A nomogram for predicting lymph node metastasis of presumed stage I and II endometrial cancer

Research

www. AJOG.org

ONCOLOGY

A nomogram for predicting lymph node metastasis of presumed stage I and II endometrial cancer Sofiane Bendifallah, MD; Anne Sophie Genin, MD; Iptissem Naoura, MD; Nathalie Chabbert Buffet, MD, PhD; Francoise Clavel Chapelon, PhD; Bassam Haddad, MD, PhD; Dominique Luton, MD, PhD; Emile Darai, MD, PhD; Roman Rouzier, MD, PhD; Martin Koskas, MD OBJECTIVE: Our objective was to develop a nomogram based on path-

RESULTS: The nomogram showed good discrimination with an area

ological hysterectomy characteristics to provide a more individualized and accurate estimation of lymph node metastasis in endometrial cancer.

under the receiver operating characteristic curve of 0.80 (95% confidence interval, 0.79 – 0.81) in the training set and 0.79 (95% confidence interval, 0.78 – 0.80) in the validation set. The nomogram was well calibrated.

STUDY DESIGN: Data from the Surveillance, Epidemiology, and End Results database for 18,294 patients who underwent hysterectomy and lymphadenectomy were analyzed. A multivariate logistic regression analysis of selected prognostic features was performed, and a nomogram to predict lymph node metastasis was constructed. A cohort of 434 patients was used for the external validation.

CONCLUSION: We developed a nomogram based on 5 clinical and

pathological characteristics to predict lymph node metastasis with a high concordance probability. Key words: endometrial cancer, lymph node metastasis, nomogram

Cite this article as: Bendifallah S, Genin AS, Naoura I, et al. A nomogram for predicting lymph node metastasis of presumed stage I and II endometrial cancer. Am J Obstet Gynecol 2012;207:197.e1-8.

E

ndometrial cancer is the most common malignancy of the female genital tract and the seventh most common cause of death from cancer in women in

From the Department of Obstetrics and Gynecology, Hospital Tenon, Assistance Publique-Hôpitaux de Paris (Drs Bendifallah, Chabbert Buffet, Darai, and Rouzier); the Department of Obstetrics and Gynecology, Hospital Bichat, Assistance Publique-Hôpitaux de Paris (Drs Naoura, Luton, and Koskas); Institut National de la Santé et de la Recherche Médicale Unité Mixte de Recherche en Santé 938, UPMC University (Drs Chabbert Buffet, Darai, Rouzier, and Koskas); and Paris Diderot University (Drs Luton and Koskas); Institut National de la Santé et de la Recherche Médicale Unité Mixte de Recherche 1018, Villejuif (Dr Clavel Chapelon); and the Department of Obstetrics and Gynecology, Centre Hospitalier Intercommunal Créteil, Créteil (Drs Genin and Haddad), France. Received Feb. 19, 2012; revised April 27, 2012; accepted June 28, 2012. The authors report no conflict of interest. Reprints: M. Koskas, MD, Department of Gynecology, Bichat University Hospital, 46 Rue Henri Huchard, 75018 Paris, France. martin. [email protected]. 0002-9378/$36.00 © 2012 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2012.06.080

Western countries.1 The vast majority of women with endometrial cancer are diagnosed with early-stage tumors that are associated with a good prognosis; however, a subgroup of women have more aggressive neoplasms and are at increased risk of relapse and death.2,3 The current surgical approach for the treatment of endometrial cancer is still debated. In particular, indications for lymph node (LN) dissection have not been established and include omitting lymphadenectomy in patients with presumed early-stage and low-grade disease, performing lymphadenectomy only in patients who are at high risk for nodal metastases and performing a complete lymphadenectomy in all uterine cancer patients irrespective of grade and depth of myometrial invasion. Several authors4-6 have suggested that complete lymphadenectomy may be associated with improved survival outcomes, particularly for patients with LN metastases. However, the retrospective nature of most of these studies has rendered their results equivocal. In contrast, the results of 2 recent randomized clinical trials showed that lymphadenectomy did not provide an overall or recurrencefree survival benefit in the early stages of

disease.7,8 These trials have been criticized for the following reasons: a limited effort with respect to the extent of dissection and lymph node evaluation, too many low-risk patients, and no direct decision on adjuvant therapy based on lymphadenectomy result. Sophisticated imaging techniques (eg, positron emission tomography/ computed tomography) offer a less invasive and morbid means of evaluating LN status.9 However, the sensitivity of such techniques is more limited than that of lymphadenectomy, which is considered the most accurate way to detect the presence of LN metastases.10 An evidencebased algorithm for surgical treatment decisions could be helpful, especially when the preoperative health status of the patient is unfavorable in terms of anesthetic status or comorbidities. Clinical and pathological variables (eg, myometrial invasion, histological type and grade) have been reported to be associated with the risk of LN metastasis. However, individually, none of these characteristics can be used to identify a subset of patients for whom LN resection is unnecessary. Prognostic tools such as nomograms that use statistical models to combine variables to obtain

SEPTEMBER 2012 American Journal of Obstetrics & Gynecology

197.e1

Research

Oncology

www.AJOG.org M ATERIALS AND M ETHODS

TABLE 1

Characteristics of the study population Variables

Training set (n ⴝ 18,294)

Validation set (n ⴝ 434)

P value

Age at diagnosis, y

.....................................................................................................................................................................................................................................

Median (mean)

62 (62.24)

66 (64.99)

Range

18-97

38-87

⬍ .001

..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

⬍ .001

Race

.....................................................................................................................................................................................................................................

White

15,801 (86.4%)

419 (96.5%)

.....................................................................................................................................................................................................................................

African American

929 (5.1%)

2 (0.5%)

1564 (8.5%)

13 (3.0%)

.....................................................................................................................................................................................................................................

Other

..............................................................................................................................................................................................................................................

⬍ .001

Tumor grade

.....................................................................................................................................................................................................................................

1

6512 (35.6%)

221 (50.9%)

2

7092 (38.8%)

153 (35.3%)

3-4

4690 (25.6%)

60 (13.8%)

..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

⬍ .001

Histologic subtype

.....................................................................................................................................................................................................................................

Adenocarcinoma

16,998 (92.9%)

399 (91.9%)

.....................................................................................................................................................................................................................................

Clear-cell

307 (1.7%)

19 (4.4%)

Papillary serous

577 (3.1%)

14 (3.2%)

Carcinosarcoma

412 (2.3%)

2 (0.5%)

Study population The mathematical model was developed using data from the Surveillance, Epidemiology, and End Results (SEER) database. All data were publicly available, deidentified, and exempt from institutional review board review. We identified patients with histologically proven endometrial cancer diagnosed between 1988 and 2007. Case listings were generated using codes specific for both clinical (age at diagnosis and race) and tumor characteristics (primary organ site, extent of disease, histologic subtype, grade differentiation, and number of regional LN examined). We included women who underwent primary operations for endometrial cancer with at least hysterectomy and extensive lymphadenectomy (more than 10 regional lymph nodes removed).12

..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

Primary site tumor characteristics

.024

.....................................................................................................................................................................................................................................

Endometrium

3936 (21.5%)

82 (18.9%)

ⱕ50%

9033 (49.4%)

198 (45.6%)

⬎50%

4039 (22.1%)

120 (27.6%)

1286 (7.0%)

34 (7.8%)

............................................................................................................................................................................................................................ ............................................................................................................................................................................................................................

.....................................................................................................................................................................................................................................

Cervical stroma invasion

..............................................................................................................................................................................................................................................

⬍ .001

FIGO 2009 stages

.....................................................................................................................................................................................................................................

I

............................................................................................................................................................................................................................

IA

12,578 (68.8%)

148 (34.1%)

IB

3378 (18.5%)

193 (44.5%)

............................................................................................................................................................................................................................ .....................................................................................................................................................................................................................................

II

............................................................................................................................................................................................................................

II

960 (5.2%)

43 (9.9%)

.....................................................................................................................................................................................................................................

III

............................................................................................................................................................................................................................

IIIA

—

—

IIIB

—

—

............................................................................................................................................................................................................................ ............................................................................................................................................................................................................................

IIIC

1378 (7.5%)

50 (11.5%)

..............................................................................................................................................................................................................................................

⬍ .001

Extent of lymphadenectomy

.....................................................................................................................................................................................................................................

Median (mean)

18.00 (20.85)

Range

10-90.0

10.00 (11.26)

.....................................................................................................................................................................................................................................

0-80.0

..............................................................................................................................................................................................................................................

FIGO, International Federation of Gynecology and Obstetrics. Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

the most accurate and reliable predictions have been adopted in several oncologic disciplines.11 The aim of this study was to develop a nomogram to predict LN status for en197.e2

dometrial cancer by combining selected clinical and pathological risk factors using a multivariate model. It could be used when hysterectomy has been performed and lymphadenectomy omitted.

American Journal of Obstetrics & Gynecology SEPTEMBER 2012

Development of the nomogram To develop a well-calibrated and exportable nomogram to predict the metastatic LN risk, we built a logistic regression model (LRM) using a training cohort (ie, a training set) of 18,294 patients, which was extracted from the SEER database using the previous criteria, and we validated the model with an independent validation cohort (ie, a validation set). Univariate and multivariate logistic regression analyses were used to test the association between the metastatic LN risk and clinicopathological characteristics. The complexity of the model was controlled using Akaike information criteria. A P ⬍ .05 was considered significant. The following variables were included in the analysis: age at diagnosis; race (white, African American, and other); histological subtype (adenocarcinoma, papillary serous, clear cell, and carcinosarcoma); grade differentiation (well ⫽ 1, moderate ⫽ 2, poor ⫽ 3-4); and primary site tumor characteristics (endometrium, 50% or less myometrial invasion, greater than 50% myometrial invasion, cervical stromal invasion). The predictive accuracy of the model was assessed in terms of its discrimination and calibration. Discrimination is the ability to differentiate between women with positive metastatic LN and

Oncology

www.AJOG.org women with negative metastatic LN, and it is measured using the receiver operating characteristic curve and summarized by the area under the curve (AUC). An AUC of 1.0 indicates perfect concordance, whereas an AUC of 0.5 indicates no relationship. Calibration (ie, the agreement between the observed outcome frequencies and the predicted probabilities) was studied using graphical representations of the relationship between the observed outcome frequencies and the predicted probabilities (calibration curves). A calibration curve can be approximated by a regression line with intercept ␣ and slope ␤. These parameters can be estimated in an LRM with the event as the outcome and the linear predictor as the only covariate. Well-calibrated models have ␣ ⫽ 0 and ␤ ⫽ 1. Therefore, a sensible measure of calibration is a likelihood ratio statistic testing the null hypothesis that ␣ ⫽ 0 and ␤ ⫽ 1. The statistic has a ␹2 distribution with 2 df (unreliability [U] statistic). We also evaluated the average (E average) and maximal errors (E maximal) between predictions and observations, which were obtained from a calibration curve.

Validation An internal validation of the accuracy estimates was performed with 200 bootstrap resamples to obtain relatively unbiased estimates. Bootstrapping allows for the simulation of the performance of the nomogram if it was applied to future patients and provides an estimate of the average optimism of the AUC. For external validation, the model was applied on a sample of 434 patients referred to as the validation set, which was developed from a single database that recorded patient data from 4 institutions: Tenon Hospital (Paris, France; 116 patients), Bichat Hospital (Paris, France; 43 patients), E3N cohort (Paris, France; 200 patients), and Creteil Hospital (Paris, France; 75 patients). Patients were included if they had available data for the components of the nomogram. In addition, to study the predictive accuracy of the model in terms of calibration, patients were clustered into deciles according to their nomogram score. For

Research

TABLE 2

Predictors of metastatic lymph nodes in multivariable analysis Variables

OR (95% CI)

Age at diagnosis, y

0.99 (0.98–0.99)

P value ⬍ .001

..............................................................................................................................................................................................................................................

Race

.028

.....................................................................................................................................................................................................................................

African American

Referent

.....................................................................................................................................................................................................................................

White

0.74 (0.59–0.94)

Others

0.89 (0.66–1.20)

..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

⬍ .001

Tumor grade

.....................................................................................................................................................................................................................................

1

Referent

.....................................................................................................................................................................................................................................

2

1.70 (1.44–2.00)

3-4

2.62 (2.21–3.11)

..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

⬍ .001

Histological subtype

.....................................................................................................................................................................................................................................

Adenocarcinoma

Referent

.....................................................................................................................................................................................................................................

Carcinosarcoma

1.347 (1.01–1.86)

Clear cell

1.89 (1.33–2.66)

Papillary serous

2.84 (2.21–3.64)

..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

⬍ .001

Primary site tumor characteristics

.....................................................................................................................................................................................................................................

Endometrium

Referent

.....................................................................................................................................................................................................................................

⬍50%

6.67 (4.35–10.00)

⬎50%

31.33 (23.48–41.00)

Cervical stroma invasion

46.13 (35.57–58.50)

..................................................................................................................................................................................................................................... ..................................................................................................................................................................................................................................... ..............................................................................................................................................................................................................................................

CI, confidence interval; OR, odds ratio. Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

each decile group, we calculated the difference between the predicted and the observed LN metastasis probability. The mean error between the predicted and the observed LN metastasis probability was the sum of the differences for each decile group divided by 10.

Other statistical tests The categorical and numerical variables were analyzed using the ␹2 test and the Student t test, respectively. Differences were considered significant at a level of P ⬍ .05. All analyses were performed using the R package with the Design, Hmisc, Design, Presence/absence (http:// lib.stat.cmu.edu/R/CRAN).

R ESULTS Patient population The overall data from the 18,294 patients in the training set and the 434 patients in the validation set were analyzed. Patient characteristics are summarized in Table 1. The populations were significantly dif-

ferent for all characteristics studied by the nomogram. The metastatic LN frequencies for the training and validation sets were 7.89% (1443 of 18,294) and 13.13% (57 of 434), respectively.

A nomogram for the prediction of metastatic lymph nodes Table 2 summarizes the multivariate logistic regression analyses. The metastatic LN risk was independently associated with age at diagnosis, race, tumor grade, histologic subtype, and the primary site invasion. The nomogram corresponding to the model is shown in Figure 1. For each patient, points were assigned for each of these 5 clinical variables, and a total score was calculated from the nomogram. The total points corresponded to a predicted metastatic LN probability. The prediction model had an AUC of 0.80 (95% confidence interval, 0.79 – 0.81) in the training set before the bootstrap technique was applied.

SEPTEMBER 2012 American Journal of Obstetrics & Gynecology

197.e3

Research

Oncology

www.AJOG.org

FIGURE 1

Nomogram predicting the probability of metastatic lymph node involvement for women with endometrial cancer

and physician decision making regarding lymphadenectomy. The predictor is called metastatic nodes (version 1) and was programmed in Java text. A Java-enabled Internet browser is required to run the applets. An example of a screen is shown in Figure 5.

C OMMENT

The probability of metastatic lymph node involvement is calculated by drawing a line to the point on the axis for each of the following variables: age, race, grade, histological subtype, and primary site invasion. The points for each variable are summed and located on the total points line. Next, a vertical line is projected from the total points line to the predicted probability bottom scale to obtain the individual probability of metastatic lymph node involvement. Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

Validation of the nomogram Discrimination First, the nomogram was internally validated using the bootstrap correction technique. The 200 repetitions of bootstrap sample corrections provided an estimated concordance probability of 0.80 (range, 0.78 – 0.82) (Figure 2, solid line). In the validation set, the discrimination accuracy of the model was 0.79 (range, 0.78 – 0.80) (Figure 2, dotted line). Calibration The predicted probability obtained from the bootstrap correction and the actual probabilities of metastatic LN in the training set are shown in the calibration plot (Figure 3). The corresponding predicted and actual probabilities of LN metastasis in the validation set are shown in the calibration plot (Figure 4). There was 197.e4

no difference between the predicted probabilities and the observed rate of metastatic LN (P ⫽ .142). Among the 10 decile groups, the difference between the predicted and observed probabilities of LN metastasis never exceeded 9% except for patients in the 10th decile group (nomogram score above 120) in which the difference reached 45% (the predicted and observed probabilities of LN metastasis were 31.0% and 76.0%, respectively) (Table 3).

Metastatic nodes We developed a computer interface that uses the prediction models described in previous text to estimate the probability of metastatic lymph node involvement in individual patients treated for endometrial cancer. This Internet-based tool may assist patient

American Journal of Obstetrics & Gynecology SEPTEMBER 2012

Cancer researchers, clinicians, and the public are increasingly interested in alternative tools such as nomograms to improve the management of women with cancer. In the current study, we have developed and validated a robust nomogram that is able to predict the risk of LN metastasis in patients with endometrial cancer. By providing predictions that are both evidence based and individualized, these estimates have the potential to improve medical decision management and to help inform the decision-making process of physicians. This tool can also be used to help select candidates for trials designed to evaluate the benefit of lymphadenectomy in endometrial cancer. Several authors have attempted to predict this risk in patients with endometrial cancers before hysterectomy. Kamura et al13 developed a formula based on the tumor diameter and the depth of myometrial invasion. In their experience, if 0.892 was chosen as the cutoff point of probability, the correct sensitivity, specificity, and accuracy were 83%, 72%, and 73%, respectively. Unfortunately, this study was based on a relatively small number of patients (175 cases), and the accuracy of the formula has not been externally validated. Similarly, Lee et al14 developed a preoperative prediction model (based on histological tumor grade, preoperative CA-125 levels, disease extent, and myometrial invasion) for identifying the low-risk group for LN metastasis in endometrial cancer. Using this model, they were able to identify a low-risk group that accounted for 57% of patients, and no nodal metastases were observed among the patients in this group. However, the model was also developed using a small sample (110 patients), and there is no evidence of its generalizability.

Oncology

www.AJOG.org In contrast, we propose a nomogram based on clinical and routinely definitive pathological characteristics of the hysterectomy specimen. This tool could generate estimates of the likelihood of metastatic lymph node involvement. We tested its general applicability in a multicenter independent population. The predictive accuracy, validity, and performance characteristics of the tool were related to a high concordance probability. In the validation set specifically, we observed an AUC of 0.79 (95% confidence interval, 0.78– 0.80) with a good calibration curve related to a mean error that never exceeded 5%. To date, the value of systematic lymphadenectomy in the early stages of endometrial cancer is controversial, and no guidelines regarding the type and extent of pelvic LN dissection have been established. The decision to perform systematic lymphadenectomy depends on both primary site tumor characteristics (ie, grade differentiation, histological subtype, deep myometrial invasion, lymphovascular space invasion, and cervical stromal involvement)4-6 and clinical parameters (ie, age and comorbidities). However, because of the findings of 2 randomized trials,7,8 the indication for systematic lymphadenectomy is likely to become increasingly less frequent, and the question of secondary lymphadenectomy (after hysterectomy) is likely to be raised more frequently. Given that the procedure is associated with low but potentially severe complication rates (eg, lymph edema, lymph cysts, deep vein thromboses, and life-threatening pulmonary emboli),15 we believe that the decision to perform lymphadenectomy should be based on an accurate and individualized risk assessment for LN metastasis. At an individual level, it is easier to decide whether secondary lymphadenectomy is suitable when the personal risk for LN metastasis is available rather than the categorization of the patient into 1 of the 3 categories already used in routine practice (low-, intermediate-, and highrisk groups).16 For example, a white women (0 points) who is 50 years old (15 points) with a preoperative grade 1 (0 points) adenocarcinoma (0 points)

Research

FIGURE 2

Receiver operating characteristic curves of the model

In the training set, the AUC after bootstrapping was 0.80 (95% CI, 0.78 – 0.82) (solid line); in the validation set, the AUC was 0.79 (95% CI, 0.78 – 0.80) (dotted line). AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic. Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

with 50% or less uterine invasion (47.5 points) in whom primary lymphadenectomy was not performed could be accurately identified with a high LN metastasis probability of 18% based on the definitive pathology results if the tumor was a grade 2 (15 points) adenocarcinoma (0 points) with greater than 50% uterine invasion (90 points). The personalized risk estimate is of particular interest in that case if substantial comorbidities are present because it informs the discussion of the benefit of the procedure based on her health status and LN metastasis risk. Before evaluation of this nomogram using preoperative tumoral characteristics, it can be used only when a hysterectomy has been performed and a lymphadenec-

tomy omitted. Apart from frail medically infirm patients in whom the surgeon may not want to perform a lymphadenectomy, at least another circumstance may discourage the surgeon to perform lymphadenectomy: when the tumor grade and/or stage have been underestimated. In such cases, our nomogram can be used to evaluate the risk of lymph node metastasis and to decide to perform secondary lymphadenectomy. The nomogram is based on definitive pathological findings. Indeed, recent data indicate that the accuracy of preoperative evaluations of myometrial invasion, tumor grade, and primary tumor extension are not adequate to determine when lymphadenectomy is required.17-19 Neubauer et al19 demonstrated that 25%

SEPTEMBER 2012 American Journal of Obstetrics & Gynecology

197.e5

Research

Oncology

www.AJOG.org

FIGURE 3

Internal calibration of the nomogram to predict metastatic lymph node involvement

The horizontal axis represents the predicted probability of metastatic lymph nodes, and the vertical axis represents the actual probability of metastatic lymph nodes. Perfect prediction would correspond to the 45-degree broken line. The solid line indicates the observed (apparent) nomogram performance. Calibration plot (P value of the U index ⫽ .09). E, difference in the predicted and calibrated probabilities between calibration and AUC (E maximum ⫽ 0.02%). Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

of patients with a preoperative grade 1 designation and 22% of patients with intraoperative assessments showed a higher grade of disease on the final pathological results. Transvaginal ultrasound has been suggested to have a high diagnostic accuracy for the prediction of cervical involvement, but it is not a reliable method for estimating the depth of myometrial infiltration.20 Frozen section analysis has recently been suggested to accurately and safely predict the low-risk group (based on myometrial invasion and grade) but only when the tumor diameter is less than 3 cm.21 Unfortunately, the authors did not evaluate the prognostic value of 197.e6

frozen section analysis to distinguish between tumors with and without myometrial invasion or to assess stromal cervical invasion, both of which are both strongly associated with an increased risk for LN metastasis in our experience. The current study has several limitations. First, the prediction of LN metastasis is accurate in all deciles of the SEER database, whereas the difference between the predicted and estimated probabilities of LN metastasis reached 45% in the 10th decile of the validation set (among the 10 decile groups, the differences between the predicted probability and observed in the validation set was lower than 5% in 6 deciles I, II, III, IV, V,

American Journal of Obstetrics & Gynecology SEPTEMBER 2012

IX). However, the difference in decile X is not likely to influence clinical practice because it concerns patients for whom the probability of LN metastasis is high and for whom the decision to perform lymphadenectomy is indisputable if the patient’s condition allows it. Similarly, in decile VII, a 12% predicted incidence of lymph node metastasis to a 21% observed incidence is high, but the real question is: in a patient with a 12% predicted incidence of lymph node metastasis, will the decision to perform a lymphadenectomy be modified if the real risk of lymph node metastasis is 21%? We believe not. For the first 5 deciles, the observed incidence of lymph node metastasis (in the training and the validation set) was always below 6%, and consequently, in such cases, lymphadenectomy cannot be recommended. On the contrary, in the validation set for the last 5 deciles, the observed incidence of lymph node metastasis was always greater than 10% and in such cases, lymphadenectomy should be recommended. In the training set for the last 3 deciles, the observed incidence of lymph node metastasis was always greater than 10%, and in such cases, lymphadenectomy should be recommended. However, the interest of the nomogram is to provide a more individualized and accurate risk for an event (lymph node metastasis in the present study), and the predicted risk should always be considered in regard to the general condition of the patient. Second, unfortunately, the SEER database does not provide data concerning lymphovascular space invasion (LVSI). LVSI is suggested to be strongly associated with lymph node metastasis, and this information could have been highly valuable to predict this risk.22 However, LVSI determination is not systematically performed by pathologists. For example, in the Efficacy of Systematic Pelvic Lymphadenectomy in Endometrial Cancer (ASTEC) study, LVSI was not stated in 20% of patients.7 Moreover, LVSI has been criticized for its subjectivity and poor reproducibility,23 and we cannot exclude that in the multivariate analysis, LVSI would remain in the final model because LVSI has been associated with

Oncology

www.AJOG.org

FIGURE 4

Calibration plot of the nomogram against the external validation set

Calibration plot (P value of the U index ⫽ .142). The predicted and observed probabilities are plotted. A perfect model would have all points on the dotted line. E, difference in the predicted and calibrated probabilities between calibration and AUC (E maximum ⫽ 16%; E average ⫽ 5.22%). Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

Research

the extension of primary tumor, depth of myometrial invasion, and histological grade.24 Third, we included only women who underwent a lymphadenectomy with at least 10 regional lymph nodes removed12 to build the most accurate nomogram. In contrast to other organs, there are no definite guidelines to guide gynecological oncologists as to what constitutes an adequate lymphadenectomy. The number of lymph nodes harvested relies on the surgical expertise and the thoroughness of the pathologist in identifying lymph nodes in the surgical specimen. For the external validation, we did not select patients with a minimal lymph node removed because we wanted to test its accuracy in everyday practice. Although the median number of nodes removed in the validation set is low, it is similar to the median number of nodes harvested in the ASTEC lymphadenectomy group.7 In conclusion, we developed a 5-variable nomogram for predicting the risk of LN metastasis after hysterectomy for endometrial carcinoma. This nomogram incorporated important prognostic factors to generate a more accurate prediction of a patient’s individualized risk. The validation of this nomogram in an

TABLE 3

Comparison between the predicted and observed lymph node metastasis probabilities among the training and validation populations Training set (n ⴝ 18,294)

Decile

Patient, n

Whole population mean

—

I

1838

II

1864

III

2588

IV V

Nomogram score 76.32

Validation set (n ⴝ 434)

Predicted probability

Observed probability

Error between predicted and observed probability

Patient, n

8.40%

8.41%

0.01%

—

0.41%

0.43%

0.02%

50

22-46

0.79%

0.64%

0.15%

52

47-61

2.16%

2.20%

0.04%

48

1465

62-71

2.93%

3.48%

0.54%

2218

72-76

3.75%

3.56%

0.19%

VI

1115

77-84

4.67%

4.84%

0.16%

34

VII

1903

85-99

6.65%

6.20%

0.45%

33

VIII

1743

100-114

11.71%

12.04%

0.33%

52

IX

2522

115-129

18.48%

18.55%

0.06%

32

X

1038

32.51%

32.21%

0.29%

33

Nomogram score 77.43

Predicted probability

Observed probability

Error between predicted and observed probability

8%

13%

5%

1%

0%

1%

30-54

2%

6%

4%

55-61

2%

0%

2%

49

62-71

3%

4%

1%

51

72-81

4%

1%

4%

82-101

7%

15%

8%

102-109

12%

21%

9%

110-124

16%

22%

6%

125-134

23%

25%

2%

31%

76%

45%

................................................................................................................................................................................................................................................................................................................................................................................

⬍21

⬍30

................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................

⬎130

⬎135

................................................................................................................................................................................................................................................................................................................................................................................

Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

SEPTEMBER 2012 American Journal of Obstetrics & Gynecology

197.e7

Research

Oncology

www.AJOG.org

FIGURE 5

Example of a screen from a computer program

Example of a screen from the computer program, Metastatic nodes, which was developed from the prediction model reported in this study. Bendifallah. Nomogram for metastasis in endometrial cancer. Am J Obstet Gynecol 2012.

independent, external, multicenter data set demonstrated its high concordance probability. This nomogram may facilitate a multidisciplinary decision-making process for clinicians and may offer better patient counseling by providing individualized information. We also hypothesize that by identifying patients who are at risk for LN metastasis, our nomogram can be used to select candidates for clinical trials that are designed to evaluate adjuvant treatment strategies. The good predictive value and the simplicity of the tool are important to ensure its applicability and spread in clinical practice. f ACKNOWLEDGMENTS The role of each author is as follows: R.R. and M.K. were responsible for the study concept and design; S.B., A.S.G., I.N., F.C.C., and B.H. were responsible for the acquisition of data; M.K., E.D., and N.C.B. performed the analysis and interpretation of data; N.C.B., F.C.C., B.H., D.L., and E.D. drafted the manuscript; S.B., M.K., and R.R. contributed to the statistical expertise; and R.R. and M.K. supervised the study.

REFERENCES 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin 2012;62: 10-29. 2. Creutzberg CL, van Putten WL, Koper PC, et al. Surgery and postoperative radiotherapy versus surgery alone for patients with stage-1 endometrial carcinoma: multicentre randomised trial. PORTEC Study Group. Post Operative radiation therapy in endometrial carcinoma. Lancet 2000;355:1404-11.

197.e8

3. Bansal N, Herzog TJ, Seshan VE, et al. Uterine carcinosarcomas and grade 3 endometrioid cancers: evidence for distinct tumor behavior. Obstet Gynecol 2008;112:64-70. 4. Cragun JM, Havrilesky LJ, Calingaert B, et al. Retrospective analysis of selective lymphadenectomy in apparent early-stage endometrial cancer. J Clin Oncol 2005;23:3668-75. 5. Katz LA, Andrews SJ, Fanning J. Survival after multimodality treatment for stage IIIC endometrial cancer. Am J Obstet Gynecol 2001; 184:1071-3. 6. Bristow RE, Zahurak ML, Alexander CJ, Zellars RC, Montz FJ. FIGO stage IIIC endometrial carcinoma: resection of macroscopic nodal disease and other determinants of survival. J Clin Oncol 2003;13:664-72. 7. Kitchener H, Swart AM, Qian Q, Amos C, Parmar MK. Efficacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009;373:125-36. 8. Benedetti Panici P, Basile S, Maneschi F, et al. Systematic pelvic lymphadenectomy vs. no lymphadenectomy in early-stage endometrial carcinoma: randomized clinical trial. J Natl Cancer Inst 2008;100:1707-16. 9. Kitajima K, Suzuki K, Senda M, et al. Preoperative nodal staging of uterine cancer: is contrast-enhanced PET/CT more accurate than non-enhanced PET/CT or enhanced CT alone? Ann Nucl Med 2011;25:511-9. 10. Geisler JP, Linnemeier GC, Manahan KJ. Pelvic and para-aortic lymphadenectomy in patients with endometrioid adenocarcinoma of the endometrium. Int J Gynaecol Obstet 2007; 98:39-43. 11. Iasonos A, Schrag D, Raj GV, Panageas KS. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol 2008;26: 1364-70. 12. Chan JK, Cheung MK, Huh WK, et al. Therapeutic role of lymph node resection in endo-

American Journal of Obstetrics & Gynecology SEPTEMBER 2012

metrioid corpus cancer: a study of 12,333 patients. Cancer 2006;107:1823-30. 13. Kamura T, Yahata H, Shigematsu T, et al. Predicting pelvic lymph node metastasis in endometrial carcinoma. Gynecol Oncol 1999; 72:387-91. 14. Lee JY, Jung DC, Park SH, et al. Preoperative prediction model of lymph node metastasis in endometrial cancer. Int J Gynecol Cancer 2010;20:1350-5. 15. Querleu D, Leblanc E, Cartron G, Narducci F, Ferron G, Martel P. Audit of preoperative and early complications of laparoscopic lymph node dissection in 1000 gynecologic cancer patients. Am J Obstet Gynecol 2006;195:1287-92. 16. Baekelandt MM, Castiglione M. Endometrial carcinoma: ESMO clinical recommendations for diagnosis, treatment and follow-up. Ann Oncol 2009;20(Suppl 4):29-31. 17. Wang X, Zhang H, Di W, Li W. Clinical factors affecting the diagnostic accuracy of assessing dilation and curettage vs frozen section specimens for histologic grade and depth of myometrial invasion in endometrial carcinoma. Am J Obstet Gynecol 2009;201:194.e1-10. 18. Case AS, Rocconi RP, Straughn JM Jr, et al. A prospective blinded evaluation of the accuracy of frozen section for the surgical management of endometrial cancer. Obstet Gynecol 2006;108:1375-9. 19. Neubauer NL, Havrilesky LJ, Calingaert B, et al. The role of lymphadenectomy in the management of preoperative grade 1 endometrial carcinoma. Gynecol Oncol 2009;112:511-6. 20. Akbayir O, Corbacioglu A, Numanoglu C, et al. Preoperative assessment of myometrial and cervical invasion in endometrial carcinoma by transvaginal ultrasound. Gynecol Oncol 2011; 122:600-3. 21. Yanazume S, Saito T, Eto T, et al. Reassessment of the utility of frozen sections in endometrial cancer surgery using tumor diameter as an additional factor. Am J Obstet Gynecol 2011;204:531.e1-7. 22. Guntupalli SR, Zighelboim I, Kizer NT, et al. Lymphovascular space invasion is an independent risk factor for nodal disease and poor outcomes in endometrioid endometrial cancer. Gynecol Oncol 2012;124:31-5. 23. Nordstrom B, Strang P, Lindgren A, Bergstrom R, Tribukait B. Carcinoma of the endometrium: do the nuclear grade and DNA ploidy provide more prognostic information than do the FIGO and WHO classifications? Int J Gynecol Pathol 1996;15:191-201. 24. Inoue Y, Obata K, Abe K, et al. The prognostic significance of vascular invasion by endometrial carcinoma. Cancer 1996;78: 1447-51.