Reprint request to: Dr Yukiharu Todo, Division of Gynecologic Oncology, National Hospital Organization, Hokkaido Cancer Center, 4-2 Kikusui, Shiroishi-ku, Sapporo 003-0804, Japan. Email: firstname.lastname@example.org
Endometrial cancer is the most common malignancy of the female genital tract in the USA, with an estimated number of 47 130 new cases in 2012. The annual number of deaths has increased from 6000 in 1997 to 8010 in 2012. Approximately 17% of patients with endometrial cancer eventually relapse and die of this disease. From this viewpoint, improvements are needed in the treatment for endometrial cancer, especially patients at high risk of a poor prognosis. On the other hand, the International Federation of Gynecology and Obstetrics (FIGO) annual report demonstrated that the survival rates of endometrial cancer have continued to increase during recent decades. This trend applies to all cases, including stage IIIC. Meanwhile, surgery has been playing the leading role in treatment of endometrial cancer, and there has been no paradigm shift except for the introduction of lymphadenectomy. Whether many patients with endometrial cancer can benefit from lymphadenectomy must be determined. Two recent randomized controlled trials showed negative effects of lymphadenectomy on prognosis,[4, 5] and many gynecologists have since declared at conferences that standard surgery for endometrial cancer does not include lymphadenectomy. However, such a declaration is an overgeneralization of the results of the randomized studies. In the present manuscript, we tried to interpret the results of these two randomized controlled trials properly, and discussed pitfalls of randomized controlled trials in surgical intervention. A surgical field to be treated is proposed in a group of patients who require lymphadenectomy. In addition, a new strategy for preventing leg edema, the most frequent complication after lymphadenectomy, is introduced.
Therapeutic Role of Lymphadenectomy: Previous Discussion
Before 2008, all studies regarding the role of lymphadenectomy in endometrial cancer were retrospective. Some studies supported a survival benefit of lymphadenectomy,[6-13] and others did not.[14-17] Kilgore et al. showed a significant survival advantage for patients who underwent multiple-node sampling compared with patients who did not undergo node sampling regardless of risk of prognosis. The mean number of nodes removed in the multiple-node sampling group was 11. Trimble et al. showed a survival benefit of lymph node sampling for patients with stage I, grade 3 disease based on data of more than 9000 women. Fanning et al. showed a potential survival benefit of lymphadenectomy for patients with intermediate-risk endometrial cancer. Cragun et al. showed a significant survival advantage for patients who had 11 or more lymph nodes removed compared with patients who did not. Lutman et al. concluded that removal of 12 or more pelvic nodes is important to improve survival of patients with stage I–II and high-risk histologic disease. Chan et al. reported that lymphadenectomy showed a survival benefit for patients with stage I, grade 3 as well as stage II–III endometrioid endometrial cancer from US nationwide cancer data on more than 39 000 cases. That study suggested that lymphadenectomy has no survival benefit for patients with stage I, grade 1–2 endometrial cancer. Abu-Rustum et al. concluded that removal of more than 10 regional lymph nodes is associated with improved survival of older patients with lower-stage disease and no adjuvant therapy or brachytherapy only. Regional lymph nodes in endometrial cancer are usually categorized into two sections: pelvic nodes and para-aortic nodes (Fig. 1). However, para-aortic lymphadenectomy has not been well valued. Mariani et al. focused on para-aortic nodes and showed a potential survival benefit of para-aortic lymphadenectomy in endometrial cancer. They reported that removal of five or more para-aortic nodes is associated with improved survival of high-risk patients.
In 2008, Benedetti-Panici et al. conducted a randomized controlled trial to assess the therapeutic effect of lymphadenectomy for the first time. In 2009, the results of the second randomized controlled trial to assess the therapeutic effect of lymphadenectomy were reported by A Study in the Treatment of Endometrial Cancer (ASTEC). Because both randomized controlled trials showed negative effects of lymphadenectomy on prognosis, many gynecologists have declared at conferences that standard surgery for endometrial cancer does not include lymphadenectomy. However, there have been many arguments on the design of the studies. In both trials, para-aortic lymphadenectomy was not valued, and only a small number of cases underwent para-aortic lymphadenectomy (Benedetti-Panici's study, 26%; ASTEC trial, not available). Therefore, the median number of lymph nodes harvested was less than that in a Japanese study (Benedetti-Panici's study, 30; ASTEC trial, 14; Japanese study, 82). Considering this limitation, appropriate interpretation of both trials would be as follows. Pelvic lymphadenectomy does not have therapeutic significance for low-risk clinical stage I endometrial cancer. Because clinical stage I includes not only low-risk patients but also intermediate- and high-risk patients, a conclusion that pelvic lymphadenectomy does not have therapeutic significance for clinical stage I endometrial cancer will cause some patients at high risk of lymph node metastasis to miss an opportunity to undergo lymphadenectomy. If lymphadenectomy has a survival benefit for intermediate- and high-risk patients, they would not be able to receive optimal treatment.
Therapeutic Role of Lymphadenectomy: Significance of Para-aortic Lymphadenectomy
In 2010, the author and colleagues presented the results of the survival effect of para-aortic lymphadenectomy (SEPAL) study in The Lancet. The study showed no survival benefit of combined pelvic and para-aortic lymphadenectomy over pelvic lymphadenectomy alone for low-risk patients. However, a significant survival benefit of para-aortic lymphadenectomy combined with pelvic lymphadenectomy was confirmed for intermediate- and high-risk patients. In that study, the number of lymph nodes harvested was 36 in the pelvic lymphadenectomy group and 82 in the combined pelvic and para-aortic lymphadenectomy group. Surgery involving para-aortic lymphadenectomy was associated with a decreased incidence of death in patients with a hazard ratio of 0.44 (95% CI = 0.30 − 0.64, P < 0.0001) compared with surgery without para-aortic lymphadenectomy. Combined pelvic and para-aortic lymphadenectomy showed a 10.6% increase in the 5-year overall survival compared with pelvic lymphadenectomy alone. Considering previous results of retrospective and prospective studies, standard primary surgery for low-risk endometrial cancer would include hysterectomy and bilateral salpingo-oophorectomy, but would not include lymphadenectomy. On the other hand, the therapeutic role of lymphadenectomy should be continuously assessed in intermediate- and high-risk endometrial cancer because combined pelvic and para-aortic lymphadenectomy might have a survival benefit. The 5-year survival rates were reportedly 79–85% for patients with stage III endometrial cancer who had been treated at tertiary hospitals in which para-aortic lymphadenectomy was routinely performed,[13, 19] but it was 65% for patients with stage III endometrial cancer and no gross swelling of para-aortic nodes who had been treated at tertiary hospitals in which para-aortic lymphadenectomy was not performed.
Many physicians regard breast cancer as a systemic disease and lymph node metastasis as an indicator that distant metastasis has already occurred. With this theory, removal of lymph nodes is not important to survival. In fact, lymphadenectomy for patients with breast cancer and lymph node metastasis is currently performed less commonly than in the past. However, it is not that simple in endometrial cancer.
The FIGO staging system was recently changed. In endometrial cancer, FIGO stage IIIC is now divided into two categories: IIIC1 and IIIC2. Para-aortic lymph node involvement is categorized as a single substage: IIIC2. The reason for this change is that previous data suggested a worse prognosis if para-aortic nodes are involved. However, the prognosis seems to depend on treatment as well as the extent of disease. In 2000, Mariani et al. showed that the 5-year overall survival for patients who had undergone para-aortic lymphadenectomy was 85% compared with 71% for patients who had not undergone para-aortic lymphadenectomy (P = 0.06) according to the data of 51 patients with lymph node metastasis. In 2007, Fujimoto et al. reported that the 5-year disease-related survival for patients who had undergone para-aortic lymphadenectomy was 69% compared with 54% for patients who had not undergone para-aortic lymphadenectomy (P = 0.19) according to the data of 63 patients with lymph node metastasis. In 2011, the author and colleagues also reported that the 5-year overall survival for 28 patients who underwent combined pelvic and para-aortic lymphadenectomy and were positive for pelvic node metastasis and negative for para-aortic node metastasis was 89% compared with 47% for 37 patients who underwent pelvic lymphadenectomy alone and were positive for lymph node metastasis, and 60% for 28 patients who underwent combined pelvic and para-aortic lymphadenectomy and were positive for para-aortic node metastasis. All of these studies suggest that lymph node metastasis is associated with a wide spectrum of prognoses for endometrial cancer and that endometrial cancer can be cured by appropriate removal of regional lymph nodes even if some lymph nodes are already affected. Although the systemic theory has a larger number of supporters in the field of breast cancer, gynecologists should not apply the systemic theory to endometrial cancer.
Differences in failure patterns between surgical procedures with and without para-aortic node dissection may support the therapeutic efficacy of para-aortic lymphadenectomy. In terms of failure sites in the SEPAL study, there was a higher rate of para-aortic node failure in the pelvic lymphadenectomy alone group than in the pelvic and para-aortic lymphadenectomy group (5.1% vs 0.6%, P = 0.0004). Para-aortic node recurrence was a failure pattern peculiar to the pelvic lymphadenectomy alone group. In addition, when only those patients who had received adjuvant chemotherapy were considered, there was still a higher rate of para-aortic node failure in the pelvic lymphadenectomy alone group than in the pelvic and para-aortic lymphadenectomy group (9.5% vs 1.3%, P = 0.0036). Adjuvant chemotherapy (cisplatin + adriamycin + cyclophosphamide/paclitaxel + carboplatin/docetaxel + carboplatin) might not be able to replace surgical removal as a treatment for metastatic lymph nodes.
Pitfalls of Randomized Surgical Trials
Although the SEPAL study supported the therapeutic efficacy of para-aortic lymphadenectomy in endometrial cancer, it was a retrospective cohort study. Thus, another prospective study is needed to validate the therapeutic effect of para-aortic lymphadenectomy. However, should such a study be a randomized controlled study? Some physicians would not participate in a randomized controlled trial in which pelvic lymphadenectomy versus combined pelvic and para-aortic lymphadenectomy is compared for patients with high-risk endometrial cancer because they think highly of the effectiveness of para-aortic lymphadenectomy and would feel uneasy about performing pelvic lymphadenectomy alone. Experienced gynecologic oncologists tend to be familiar with para-aortic lymphadenectomy and its benefits and may decline participation in such a randomized controlled trial. Conversely, doctors with limited experience may be assigned the task of performing para-aortic lymphadenectomy, and the desired outcome may not be achieved because of the doctors' inadequate experience. Both scenarios give rise to a situation in which quality control of treatment might be reduced in the para-aortic lymphadenectomy group. Nonparticipation of such physicians would result in a selection bias in randomized studies, and conclusions based on the results of such studies would not be reliable. A randomized controlled study is, therefore, not the best format with which to demonstrate the full benefits of lymphadenectomy. A prospective cohort study may be a reasonable option for the trial and may even be the most appropriate method with which to assess the therapeutic significance of lymphadenectomy for high-risk patients.
Extent of Para-aortic Lymphadenectomy
Although the addition of para-aortic lymphadenectomy to pelvic lymphadenectomy might have a survival benefit in intermediate- and high-risk endometrial cancer, it is not actually quite that simple. Para-aortic lymph nodes are usually categorized into the areas above and below the inferior mesenteric artery (Fig. 1). Some sentinel node mapping studies in endometrial cancer showed that more than half of para-aortic nodes identified as sentinel were located above the inferior mesenteric artery.[26, 27] In 2008, Mariani et al. showed that 77% of patients with para-aortic lymph node metastasis harbor disease above the inferior mesenteric artery. In 2010, Fotopoulou et al. showed that 54% of patients with stage IIIC and 70% of patients with para-aortic node metastasis harbor disease above the inferior mesenteric artery. They indicated the need for systematic lymphadenectomy including the pelvic and para-aortic area up to the renal vessels. In the SEPAL study, almost all patients in the pelvic and para-aortic lymphadenectomy group underwent removal of the section above the inferior mesenteric artery. Combined pelvic and para-aortic lymphadenectomy without removal of the area between the inferior mesenteric artery and renal veins might be insufficient to improve survival for patients with high-risk endometrial cancer. In a future study, to assess the therapeutic role of lymphadenectomy, information on the surgical field should be obtained and reported. Like Piver's classification of hysterectomy in cervical cancer, it is desirable that worldwide classification is defined.
Complications with or without Para-aortic Lymphadenectomy
Although selective lymphadenectomy does not increase morbidity,[30, 31] systematic lymphadenectomy requires a longer operation time and increases postoperative complications. In Benedetti-Panici's trial, both early and late postoperative complications occurred significantly more frequently in patients who had undergone pelvic lymphadenectomy (31% in the lymphadenectomy arm and 14% in the no-lymphadenectomy arm, P = 0.0001). Similarly, in the ASTEC trial, more patients in the lymphadenectomy arm developed postoperative complications than in the no-lymphadenectomy arm. Kodama et al. reported that the most frequent complication was lymphedema and that the second most frequent complication was ileus. Considering these facts, systematic lymphadenectomy should not be performed for patients who have little possibility of receiving benefits from lymph node dissection. We then need to consider what kind of surgery should be performed for patients other than those at low risk. The SEPAL study suggested that combined pelvic and para-aortic lymphadenectomy has a survival benefit for intermediate- and high-risk patients. This leads to the question of whether para-aortic lymphadenectomy itself increases perioperative complications, which has not been well discussed. The author and colleagues conducted a retrospective analysis to compare morbidity rates with and without para-aortic lymphadenectomy. The incidence rate of leg edema that originates in systematic lymphadenectomy is reportedly about 25%.[34, 35] Our study also showed that the most frequent complication was leg edema and that its incidence rate was 26%. In addition, that study showed that para-aortic lymphadenectomy did not increase leg edema. The incidence rate of ileus that originates in systematic pelvic and para-aortic lymphadenectomy is reportedly 8–50% in patients with a malignant gynecologic tumor.[36-38] Fujita et al. reported that postoperative ileus occurred in 13% of patients who underwent para-aortic lymphadenectomy, but severe ileus occurred in only 1.4% of patients who underwent para-aortic lymphadenectomy. They concluded that para-aortic lymphadenectomy is a feasible and safe operative procedure. Fagotti et al. conducted a prospective study to elucidate the incidence of ileus in patients who had undergone para-aortic lymphadenectomy. They reported that postoperative ileus occurred in 50% of patients who had undergone para-aortic lymphadenectomy and that severe ileus occurred in 11% of patients who had undergone para-aortic lymphadenectomy. However, all patients with severe ileus recovered with only conservative management. The author and colleagues showed that the incidence rate of ileus after combined pelvic and para-aortic lymphadenectomy was 12%, but severe ileus occurred in only 1.4% of patients. That study showed a significant increase in postoperative ileus secondary to the addition of para-aortic lymphadenectomy, but no significant difference in severe ileus with or without para-aortic lymphadenectomy (1.4% vs 0.7%, P = 0.58). Based on the results of these studies, the author emphasizes that postoperative ileus after para-aortic lymphadenectomy occurs with a relatively high incidence rate, but is manageable.
In addition, the author and colleagues showed that para-aortic lymphadenectomy did not increase postoperative thrombosis, lymphocyst development, or intraoperative organ injury and concluded that para-aortic lymphadenectomy is a safe operative procedure when performed by experienced surgeons in tertiary centers.
New Strategy for Preventing Leg Edema
As described in the previous chapter, leg edema is the most frequent complication after lymphadenectomy.[32, 33] Postoperative leg edema is a serious complication and a chronic disease that lasts a lifetime in most cases. Sentinel lymph node navigation surgery is a promising treatment for lowering the frequency of postoperative leg edema in patients with cervical carcinoma of the uterus. However, it is unknown whether sentinel lymph node navigation surgery can be applied in endometrial cancer. According to the results of previous studies, adjuvant radiation therapy and number of resected lymph nodes are risk factors for leg edema.[34, 35, 41] A new strategy for decreasing the incidence of postoperative leg edema has been proposed. Abu-Rustum suggested that removal of circumflex iliac nodes distal to the external iliac nodes (CINDEIN) is likely to be a factor contributing to the risk of postoperative leg edema. CINDEIN, which are the most distal external iliac lymph nodes (Fig. 1), have been called circumflex iliac nodes, distal external iliac lymph nodes, and suprainguinal nodes[27, 44] in the literature, and they commonly comprise an enlarged lobular conglomerate of lymphatic channel-rich nodal and adipose tissue. There are three routes of lymphatic spread in endometrial cancer. The first route is from the fundus toward the adnexa and infundibulopelvic ligaments to the para-aortic nodes. The second route is from the lower and middle thirds of the uterus in the base of the broad ligaments toward the lateral pelvic wall. The third route is along the round ligaments to the CINDEIN. However, this third route might be a minor site in terms of metastasis. Matsumoto et al. reported that CINDEIN metastasis occurred the least often among all metastatic sites in endometrial cancer. Niikura et al. reported that CINDEIN can be detected as sentinel lymph nodes in endometrial cancer, but the rate of detection was only 3.6%.
In 2010, the author and colleagues showed that removal of CINDEIN was an independent risk factor for leg edema after lymphadenectomy as well as adjuvant radiation therapy and resection of more than 31 lymph nodes based on data of patients with endometrial cancer. In 2011, we also showed that removal of CINDEIN was associated with an increased incidence of leg edema based on data of patients with cervical cancer. Hareyama et al. showed an efficacy of CINDEIN-sparing lymphadenectomy for reducing postoperative leg edema based on data of more than 300 patients. Although elimination of CINDEIN dissection may be helpful in reducing the incidence of leg edema, CINDEIN are regional lymph nodes of endometrial cancer. The author and colleagues investigated the metastatic incidence of CINDEIN in endometrial cancer. The incidence rate of CINDEIN metastasis was 10% in stage IIIC (node-positive) cases. However, the incidence rate was only 1.7% among all cases of endometrial cancer. Matsumoto et al. reported that these incidences were 14.8% in node-positive cases and 3.8% in all cases. Although the incidence in all cases in their study was slightly higher than that in our study, low-risk cases might not have been included in their study. From the viewpoint of CINDEIN metastasis in endometrial cancer, we do not believe that elimination of CINDEIN dissection can be performed in all patients with endometrial cancer. The author also showed that high-risk histology results (grade 3 endometrioid cancer or non-endometrioid cancer) and pelvic node metastasis were independent risk factors for CINDEIN metastasis. Removal of CINDEIN can be eliminated in patients with G1/G2 endometrial cancer. If possible, CINDEIN should be preserved in patients with endometrial cancer because preservation of CINDEIN might result in a reduction in the incidence of postoperative lower-extremity lymphedema.