Complete Pelvic Lymphadenectomy in Patients with Clinical Early, Grade I and II Endometrioid Corpus Cancer

Discussion

The controversy regarding lymphadenectomy in endometrial cancer has been mainly focused on the management of patients with clinical early disease and favorable histological characteristics. At present, routine lymphadenectomy is not generally considered justifiable for all these patients (4). It has been shown that increasing tumor grade, depth of myometrial invasion and tumor size greater than 2 cm are strong predictors of nodal involvement (3, 13, 14). Therefore, many surgeons use intraoperative pathological assessment to identify patients at risk of nodal metastasis and avoid unnecessary lymphadenectomies. However, the ability of these methods to accurately detect high-risk patients remains limited. Case et al. reported diagnoses of myometrial invasion and tumor grade based on intraoperative frozen section were upstaged in 28% and upgraded in 38% of the cases, respectively, in the final histology (15). Other clinicians have attempted to discriminate high-risk patients using information taken from the preoperative histology and imaging techniques. However, neither have these strategies proved accurate in selecting patients with adequate safety (16, 17). Consequently, surgical staging remains the only precise method for determining lymph node metastasis and several authors suggest that comprehensive staging should be considered for all patients with endometrial cancer (18). Yet, even in centers advocating lymph node assessment, recommendations for lymphadenectomy are inconsistent and clinical practices vary from sampling alone to complete lymphadenectomy (19).

In this study, we explored the risk of pelvic nodal involvement in a prospective cohort of patients with a preoperative diagnosis of clinical early grade I and II endometrioid cancer. All patients underwent complete pelvic lymphadenectomy and the occurrence of nodal metastasis was analyzed in relation to tumor grade and depth of myometrial invasion. The preoperative grade did not correlate with final grade in 32.4% of the cases and was upgraded in 24.1%. The postoperative histology showed that 9 patients (8.3%) had grade III tumors, of whom 3 presented with grade I (4.5%) and 6 (14.3%) with grade II lesions preoperatively. Moreover, 10 patients (9.3%) had tumors of non-endometrioid histology, including 1 undifferentiated and 2 clear cell carcinomas. These findings are in accordance with the data reported in previous studies. Frumovitz et al. found that 23% of endometrioid grade I and II tumors combined were upgraded in the final histology, with 5.9% grade III carcinomas (20). Similarly, in the study by Sanjuan et al., 7.9% of patients presenting with grade I or II disease had grade III tumors in the post-operative histology (21). These data show that about 10% of patients presenting with early disease of favorable histological characteristics harbor aggressive tumors.

In 23 of our patients (21.3%), deep (>50%) myometrial invasion was demonstrated in the final pathological evaluation. Thirteen of these patients had presented with grade I and 10 with grade II tumors. Ben-Shachar et al. also reported that 25% of patients presenting with endometrioid grade I tumors had deep myometrial invasion with or without extrauterine disease (6).

A total of 11 patients (10.2%) had pelvic nodal involvement as documented in the final pathology. Four of them had presented with grade I and seven with grade II tumors. According to these findings, the probability of pelvic lymph node metastasis in patients presenting with clinical early endometrioid grade I or II disease is 6% and 16.7% respectively. Another 10 node-negative patients were finally upstaged due to metastases found in other sites. The overall 19.4% stage migration at surgery in this series is similar to the 21-23% reported in previous studies, in which more cases of non-endometrioid histology and grade III tumors were included (5, 7, 13).

In our modest series, increasing tumor grade and myometrial invasion was associated with higher risk of pelvic nodal involvement. This finding has been well demonstrated in the seminal Gynecologic Oncology Group (GOG) studies (3, 13) and in the recent large study by Chi and co-workers on the incidence of pelvic lymph-node metastasis based on tumor grade and depth of myometrial invasion (7). None of our patients with final grade I tumors invading only the inner half of the myometrium had nodal metastasis, while one out of seven with deeper invasion (14.3%) was node positive. In the GOG study, a lower but comparable incidence of pelvic lymph node metastasis was found (2/18=11%) in patients with grade I tumors invading the outer third of the myometrium (3). However, in the series of Chi et al., no patient with these characteristics had nodal involvement (7). A probable explanation for this difference lies in the small number of patients included in this category (5 in the study of Chi et al. and 7 in ours). Similarly, our data on the risk of nodal metastasis in patients with grade III tumors are based on a very small sample size and should be disregarded.

The occurrence of pelvic nodal involvement in patients with final grade II lesions (overall=15.9%, invasion of inner half=12.1%, invasion of outer half=27.3%) was higher than that reported for this group in previous studies (3, 7, 14). This difference is considerable given the comparable percentage of patients with deep myometrial invasion in this category, included in the present as well as in the previously published series: For tumors invading the outer half, 25% in the present study and 22.5% in the Chi et al. series (7) and for tumors invading the middle and outer third, 24% and 10% respectively in the GOG study (3). Furthermore, Mariani et al. reported no incidence of pelvic nodal involvement in 59 patients with grade I and grade II tumors smaller than 2 cm invading the inner half of the myometrium and suggested simple hysterectomy for these patients in the absence of extrauterine macroscopic disease (14).

The small sample size of this study might be a sufficient explanation for the above differences. However, the extent of lymphadenectomy may also have an impact regarding the increased incidence of nodal involvement. In a recent study of 11,443 patients, Chan et al. examined the probability of detecting at least one positive lymph node in relation to the number of nodes removed. They found that sampling of up to 5 nodes diagnosed only 27.6% of patients with nodal metastasis and the chance of detecting at least one positive node increased with the number of lymph nodes recovered, needing more than 20 nodes to ensure adequate assessment of the nodal status (12). In the GOG series, various types of lymphadenectomy were performed and no minimal number of nodes removed was defined for including patients in the study (3). In the study of Chi et al., patients with at least 8 nodes examined were included and the mean number of lymph nodes recovered was 15 (7). Finally, in the Mariani et al. series, lymphadenectomy was considered as the removal of at least one lymph node (14). In our study, the cut-off limit for patient inclusion was 15 nodes and the mean number of nodes examined was 23, which may have attributed to finding a higher incidence of pelvic nodal involvement. Notably, in 5 out of 11 patients having nodal metastasis in this study, a single positive node was found in the final pathology.

While adequate lymph node dissection is important to ensure the absence of nodal metastasis, the number of lymph nodes that need to be removed in order to define a complete lymphadenectomy has not been established. However, several authors have argued that pelvic lymphadenectomy performed routinely may properly guide adjuvant treatment decisions, reducing the costs and potential complications of pelvic radiotherapy, based on the presence of uterine risk factors (22-24).

It has to be acknowledged that our data suffer the limitations of small sample size and a high rate of discordance between pre- and postoperative tumor grading. Despite these limitations, these findings show that among patients presenting with clinical early endometrial cancer with favorable histological characteristics, about 1 in every 10 harbor tumor of aggressive histology and 1 in every 5 will have a surgical stage higher than I due to metastasis in the pelvic lymph nodes or other extrauterine sites. Complete pelvic lymphadenectomy may increase the chances of detecting single nodal metastasis and should be considered in all patients undergoing lymph node dissection for endometrial cancer.