lymph nodes removed ( Figs. 5.15 and 5.16 ). Lymph node removal resulted in a better survival rate than for those without lymph node removal plus postoperative radiation. Cragun and colleagues reported that patients with grade III or poorly differentiated cancers with more than 11 lymph nodes removed had an improved survival (HR, 0.25) and PFS (HR, 0.26) compared with those with fewer than 11 lymph nodes removed. However, the number of lymph nodes removed was not predictive of survival outcome in those with grade I and II tumors. This association between lymphadenectomy and improved survival remained when controlling for adjuvant radiation treatment, emphasizing the impact of extensive lymph node dissection. A retrospective review by Chan and colleagues of more than 12,000 patients showed an increased 5-year survival rate for the intermediate- and high-risk group that underwent an extensive lymph node dissection. However, there was no benefit of nodal dissection seen in the low-risk group.

Removal of nodes involved by tumor has been supported as a therapeutic option. In a study reported by Havrilesky et al., 91 patients were identified with stage IIIc disease. There were 39 with microscopic involvement of the lymph nodes and 52 with grossly enlarged nodes. After surgery, 92% received some type of adjuvant therapy with 85% receiving radiation therapy. The survival rate (5 years) was 58% in the 39 with microscopic lymph nodes, 48% in 41 patients with grossly positive lymph nodes completely resected, and only 22% in the 11 with unresected lymph nodes. The authors believed that these data suggested a therapeutic benefit for lymphadenectomy. Similarly, in a small study of 41 patients, Bristow and coworkers noted disease-free survival (DFS) was improved if patients with bulky adenopathy underwent complete resection of involved nodes compared with patients who had gross residual disease in lymph nodes remaining after surgery (37.5 vs. 8.8 months; P = .006).

Onda and colleagues carried out thorough pelvic and paraaortic lymphadenectomies on 173 patients with stage I to III endometrial cancer. The average numbers of lymph nodes removed were 38 pelvic and 29 paraaortic. There were 30 patients (17%) with positive nodes: 10 to pelvis only, two paraaortic only, and 18 with metastasis to both pelvic and paraaortic nodes. Selected patients received radiation therapy with extended fields, combination chemotherapy, or both. In the 10 patients with only pelvic metastasis, the 5-year survival rate was 100%; it was 75% in those with paraaortic involvement. The authors suggest that, although postoperative treatment may contribute to these excellent results, systematic pelvic and PALA was a contributing factor.

One of the important effects of lymph node dissection is that it identifies that most patients with negative lymph nodes are at very low risk for recurrence. As such, lymph node dissection has allowed for modifications of postoperative therapy away from pelvic radiation. In one study, Mohan and associates evaluated 159 stage I patients who had full PLA and received vaginal brachytherapy rather than the more traditional pelvic radiation. In general, vaginal brachytherapy is a quicker therapy (1 to 3 days vs. 5 to 6 weeks for pelvic radiation) and is associated with fewer side effects. In the Mohan study, the 15-year OS was 92%, and the recurrence rate was 4%, all at distant sites. Podratz and associates reviewed four studies that used thorough lymphadenectomies in moderate- and high-risk patients who did not receive postoperative radiation therapy. There were 20 recurrences (7%) in 305 patients; only five recurrences were local or regional, with four being in the vagina. Those four did not receive postoperative brachytherapy but were salvaged with subsequent radiation.

The disadvantages of a lymph node dissection include that the performance of lymph node dissection requires a specially trained surgeon, has the potential to increase complications, as previously discussed, and has not been prospectively validated to improve outcomes. Two phase III trials compared the use of lymph node dissection or not in patients with endometrial cancer with no identifiable improved outcome with lymph node dissection.

The Cochrane Database concluded that there was no overall statistical difference between PFS and OS in patients who had a lymphadenectomy versus those who did not but did report more morbidity for those patients who had a lymphadenectomy. Cochrane only considered the ASTEC and the Panici trial in reaching this conclusion.

Patients with stage II carcinoma of the endometrium, because of extension of disease into the endocervix, will have a greater propensity for lymph node metastasis. For example, in the GOG 33 protocol, 16% of patients with cervical involvement had positive pelvic lymph nodes versus 8% with a fundal location of tumor. Even in patients with occult cervical involvement, 19% of patients had lymph node metastases. In addition, parametrial and vaginal involvement are thought to be more common when the cervix is involved. As such, surgery in the form of radical hysterectomy and lymphadenectomy can be considered when there is gross cervical involvement with tumor. Because many cases of cervical involvement are occult and only recognized after surgery, simple hysterectomy with lymph node sampling (SLN algorithm) appears to be adequate surgery in most cases. Postoperative radiation therapy can be planned, depending on surgical-pathologic findings, including use of pelvic radiation or VCB, both, or neither.

Currently, a minimally invasive approach to endometrial cancer staging is advocated given similar oncologic outcomes with reduced intra- and perioperative morbidity. A large GOG study, led by Walker and colleagues (GOG LAP2), compared laparoscopy versus laparotomy in the surgical management of endometrial cancer. Patients were randomly assigned between laparoscopic versus open laparotomy surgical staging. Laparoscopy was initiated and completed without conversion in 74%. Conversion from laparoscopy to laparotomy was secondary to poor visibility in 15%, metastatic cancer in 4%, and bleeding in 3%. Patients randomly assigned to undergo laparoscopy had significantly fewer moderate to severe postoperative (14% versus 21%) complications and similar rates of intraoperative complications. Length of hospital stay was significantly shorter for those randomized to undergo laparoscopy (median 3 days vs. 4 days), but operative time was significantly longer (median, 204 vs. 130 minutes). Pelvic and paraaortic nodes were removed in 92% of patients undergoing laparoscopy and 96% of patients undergoing laparotomy, and cytology was performed in 96% versus 98%. Neither treatment arm demonstrated an improved ability to detect metastatic disease. Quality-of-life evaluation found a better body image and return to normal activities for the patients undergoing laparoscopy. Laparoscopic surgical staging is feasible and safe for patients with uterine cancer and results in fewer complications and shorter hospital stays. Recurrence rates were lower than anticipated; the estimated 3-year recurrence rate was 11% with laparoscopy and 10% with laparotomy, and the estimated 5-year OS rate was almost identical in both arms (90%).

Today, advancements of laparoscopic surgery, including use of a robotic surgical platform, have permitted minimally invasive approaches with a reduction in conversion rates to open laparotomy. Robotic surgery offers three-dimensional graphics, visual magnification, improved ergonomics, masking of operating physician tremor, and greater dexterity of surgical instruments. Robotic surgery has expanded the application of laparoscopy, especially in obese patients or patients in whom a vaginal component is more difficult. Although the majority of studies comparing conventional laparoscopy to robotic assisted laparoscopic surgery in the endometrial cancer space have been retrospective, small randomized prospective trials have been performed. Multiple retrospective trials suggest similar oncologic outcomes. Mäenpää and colleagues randomly assigned 101 endometrial cancer patients to conventional laparoscopy and robotic assisted laparoscopic surgery. The robotic approach was associated with a significant reduction in operative time, total time spent in the operative theater, and in conversions to laparotomy (Maenpaa, et al., 2016). Certainly, cost considerations must be accounted for given the large upfront capital costs associated with acquisition of robotic surgical equipment.

Increasing data suggest that even in grade I disease, as noted on endometrial biopsy, a significant number of patients have full surgical staging findings that would have an impact on further therapy. Ben-Shachar and colleagues, in evaluating 181 grade I endometrial cancers, found that 19% had grade change on hysterectomy specimen, 11% had extrauterine disease, 4% had lymph node metastasis, and 26% on final evaluation had high-risk intrauterine factors. Of note, the authors believed because of full surgical staging that 12% needed and received adjuvant therapy, and 17% who may have received postoperative treatment did not based on full surgical findings. Geisler and associates found in 349 patients that of those with grade I lesions, 16% had positive nodes and 3% had positive paraaortic nodes only. Of all positive nodes, 31% occurred in grade I lesions. As a result of these and other studies, it is advocated that all patients with endometrial cancer should have the benefit of surgical staging, which includes peritoneal cytology, TAH, and BSO, with SLN dissection. Obviously, complete evaluation of the entire peritoneal cavity and its contents should be performed, and any suspicious areas should be pathologically evaluated.

Radiation therapy

Radiation therapy remains a cornerstone of adjuvant therapy for endometrial cancer and has evolved considerably over time. Historically, nearly all patients received some form of preoperative or postoperative radiation therapy given the perception that risk for recurrence was high. Whereas GOG 33 defined relationships between depth of invasion and tumor grade and extrauterine spread after surgical staging, it also showed that most patients were not at risk for extrauterine disease spread. As such, the routine use of pelvic radiation therapy became increasingly questioned. To date, four large prospective randomized studies have compared external beam pelvic radiation with observation or vaginal brachytherapy in women with endometrial cancer ( Table 5.17 ). Aalders and colleagues compared VCB with or without pelvic radiation in 540 patients with early-stage endometrial cancer. Patients did not undergo lymph node dissection. The addition of pelvic radiation therapy reduced local failures (2% vs. 7%), but there was no difference in 5-year survival between groups (89% pelvic, 91% brachytherapy). Onsrud et al. updated the long-term follow-up for these patients. After a median 20.5 years of follow-up, no statistically significant difference was revealed in OS between treatment groups. However, women younger than age 60 years had significantly higher mortality rates after external-beam radiotherapy (EBRT) (HR, 1.36; 95% CI, 1.06 to 1.76) than the control group. The risk of secondary cancer increased after EBRT, especially in women younger than age 60 years (HR, 2.02; 95% CI, 1.30 to 3.15). The authors concluded that there was no survival benefit of external pelvic radiation in early-stage endometrial carcinoma. In women younger than age 60 years, pelvic radiation decreased survival and increased the risk of secondary cancer. Thus, adjuvant EBRT should be used with caution, especially in women with long life expectancies.

In the Dutch PORTEC trial, Creutzberg et al. enrolled 714 patients with grade I lesions with greater than 50% myometrial invasion, grade II lesions with any amount of invasion, or grade III with less than 50% invasion. Patients were randomly assigned postoperatively to receive either external radiation or observation. None of the patients were surgically staged, and all histologies were eligible. In the 654 eligible for follow-up, local and regional recurrences were less frequent in the radiation therapy group (4% vs. 14%) than in the observational group, but the 5-year survival rates were similar, 81% versus 85%, respectively. As published by Creutzberg et al., after 15 years, there remained no survival advantage and a trend for long-term risk of second cancers.

The GOG performed a phase III trial, GOG protocol 99, of surgery with or without adjuvant external beam pelvic radiation in intermediate-risk endometrial ACs. These included all women with any degree of myometrial invasion, any grade, and no evidence of lymph node metastasis (stage IB, IC, IIA occult, and IIB occult). All patients were required to have surgical staging with histologic evaluation of the lymph nodes. A total of 202 patients received no radiation (no adjuvant therapy [NAT]), and 190 received pelvic radiation. The median follow-up period was 69 months. The reported recurrence rate was 15% in the NAT group compared with 6% to 8% in the radiation group ( P = .007). Local recurrences were 9% versus 2%, respectively. The OS at 48 months was 86% for NAT and 92% for radiation, with intercurrent diseases accounting for half or more of the deaths in both groups. Deaths from disease were 8% in both the NAT versus radiation group. Grade III and IV toxicity by treatment was 5% versus 14% in the NAT and radiation, groups, respectively. Twelve of the 13 women who had isolated vaginal recurrences, and were initially randomized to the NAT arm were treated with radiation, and five have died of disease.

Importantly, GOG 99 allowed for the identification of a H-IR subgroup, in a post-hoc manner, that appeared to derive the greatest benefit from pelvic radiation. In the H-IR cohort, the recurrence rate was 27% in the observation arm, versus 13% after pelvic radiation. All other eligible patients were considered low–intermediate risk and had recurrence rates less than 1% to 3% regardless of the use of radiation therapy. H-IR was defined using a combination of age, grade, presence of LVSI, and depth of invasion.

Data from a pooled trial of the Medical Research Council (MRC), ASTEC and National Cancer Institute of Canada Clinical Trials Group Endometrial Cancer 5 (NCIC CTG (EN.5) randomized trials also showed no survival advantage for adjuvant EBRT in the treatment of endometrial cancer (5-year survival rate ~84% both groups). In this study including 906 patients randomly assigned to pelvic radiation or observation, 30% of patients underwent lymphadenectomy, and 50% received vaginal brachytherapy. Pelvic radiation reduced local failures (6% with observation vs. 3% after pelvic radiation). In a subsequent Cochrane review on the use of adjuvant radiation therapy for stage I endometrial cancer, the authors concluded that there is no survival advantage with routine use of pelvic radiation therapy and that benefits must be balanced against the toxicity and morbidity of therapy.

Essentially all studies evaluating pelvic radiation therapy show improved local control rates compared to observation. The local control, however, is largely a result of reduction in vaginal cuff recurrences. A considerable amount of data has been collected to evaluate vaginal recurrence and survival rate with surgery alone or combined therapy when mainly preoperative application of brachytherapy and surgery were used. Data have also been evaluated concerning the grade of the tumor ( Table 5.18 ) and, in some instances, the depth of myometrial involvement ( Table 5.19 ). In patients who had postoperative radiation, there appeared to be a lower incidence of vaginal vault recurrences, although there does not appear to be much difference in the grade I and grade II lesions. Vaginal vault recurrence did not appear to affect survival.

Results of a randomized prospective trial comparing adjuvant EBRT versus vaginal brachytherapy in 427 patients with intermediate-risk early-stage endometrial cancer have been reported (PORTEC-2). No differences in recurrence rate (vaginal failure 2% with pelvic radiation, 0.9% with vaginal brachytherapy) or OS (3-year survival rate ~90% both groups) were observed between the two treatment arms, suggesting that patients with intermediate-risk early-stage endometrial cancer can be treated with adjuvant vaginal brachytherapy alone. Although these findings have led many clinicians in the United States to move away from the routine use of EBRT in early-stage endometrial cancer, it is important to note that in the patients who did not undergo a lymph node dissection, pelvic failure rates were higher in the vaginal brachytherapy group (3.5% vs. 0.7%), and the study excluded higher-risk patients such as those with grade III or deeply invasive tumors. An additional study comparing pelvic radiation followed by VCB to VCB alone for “medium-risk” endometrial cancer by Sorbe et al. showed results similar to PORTEC-2, demonstrating that pelvic radiation reduced locoregional relapse rates (1% vs. 5%, P = .013) with no difference in OS. Sorbe et al. also reported a randomized trial of vaginal brachytherapy versus no further treatment in patients with low-risk endometrial cancer (grade I or II endometrioid cancer with <50% invasion). There was no significant reduction in vaginal recurrence in the group receiving brachytherapy (3% vs. 1%).

GOG protocol 249 was designed to determine whether adjuvant chemotherapy could improve cancer outcomes in patients with H-IR endometrial cancer ( ). This was based on the observation that both PORTEC-1 and GOG protocol 99 reported distant failure rates of 20% to 30% in high-risk patients. Eligible patients had International Federation of Gynecology and Obstetrics (2009) stage I endometrioid histology with H-IR criteria, stage II disease, or stage I to II serous or clear cell tumors. A total of 601 patients were randomly assigned between RT (45 to 50.4 Gy over 5 weeks) or VCB followed by intravenous paclitaxel 175 mg/m ² (3 hours) plus carboplatin (area under the curve [AUC], 6) every 21 days for three cycles ( ). With a median follow-up of 53 months, the 60-month RFS was 0.76 for RT and 0.76 for brachytherapy followed by chemotherapy (HR 0.92; 90% CI 0.69 to 1.23). Furthermore, no difference in 60-month OS was identified between study arms (HR 1.04; 90% CI 0.71 to 1.52). Pelvic or paraaortic nodal recurrences were more common in the chemotherapy arm, 9% versus 4%. Ironically, patients with serous cancer accounted for 15% of patients enrolled on GOG 249, and although the expectation was that chemotherapy may benefit patients with this high-risk histology, there was no significant difference in PFS or OS when chemotherapy was combined with VCB.

Two additional pivotal clinical trials looked to better define the utility of radiation and chemotherapy in the management of locally advanced endometrial cancer. GOG protocol 258 was developed in an effort to address local recurrences, reported to occur in up to 20% of patients treated on the chemotherapy arm of GOG 122 ( ). The experimental arm was supported by prior studies, including RTOG-9708, which examined adjuvant cisplatin 50 mg/m ² intravenously on days 1 and 29 plus volume-directed radiation therapy (45 Gray ± brachytherapy) followed by paclitaxel 175 mg/m ² and carboplatin AUC 5 every 21 days for four cycles. GOG 258 enrolled a total of 813 patients with stage III or IVa endometrioid endometrial cancer, as well as stage I or II serous or clear cell carcinoma with positive washings. Patients were randomized to combined chemoradiation which consisted of cisplatin 50 mg/m ² on days 1 and 29 together with volume-directed external beam radiation therapy, followed by carboplatin AUC of 5 to 6 plus paclitaxel at a dose of 175 mg/m ² every 21 days for four cycles, with GCSF support or a chemotherapy-only regimen consisting of carboplatin AUC of 6 plus paclitaxel at a dose of 175 mg/m ² every 21 days for six cycles. In the chemoradiotherapy group, external beam radiation therapy was delivered to the pelvis with or without paraaortic fields. The planned total dose was 45 Gy in 25 fractions at 180 cGy per fraction. Intensity modulated radiotherapy and vaginal brachytherapy were allowed only in the chemoradiotherapy group ( ). Nearly all enrolled patients (98%) had no gross residual disease, and 94% underwent some degree of lymph node assessment. At 60 months, the Kaplan–Meier estimate of the percentage of patients alive and relapse free was 59% (95% CI, 53 to 65) in the chemoradiotherapy group and 58% (95% CI, 53 to 64) in the chemotherapy only group (HR 0.90; 90% CI, 0.74 to 1.10). Chemoradiotherapy was associated with a lower 5-year incidence of vaginal recurrence (2% vs. 7%; HR 0.36; 95% CI, 0.16 to 0.82) and pelvic and paraaortic lymph node recurrence (11% vs. 20%; HR 0.43; 95% CI, 0.28 to 0.66) than chemotherapy alone, but distant recurrence was more common in association with chemoradiotherapy (27% vs. 21%; HR 1.36; 95% CI, 1.00 to 1.86). Furthermore, assessment of quality-of-life endpoints suggested both acute and long-term toxic effects of combination chemoradiation therapy, suggesting that chemotherapy alone was the most appropriate adjuvant regimen in this patient population.

In an analogous manner, PORTEC-3 was designed to compare adjuvant chemotherapy during and after radiotherapy (chemoradiotherapy) versus pelvic radiation alone in women with high-risk endometrial cancer. Eligible women had stage I, grade III endometrioid carcinoma with deep myometrial invasion or lymph-vascular space invasion (or both), stage II or III endometrioid AC, or stage I to III with serous or clear cell histology. Nearly half of the women enrolled on this trial had stage I or II disease. Patients were randomly assigned (1:1) to receive radiotherapy alone (48.6 Gy in 1.8 Gy fractions given on 5 days per week) or radiotherapy and chemotherapy (consisting of two cycles of cisplatin 50 mg/m ² given during radiotherapy, followed by four cycles of carboplatin AUC5 and paclitaxel 175 mg/m ² ) ( ). A total of 686 women were enrolled, with investigators showing an improvement in 5-year failure free survival in the chemoradiotherapy arm (HR 0.71, 95% CI 0.53 to 0.95; P = .022), but no significant improvement in OS (HR 0.76, 95% CI 0.54 to 1.06; P = .11) ( ). Grade 3 or worse adverse events were significantly more common in the chemoradiotherapy arm. When specifically examining patients with stage III disease, the investigators showed a more pronounced FFS advantage (HR 0.66, 95% CI 0.45 to 0.97) with a non-significant trend toward improved OS (HR 0.71, 95% CI 0.45 to 1.11).

An updated post-hoc survival analysis of PORTEC-3, completed at a median follow-up of nearly 73 months suggested both a significant OS and failure-free survival advantage with chemoradiotherapy ( ). This advantage appeared to be most pronounced in patients with stage III disease, as well as those with serous histology. Taken together, the results of GOG 258 as well as PORTEC-3 highlight the importance of chemotherapy in the management of patients with stage III endometrial carcinoma. Conversely, the results of GOG 249 and PORTEC-3 suggest that the addition of chemotherapy to radiation in patients with stage I or II disease may confer limited benefit with additional treatment-related toxicity.