CHAPTER 4 PARAAORTIC DISSEMINATION IN ENDOMETRIAL CANCER

CHAPTER 4 PARAAORTIC DISSEMINATION IN ENDOMETRIAL CANCER New surgical guidelines (Table 1) were instituted at Mayo Clinic on January 1, 2004. A recent prospective analysis of 252 consecutive patients who had surgery at Mayo Clinic, Rochester, MN during a 21-month period (January 2004 to September 2005) identified a 15% rate of positive lymph nodes. Paraaortic lymph nodes were involved in approximately 60% of patients (Mayo Clinic Unpublished data 2005). Similarly, in a previous retrospective study from Mayo Clinic (Mariani 2004), we observed that 47% of patients with positive pelvic lymph nodes had either positive paraaortic lymph nodes or subsequently experienced a paraaortic recurrence. Moreover, it has also been reported that approximately 55% to 57% of patients with positive lymph nodes have tumor in the paraaortic area (McMeekin 2001, Mariani 2001b). These observations should guide recommendations for and the extent of adjuvant radiotherapy in both definitively staged and inadequately staged patients. If external beam radiotherapy is selected as adjuvant therapy to decrease pelvic sidewall recurrence in non-staged patients, the above observations would suggest that the para-aortic area should routinely be included in the field of treatment. The following manuscript will identify predictors of paraaortic dissemination, with special emphasis on the fact that approximately half of the patients with detectable disease in the pelvic lymph nodes will also experience paraaortic disease. As noted above, this information will have implications for the postoperative treatment. 11

Table 1. Surgical guidelines for the treatment of endometrial cancer at Mayo Clinic, Rochester (2004-2006) (From Mariani et al. [2004b]) Treatment Hysterectomy, bilateral salpingo-oophorectomy, peritoneal cytology, bilateral pelvic and paraaortic lymphadenectomy (up to renal vessels) 1. Can omit lymphadenectomy if: (A) All the following: no myometrial invasion, endometrioid, no evidence of tumor outside the corpus (independently of grade or tumor diameter) (B) All the following: endometrioid, grade 1 or 2, 0 < myometrial invasion < 50%, tumor diameter 2 cm, no evidence of tumor outside the corpus 2. If non-endometrioid (serous, clear cells), add complete omentectomy, appendectomy, peritoneal biopsies (11 pairs: cul-de-sac, bladder peritoneum, R diaphragm, R/L colic gutters, R/L pelvic sidewall, small-/large-bowel serosa and mesentery, any suspicious area) Notes: R, Right; L, Left. 12

Gynecologic Oncology 92 (2004) 833 838 www.elsevier.com/locate/ygyno Endometrial carcinoma: paraaortic dissemination $ Andrea Mariani, a Gary L. Keeney, b Giacomo Aletti, c Maurice J. Webb, a Michael G. Haddock, d and Karl C. Podratz a, * a Section of Gynecologic Surgery, Mayo Clinic, Rochester, MN 55905, USA b Division of Anatomic Pathology, Mayo Clinic, Rochester, MN 55905, USA c Department of Mathematics, University of Milano, Milan, Italy d Division of Radiation Oncology, Mayo Clinic, Rochester, MN 55905, USA Received 11 June 2003 Abstract Objective. The objective of our study was to identify pathologic factors predictive of tumor dissemination to paraaortic lymph nodes (LNs) in endometrial carcinoma. The identification of the risk factors may potentially facilitate selection of patients for radical surgery or radiotherapy directed to the paraaortic area (PAA). Methods. The study population was a cohort from 612 consecutive patients with endometrial cancer surgically managed at our institution over a 10-year period. Tumor dissemination to the PAA was identified by selecting those patients who had either paraaortic LNs positive for disease at the time of primary surgery or those who subsequently experienced paraaortic failure or both (n = 41; the PA mets subgroup). Therefore, patients for whom no information was available about the status of paraaortic LNs but who had received adjuvant irradiation to the PAA and those for whom information was not available about sites of recurrent disease were excluded from the analysis, leaving 566 patients to compose the study population. Results. On the basis of univariate analysis, numerous pathologic variables were significantly ( P V 0.01) associated with PA mets. However, logistic regression analysis identified only two independent factors predictive of PA mets: positive pelvic LNs ( P < 0.001, OR = 5.00) and lymphovascular invasion (LVI) ( P = 0.01, OR = 1.99). Notably, only 2% of patients with negative pelvic LNs had PA mets compared with 47% of those with positive pelvic LNs ( P < 0.001). When both pelvic LNs and LVI were negative, only 0.8% of the patients had PA mets compared with 31% of patients for whom at least one of the two variables was positive ( P < 0.001). Conclusion. Positive pelvic LNs and LVI identify a subgroup of high-risk patients (approximately one sixth of the overall population) who potentially may benefit from formal lymphadenectomy or adjuvant therapy or both directed to the PAA. Furthermore, with 47% of patients with positive pelvic LNs having PA mets, unstaged patients at risk for pelvic LN involvement should be considered candidates for both pelvic and paraaortic external beam radiotherapy or surgical restaging. D 2004 Elsevier Inc. All rights reserved. Keywords: Adjuvant therapy; Endometrial cancer; Extended-field radiotherapy; Lymphovascular invasion; Myometrial invasion; Paraaortic lymphadenectomy; Paraaortic failure; Positive lymph nodes Introduction In the United States, endometrial adenocarcinoma is the most common malignancy of the female genital tract. It has been estimated that during calendar year 2003, 40,100 new cases and 6800 deaths will be attributed to this disease [1]. $ Presented as a poster at the 33rd Annual Meeting of the Society of Gynecologic Oncologists, Miami Beach, FL, March 16 to 20, 2002. There is no conflict of interest with this manuscript. * Corresponding author. Section of Gynecologic Surgery, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. E-mail address: podratz.karl@mayo.edu (K.C. Podratz). In 1988, the staging system of endometrial cancer was changed from clinical to surgical [2]. This decision implied that both pelvic and paraaortic lymph nodes (LNs) should be assessed in all patients with endometrial cancer [3]. However, we recently identified a subgroup of patients with lowrisk endometrial cancer, in which both pelvic and paraaortic lymphadenectomies may be safely avoided [4]. Furthermore, we observed that paraaortic lymphadenectomy may have a therapeutic value in a selected group of patients with positive LNs [5]. Therefore, the identification of patients at high risk for paraaortic LN invasion or paraaortic relapse would significantly increase the diagnostic and therapeutic 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2003.11.032

834 A. Mariani et al. / Gynecologic Oncology 92 (2004) 833 838 indexes for paraaortic lymphadenectomy. Furthermore, the identification of risk factors for metastases to the paraaortic LNs would allow initiation of potentially curative adjuvant therapy to the paraaortic area (PAA) [6,7]. The objective of our study was to identify pathologic factors predictive of tumor dissemination to the paraaortic LNs in patients who have endometrial carcinoma. The identification of the above risk factors may potentially facilitate the selection of patients for radical surgery or radiotherapy (or both) directed to the PAA. Materials and methods From 1984 to 1993, 815 patients with endometrial cancer were managed surgically at Mayo Clinic (Rochester, MN). We retrieved their records from the database and selected 612 patients with epithelial endometrial cancer who satisfied the following inclusion criteria: (1) treatment included hysterectomy and removal of existing adnexal structures and (2) no other malignancy was diagnosed within 5 years before or after the diagnosis of endometrial cancer (except for carcinoma in situ or skin cancer other than melanoma). Of the 612 study patients, information for most of the prognostic variables was not available for two, who therefore were not included in the study. Moreover, we reasoned that tumor dissemination to the PAA may be identified selecting those patients who had either paraaortic LNs positive for disease at the time of primary surgery or those who subsequently experienced paraaortic relapse or both (n = 41; the PA mets ). To eliminate a possible bias, we excluded from analysis 34 patients for whom no information was available about the status of paraaortic LNs but who received adjuvant irradiation to the PAA (also, no information was available about the site of recurrence in 1 of these patients). Furthermore, we excluded 10 patients for whom no information was available about the site or sites of recurrent disease. Therefore, the present analysis was conducted on a group of 566 patients. Staging was defined according to the International Federation of Obstetricians and Gynecologists (FIGO) surgical staging system [2]. For patients who received treatment before 1988, stage was determined retrospectively from surgical and pathologic assessment. Histologic classification was performed according to the World Health Organization classification [8]. Architectural grading was based on the degree of glandular differentiation in accordance with the FIGO guidelines [2]. Lymphovascular invasion (LVI) was considered to be present when tumor cells were within or attached to the wall of a capillary-like space. Hyperplasia associated with the tumor was diagnosed when an area of hyperplasia was found in the same sample adjacent to the tumor. Primary tumor diameter was measured macroscopically by the pathologist on fresh tissue. The method for measuring the tumor has been described elsewhere [4]. All hematoxylin- and eosin-stained sections of the tumor were reviewed retrospectively by one of us (G.L.K.) to confirm the original diagnosis of adenocarcinoma and to determine FIGO grade, histologic subtype, presence of tumor-associated hyperplasia, and LVI. All surgical procedures were the responsibility of a gynecologic oncologist. We defined lymph node dissection as the removal of at least one pelvic or paraaortic (or both) LN. LN dissection was usually performed in patients considered by the surgeon, from an intraoperative analysis of histologic grade of the tumor and depth of myometrial invasion determined from frozen section tissue review, to be at risk for LN metastasis. Paraaortic LN dissection usually was performed if intraoperative risk factors (such as deep myometrial invasion or presence of extrauterine disease) were identified by the pathologist. We considered LNs as grossly positive when they were described as palpable or suspicious in the surgical report. Postoperative adjuvant radiotherapy consisted of external pelvic, paraaortic, or abdominal irradiation or vaginal brachytherapy or some combination of these. Irradiation of the PAA was defined as extended-field radiotherapy. Occasionally, oral megestrol acetate was prescribed or adjuvant chemotherapy was administered. When follow-up information about survival and recurrence was not sufficiently detailed in the clinical records, information was obtained from death certificates and letters or telephone calls to patients and family physicians. We defined paraaortic relapse as tumor relapse in the PAA as the primary site of failure (eventually associated with other sites of relapse). For statistical purposes, endometrioid, endometrioid with squamous differentiation, and adenosquamous tumors were considered together. Grades 1 and 2 lesions were combined and compared with grade 3 lesions. The choice of 2 cm as a determinant for the analysis of primary tumor diameter was based on our previous experience [4]. Statistical analysis was performed with the Fisher exact test, v 2 analysis, and logistic regression analysis. Differences were considered statistically significant at P < 0.05. SAS System 6.10 statistical software was used for the analysis. Results The mean age (Fstandard deviation [SD]) of the 566 study patients was 64.7 F 10.6 years (range, 22 90 years), and the mean body mass index (FSD) was 30.8 F 8.4 kg/m 2 (range, 16.0 65.5 kg/m 2 ). Preoperative therapy was administered to 11 patients (2%): 8 received external radiotherapy and 3 chemotherapy. Pelvic LN dissection was performed in 337 patients (60%) and paraaortic dissection in 93 (16%) (overall, 342 patients had pelvic or paraaortic [or both] LN dissection). The number and percentage of patients undergoing pelvic or paraaortic lymphadenectomy (or both), stratified by grade, are summarized in Table 1. The mean number of LNs

A. Mariani et al. / Gynecologic Oncology 92 (2004) 833 838 835 Table 1 Number and percentages of patients undergoing pelvic (PLN) and paraaortic lymphadenectomy (PALN) stratified by grade PLN PALN Grade 1 2 (n = 458), no. of patients (%) Grade 3 (n = 108), no. of patients (%) 176 (38) 48 (44) + 4 (1) 1 (1) + 219 (48) 30 (28) + + 59 (13) 29 (27) harvested was 16.0 pelvic nodes (range, 1 55) and 6.2 paraaortic nodes (range 1 43). In particular, 228 (40%) patients had 10 or more pelvic LNs harvested, and 49 patients (9%) had 5 or more paraaortic LNs removed. Positive LNs were identified in 54 patients: positive pelvic LNs in 49 and positive paraaortic LNs in 19 (14 had both positive pelvic and paraaortic LNs). Of the five patients who had only positive paraaortic LNs, three had no information available about pelvic LN status and the other two had positive paraaortic LNs and negative pelvic LNs. Both patients who had negative pelvic and positive paraaortic LNs had stage IV tumor, with disease in the upper abdomen. Positive paraaortic LNs were described as palpable in 19 patients (3%). At the time of the operation, 11 of these patients had paraaortic LN dissection and metastases were found in the PAA in 8 (73%) of them. On the basis of adverse prognostic factors, adjuvant radiotherapy was administered to 184 patients (33%), hormone therapy to 28 (5%), and adjuvant systemic cytotoxic therapy to 18 (3%). Of the 184 patients who received adjuvant radiotherapy, 170 had external beam radiotherapy (49 had associated vaginal brachytherapy), 11 had intraperitoneal administration of phosphorus 32, and 3 had adjuvant vaginal brachytherapy only. Of the 170 patients who had adjuvant external beam radiotherapy, all received pelvic radiotherapy, 13 received extended-field radiotherapy, and 51 received whole abdominal irradiation. The median dose of irradiation administered to the pelvis, PAA, and the whole abdomen was 5040, 4500, and 2550 cgy, respectively. Table 2 Associations between paraaortic failures (PAF), positive paraaortic lymph nodes (PALN), and different types of therapies directed to the paraaortic area a Therapy PALN PAF No. of patients PALD only (n = 81) Pos No 4 (1 CT, 1 HT) Pos Yes 3 (2 HT) Neg No 73 (4 HT) Neg Yes 1 PALD + PART (n = 12) Pos No 9 (2 CT) Pos Yes 3 (1 HT) PART only (n = 1) NA Yes 1 No PALD and no PART NA No 452 (13 CT, 16 HT) (n = 472) NA Yes 20 (2 CT, 4 HT) CT, chemotherapy; HT, hormone therapy; NA, not available; Neg, negative; Pos, positive. a That is, paraaortic lymphadenectomy (PALD) and paraaortic radiotherapy (PART). Table 3 Identification of independent risk factors for either positive paraaortic lymph nodes or paraaortic failure or both according to regression analysis Independent risk factor Odds ratio P value 95% CI Positive pelvic node(s) 5.00 <0.001 2.78 9.01 Lymphovascular invasion 1.99 0.01 1.15 3.65 CI, confidence interval. After a median follow-up of 74 months (range, 0 155 months), 28 relapses were documented in the PAA (5%). The median time to paraaortic relapse was 12 months. Of the 41 patients with PA mets, 19 had positive paraaortic LNs (six of them also had a subsequent relapse in the PAA) and the other 22 had relapse in the PAA (one had negative paraaortic nodes, and no information was available about the status of the PAA for 21). Of the 41 patients with PA mets, only 12 (29%) had palpable paraaortic LNs at the time of the operation. The associations between paraaortic failures, positive paraaortic LNs, and the different types of therapies directed to the PAA (i.e., paraaortic lymphadenectomy and paraaortic radiotherapy) are summarized in Table 2. On the basis of univariate analysis, the following clinical and pathologic factors were significantly ( P V 0.01) associated with PA mets: stage IV disease, cervical stromal invasion, adnexal involvement, myometrial invasion z 66%, tumor diameter > 2 cm, tumor invasion of the entire uterine cavity, positive peritoneal cytology, positive pelvic LNs, histologic grade 3, nonendometrioid histologic subtype, and LVI. Age, body mass index, and hyperplasia associated with the tumor were not significant ( P > 0.05) predictors of PA mets on univariate analysis. However, logistic regression analysis, using the stepwise forward inclusion variable selection method, identified only two independent factors predictive of PA mets: positive pelvic LNs and LVI (Table 3). Notably, 2% of patients with negative pelvic LNs had PA mets compared with 47% of those with positive pelvic LNs. Moreover, 2% of patients without LVI had PA mets, compared with 33% with LVI ( P < 0.001) (Tables 4 and 5). Patients with positive pelvic LNs or LVI (or both) represented 17% (n = 98; 15 had LVI without assessment of pelvic LNs) of the overall population of 566 patients with endometrial Table 4 Presence of either positive paraaortic lymph nodes or paraaortic relapse or both (PA mets) according to lymphovascular invasion (LVI) and pelvic lymph node status Characteristic Patients % PA mets P value No. % a Pelvic lymph nodes <0.001 Negative 288 85 2 Positive 49 15 47 Missing 229 LVI <0.001 Yes 83 15 33 No 461 85 2 Missing 22 a Percentages excluded missing cases.

836 A. Mariani et al. / Gynecologic Oncology 92 (2004) 833 838 Table 5 Positive predictive value (PPV) and negative predictive value (NPV) of pelvic lymph node status and lymphovascular invasion (LVI) for the prediction of positive paraaortic nodes or paraaortic relapse or both a cancer included in this study and 25% (n = 83) of the 337 patients for whom information was available about both pelvic LN status and LVI. Considering only this subgroup of 337 patients, we observed that in the presence of negative pelvic LNs and no LVI, only 2 of 254 patients (0.8%) had PA mets compared with 26 of 83 patients (31%) for whom at least one of these two variables was positive ( P < 0.001) (Table 5). Specifically, 8 of 49 patients (16%) with only one variable positive and 18 of 34 (53%) with both variables positive had PA mets. Furthermore, of the 49 patients with only one variable positive, 3 of 34 patients (9%) with negative pelvic nodes but with LVI had PA mets, compared with 5 of 15 (33%) with positive pelvic nodes but without LVI. Discussion % of overall population PPV (%) NPV (%) LVI 20 31 97 Positive pelvic LNs 15 47 98 Present study criteria b 25 31 99 LNs, lymph nodes. a Considers the population of 337 patients with information available for both the two variables. b At least one of the following two variables: LVI, positive pelvic LNs. The reported percentage of patients with corpus cancer and documented tumor in the PAA has ranged from 0% to 19% [6,9 17]. The recognized differences among the studies in the literature probably reflect different study populations and the extent of the lymphadenectomies. It has been suggested that paraaortic lymphadenectomy should be performed in all patients who have endometrial cancer [10,11,18]. However, because of the severe medical comorbid conditions that occur with corpus cancer in many patients [19] and the potential increase in operative times [5], blood loss [20], and surgical complications [21] associated with LN dissection, paraaortic LNs usually are not routinely assessed in all patients with endometrial cancer [12,22,23]. We recently suggested that, for diagnostic purposes, pelvic and paraaortic lymphadenectomy may be safely avoided in only a selected population at low risk for LN metastases (i.e., endometrioid, grades 1 or 2 tumor, with less than 50% myometrial invasion, and tumor diameter V2 cm) [4]. Moreover, we reported that paraaortic lymphadenectomy may have a therapeutic value in patients with positive pelvic or paraaortic LNs [5]. It is likely that the therapeutic value of radical surgery in the PAA is confined to patients at risk for paraaortic dissemination. Moreover, the identification of predictors of tumor dissemination to the PAA may allow the selection of patients with no information available about the status of the paraaortic LNs who may potentially benefit from adjuvant extended-field irradiation [6,7]. As noted in Materials and methods, we reasoned that patients who had documented PA mets or those who subsequently experienced a paraaortic recurrence (PA mets) potentially may derive diagnostic or therapeutic (or both) benefit from paraaortic lymphadenectomy or may potentially benefit from extended-field irradiation. We eliminated from the present analysis patients for whom no information was available about paraaortic LN status but who received adjuvant radiotherapy to the PAA. This was done to exclude the possible bias of a potentially therapeutic procedure [6,7] administered to a subgroup of at-risk patients. Histologic grade, depth of myometrial invasion, cervical invasion, adnexal metastasis, and tumor size have been reported to be significant predictors of paraaortic LN metastases [7,9,11,16,17]. In addition, we recently reported that paraaortic relapses are correlated significantly with LN status [24]. The frequency of paraaortic relapse reported is between 0% and 17% [7,13,22]. Potentially, paraaortic relapse can be minimized by performing formal paraaortic lymphadenectomy [5] and administering adjuvant paraaortic irradiation [6,7] or chemoradiation [15]. In the present study, we identified several clinical and pathologic factors associated with paraaortic dissemination in patients with endometrial cancer. However, positive pelvic LNs and LVI were the only independent predictors of paraaortic dissemination based on logistic regression analysis (Table 3). With regard to the presence of positive pelvic LNs, we observed a 47% rate of paraaortic dissemination in the subgroup of patients with metastasis to the pelvic LNs (Tables 4 and 5). Pelvic node metastases generally have been reported to be highly predictive of paraaortic LN involvement, with a rate of paraaortic positivity ranging from 32% to 80% in the different series of patients with positive pelvic LNs [9 11,14 17]. The various pathologic patterns of pelvic LN metastases [25] and the number of positive pelvic LNs [16,18,25] can significantly influence paraaortic LN status. In our study, pelvic LN status had both a modest positive predictive value for paraaortic dissemination and a high negative predictive value (Table 5). In fact, only 2% of patients with negative pelvic LNs had PA mets in our series (Tables 4 and 5). Although a direct route of lymphatic spread from the corpus uteri to the PAA via the infundibulopelvic ligament has been hypothesized from anatomical and autopsy studies [26,27], metastases to the paraaortic LNs without involvement of pelvic LNs is uncommon in most surgical series; this has been reported to range from 0% to 6% of cases [9 11,13 16]. Collectively, the frequency of paraaortic nodal involvement in the absence of pelvic node metastases approximates 2% (Table 6); this was again confirmed in the present study. With regard to LVI, it is well recognized that the presence of tumor cells within capillary-like spaces is a strong predictor of both LN metastasis [9,11,16,17,28] and lym-

A. Mariani et al. / Gynecologic Oncology 92 (2004) 833 838 837 Table 6 Literature review of the prevalence of positive aortic nodes and negative pelvic nodes (PANNPN) in patients with corpus cancer according to the overall population with endometrial cancers (Tot. Pt.) and the subgroup of patients with positive lymph nodes (Tot. Pos. LNs) Study Stage PANNPN/ Tot. Pt. % PANNPN/ Tot. Pos. LNs Morrow et al. [9] Clinical I II 18/802 2 18/111 16 Yokoyama et al. [11] I III 4/63 6 4/18 22 Onda et al. [15] I III 2/173 1 2/30 7 Larson et al. [10] I IV 0/50 0 0/10 0 Ayhan et al. [14] Clinical I 6/209 3 6/36 17 Fanning et al. [13] I III 0/60 0 0/5 0 Hirahatake et al. [16] I IV 2/200 1 2/42 5 Present series I IV 2/90 a 2 2/51 a 4 Total 34/1647 2 34/303 11 a We excluded the three patients who had positive paraaortic nodes but for whom no data were available about the pelvis. phatic relapse [24]. Specifically, it has been reported that 50% to 90% of patients who have positive paraaortic LNs have demonstrable LVI in the primary tumor [11,28]. Furthermore, LVI has prognostic importance even after stratifying for LN status [28]. Moreover, LVI is correlated significantly with microscopic node metastases identified with immunohistochemical staining in patients declared to have negative pelvic LNs from routine histologic assessments [29]. These observations may explain in part the occurrence of positive paraaortic LNs (albeit uncommon) with negative pelvic LNs. Because both LVI and positive pelvic LNs are the strongest predictors of tumor invasion of the PAA, these two variables may be potentially used for deciding whether extended-field irradiation may be indicated as adjuvant therapy in patients for whom information about the status of the paraaortic LNs is not available. In our analysis, disease was either detected via staging or as subsequent relapse in the PAA in 47% of patients with positive pelvic LNs (Table 5), with an odds ratio of 5.00 (Table 3). For this reason, patients with positive pelvic LNs in the absence of paraaortic node dissection should be informed of this potential risk and counseled about potential benefits of extended-field irradiation and its associated morbidity versus the risk of relapse. The presence of LVI may be used as an adjunct for a better risk assessment. In general clinical practice, pelvic lymphadenectomy is not frequently performed in patients with endometrial cancer. Following such management, whole pelvic external beam radiotherapy is generally administered according to uterine pathologic risk factors [30]. However, advocates of this treatment scheme do not take into account that approximately half of the patients with positive pelvic LNs will also have disease in the PAA (Table 5). Hence, this type of management will lead to overtreatment of a significant percentage of patients who do not have occult pelvic sidewall disease [31] and, based on our data (Table 5), undertreatment of nearly % half of the patients with pelvic nodal involvement who have associated metastatic disease in the PAA. Conclusion Positive pelvic LNs and LVI identify a subgroup of highrisk patients (approximately one-sixth of the overall population) who potentially might benefit from formal lymphadenectomy or adjuvant therapy or both directed to the PAA. Furthermore, with 47% of patients with positive pelvic LNs having PA mets, patients with unstaged disease at risk for pelvic LN involvement should be considered candidates for both pelvic and paraaortic external beam radiotherapy or for surgical restaging. Acknowledgments Supported by the Mayo Cancer Center (P30CA15083) and the Rochester Research Committee, Mayo Foundation. References [1] Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer statistics, 2003. CA Cancer J Clin 2003;53:5 26. [2] FIGO stages 1988 revision: definitions of the clinical stages in carcinoma of the vulva. Gynecol Oncol 1989;35:125 6. [3] Barnes MN, Kilgore LC. Complete surgical staging of early endometrial adenocarcinoma: optimizing patient outcomes. Semin Radiat Oncol 2000;10:3 7. [4] Mariani A, Webb MJ, Keeney GL, Haddock MG, Calori G, Podratz KC. Low-risk corpus cancer: is lymphadenectomy or radiotherapy necessary? Am J Obstet Gynecol 2000;182:1506 19. [5] Mariani A, Webb MJ, Galli L, Podratz KC. Potential therapeutic role of para-aortic lymphadenectomy in node-positive endometrial cancer. Gynecol Oncol 2000;76:348 56 (doi:10.1006/gyno.1999.5688). [6] Rose PG, Cha SD, Tak WK, Fitzgerald T, Reale F, Hunter RE. Radiation therapy for surgically proven para-aortic node metastasis in endometrial carcinoma. Int J Radiat Oncol Biol Phys 1992;24: 229 33. [7] Hicks ML, Piver MS, Puretz JL, Hempling RE, Baker TR, McAuley M, et al. Survival in patients with paraaortic lymph node metastases from endometrial adenocarcinoma clinically limited to the uterus. Int J Radiat Oncol Biol Phys 1993;26:607 11. [8] Scully RE, Bonfiglio TA, Kurman RJ, Silverberg SG, Wilkinson EJ. Histological typing of female genital tract tumours. 2nd ed. Berlin: Springer-Verlag; 1994. p. 13 8. [9] Morrow CP, Bundy BN, Kurman RJ, Creasman WT, Heller P, Homesley HD, et al. Relationship between surgical-pathological risk factors and outcome in clinical stage I and II carcinoma of the endometrium: a Gynecologic Oncology Group study. Gynecol Oncol 1991;40:55 65. [10] Larson DM, Johnson KK. Pelvic and para-aortic lymphadenectomy for surgical staging of high-risk endometrioid adenocarcinoma of the endometrium. Gynecol Oncol 1993;51:345 8 (doi:10.1006/gyno. 1993.1301). [11] Yokoyama Y, Maruyama H, Sato S, Saito Y. Indispensability of pelvic and paraaortic lymphadenectomy in endometrial cancers. Gynecol Oncol 1997;64:411 7 (doi:10.1006/gyno.1996.4573). [12] Belinson JL, Lee KR, Badger GJ, Pretorius RG, Jarrell MA. Clinical stage I adenocarcinoma of the endometrium Analysis of recurrences

838 A. Mariani et al. / Gynecologic Oncology 92 (2004) 833 838 and the potential benefit of staging lymphadenectomy. Gynecol Oncol 1992;44:17 23. [13] Fanning J, Nanavati PJ, Hilgers RD. Surgical staging and high dose rate brachytherapy for endometrial cancer: limiting external radiotherapy to node-positive tumors. Obstet Gynecol 1996;87:1041 4. [14] Ayhan A, Tuncer ZS, Tuncer R, Yuce K, Kucukali T. Tumor status of lymph nodes in early endometrial cancer in relation to lymph node size. Eur J Obstet Gynecol Reprod Biol 1995;60:61 3. [15] Onda T, Yoshikawa H, Mizutani K, Mishima M, Yokota H, Nagano H, et al. Treatment of node-positive endometrial cancer with complete node dissection, chemotherapy and radiation therapy. Br J Cancer 1997;75:1836 41. [16] Hirahatake K, Hareyama H, Sakuragi N, Nishiya M, Makinoda S, Fujimoto S. A clinical and pathologic study on para-aortic lymph node metastasis in endometrial carcinoma. J Surg Oncol 1997;65: 82 7. [17] Creasman WT, Morrow CP, Bundy BN, Homesley HD, Graham JE, Heller PB. Surgical pathologic spread patterns of endometrial cancer. A Gynecologic Oncology Group study. Cancer (Suppl.) 1987;60: 2035 41. [18] McMeekin DS, Lashbrook D, Gold M, Scribner DR, Kamelle S, Tillmanns TD, et al. Nodal distribution and its significance in FIGO stage IIIc endometrial cancer. Gynecol Oncol 2001;82:375 9 (doi:10.1006/gyno.2001.6278). [19] Orr Jr JW, Holimon JL, Orr PF. Stage I corpus cancer: is teletherapy necessary? Am J Obstet Gynecol 1997;176:777 88. [20] Berclaz G, Hanggi W, Kratzer-Berger A, Altermatt HJ, Greiner RH, Dreher E. Lymphadenectomy in high risk endometrial carcinoma stage I and II: no more morbidity and no need for external pelvic radiation. Int J Gynecol Cancer 1999;9:322 8. [21] Cliby WA, Clarke-Pearson DL, Dodge R, et al. Acute morbidity and mortality associated with selective pelvic and para-aortic lymphadenectomy in the surgical staging of endometrial adenocarcinoma. J Gynecol Tech 1995;1:19 26. [22] Kim YB, Niloff JM. Endometrial carcinoma: analysis of recurrence in patients treated with a strategy minimizing lymph node sampling and radiation therapy. Obstet Gynecol 1993;82:175 80. [23] Faught W, Krepart GV, Lotocki R, Heywood M. Should selective paraaortic lymphadenectomy be part of surgical staging for endometrial cancer? Gynecol Oncol 1994;55:51 5 (doi:10.1006/gyno. 1994.1246). [24] Mariani A, Webb MJ, Keeney GL, Aletti G, Podratz KC. Predictors of lymphatic failure in endometrial cancer. Gynecol Oncol 2002; 84:437 42 (doi:10.1006/gyno.2001.6550). [25] Mariani A, Webb MJ, Rao SK, Lesnick TG, Podratz KC. Significance of pathologic patterns of pelvic lymph node metastases in endometrial cancer. Gynecol Oncol 2001;80:113 20 (doi:10.1006/ gyno.2000.6050). [26] Burke TW, Levenback C, Tornos C, Morris M, Wharton JT, Gershenson DM. Intraabdominal lymphatic mapping to direct selective pelvic and paraaortic lymphadenectomy in women with high-risk endometrial cancer: results of a pilot study. Gynecol Oncol 1996;62:169 73 (doi:10.1006/gyno.1996.0211). [27] Henriksen E. The lymphatic dissemination in endometrial carcinoma. A study of 188 necropsies. Am J Obstet Gynecol 1975;123:570 6. [28] Inoue Y, Obata K, Abe K, Ohmura G, Doh K, Yoshioka T, et al. The prognostic significance of vascular invasion by endometrial carcinoma. Cancer 1996;78:1447 51. [29] Fishman A, Klein A, Zemer R, Zimlichman S, Bernheim J, Cohen I, et al. Detection of micrometastasis by cytokeratin-20 (reverse transcription polymerase chain reaction) in lymph nodes of patients with endometrial cancer. Gynecol Oncol 2000;77:399 404 (doi:10.1006/ gyno.2000.5781). [30] Maggino T, Romagnolo C, Landoni F, Sartori E, Zola P, Gadducci A. An analysis of approaches to the management of endometrial cancer in North America: a CTF study. Gynecol Oncol 1998;68:274 9 (doi:10.1006/gyno.1998.4951). [31] Podratz KC, Mariani A, Webb MJ. Staging and therapeutic value of lymphadenectomy in endometrial cancer (editorial). Gynecol Oncol 1998;70:163 4 (doi:10.1006/gyno.1998.5150).