Resection of lymph node metastases influences survival in stage IIIC endometrial cancer

Gynecologic Oncology 99 (2005) 689 – 695 www.elsevier.com/locate/ygyno

Resection of lymph node metastases influences survival in stage IIIC endometrial canceri Laura J. Havrilesky a,*, Janiel M. Cragun a, Brian Calingaert b, Ingrid Synan a, Angeles Alvarez Secord a, John T. Soper a, Daniel L. Clarke-Pearson a, Andrew Berchuck a a

Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Duke University Medical Center, Box 3079, Durham, NC 27710, USA b Biostatistics and Bioinformatics, Duke University Medical Center, Box 3079, Durham, NC 27710, USA Received 15 April 2005 Available online 29 August 2005

Abstract Objective. Surgical staging of endometrial cancer identifies those patients with microscopic metastatic disease most likely to benefit from adjuvant therapy and may also confer therapeutic benefit. Our objective was to compare survival of patients who underwent resection of grossly positive lymph nodes (LN) to those with microscopically positive LN. Methods. Patients had stage IIIC endometrial cancer with pelvic and/or aortic LN metastases and underwent surgery between 1973 and 2002. Exclusion criteria included pre-surgical radiation and second primary cancer. Survival was analyzed using Kaplan – Meier method and Cox proportional hazards model. Results. Mean age of 96 patients with stage IIIC endometrial cancer was 64. There were 45 cases with microscopic LN involvement and 51 with grossly enlarged LN. Overall, 41% had disease in aortic LN, which in 18% represented isolated aortic LN metastasis. Adjuvant therapies were given to 92% of patients (85% radiotherapy, 10% chemotherapy, 10% progestins). Among those with grossly involved LN, 86% were completely resected. Five-year disease-specific survival (DSS) was 63% in 45 patients with microscopic metastatic disease compared to 50% in 44 patients with grossly positive LN completely resected and 43% in 7 with residual macroscopic disease. In multivariable analyses, gross nodal disease not debulked (HR = 6.85, P = 0.009), serosal/adnexal involvement (HR = 2.24, P = 0.036), diagnosis prior to 1989 (HR = 4.33, P < 0.001), older age (HR = 1.09, P < 0.001), and >2 positive lymph nodes (HR = 3.12, P = 0.007) were associated with lower DSS. Conclusion. Grossly involved LN can often be completely resected in patients with stage IIIC endometrial cancer. These retrospective data provide evidence suggestive of a therapeutic benefit for lymphadenectomy in endometrial cancer. D 2005 Elsevier Inc. All rights reserved. Keywords: Stage IIIC endometrial cancer; Lymphadenectomy

Introduction Endometrial cancer is the most common gynecologic malignancy with 40,100 new cases and 6800 deaths reported each year [1]. Since the establishment of surgical staging criteria by FIGO in 1988, complete staging has included abdominal exploration, pelvic peritoneal cytology, i

Focused plenary presentation, Society of Gynecologic Oncologists 36th Annual Meeting, Miami, FL, USA, March 2005. * Corresponding author. Fax: +1 919 684 8719. E-mail address: [email protected] (L.J. Havrilesky). 0090-8258/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2005.07.014

hysterectomy, bilateral salpingo-oophorectomy, and pelvic and aortic selective lymphadenectomy. Patients with FIGO stage IIIC disease have pelvic or aortic lymph node metastasis, with or without positive peritoneal cytology, adnexal, cervical, or vaginal metastasis. Post-operative treatment of stage IIIC disease usually consists of adjuvant radiotherapy, chemotherapy, or both, and 5-year survival varies from 35 to 84% in reported studies [2– 6]. Several prior studies have suggested a therapeutic benefit to the performance of selective lymphadenectomy for endometrial cancer [7 –10]. We previously reported that the number of lymph nodes removed was an independent

690

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695

predictor of survival in apparent stage I endometrial cancer [10]. The benefit of debulking gross nodal metastasis has not been well established for this disease. The goals of this study were to assess predictors of disease-specific survival (DSS) and to determine whether debulking grossly metastatic disease affected outcomes among patients with FIGO stage IIIC endometrial cancer.

Materials and methods Following Institutional Review Board approval, the Duke University Tumor Registry was reviewed to identify all patients with endometrial cancer who received primary surgical treatment at Duke University Medical Center between 1973 and 2002. Inclusion criteria were an initial procedure that included hysterectomy and selective lymphadenectomy (pelvic or pelvic and aortic) and a FIGO stage of IIIC. The decision to perform lymphadenectomy and the extent of lymphadenectomy varied by surgeon preference and was typically decided with consideration of the grade, depth of invasion, size and location of the primary tumor at the time of hysterectomy, as well as each patient’s medical comorbidities. At our institution, selective pelvic lymphadenectomy for endometrial cancer typically involves removing lymphatic tissue from the anterior and medial surfaces of the iliac vessels and from the obturator fossa superior to the obturator nerve. Aortic selective lymphadenectomy usually consists of removal of precaval and lower right and left aortic lymphatic tissue to the level of the inferior mesenteric artery. Exclusion criteria included preoperative radiation, coexisting second primary gynecologic cancer, and primary surgery not performed at our institution. Clinical data were abstracted by review of patient charts. Operative reports were reviewed for gross evidence of metastatic disease and extent of residual disease. Occult metastatic disease was defined as no suspicious nodes or other suspicious extrauterine disease seen at laparotomy. Lymph nodes were defined as suspicious or not suspicious for metastasis based on the surgeon’s clinical impression described in the operative report. Gross residual disease was defined as disease not debulked with at least one dimension of 2 cm or greater. Original pathology reports were reviewed for histologic type, depth of myometrial invasion, FIGO grade, FIGO stage, number, location, and status of retroperitoneal nodes, peritoneal cytology report, and sites of metastasis. The Kaplan – Meier method was used to generate survival curves and calculate overall survival (OS) and DSS [11]. The log-rank test was used to assess differences in survival rates [12]. Stratified analyses were performed to examine possible effect modification by clinical risk factors (gross residual disease, presence of positive aortic node, and histology). Univariate and multivariable analyses were performed using the Cox proportional hazards model [13] to identify

prognostic factors independently associated with survival. Models were developed using stepwise forward selection, retaining variables significant at the two-sided alpha = 0.05 level. Variables evaluated in the multivariable model were number of positive nodes (analyzed as both a continuous and a dichotomous [divided at the median] variable), depth of myometrial invasion (inner/outer), histologic type (papillary serous/clear cell vs. other), grade (<3 vs. 3), adjuvant radiation, adjuvant chemotherapy, year of surgery (dichotomous, divided at 1988 –1989), age at surgery (continuous), race (white vs. other), aortic nodes removed (yes vs. no), serosa, adnexae, or peritoneal cytology (any positive vs. all negative), grossly visible nodal disease (3 categories: no [reference group], yes-debulked to microscopic residual, yesnot completely debulked), and presence of metastasis in aortic lymph nodes. For continuous variables, quadratic and cubic terms were assessed. All statistical analyses were performed using SAS Version 8.2 (Statistical Analysis Software, Cary, NC).

Results Of 1656 patients treated for endometrial cancer between 1973 and 2002, 883 underwent full surgical staging including hysterectomy and selective lymphadenectomy. 112 (12.7%) patients who underwent hysterectomy and selective lymphadenectomy were diagnosed with stage IIIC endometrial cancer. Of these, 10 were excluded due to preoperative radiotherapy; 5 were excluded for simultaneously diagnosed co-existing primary gynecologic cancers; and 1 was excluded because her primary surgery was at another institution. Of the remaining 96 patients with stage IIIC endometrial cancer, the mean age was 64 years (range 30– 93), and 64% were Caucasian. The median number of lymph nodes removed was 19; median number of pelvic nodes removed was 13.5 (interquartile range 9 –20), and median number of aortic nodes removed was 4 (interquartile range 1 –7). The median number of lymph nodes containing metastases was 2 (interquartile range 1 –4). Clinical and pathologic characteristics are listed in Table 1. Papillary serous or clear cell histologic types were identified in 18/96 (19%) patients. Poorly differentiated tumors (55/96, 59%) and outer 2 myometrial invasion (63/96, 72%) were present in the majority of patients. There were three patients (3%) with grade 1 tumors invading less than 1/2 of the myometrial thickness; all three had clinically suspicious lymph nodes. Four patients had no myometrial invasion, of whom two had grade 2 and two had grade 3 tumors. Eighty patients (83%) had both pelvic and aortic lymph nodes resected; of these, 42 (52%) had isolated pelvic node metastasis, 14 (18%) had isolated aortic node metastasis, and 24 (30%) had both pelvic and aortic node metastasis. Metastatic disease was occult in 45 (47%) patients. Grossly enlarged nodes containing metastasis were identified in 51 (53%) patients. Among patients with grossly enlarged

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695 Table 1 Clinical and pathologic characteristics Characteristic Age at surgery Year of surgery 1973 – 1988 1989 – 2002 FIGO grade (n = 94) 1 2 3 Depth of myometrial invasion (n = 88) Inner 1/2 Outer 1/2 Cervical involvement (n = 93) Negative Positive Histologic type (n = 95) Endometrioid or mucinous Papillary serous or clear cell Race (n = 94) Caucasian African-American Other Pelvic node status (n = 94) Negative Positive Aortic node status (n = 82) Negative Positive Adjuvant radiotherapy (n = 96) No Yes Adjuvant chemotherapy (n = 96) No Yes Adjuvant progestins (n = 96) No Yes

n (%)

Mean

Range

64.1

30 – 93

41 (43) 55 (57) 5 (5) 34 (36) 55 (59)

‘

25 (28) 63 (72) 63 (68) 30 (32) 77 (81) 18 (19) 60 (64) 31 (33) 3 (3) 14 (15) 80 (85) 42 (51) 40 (49) 14 (15) 82 (85) 86 (90) 10 (10) 86 (90) 10 (10)

lymph nodes, 44/51 (86%) were debulked to microscopic residual, while 7/51 (14%) had macroscopic residual disease at the conclusion of surgery. Adjuvant radiotherapy was administered to 82/96 (85%) patients. Ten of 96 (10%) received chemotherapy, while 10/96 (10%) received adjuvant hormonal treatment. Seventy (73%) received adjuvant radiotherapy alone, while seven (7%) received adjuvant chemotherapy and radiotherapy, five (5%) received adjuvant radiotherapy and hormonal therapy, five (5%) received adjuvant hormonal therapy alone, and three (3%) received adjuvant chemotherapy alone. Radiotherapy usually consisted of 45 to 50 Gy external beam pelvic radiotherapy with or without a 45 Gy extended aortic field boost. Adjuvant chemotherapy regimens included cisplatin alone (3/10 patients), cisplatin and doxorubicin (5/10), carboplatin alone (1/10), and liposomal doxorubicin (1/10). Hormonal therapy consisted of megestrol acetate (5 patients) or medroxyprogesterone acetate (5 patients) at varying dosages and for varying time periods depending upon patient tolerance and status of disease.

691

Median follow up among censored patients in the OS and DSS analyses were 58 and 50 months respectively. Using Kaplan– Meier analysis, 5-year OS and DSS were 47% and 55%, respectively (Fig. 1). Five-year DSS was 63% among 55 patients with occult metastatic disease, 50% among 44 patients with gross metastatic nodal disease debulked to microscopic residual, and 43% among 7 patients with gross residual nodal disease following primary surgery (Fig. 2, P = 0.247). Among the 7 patients in whom gross nodal disease could not be debulked, four died of their cancer within 2 years of diagnosis, one died of unknown causes 8 years following diagnosis, and two were alive without evidence of disease 21 months and 8 years following diagnosis. Patients with at least one positive aortic node (n = 40) had a 5-year DSS of 49%, compared to 60% in patients who had isolated pelvic node metastasis (n = 42) ( P = 0.150). Patients with papillary serous or clear cell histologic type (n = 18) had poorer DSS compared to patients with low risk histologic types (n = 77) P = 0.011, with DSS at 5 years of 34% and 59%, respectively. Patients who had surgery prior to 1989 had significantly worse DSS compared to those whose surgery was performed after 1989 ( P = 0.001, 5-year survival 36% vs. 74%). Median follow up was different between the two time periods (103 months among patients diagnosed prior to 1989, 48 months among patients diagnosed from 1989 forward). We performed univariate and multivariable analyses to investigate which clinical and pathological variables were associated with DSS and OS (Table 2). In univariate analyses, the following variables were associated with lower DSS and OS: diagnosis prior to 1989; older age; more than 2 lymph nodes positive; papillary serous or clear cell histologic type; and either serosa, adnexae, or peritoneal cytology positive. Because 10 subjects were missing data on peritoneal cytology (either not performed or not recorded), we also tested the significance of positive serosa or adnexae. This variable was significant for DSS but not OS. We used serosa/adnexae in the multivariable analyses in order to retain the 10 subjects with missing values for peritoneal cytology. The number of positive lymph nodes was also significantly associated with DSS and OS when analyzed as a continuous variable and had a complex relationship including linear, quadratic, and cubic terms. In multivariable analyses, the following variables were associated with lower DSS: older age (HR 1.09, P < 0.001), diagnosis prior to 1989 (HR 4.33, P < 0.001), serosa or adnexa positive (HR 2.24, P = 0.036), greater than 2 positive lymph nodes (HR 3.12, P = 0.007), and gross residual disease not debulked at surgery (HR 6.85, P = 0.009). Following parameter adjustment of the multivariable model for direct comparison between groups, patients whose grossly positive nodes were not completely debulked had lower DSS compared to those whose grossly positive nodes were debulked to micro-

692

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695

Fig. 1. Disease-specific survival among 96 patients with stage IIIC endometrial cancer.

scopic residual (HR = 5.00, P = 0.033). The number of positive nodes was significantly associated with DSS when analyzed as a dichotomous variable or as a continuous variable. However, as a continuous variable, this again had a complex relationship which involved quadratic and cubic terms. All variables associated with lower DSS were also associated with lower OS with the exception of serosa/adnexa positive, which was not included in the multivariable model for OS because it was not significant during stepwise selection. The most common sites of failure were distant. Recurrence occurred at multiple sites in 11% of patients, pulmonary in 10%, other distant sites in 8%, and pelvic or

aortic node distributions in 6%. Isolated pelvic or vaginal vault recurrences were relatively rare (5%). Among 69 patients who received adjuvant radiotherapy alone, sites of recurrence included: pelvis or vault (4%), abdomen (3%), aortic nodes (3%), lung (12%), liver (1.5%), another distant site (12%), and multiple sites (10%). Among seven patients receiving a combination of adjuvant radiotherapy and chemotherapy, there was one distant recurrence and one multiple site recurrence. Among five patients receiving adjuvant radiotherapy and hormonal therapy, there was one pulmonary recurrence. Among five patients who received adjuvant hormonal therapy alone, there was one vault recurrence and one pulmonary recurrence. Among three

Fig. 2. Disease-specific survival among patients with FIGO stage IIIC endometrial cancer stratified according to residual nodal disease (microscopic nodal disease [n = 45], macroscopic nodal disease completely debulked [n = 44], macroscopic residual disease [n = 7]).

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695

693

Table 2 Results of Cox proportional hazards model n

Overall survival Parameter Estimate

Univariate analysis Age at surgery (years, continuous) Diagnosis before 1989 Papillary serous or clear cell >2 positive lymph nodes Number of positive lymph nodes linear quadratic cubic Serosa/adnexa/cytology positive Peritoneal cytology positive Adjuvant radiotherapy Serosa/adnexa positive Grossly positive nodes None (reference group) Macroscopic, debulked Macroscopic, not debulked Positive aortic node (yes vs. no) Aortic nodes removed (yes vs. no) Adjuvant chemotherapy Race Caucasian Outer 1/2 myometrial invasion Poorly differentiated Total nodes removed (continuous)

Disease-specific survival Hazard

P value

SE

Ratio

(95% CI)

0.048 0.958 0.846 0.652

0.014 0.299 0.319 0.286

1.05 2.61 2.33 1.92

(1.02 – 1.08) (1.45 – 4.68) (1.25 – 4.36) (1.10 – 3.36)

0.001 0.001 0.008 0.023

89 86 95 93

0.692 0.100 0.004 0.647 0.598 0.568 0.365

0.319 0.049 0.002 0.300 0.342 0.369 0.308

1.91 1.82 0.57 1.44

45 44 7 82 96 96 94 88 94 96

0.053 0.549 0.302 0.446 0.454 0.217 0.159 0.117 0.000

0.298 0.493 0.299 0.471 0.598 0.302 0.348 0.288 0.012

1.00 1.05 1.73 1.35 1.56 0.64 0.81 1.17 1.12 1.00

95 96 95 96 96

Multivariable model (n = 95 overall survival, n = 92 Age at surgery (years, continuous) 0.066 Diagnosed before 1989 1.157 >2 positive lymph nodes 0.934 Serosal/adnexal involvement Grossly positive nodes None Macroscopic, debulked 0.118 Macroscopic, not debulked 1.560

Parameter

Hazard

P value

SE

Ratio

(95% CI)

0.048 1.086 0.891 0.946

0.016 0.349 0.364 0.330

1.05 2.96 2.44 2.57

(1.02 – 1.08) (1.49 – 5.88) (1.20 – 4.98) (1.35 – 4.92)

0.003 0.002 0.014 0.004

(1.06 – 3.44) (0.93 – 3.56) (0.27 – 1.17) (0.79 – 2.63)

0.030 0.040 0.035 0.031 0.081 0.123 0.235

0.758 0.100 0.004 0.868 0.505 0.592 0.736

0.362 0.054 0.002 0.351 0.413 0.419 0.339

2.38 1.66 0.55 2.09

(1.20 – 4.74) (0.74 – 3.72) (0.24 – 1.26) (1.07 – 4.05)

0.036 0.063 0.060 0.013 0.222 0.158 0.030

(reference) (0.59 – 1.89) (0.66 – 4.55) (0.75 – 2.43 (0.62 – 3.94) (0.20 – 2.05) (0.45 – 1.45) (0.59 – 2.32) (0.64 – 1.98) (0.98 – 1.02)

0.860 0.266 0.312 0.344 0.448 0.473 0.648 0.683 0.989

0.445 0.797 0.498 0.103 0.630 0.143 0.162 0.144 0.003

0.347 0.565 0.350 0.480 0.729 0.355 0.408 0.334 0.014

1.00 1.56 2.22 1.65 1.11 0.53 0.87 1.18 1.16 1.00

(reference) (0.79 – 3.08) (0.73 – 6.71) (0.83 – 3.27) (0.43 – 2.84) (0.13 – 2.22) (0.43 – 1.74) (0.53 – 2.62) (0.60 – 2.22) (0.98 – 1.03)

0.200 0.158 0.155 0.830 0.387 0.688 0.691 0.666 0.856

disease-specific survival) 0.017 1.07 (1.03 – 1.11) 0.340 3.18 (1.63 – 6.19) 0.343 2.55 (1.30 – 4.99)

<0.001 <0.001 0.007

0.083 1.466 1.137 0.809

0.021 0.436 0.423 0.385

1.09 4.33 3.12 2.24

(1.04 – 1.13) (1.84 – 10.19) (1.36 – 7.15) (1.05 – 4.78)

<0.001 <0.001 0.007 0.036

0.315 1.924

0.428 0.740

1.00 1.37 6.85

(reference) (0.59 – 3.17) (1.61 – 29.18)

0.461 0.009

0.346 0.568

1.00 1.13 4.76

(reference) (0.57 – 2.21) (1.57 – 14.48)

patients receiving chemotherapy alone, two developed recurrence at multiple sites. Adverse events occurred in 24% of patients (Table 3). Surgical adverse events included small bowel obstruction requiring surgery (3%), return to the operating room within 30 days of surgery (3%), genitourinary obstruction (2%), wound dehiscence (1%), and lymphocyst requiring drainage (1%). Medical adverse events included small bowel obstruction treated medically (3%), deep venous thrombus (3%), congestive heart failure (2%), pulmonary embolus (1%), myocardial infarction (1%), pneumonia (1%), and pyelonephritis (1%).

Discussion In this large single institution study of 96 patients with stage IIIC endometrial cancer, we investigated clinical and pathological factors associated with nodal metastasis and survival. While the majority of patients with nodal meta-

Estimate

0.7329 0.006

stasis had either deep myometrial invasion or a poorly differentiated tumor, 3/96 (3%) of patients with nodal metastasis had grade 1 tumors with less than (1/2) myometrial invasion. Interestingly, all three patients had clinically suspicious lymph nodes at surgery. In a large GOG pilot study, patients with grade 1, superficially invasive cancers had a 2 –5% risk for nodal metastasis [14]. Our finding underscores the small but significant risk of lymph node metastasis in patients with ‘‘low risk’’ endometrial cancer and suggests the potential importance of lymph node assessment. Our 5-year overall and DSS rates of 47% and 55% are consistent with several prior published reports [2,5,6,15]. McMeekin et al. reported an overall 5-year survival rate of 65% for patients with stage IIIC endometrial cancer; the series was smaller (n = 47), collected over a more recent time interval (1989 – 1998), and had a slightly shorter median follow up time (37 months) than the current study [4]. Onda et al. and Katz et al. have reported superior outcomes among small series of patients with stage IIIC

694

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695

Table 3 Adverse events Number

(%)

All adverse events

23

(24.0)

Medical adverse events

14

(14.6)

3 3 2 1 1 1 1 5

(3.1) (3.1) (2.1) (1.0) (1.0) (1.0) (1.0) (5.2)

SBO medically treated Deep venous thrombosis Congestive heart failure Pneumonia Pulmonary embolus Acute myocardial infarction Pyelonephritis Other Surgical adverse events SBO requiring surgery Return to OR within 30 days Urinary obstruction Wound dehiscence Lymphocyst requiring drainage Other

16

(16.6)

3 3 2 1 1 7

(3.1) (3.1) (2.1) (1.0) (1.0) (7.3)

SBO: small bowl obstruction. OR: operating room.

disease using a combination of aggressive surgical treatment (complete pelvic and aortic lymphadenectomy to the renal vessels on all patients) and adjuvant radiotherapy with or without chemotherapy [3,16]. The present series is the largest reported single institution study of this subgroup and encompasses patients treated over a 29-year period. Although it is difficult to meaningfully compare the results from different single-institution studies, it is our hope that outcomes among patients with stage IIIC disease are continuing to improve as the appropriate adjuvant regimen for this group of patients is clarified by randomized controlled trials. We found that lymph nodes containing metastatic disease were grossly enlarged or suspicious in 53% of patients. Prior studies similarly reported suspicious lymphadenopathy in 39 – 63% of patients with stage IIIC disease [2,4,15]. Patients with macroscopic disease in lymph nodes had lower DSS (48%) than those with occult nodal disease (63%), and those in whom gross nodal disease could not be completely debulked had the lowest DSS at 5 years (43%). In multivariable analysis, a greater number of positive nodes (HR 3.12, P = 0.007) and the presence of gross residual disease not debulked following surgery (HR 6.85, P = 0.009) were significant risk factors for death of disease. While the value of surgical debulking is well established for ovarian cancer, less data are available regarding its utility for endometrial cancer. Bristow et al. previously reported that complete resection of macroscopic nodal disease was a significant predictor of DSS among 41 patients from a single institution with surgical stage IIIC endometrial cancer [2]. Mariani et al. found that patients with stage IIIC disease who had undergone aortic lymphadenectomy had improved survival

compared to those who had not [15]. In addition, several authors have reported that improved outcomes are associated with ‘‘optimal’’ debulking of gross intraperitoneal (stage IV) endometrial cancer [17 – 19]. Our data therefore add to the growing body of literature regarding the potential value of both lymphadenectomy and debulking surgery for endometrial cancer. Older age at the time of diagnosis and diagnosis prior to 1989 were associated with poor prognosis using multivariable analysis. Age is a consistent predictor of poor outcome in prior studies of stage IIIC endometrial cancer [2, 4]. Patients in our study treated prior to 1989 had a 5-year DSS of 36% compared to 74% for patients treated from 1989 forward. Bristow et al. similarly reported improved outcomes among patients treated in the latter years of their series [2]. We chose 1988 as a cutoff because the FIGO staging system changed during that year. It is possible that differences in median follow up between the two time periods (103 months prior to 1989, 48 months from 1989 forward) also affected the calculated DSS. It is also possible that the improvement in outcomes in the latter years of the series is related to advancements in the adjuvant therapies administered, such as techniques and equipment used for radiotherapy as well as the more frequent and updated use of chemotherapy regimens. Although the optimal adjuvant treatment for stage IIIC endometrial cancer has yet to be determined, recent studies suggest that patients with advanced endometrial cancer have improved survival with the administration of expanded chemotherapy regimens [20,21]. Metastasis to adnexa (24%), uterine serosa (13%), or peritoneal cytology (17%) was frequent in our series. We found poorer DSS among patients with positive adnexa, serosa, or peritoneal cytology, compared to patients with metastasis confined to the lymph nodes (P = 0.013), with 5-year DSS being 49% and 66% respectively. In multivariable analyses, adnexal or serosal involvement retained significance as a prognostic factor (HR 2.24, P = 0.036) for DSS. Serosal metastasis, adnexal metastasis, and positive cytology are all established risk factors for endometrial cancer and have previously been associated with lower survival in series of patients with stage IIIC disease [2,4]. In conclusion, our data suggest that complete debulking of gross nodal metastasis is associated with improved DSS among patients with stage IIIC endometrial cancer, and thus there may be value in performing this procedure in a systematic and thorough manner when feasible. The small number of patients left with residual disease in this series limits our power to demonstrate this conclusively. The potential benefit of debulking procedures should be weighed against each patient’s medical fitness for extensive surgical procedures. These data add to the growing literature documenting the possible therapeutic benefit of lymphadenectomy in the management of patients with endometrial cancer.

L.J. Havrilesky et al. / Gynecologic Oncology 99 (2005) 689 – 695

References [1] Jemal A, Murray T, Samuels A, et al. Cancer statistics, 2003. CA Cancer J Clin 2003;53(1):5 – 26. [2] Bristow RE, Zahurak ML, Alexander CJ, et al. FIGO stage IIIC endometrial carcinoma: resection of macroscopic nodal disease and other determinants of survival. Int J Gynecol Cancer 2003;13(5): 664 – 72. [3] Onda T, Yoshikawa H, Mizutani K, et al. Treatment of node-positive endometrial cancer with complete node dissection, chemotherapy and radiation therapy. Br J Cancer 1997;75(12):1836 – 41. [4] McMeekin DS, Lashbrook D, Gold M, et al. Analysis of FIGO Stage IIIc endometrial cancer patients. Gynecol Oncol 2001;81(2): 273 – 8. [5] Hirahatake K, Hareyama H, Sakuragi N, et al. A clinical and pathologic study on para-aortic lymph node metastasis in endometrial carcinoma. J Surg Oncol 1997;65(2):82 – 7. [6] Greven KM, Curran WJ, Whittington R Jr, et al. Analysis of failure patterns in stage III endometrial carcinoma and therapeutic implications. Int J Radiat Oncol Biol Phys 1989;17(1):35 – 9. [7] Kilgore LC, Partridge EE, Alvarez RD, et al. Adenocarcinoma of the endometrium: survival comparisons of patients with and without pelvic node sampling. Gynecol Oncol 1995;56(1):29 – 33. [8] Fanning J. Long-term survival of intermediate risk endometrial cancer (stage IG3, IC, II) treated with full lymphadenectomy and brachytherapy without teletherapy. Gynecol Oncol 2001;82(2): 371 – 4. [9] Mohan DS, Samuels MA, Selim MA, et al. Long-term outcomes of therapeutic pelvic lymphadenectomy for stage I endometrial adenocarcinoma. Gynecol Oncol 1998;70(2):165 – 71. [10] Cragun J, Havrilesky LJ, Calingaert B, Synan I, Secord AA, Soper JT, et al. Retrospective analysis of selective lymphadenectomy in apparent early-stage endometrial cancer. J Clin Oncol 2005;23(16).

695

[11] Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457 – 81. [12] Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22(4): 719 – 48. [13] Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187 – 220. [14] Creasman WT, Morrow CP, Bundy BN, et al. Surgical pathologic spread patterns of endometrial cancer. A Gynecologic Oncology Group Study. Cancer 1987;60(8 Suppl):2035 – 41. [15] Mariani A, Webb MJ, Galli L, et al. Potential therapeutic role of paraaortic lymphadenectomy in node-positive endometrial cancer. Gynecol Oncol 2000;76(3):348 – 56. [16] Katz LA, Andrews SJ, Fanning J. Survival after multimodality treatment for stage IIIC endometrial cancer. Am J Obstet Gynecol 2001;184(6):1071 – 3. [17] Chi DS, Welshinger M, Venkatraman ES, et al. The role of surgical cytoreduction in Stage IV endometrial carcinoma. Gynecol Oncol 1997;67(1):56 – 60. [18] Bristow RE, Zerbe MJ, Rosenshein NB, et al. Stage IVB endometrial carcinoma: the role of cytoreductive surgery and determinants of survival. Gynecol Oncol 2000;78(2):85 – 91. [19] Goff BA, Goodman A, Muntz HG, et al. Surgical stage IV endometrial carcinoma: a study of 47 cases. Gynecol Oncol 1994; 52(2):237 – 40. [20] Fleming GF, Brunetto VL, Cella D, et al. Phase III trial of doxorubicin plus cisplatin with or without paclitaxel plus filgrastim in advanced endometrial carcinoma: a Gynecologic Oncology Group Study. J Clin Oncol 2004;22(11):2159 – 66. [21] Randall ME, Brunetto G, Muss H, Mannell RS, Spirtos N, Jeffrey F, et al. Whole abdominal radiotherapy versus combination doxorubicin – Cisplatin chemotherapy in advanced endometrial carcinoma: a randomized phase III trial of the Gynecologic Oncology Group. Proc Am Soc Clin Oncol 2003;22:2.