Introduction
Many cases of endometrial cancer are detected at an early stage. Of all detected cases, 71.9% are diagnosed at stage I and 6.0% at stage II. Advanced cancers with extrauterine lesions are rare, with only 13.3% diagnosed at stage III and 7.5% at stage IV [
1]. The standard treatments for endometrial cancer include total hysterectomy, bilateral salpingooophorectomy, and regional lymphadenectomy. In Japan, regional lymphadenectomy is performed in approximately 60-70% of patients with endometrial cancer [
2]; of these, only a few cases have lymph node (LN) metastasis. Regional LN metastasis is one of the most important prognostic factors in endometrial cancer. If a LN is positive for metastasis, it is diagnosed as stage IIIc or higher, making the diagnostic significance of LN dissection abundantly clear. However, given that many cases show negative LN metastases, there is still a debate on the benefits of its therapeutic significance. Despite the existence of previous reports, the significance of regional LN dissection remains controversial, and there is no consensus within the medical community on its relevance.
The Japan Society of Obstetrics and Gynecology (JSOG) has a gynecological cancer registry (GCR) that records clinicopathological factors, treatment methods, and prognostic outcomes for several endometrial cancer cases in Japan [
3]. Here, we analyzed the clinicopathological factors and treatment outcomes using data from the GCR of JSOG to examine the impact of regional LN dissection on endometrial cancer.
Discussion
In this retrospective study of a relatively large sample size, we analyzed the profiles of 34,575 patients. The distribution of surgical stage and histological type of the subjects was similar to that of the general population, based on the cancer registry data. Unlike most Western countries, the majority of the postoperative treatments of the subjects consisted of chemotherapy, rather than radiation therapy, which is one of the hallmarks of this study.
In this study, the group that underwent regional LN dissection had a better prognosis than the non-dissected group. When surgical staging was taken into consideration, the LN dissected group had a better prognosis than the non-dissected group in all surgical stages, except for stage Ia. Considering the surgical stage and histological type, the LN dissected group had a better prognosis than the non-dissected group at all surgical stages, except for stage Ia G1 and stage Ia G2. Therefore, regional LN dissection may improve the prognosis of endometrial cancers, except for patients at stages Ia G1 and Ia G2. The prognosis was significantly better in the LN dissected group without AC than in the non-dissected group with AC in high-risk cases. The multivariate analysis also showed that the HR for LN dissection was much lower than that for AC, suggesting that the impact of LN dissection on prognosis was greater than that of AC. Based on this finding, we do not recommend inadequate replacement of LN dissection with AC. There may be several reasons why the OS rate did not improve when chemotherapy was administered instead of lymphadenectomy. The possible reasons are as follows: 1) institutions that could not perform lymphadenectomy may have provided poor quality of surgery, and chemotherapy alone did not improve the prognosis sufficiently; 2) lymphadenectomy was not performed, making accurate staging impossible and resulting in possible underestimation of patients with stage IIIc or higher; and 3) there were patients with poor general condition who could not undergo lymphadenectomy owing to complications.
There are several studies on the therapeutic significance of regional LN dissection for endometrial cancer. Retrospective studies have reported that LN dissection significantly improved the prognosis of patients with stage I endometrioid carcinoma G3, invasion of more than half of the myometrium, or stage II endometrioid carcinoma [
5,
6]. The ASTEC trial, a prospective study that investigated the significance of pelvic lymphadenectomy (PLN), compared 686 patients with LN dissection and 683 who did not undergo dissection; the 5-year OS rate of the patients was 80% and 81%, respectively, with no significant difference between the 2 groups [
7]. In this study, adjuvant radiotherapy was performed in fewer than 10% of the patients in the low-risk group and about half of the patients in the intermediate- and high-risk groups. Another prospective Italian study compared 264 LN dissection groups and 250 non-dissection groups; the researchers found a 5-year OS rate of 85.9% and 90.0%, respectively, with no significant between-group differences [
8]. However, these trials included a large number of low-risk patients with LN metastasis, and the median number of dissected LNs in the ASTEC trial, which was 12, might have been too small. Furthermore, the SEPAL study, a retrospective cohort study conducted in Japan, compared 325 patients who underwent PLN alone with 346 patients who underwent PLN along with para-aortic lymphadenectomy (PAN). There was no significant difference in the recurrence-free survival (RFS) or OS rate in the low-risk group. However, in the intermediate- and high-risk groups, the 5-year RFS rate was 64.8% and 80.7%, and the 5-year OS rate was 72.6% and 83.2%, in the groups that underwent PLN alone and PLN+PAN, respectively. The RFS and OS rates were, therefore, significantly better in the PLN+PAN group than in the PLN alone group [
9]. Multivariate analysis revealed LN dissection to be an independent prognostic factor, with an effect on improving prognosis only when patients with a high-risk of LN metastasis were assessed.
The results of our study support the findings of the SEPAL study. Because chemotherapy is used in most cases in Japan as an adjuvant treatment [
1,
2,
10], the results may differ from those reported in Europe and the United States, where adjuvant radiation therapy is administered [
11,
12]. The control of pelvic lesions by adjuvant radiation therapy was replaced by regional LN dissection, and further control of systemic lesions by chemotherapy may lead to a prolonged prognosis [
13].
The strengths of this study are that it included a large number of cases and that most institutions in Japan were included in the registry of all the cases, thereby reducing the bias among institutions or cases. Conversely, the limitations are that this study is retrospective, the data required for detailed analysis were insufficient, the prognostic data included only OS, and the following biases could not be excluded. The first bias is that the non-dissected group was likely to include cases with poor prognosis for other reasons. Potential high-risk patients, such as those with severe complications and the elderly, tended to undergo no systematic lymphad-enectomy, which may have contributed to the shortened OS in the non-dissected groups. Although patients in the non-dissected group were significantly older, the median age was almost the same (58 and 59 years old in the dissected and non-dissected groups, respectively), and a difference with a significant impact on OS was unlikely.
The second bias could be latent LN metastases in the non-dissected group. LN metastases may be missed in stage I to stage IIIb patients who have not undergone systematic LN dissection. Such false-negative cases may contribute to the shortened OS in the non-dissected group. However, in Japan, many patients undergo magnetic resonance imaging and computed tomography before surgery, and it is unlikely that the apparent LN metastasis was overlooked during imaging. The third bias is that the enrollment data lack qualitative assurance of the procedure of hysterectomy and LN dissection. Because the details of hysterectomy and definition of LN dissection have not been determined, LN biopsy may have been registered as LN dissection, and the effects of such discrepancies on OS cannot be denied. However, these factors might reduce any improvements in the prognosis of LN dissection cases, and a strict registration of surgical procedures can lead to further differences.
In consideration of the above points, this study suggests that LN dissection may have a prognostic effect on endometrial cancer. The elimination of as much bias as possible is essential for an accurate evaluation of therapeutic outcomes, and a phase 3 randomized clinical trial is deemed necessary.