Therapeutic role of lymph node resection in endometrioid corpus cancer

A study of 12,333 patients

Authors

  • John K. Chan MD,

    Corresponding author
    1. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford Cancer Center, Stanford, California
    • Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford Cancer Center, 875 Blake Wilbur Dr., MC 5827, Stanford, CA 94305
    Search for more papers by this author
    • Fax: (650) 723-7448

  • Michael K. Cheung BA,

    1. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford Cancer Center, Stanford, California
    Search for more papers by this author
  • Warner K. Huh MD,

    1. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Alabama at Birmingham, Birmingham, Alabama
    Search for more papers by this author
  • Kathryn Osann PhD,

    1. Division of Hematology and Oncology, Department of Medicine, Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, Orange, California
    Search for more papers by this author
  • Amreen Husain MD,

    1. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford Cancer Center, Stanford, California
    Search for more papers by this author
  • Nelson N. Teng MD, PhD,

    1. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford Cancer Center, Stanford, California
    Search for more papers by this author
  • Daniel S. Kapp MD, PhD

    1. Division of Radiation Therapy, Department of Radiation Oncology, Stanford University School of Medicine, Stanford Cancer Center, Stanford, California
    Search for more papers by this author

Abstract

BACKGROUND.

The purpose of the current study was to determine the potential therapeutic role of lymphadenectomy in women with endometrioid corpus cancer.

METHODS.

Demographic and clinicopathologic information were obtained from the Surveillance, Epidemiology, and End Results Program between 1988–2001. Data were analyzed using Kaplan-Meier methods and Cox proportional hazards regression.

RESULTS.

In all, 12,333 women (median age, 64) underwent surgical staging with lymph node assessment, including 9,009, 1,211, 1,223, and 890 with Stage I-IV disease. Over the time intervals 1988–1992, 1993–1997, and 1998–2001, the percentage of patients undergoing lymph node staging increased from 22.6%, 29.6%, to 40.9% (P < .001). In the intermediate/high-risk patients (Stage IB, Grade 3; Stage IC and II-IV, all grades), a more extensive lymph node resection (1, 2–5, 6–10, 11–20, and >20) was associated with improved 5-year disease-specific survivals across all 5 groups at 75.3%, 81.5%, 84.1%, 85.3%, and 86.8%, respectively (P < .001). For Stage IIIC-IV patients with nodal disease, the extent of node resection significantly improved the survival from 51.0%, 53.0%, 53.0%, 60.0%, to 72.0%, (P < .001). However, no significant benefit of lymph node resection in low-risk patients could be demonstrated (Stage IA, all grades; Stage IB, Grades 1 and 2 disease; P = .23). In multivariate analysis, a more extensive node resection remained a significant prognostic factor for improved survival in intermediate/high-risk patients after adjusting for other factors including age, year of diagnosis, stage, grade, adjuvant radiotherapy, and the presence of positive nodes (P < .001).

CONCLUSIONS.

The findings of the current study suggest that the extent of lymph node resection improves the survival of women with intermediate/high-risk endometrioid uterine cancer. Cancer 2006. © 2006 American Cancer Society.

As the most common gynecologic malignancy, over 40,000 new cases of endometrial cancer are diagnosed each year.1 The number of deaths from uterine cancer has increased significantly since 1987.1, 2 One of the potential challenges for defining the most effective treatment of endometrial cancer arises from inconsistencies in the surgical staging and, in particular, the role of lymph node assessment. For example, the Gynecologic Oncology Group recently completed a trial on intermediate-risk endometrial cancer that required study participants to undergo a “selective” bilateral pelvic and para-aortic lymphadenectomy.3 In contrast, the National Comprehensive Cancer Network advocates lymph node “dissection” rather than nodal “sampling” in patients undergoing primary surgical management of endometrioid uterine cancer.4 Furthermore, the surgical treatment of endometrial cancer in academic and community hospitals varies from performing lymph node sampling only in those at high risk for nodal metastases to doing complete lymphadenectomies in all uterine cancer patients, irrespective of grade and depth of myometrial invasion.

Some retrospective studies from single institutions have demonstrated a benefit in lymphadenectomy, whereas others have found no overall survival advantage.5–9 The value of lymph node dissection is particularly difficult to demonstrate, given the overall favorable outcome in the majority of uterine cancer patients. Moreover, it is prudent to avoid overtreatment of low-risk patients even though the morbidity associated with lymph node dissection is low.6 To address this issue, the disease-specific survival outcomes of 12,333 endometrioid corpus cancer patients who underwent a lymph node assessment were analyzed to determine the potential therapeutic benefit of lymph node resection.

MATERIALS AND METHODS

Demographic, clinicopathologic, treatment, and survival information of women diagnosed with endometrioid corpus cancer during the period from January 1, 1988, through December 31, 2001, were extracted with permission from the Surveillance, Epidemiology, and End Results (SEER) program of the US National Cancer Institute. These data are representative of approximately 14% of the US population and are reported from 12 population-based registries including San Francisco-Oakland, Connecticut, metropolitan Detroit, Hawaii, Iowa, New Mexico, Seattle (Puget Sound), Utah, metropolitan Atlanta, Alaska, San Jose-Monterey, and Los Angeles.10

Of 15,689 patients who underwent a surgical staging procedure and lymph node assessment, 1,473 with uterine papillary serous carcinoma, 391 with clear cell carcinoma, 1,461 with sarcomas, and 31 with rare histologic cell types were excluded. Patients who did not undergo a staging procedure with a lymph node sampling were excluded because the reasons for not performing lymph node dissections are unclear in this heterogeneous group of patients.

Of the 12,333 women included in this analyses, patients were divided into 5 groups based on the extent of their lymph node resection (1 node, 2–5 nodes, 6–10 nodes, 11–20 nodes, and >20 nodes). Each nodal group consisted of more than 800 patients. To better characterize the patient population, the race classifications of the SEER program were categorized into four groups: Caucasians, African-Americans, Asians, and Others. Asians were defined as Chinese, Japanese, Korean, Vietnamese, and Filipina. All other race and ethnicity classifications were defined as Others.

To analyze trends in the study cohort and to determine 5-year disease-specific survival, chi-square tests and Kaplan-Meier analysis with log-rank tests were performed for the nodal groups defined above. The outcome of interest was death from endometrial cancer and time to death was censored in women who died from causes other than uterine cancer. The Cox proportional hazards model was used to investigate the significance of extensive lymph node resection after accounting for other patient characteristics including age, race, year of diagnosis, stage, grade, adjuvant radiotherapy, and number and percent of positive nodes. To meet proportionality assumptions, the extent of lymph node resection was divided into 3 groups (1–10 nodes, 11–20 nodes, >20 nodes). Multivariate analyses were conducted separately for low-risk (Stage IA, all grades; Stage IB, Grades 1 and 2), intermediate/high-risk (Stage IB, Grade 3, Stage IC-IV, all grades), and high-risk (Stage IIIC-IV with positive nodes) patients. Covariates including interactions were included in the models based on restricted chi-square tests. Kaplan-Meier curves are presented for the three lymph node groups used in the multivariate analysis. Two-tailed tests at P-values less than 0.05 were considered significant. All data were analyzed using Intercooled STATA (v. 8.0; STATA, College Station, TX) and SAS (v. 6.12; SAS, Cary, NC).

RESULTS

From 1988 to 2001, 12,333 patients were diagnosed with endometrioid uterine cancer and underwent a surgical staging procedure including lymph node (pelvic, periaortic, or both) assessment and stage assignment based on the 1988 International Federation of Gynecology and Obstetrics staging system. The patient demographics based on the extent of node dissection are depicted in Table 1. Over the 14-year period, the percent of patients who underwent lymph node assessments increased from 22.6%, 29.6%, to 40.9% for the time intervals 1988–1992, 1993–1997, and 1998–2001, respectively (P < .001). In addition, the median number of nodes resected increased from 7 to 9 to 12 over the same time periods (P < .001).

Table 1. Patient Demographics
 Total (n = 12,333)1 Lymph node (n = 879)2–5 Lymph nodes (n = 3068)6–10 Lymph nodes (n = 2631)11–20 Lymph nodes (n = 3222)>20 Lymph nodes (n = 2533)
  • Percent of patients in columns for a given parameter.

  • *

    Age at diagnosis in years.

  • Asians were defined arbitrarily as Chinese, Japanese, Korean, Vietnamese, and Filipina.

  • Others were defined as all other race/ethnicity parameters.

Median age at diagnosis, y*646565646462
Median year of diagnosis199719951996199719971998
Race (%)
White10,732755 (85.9)2652 (86.5)2300 (87.5)2822 (87.5)2203 (87.0)
African-Americans60958 (6.6)165 (5.4)143 (5.4)150 (4.7)93 (3.7)
Asians71442 (4.8)173 (5.6)140 (5.3)176 (5.5)183 (7.2)
Other27824 (2.7)78 (2.5)48 (1.8)74 (2.3)54 (2.1)

Pathologic and adjuvant treatment data associated with the extent of lymph node resection are shown in Table 2. Data on adjuvant chemotherapy and hormonal therapy were not available for analysis. Although the use of adjuvant radiotherapy did not differ significantly between the 5 nodal groups (P = .27), there was a trend toward performing adjuvant radiation more often for those who underwent a less extensive lymph node resection.

Table 2. Pathologic and Adjuvant Radiation Data*
 Total (n = 12,333)1 Lymph node (n = 879) No. (%)2–5 Lymph nodes (n = 3068) No. (%)6–10 Lymph nodes (n = 2631) No. (%)11–20 Lymph nodes (n = 3222) No. (%)>20 Lymph nodes (n = 2533) No. (%)Chi-square test P
  • Percent of patients in columns for a given parameter.

  • Two-tailed tests at P less than .05 were considered significant.

  • *

    Stage and grade based on 1998 FIGO classification.

  • P from chi-square test of differences between nodal groups by stage, grade, and radiation.

Stage of disease
 Stage I9009602 (68.4)2312 (75.4)1964 (74.6)2363 (73.3)1768 (69.8) 
 Stage II121177 (8.8)258 (8.4)251 (9.5)354 (11.0)271 (10.7)<.001
 Stage III122386 (9.8)291 (9.5)253 (9.6)301 (9.3)292 (11.5) 
 Stage IV890114 (12.0)207 (6.7)163 (6.3)204 (6.4)202 (8.0) 
Grade of disease
 Grade 13703267 (30.4)1006 (32.8)847 (32.2)943 (29.3)640 (25.2) 
 Grade 24955348 (39.6)1211 (39.5)1071 (40.7)1283 (39.8)1042 (41.0)<.001
 Grade 33224234 (26.6)737 (24.0)600 (22.8)874 (27.1)779 (30.8) 
 Unknown45130 (3.4)114 (3.7)113 (4.3)122 (3.8)72 (2.8) 
Radiation
 No radiation7738528 (60.1)1895 (61.8)1643 (62.4)2038 (63.2)1634 (64.5) 
 Radiation4450337 (38.3)1138 (37.1)955 (36.3)1152 (35.8)868 (34.3).27
 Unknown14514 (1.6)35 (1.1)33 (1.3)32 (1.0)31 (1.2) 

The 5-year disease-specific survivals in patients with Stage I, II, III, and IV disease were 95.5%, 90.4%, 73.9%, and 53.3%, respectively. In the overall study group, women who were ≤65 years of age had a significant survival advantage compared with those >65 years: 93.5% versus 86.2% (P < .001). Caucasians had a 5-year survival of 90.0% compared with 85.9% and 92.0% in African-Americans and Asians, respectively (P = .01). Furthermore, those with Grade 1 tumors had a statistically significant higher survival at 97.7% compared with 91.9% and 78.7% of those with Grade 2 and 3 disease (P < .001).

Of the Stage I patients, women who had 1 node, 2–5 nodes, 6–10 nodes, 11–20 nodes, and >20 nodes removed had corresponding improved 5-year disease-specific survival rates of 92.8%, 95.4, 95.9, 95.3, and 96.5%, respectively (P = .005). Of the Stage I patients with myometrial invasion <50%, a more extensive lymph node resection was associated with improved 5-year survivals from 94.9%, 95.8%, 95.6%, 96.3%, to 97.3%, although these differences were not statistically significant (P = .46). However, in those with ≥50% myometrial invasion, there was a statistically significant improvement in 5-year survival from 81.1%, 90.4%, 93.3%, 90.5%, to 94.3% (P < .001). Moreover, across all stages higher numbers of nodes resected were associated with an improvement in survival in women with both Grade 2 and Grade 3 disease from 87.1%, 91.2%, 92.9%, 91.9%, to 93.7% (P = .001) and 65.2%, 75.6%, 77.5%, 81.9%, to 84.1% (P < .001), respectively.

Women with intermediate or high-risk disease, defined as Stage IB Grade 3, Stage IC-IV all grades, appeared to benefit from an extensive lymph node resection (1, 2–5, 6–10, 11–20, >20) with an improved 5-year survival of 75.3%, 81.5%, 84.1%, 85.3%, and 86.8%, respectively (P < .001). Conversely, no statistically significant benefit of extensive lymph node resection in those with low-risk disease could be demonstrated (Stage IA, all grades; Stage IB, Grades 1 and 2). Kaplan-Meier survival curves are presented for patients in three different nodal groups based on disease risk (Fig. 1).

Figure 1.

(A) Kaplan-Meier disease-specific survival by number of lymph nodes removed in low-risk patients (Stage IA, all grades; Stage IB, Grades 1 and 2). (B) Kaplan-Meier disease-specific survival by number of lymph nodes removed in intermediate/high-risk patients (Stage IB, Grade 3; Stage IC-IV, all grades). (C) Kaplan-Meier disease-specific survival by number of lymph nodes removed in patients with nodal disease (Stage IIIC-IV).

To further elucidate the potential benefits of extensive lymph node resection, Stage IIIC-IV patients with nodal disease were analyzed to determine the role of extensive node resection in this high-risk group of patients. Of 1,221 such patients, the median number of nodes resected was 11 (range, 1–90). Patients who had 1 node removed had a survival of only 51.0% compared with 53.0%, 53.0%, 60.0%, and 72.0% in those who had 2–5, 6–10, 11–20, and >20 nodes removed, respectively (P < .001; Table 3). Furthermore, the survival advantage of removing more nodes persists in patients who had only 1 metastatic node (P = .001), 2–5 metastatic nodes (P < .001), and >5 metastatic nodes (P < .001). Figure 1C depicts the survival of Stage IIIC-IV patients with positive nodes after removing 1–10, 11–20, and >20 nodes. In a subset analysis of only Stage IIIC patients, the median number of lymph nodes taken was 12. The 5-year disease-specific survival of Stage IIIC patients with <12 and ≥12 nodes removed was 67.0% versus 73.4% (P = .04). To minimize the potential sampling bias of examining 1 versus multiple nodes, an additional analysis was performed on the percentage of positive nodes rather than the number of nodes. This analysis found that women with ≤5% positive nodes had a significant survival advantage over those with 5–10%, 10–20%, and >20% positive nodes with a corresponding 5-year disease-specific survival of 84% versus 73%, 62%, and 51%, respectively (P < .001).

Table 3. Five-Year Disease-Specific Survival Analysis Based on Stage and Grade*
 No.Total % (SE)1 Lymph node % (SE)2–5 Lymph nodes % (SE)6–10 Lymph nodes % (SE)11–20 Lymph nodes % (SE)>20 Lymph nodes % (SE)Log-Rank
  • Myo. Inv., myometrial invasion; SE, standard error.

  • *

    Stage and grade based on 1998 FIGO classification.

  • Including only those with positive lymph node disease.

  • Includes all stages of disease; 451 patients with unknown grade of disease were excluded.

Stage of disease
Stage I900995.5 (0.3)92.8 (1.2)95.4 (0.5)95.9 (0.6)95.3 (0.6)96.5 (0.6)0.005
Myo. Inv. <50%482696.0 (0.4)94.9 (1.4)95.8 (0.7)95.6 (0.8)96.3 (0.6)97.3 (0.7)0.46
Myo. Inv. ≥50%167291.0 (0.9)81.1 (4.0)90.4 (1.7)93.3 (1.7)90.5 (1.8)94.3 (1.8)<0.001
Stage II121190.4 (1.0)84.9 (4.2)85.9 (2.5)93.8 (1.8)93.0 (1.8)0090.5 (2.3)0.028
Stage IIIC-IV122159.7 (1.8)51.0 (5.9)53.0 (3.7)53.0 (4.5)60.0 (3.6)72.0 (3.3)<0.001
1 positive lymph node52868.0 (2.6)51.3 (5.9)63.9 (5.2)71.1 (6.7)74.9 (5.1)82.6 (4.6)<0.001
2–5 positive lymph nodes54755.1 (2.8)42.9 (5.0)46.0 (6.3)57.9 (5.3)71.9 (5.2)<0.001
>5 positive lymph nodes14646.4 (5.1)20.9 (11.4)33.8 (8.6)60.5 (7.2)<0.001
Grade of disease
 Grade 1370397.7 (0.3)97.6 (1.1)97.8 (0.6)97.9 (0.6)97.9 (0.5)97.1 (0.9)0.68
 Grade 2495591.9 (0.5)87.1 (2.0)91.2 (1.0)92.9 (1.0)91.9 (1.0)93.7 (1.0)0.001
 Grade 3322478.7 (0.9)65.2 (3.4)75.6 (1.8)77.5 (2.0)81.9 (1.6)84.1 (1.7)<0.001

Patients who did not undergo a lymphadenectomy were excluded from this study because this group represents a heterogeneous cohort of patients. For example, many of these women may have had low-grade tumors without myometrial invasion or possible medical comorbidities and surgical complications that precluded the surgeons from performing a lymphadenectomy. An analysis was performed to determine the clinicopathologic factors and outcomes of the unstaged compared with staged patients. In Stage I disease, the proportion of women with Grade 1 and superficially myoinvasive cancers was in fact higher in those who did not receive a lymphadenectomy (54.8% vs. 34.7%, P < .001 for Grade 1 disease; 31.6% vs. 21.8%, P < .001 for superficial invasion) compared with those who underwent a lymphadenectomy. Unstaged patients with early and advanced stage disease had survivals of 95.5% and 41.1% compared with 94.8% and 64.3% in staged patients. Although there was no statistical difference in survival of the Stage I patients, those with Stage II, III, and IV disease without lymph node sampling had a significantly worse survival (82.2% vs. 90.4%, 63.5% vs. 73.9%, and 28.2% vs. 53.3%, respectively) compared with the staged patients.

After adjusting for independent prognostic factors including age, race, year of diagnosis, stage, grade, adjuvant radiation therapy, and the presence of positive lymph nodes in multivariate analysis, a more extensive lymph node resection remained as a significant independent prognostic factor for improved disease-specific survival (P < .001). With all other prognostic factors held constant, the hazard ratios for the removal of 10–20 and >20 lymph nodes relative to 1–10 nodes were 0.80 (95% confidence interval [CI]: 0.73–0.89) and 0.65 (95% CI: 0.58–0.71) for intermediate/high-risk patients and 0.77 (95% CI: 0.65–0.91) and 0.60 (95% CI: 0.50–0.70) for Stage IIIC-IV patients with positive nodes (Table 4). Furthermore, after controlling for the number of metastatic lymph nodes and percent positive nodes of all resected nodes in high-risk (Stage III-IV) patients, the extent of the number of benign nodes removed remained as an independent prognostic factor for improved survival (P < .001). Race and year of diagnosis were not significantly associated with survival after adjusting for other covariates in multivariate analysis. However, the limited follow-up of more recently diagnosed women may explain the lack of statistical significance for year of diagnosis.

Table 4. Relative Hazards Rates for Disease-Specific Mortality Associated with Lymph Node Resection*
 Relative hazards rateRelative hazards rate (95% CI)
1–10 Lymph nodes11–20 Lymph nodes>20 Lymph nodes
  • CI, confidence interval.

  • *

    Stage and grade based on 1998 FIGO classification.

  • Adjusted for age at diagnosis, grade, and radiation.

  • Adjusted for age at diagnosis, grade, stage, radiation, number and percent of positive lymph nodes.

  • §

    Adjusted for age at diagnosis, grade, stage, radiation, number and percent of positive lymph nodes.

Low-risk
 Stage IA (all grades), and Stage IB (Grades 1 and 2)1.001.12 (0.87–1.45)1.26 (0.98–1.63)
Intermediate/high-risk
 Stage IB (Grade 3), and Stage IC-IV (all grades)1.000.80 (0.73–0.89)0.65 (0.58–0.71)
Patients with lymph node disease
 Stage IIIC-IVwith positive lymph nodes only§1.000.77 (0.65–0.91)0.60 (0.50–0.70)

DISCUSSION

The observation that the number of pathologically negative nodes is correlated with survival has been shown in melanoma, breast, and rectal cancers.11–15 Previous studies from single institutions have demonstrated that a more extensive lymph node sampling may contribute to the survival of women with endometrial cancer.5, 7, 16, 17 Recently, Cragun et al.6 showed that patients with poorly differentiated Stage I and IIA endometrial cancers having more than 11 pelvic nodes removed had an improved overall and progression-free survival compared with those with 11 or fewer nodes resected. However, those authors were unable to demonstrate similar findings in women with Grade 1 and 2 tumors. Some of these prior studies have failed to include details regarding the extent of nodal resections based on the number of lymph nodes removed due to the small study group size, whereas others have reported their data on node counts based on an arbitrary cutoff.5, 6 Furthermore, many reports have included a heterogeneous cohort of clinically staged patients and those with high-risk histologic cell types.

The current report is the largest series to date on endometrioid uterine cancer patients showing a survival benefit in subsets of patients who had a more extensive lymph node resection. Due to the large numbers of Stage I and II patients (more than 10,000), the study cohort could be divided into 5 groups, according to the number of nodes removed, with more than 800 patients in each group. Furthermore, the strength in numbers allowed the analysis of specific subgroups based on stage and grade of disease. As such, this study was able to demonstrate a significant improvement in survival associated with the extent of lymphadenectomy even in those with Grade 2 tumors and a trend toward improvement in Stage IB disease (Table 3). These findings have not been reported in other series, perhaps due to their limited sample sizes.

In a large population-based study, Trimble et al.9 also showed that a more extensive lymph node dissection was associated with an increased survival among patients with Stage I and Grade 3 disease, but not in Grade 1 or 2 tumors. The authors recognized that their study was limited by the low median node count of 7 and the lack of information on postoperative adjuvant radiotherapy. Compared with the Trimble et al.9 series, the current study provides an analysis on an additional 7,017 patients with Stage I and II disease. The 5 nodal groups in this report convincingly demonstrate that the survival advantage was associated with the extent of lymphadenectomy. Furthermore, the results of this study showed that the use of adjuvant radiotherapy was not different among the 5 nodal groups suggesting that adjuvant radiotherapy did not explain the better survival observed in those who underwent more extensive nodal resections. Women with Stage IB Grade 2 disease or less in this series do not have any demonstrable therapeutic benefit by the removal of regional lymph nodes. Clearly, it is difficult to demonstrate a survival advantage in low-risk disease, particularly when the survival of Stage IA and IB patients reaches over 98% and 96%, respectively. In patients with Grade 1 cancers, the risk of extrauterine disease is relatively low. However, Ben-Shachar et al. showed that up to 19% of Grade 1 cancers on biopsy were subsequently upgraded on final pathology.18 Furthermore, these authors found that lymphadenectomy significantly impacted postoperative treatment decisions in 29% of the cases. Although CA125 can predict for extrauterine spread with a sensitivity 30% to 70%,19, 20 its use to predict the need for surgical staging remains controversial. Dotters21 performed a prospective study and found that CA125 test correctly identified only 50% of those requiring lymphadenectomy. Thus, those authors concluded that CA125 lacked the sensitivity to identify those who require staging procedures. Similarly, Koper et al.22 showed that the sensitivity of CA125 in identifying patients who needed a lymphadenectomy was only 17%. Therefore, without the explicit ability to prove that disease is confined to the uterus or confirm histologic grade preoperatively or during surgery, retroperitoneal surgical and histologic evaluation for staging intent may still have merit.

Because advanced uterine cancer comprises less than 20% of patients with uterine cancer, the overall benefit of systematic lymphadenectomy in node-positive endometrial cancer has been difficult to demonstrate.2 This report evaluated 1,221 women with nodal disease who underwent a limited to an extensive lymph node resection. The extent of lymph node resection improved the survival of these patients even after controlling for the number of nodal metastases (Table 3). These findings suggest that the improvement in survival associated with the extent of lymph node resection may not simply be due to stage migration.

The improvement in survival associated with extent of lymph node resection may be explained by several mechanisms. First, with an increasing number of nodes removed, there is a higher statistical probability of obtaining a sufficient number of nodes to adequately stage patients. Thus, the survival difference observed in this study may simply be due to comparing patients with true Stage I disease after a thorough staging procedure with inaccurately staged patients with true Stage IIIC disease. Second, a thorough lymph node resection may improve the patient's survival by removing micrometastatic disease within the node that was not detected on routine hematoxylin and eosin analyses.23 Thus, clonogenic tumor cells can potentially develop into macrometastatic nodal disease that initially would have been considered negative on pathologic examination. The results of this study showing the benefits of complete lymphadenectomy in patients with node-positive disease (Stage IIIC) support this hypothesis.

It remains to be determined if there exists an absolute number of nodes removed that clearly defines a therapeutic lymphadenectomy. This report divided the study group into 5 different nodal groups to demonstrate that the number of nodes resected is merely a surrogate for showing the benefits of nodal resection over sampling. In multivariate analysis, the hazard ratio declined significantly in a stepwise fashion with increasing extensiveness of lymph node resection after adjusting for other independent prognostic factors (Table 4). Although nodal count is a reflection of the extensiveness of lymph node dissection, it may also depend on various factors such as comprehensiveness of pathologic analysis, surgical expertise, and anatomical variations among patients. Although the current study and many other retrospective studies have demonstrated a survival benefit in those who had a more extensive lymphadenectomy, there are certain patients in whom lymph node sampling or lymphadenectomy may not be feasible due to comorbid factors, blood loss, or body habitus. Notwithstanding these predisposing factors, Cragun et al.6 showed that lymph node resection is associated with adverse events, with the most common medical and surgical complication being prolonged ileus and lymphocyst formation, respectively.

This study was limited by the lack of information on surgeon subspecialty, central pathology review, adjuvant hormonal and chemotherapy, time to recurrence, subsequent surgical and medical therapies, and surgical morbidity. The SEER database does not report detailed information on the depth of myometrial invasion in those who have positive nodes. Additionally, there were limited data on the specific laterality and location of nodal resection (pelvic vs. periaortic). However, a recent study of by McMeekin et al.24 found no significant survival difference between patients who had positive periaortic lymph nodes versus those with positive pelvic lymph nodes. Patients with more extensive nodal disease and advanced stage cancers may only receive a single lymph node sampling rather than a full lymphadenectomy. Thus, a limited node sampling may serve as a surrogate marker for this small subset of patients with poor prognostic cancers. If this hypothesis is true, there would be a significant number of patients with Stage IIIC (nodal disease) that would have only received a single node resection. However, only 5.9% of patients diagnosed with Stage IIIC disease received a single lymph node resection in this study. In addition, this explanation cannot account for the survival benefit associated with an extensive lymph node dissection in the women with intermediate risk. Moreover, the benefit of an extensive lymph node resection was demonstrated on multivariate analysis as an independent prognostic factor even in the presence of nodal metastases. Although information on the use of chemotherapy was not available, it is unlikely chemotherapy was administered to the nearly 80% of the patients in this study because there is no evidence that adjuvant chemotherapy improves the survival of patients with Stage I-II disease. Thus, the use of chemotherapy is unlikely to bias the results.

The current study consists of an unselected group of patients spanning 12 US regions, thus minimizing selection and surveillance biases often associated with clinical trials and studies from single academic institutions. In order to further investigate the benefit of lymph node resection in endometrioid corpus cancer, one would need to design a large, prospective randomized trial with standardized surgical staging and pathology review. However, given the well-known biases within the specialty, surgeons and patients may not be willing to subject themselves to a randomized trial involving “no nodal resection” or “limited nodal resection” arms.

Surgical treatment for endometrioid uterine cancer needs to be standardized, particularly in those with intermediate or high-risk disease. Because nearly two-thirds of such patients have Grade 2 or 3 disease and may benefit from a thorough lymphadenectomy, surgeons with the technical expertise in performing this procedure should be involved in their care.2, 7 Finally, a detailed documentation of regional nodal involvement can potentially circumvent or direct the need for postoperative adjuvant treatment.

Acknowledgements

We thank Dr. Philip Lavori, PhD (Stanford Cancer Center), for statistical consultation and helpful comments on the article and Stella K. Cheung, (Stanford University), for graphics and technical assistance. Dr. John K. Chan is independent of any commercial funder, has had full access to all the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Ancillary