The effect of para-aortic lymphadenectomy on survival in early stage endometrioid grade 3 and non-endometrioid endometrial cancers

Abstract

Background

The prognostic and therapeutic significance of para-aortic lymphadenectomy remains controversial due to conflicting evidence in the literature. There are limited studies specifically examining patients with grade 3 endometrioid and non-endometrioid histologies. In this study, we aimed to investigate the effect of adding para-aortic lymphadenectomy to pelvic lymphadenectomy on survival outcomes in patients with early-stage uterine-confined grade 3 endometrioid and non-endometrioid endometrial carcinoma.

Methods

This retrospective cohort study included 182 patients with FIGO 2009 stage I–II, pathologically node-negative grade 3 endometrioid or non-endometrioid endometrial cancers treated at a single tertiary center between 1995 and 2025. Patients underwent either pelvic lymphadenectomy (PLND) alone 85 (46.7%) or combined pelvic and para-aortic lymphadenectomy (PPLND) 97 (53.3%). Clinicopathological data and survival outcomes were compared. Kaplan–Meier survival curves, log-rank tests, and Cox proportional hazards models were used to assess disease-free survival (DFS) and overall survival (OS).

Results

The median duration of follow-up in the overall cohort was 60.5 months, while the median time to recurrence among patients who experienced disease relapse was 52.0 months. No significant difference was observed in DFS or OS between patients who underwent PLND alone and those who underwent PPLND. The hazard ratio (HR) for DFS was 0.88 (95% CI: 0.47–1.64, p = 0.690), and similarly, the HR for OS was 0.98 (95% CI: 0.60–1.62, p = 0.941). In univariate and multivariate Cox regression analysis, non-endometrioid histology was associated with worse DFS and OS (p:<0.001, p:0.04). While there was no significant difference in DFS in patients aged ≥ 60 years, a three-fold worse OS outcome was found (HR = 3.08; 95% CI: 1.59–5.99; p < 0.001). Other clinicopathological parameters, such as FIGO stage, depth of myometrial invasion, presence of lymphovascular invasion (LVSI), cytology results, surgical approach, and extent of lymphadenectomy, did not show a significant association with DFS or OS in multivariate analysis (all p > 0.05).

Conclusions

In patients with early-stage, node-negative grade 3 endometrioid and non-endometrioid endometrial cancers, the addition of para-aortic lymphadenectomy to pelvic dissection did not confer a survival advantage. Moreover, older age (≥ 60 years) was associated with significantly worse overall survival.

Introduction

Endometrial cancer (EC) is the most common gynecologic malignancy among women, and its global incidence has been steadily increasing. This rise has been largely attributed to aging populations, declining fertility rates, the widespread use of tamoxifen, and the global obesity epidemic [1].

The 2021 ESGO/ESTRO/ESP and the recently published 2025 ESGO/ESTRO/ESP guidelines adopted a fully molecular-based classification system. In this revised model, four molecular subtypes (POLE-ultramuted (POLEmut), mismatch repair-deficient (MMRd), p53-aberrant (p53abn), and no specific molecular profile (NSMP)) were integrated with disease stage and additional biological features to define low, intermediate, high-intermediate, high, and indeterminate risk categories [2, 3]. In the classification first proposed by Bokhman in 1983, endometrial carcinoma was categorized based on its biological behavior and clinical course into two distinct types: estrogen-dependent endometrioid adenocarcinomas (Type I) and estrogen-independent adenocarcinomas (Type II) [4]. Type I tumors are estrogen-dependent endometrioid adenocarcinomas, typically characterized by low grade, early stage, and indolent clinical behavior. In contrast, Type II tumors are estrogen-independent non-endometrioid carcinomas that usually occur in older and thinner women and are commonly associated with serous or clear cell histologic subtypes [5]. As knowledge of the pathological and molecular characteristics of endometrial cancer has advanced, tumors with aggressive histologic features—including grade 3 endometrioid carcinomas and non-endometrioid subtypes such as serous, clear cell, undifferentiated, and carcinosarcomatous histologies—have been reclassified as stage IIC disease under the updated 2023 staging system of the International Federation of Gynecology and Obstetrics (FIGO). This revision reflects the recognition that these high-grade tumors, although confined to the uterus, demonstrate a significantly higher risk of recurrence and adverse outcomes, thereby warranting their distinction from low-grade stage I tumors in both clinical management and prognostic assessment [6].

Assessment of lymph node status plays a critical role in determining the need for adjuvant treatment, and thus, systematic lymphadenectomy—encompassing both pelvic and para-aortic lymph node dissection (PPLND)—has become a key component of the primary surgical staging approach in endometrial cancer management [3, 7]. Although sentinel lymph node (SLN) biopsy has recently emerged as a reliable alternative for nodal assessment, particularly in low- and intermediate-risk cases, systematic lymphadenectomy remains the standard for high-risk patients due to its superior staging accuracy [7]. The SEPAL study demonstrated that combined PPLND was associated with improved survival in patients with intermediate- and high-risk endometrial cancer, when compared to pelvic lymphadenectomy (PLND) alone [8]. However, it remains unclear whether this survival benefit was due to the removal of clinically positive para-aortic nodes (debulking effect) or a true therapeutic impact of the procedure. Supporting this uncertainty, other studies have shown no significant survival advantage with para-aortic lymphadenectomy in early-stage, node-negative patients with grade 3 endometrioid or non-endometrioid histologies [9–11]. The prognostic and therapeutic relevance of para-aortic lymphadenectomy remains controversial, with conflicting evidence in the literature, particularly for patients with clinically node-negative disease [12]. Furthermore, para-aortic lymph node dissection is associated with longer operative times, increased intraoperative blood loss, and higher rates of postoperative complications, including transfusion requirements, which raises concerns about its routine application [13].

Importantly, limited studies have specifically addressed patients with grade 3 endometrioid and non-endometrioid histologies—subtypes characterized by the highest rates of nodal metastasis and responsible for approximately 40% of endometrial cancer-related deaths [14–16]. These high-risk subtypes may have distinct biological behavior and patterns of spread that could influence the therapeutic value of lymphadenectomy.

In this study, we aimed to investigate the impact of PLND alone versus combined PPLND on oncologic outcomes in patients with grade 3 endometrioid and non-endometrioid endometrial carcinoma.

Materials and methods

Patients

This retrospective study included patients who underwent staging surgery at Bursa Uludag University Faculty of Medicine Hospital between January 1995 and May 2025 and were pathologically diagnosed with FIGO 2009 stage I–II, grade 3 endometrioid or non-endometrioid endometrial cancer. The study was conducted in accordance with the Declaration of Helsinki and approved by the Bursa Uludag University Clinical Research Ethics Committee (Protocol No. 2025 − 800/14–19). Due to the retrospective nature of the study, the ethics committee exempted the requirement for informed consent.

Inclusion criteria

Inclusion criteria comprised patients with histopathologically confirmed grade 3 endometrioid or non-endometrioid endometrial carcinomas who underwent PPLND, had FIGO 2009 stage I–II disease, and demonstrated no lymph node metastasis on final pathological assessment.

Exclusion criteria

Exclusion criteria included patients with grade 1–2 endometrioid histology, evidence of lymph node metastasis, FIGO 2009 stage III or higher disease following surgical staging, presence of synchronous malignancies, absence of lymph node evaluation during surgery, or administration of neoadjuvant treatment.

Diagnostic criteria and grouping criteria for patients

Staging surgery included total hysterectomy, bilateral salpingo-oophorectomy, pelvic lymphadenectomy and/or para-aortic lymphadenectomy, peritoneal lavage cytology, and omental biopsy in patients with non-endometrioid histology. Para-aortic lymphadenectomy was performed selectively based on the patient’s risk profile or the surgeon’s clinical judgment. The decision not to perform para-aortic lymphadenectomy was primarily based on intraoperative and patient-related factors that could increase surgical risk or morbidity. Specifically, para-aortic dissection was avoided in patients who were unable to tolerate prolonged surgery, had significant comorbidities, or presented with high body mass index (BMI), that could increase operative difficulty and postoperative complications. These considerations guided the surgical team to limit the extent of lymphadenectomy in order to ensure patient safety and minimize perioperative risk.

Surgical procedures were performed laparoscopically or via laparotomy. PLND included removal of the external, internal, and common iliac chains, as well as lymph nodes in the obturator region. Para-aortic dissection was performed covering the area around the vena cava, the aortocaval area and the left para-aortic regions, and up to the level of the renal vein as the upper limit. All patients underwent preoperative imaging with computed tomography (CT) or positron emission tomography combined with CT (PET/CT) to assess lymph node metastases. In patients considered clinically early-stage and node-negative based on imaging, the aim of lymphadenectomy is to identify occult metastatic lymph nodes that cannot be detected radiologically. Patients were divided into two groups: those who underwent only PLND and those who underwent PPLND.

Data collection

Clinical and pathological data obtained from patient files: The criteria included variables such as age at diagnosis, BMI, preoperative CA-125 level, FIGO stage (2009), tumor histology, tumor size, depth of myometrial invasion, lymphovascular invasion (LVSI) status, peritoneal lavage cytology results, recurrence and recurrence site, surgical methods, and adjuvant therapy status.

Decisions regarding adjuvant therapy were made by a multidisciplinary tumor council, taking into account the clinical and pathological characteristics of the patients. Treatment approaches were implemented in accordance with the NCCN and ESGO guidelines in effect at the time. The decision to administer radiotherapy and the dose/scope to be administered were determined based on uterine risk factors such as depth of myometrial invasion, LVSI, cervical stromal involvement, and tumor stage. Addition of chemotherapy (mostly a carboplatin and paclitaxel regimen) was preferred in cases of high-risk histological subtypes (serous, clear cell, carcinosarcoma, mixed type) with myometrial invasion or in cases of positive peritoneal cytology [2, 3, 7]. Patients with serious comorbidities (e.g., renal or liver dysfunction) or who refused treatment were not treated with adjuvant therapy. Postoperative follow-up was conducted every 3–6 months during the first 2 years and every 6 months thereafter up to 5 years. Follow-up of patients who completed their 5th year continued annually. Further imaging with CT or PET/CT was performed in the presence of clinical findings suggesting recurrence or metastasis. Findings confirmed by imaging methods according to RECIST 1.1 criteria or by biopsy were considered disease recurrence. Recurrences were classified as local (vaginal, pelvic lymph nodes, bladder, rectum), distant (mid or upper abdominal cavity, para-aortic lymph nodes, lung, liver, bone, brain), or both if present simultaneously. Disease-free survival (DFS) was defined as the time from surgery to disease recurrence or last follow-up date; overall survival (OS) was defined as the time from surgery to cancer-related death or last follow-up date.

Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY, USA). Comparisons between continuous variables were made using the Mann–Whitney U test, and categorical variables were compared using the Pearson chi-square test based on cell counts. Descriptive statistics were presented as median (range) for continuous variables and number (percentage) for categorical variables. A two-sided p-value < 0.05 was considered statistically significant in all analyses.

Survival outcomes, including disease-free survival and overall survival, were estimated using the Kaplan–Meier method, and differences between survival curves were evaluated using the log-rank test. To identify independent prognostic factors affecting DFS and OS, Cox proportional hazards regression analyses were performed. Initially, univariate Cox regression was applied to assess the impact of each clinical and pathological parameter on survival outcomes. Covariates reaching a threshold significance of p < 0.05 on univariable analysis were selected for inclusion in multivariable models. Results were reported as hazard ratios (HRs) with corresponding 95% confidence intervals (CIs).

Results

Study population and flow chart

A total of 288 patients diagnosed with clinically early-stage non-endometrioid and endometrioid grade 3 endometrial cancer were evaluated. After applying strict eligibility criteria, 40 patients were excluded from the analysis due to the lack of lymph node assessment. Sixty-six patients who had undergone staging surgery and were found to have lymph node metastasis were also excluded. A total of 182 patients who were surgically staged and had pathological FIGO 2009 stage I–II disease were included in the study. Of these, 85 patients (46.7%) underwent PLND only, while 97 patients (53.3%) underwent combined PPLND. The patient selection flowchart for the study cohort of stage I–II non-endometrioid and endometrioid grade 3 endometrial cancer is shown in Fig. 1.

Fig. 1.

Patient selection flowchart for the study cohort of stage I-II non-endometrioid and endometrioid grade 3 endometrial cancer

Demographics and baseline characteristics

The median duration of follow-up in the overall cohort was 60.5 months, while the median time to recurrence among patients who experienced disease relapse was 52.0 months. Table 1 presents the clinical and pathological characteristics of the 182 eligible patients, stratified by the type of lymphadenectomy performed. The median age of the cohort was 65 years (range: 38–90). Patients were predominantly aged ≥ 60 years (71.9%, n = 131). Age distribution did not differ significantly between PLND and PPLND groups (p = 0.121). The median BMI of the entire cohort was 34.2 kg/m² (range: 20.0–63.0). Patients in the PLND group had significantly higher BMI (37.7 kg/m²) compared to those in the PPLND group (33.0 kg/m²; p = 0.003). The median preoperative CA-125 level was 16.0 U/mL across the cohort and did not differ significantly between the groups (p = 0.897). The median tumor size was 3.5 cm (range: 0.4–15.3), and tumors were significantly larger in the PPLND group (4.0 cm) than in the PLND group (3.2 cm; p = 0.013). A significant difference was observed in the number of lymph nodes removed between patients who underwent PLND and those who underwent combined PPLND. The median number of pelvic lymph nodes removed was 20 (range: 1–64), while the median number of para-aortic nodes was 13 (range: 1–46). The median total number of nodes removed (pelvic + para-aortic) was 35 (range: 4–94). Statistical analysis demonstrated that the total number of harvested lymph nodes was significantly higher than the number removed by pelvic dissection alone (p < 0.001). In terms of histology, 46.7% (n = 85) of tumors were grade 3 endometrioid carcinomas, and 53.3% (n = 97) were non-endometrioid types, including serous (18.1%), clear cell (7.1%), carcinosarcoma (12.1%), mixed (3.3%), and other rare subtypes (12.6%). The distribution of histologic subtypes did not significantly differ between groups (p = 0.591).

Table 1.

Comparison of clinical and pathological characteristics between PLND and PPLND groups

	Total (n = 182)	PLND (n = 85)	PPLND (n = 97)	p a value
Age, y, median (range), n, (%)	65 (38–90)	64.9 (38–79)	65 (44–90)	0.121
<60 years	51 (28.1%)	29 (34.1%)	22 (22.7%)
≥60 years	131 (71.9%)	56 (65.9%)	75 (77.3%)
BMI, kg/m2, median (range)	34.2 (20.0–63.0)	37.7 (20.0–63.0)	33.0 (21.0–55.4)	0.003
Pre-operative CA-125 (U/mL), median (range)	16.0 (2.8–400.0)	16.0 (2.8–400.0)	16.3 (4.0–380.0)	0.897
Tumor size, cm, median (range)	3.5 (0.4–15.3)	3.2 (0.4–9.0)	4.0 (0.5–15.3)	0.013
No. of Lymph node harvested, median (range)	35 (4.0–94.0)	20 (1.0–64.0)	35 (4.0–94.0)	< 0.001
Histological types, n, (%)				0.591b
Grade 3 endometrioid	85 (46.7%)	42 (49.4%)	43 (44.3%)
Non-endometrioid	97 (53.3%)	43 (50.6%)	54 (55.7%)
Serous	33 (18.1%)	12 (14.1%)	21 (21.6%)
Clear cell	13 (7.1%)	2 (2.4%)	11 (11.3%)
Carcinosarcoma	22 (12.1%)	12 (14.1%)	10 (10.3%)
Mixed	6 (3.3%)	3 (3.5%)	3 (3.1%)
Othersc	23 (12.6%)	14 (16.5%)	9 (9.3%)
FIGO staged, n, (%)				0.197
IA	107 (58.8%)	44 (51.8%)	63 (64.9%)
IB	53 (29.1%)	29 (34.1%)	24 (24.7%)
II	22 (12.1%)	12 (14.1%)	10 (10.3%)
Depth of myometrial invasion, n, (%)				0.041
None	22 (12.1%)	7 (8.2%)	15 (15.5%)
< 50	97 (53.3%)	41 (48.2%)	56 (57.7%)
≥ 50	63 (34.6%)	37 (43.5%)	26 (26.8%)
LVSI, n, (%)				0.916
Present	51 (28.0%)	23 (27.1%)	28 (28.9%)
Absent	131 (72.0%)	62 (72.9%)	69 (71.1%)
Peritoneal washing cytology, n, (%)		78 (91.8%) 7 (8.2%)	94 (96.9%) 3 (3.1%)	0.233
Negative	172 (94.5%)
Positive	10 (5.5%)
Methods of surgery, n, (%)				0.898
Laparoscopy	24 (13.2%)	12 (14.1%)	12 (12.4%)
Laparotomy	158 (86.8%)	73 (85.9%)	85 (87.6%)
Adjuvant treatment, n, (%)				0.005
None	16 (8.8%)	13 (15.3%)	3 (3.1%)
Radiotherapy onlye	14 (7.7%)	6 (7.1%)	8 (8.2%)
Chemotherapy only	94 (51.6%)	47 (55.3%)	47 (48.5%)
Chemotherapy and radiotherapy	58 (31.9%)	19 (22.4%)	39 (40.2%)
Recurrence, n (%)				0.309
Yes	42 (23.1%)	23 (27.1%)	19 (19.6%)
No	140 (76.9%)	62 (72.9%)	78 (80.4%)
Recurrence site, n, (%)	42 (100%)			0.040
Local	20 (47.6%)	9 (39.1%)	11 (57.9%)
Distant	17 (40.5%)	13 (56.5%)	4 (21.1%)
Local + Distant	5 (11.9%)	1 (4.3%)	4 (21.1%)
Death, n (%)	64/182 (35.2%)	35/85 (41.2%)	29/97 (29.9%)	0.151
Cause of death, n (%)				0.681
Cancer-related	36 (56.2%)	21 (60.0%)	15 (51.7%)
All other causes	28 (43.8%)	14 (40.0%)	14 (48.3%)

PLND pelvic lymphadenectomy, PPLND pelvic and para-aortic lymphadenectomy, n number, BMI body mass index, FIGO International Federation of Gynecology and Obstetrics, CA-125 cancer antigen 125, LVSI lymph-vascular space invasion

^ap-values of age, BMI, CA-125, tumor size and no. of Lymph node harvested were determined by Mann-Whitney U test and the others were determined by Chi-squared test

^bStatistical analysis was performed for grade 3 endometrioid adenocarcinoma versus non-endometrioid carcinoma

^cVilloglandular adenocarcinoma, squamous cell carcinoma, undifferentiated and dedifferentiated carcinoma were included

^dThe disease stage was based on the 2009 FIGO staging system

^eRadiotherapy included external beam radiotherapy (EBRT) only, vaginal brachytherapy only, or EBRT plus vaginal brachytherapy

According to FIGO 2009 staging, 58.8% of patients were stage IA, 29.1% were stage IB, and 12.1% were stage II. Stage distribution was not significantly different between the two groups (p = 0.197). Myometrial invasion depth showed significant variation (p = 0.041): ≥50% invasion was more frequent in the PLND group (43.5%) than in the PPLND group (26.8%).

LVSI was detected in 28.0% (n = 51) of patients, with similar rates between groups (p = 0.916). Peritoneal washing cytology was positive in 5.5% (n = 10) of patients and did not differ significantly between groups (p = 0.233). Most surgeries were performed via laparotomy (86.8%, n = 158). The proportion of laparoscopic versus open procedures was similar in both groups (p = 0.898). Adjuvant treatment strategies varied significantly by lymphadenectomy extent (p = 0.005). While the proportion of patients receiving no adjuvant therapy or radiotherapy alone was higher in the PLND group, patients in the PPLND group more frequently received combined chemotherapy and radiotherapy (40.2% vs. 22.4%). Recurrence occurred in 23.1% (n = 42) of the total cohort, with a higher rate observed in the PLND group (27.1%) than in the PPLND group (19.6%), although this difference was not statistically significant (p = 0.309). Among recurrences, the distribution of local vs. distant failures differed significantly (p = 0.040), with local recurrences being more common in the PPLND group. A total of 64 patients (35.2%) died during follow-up, including 36 cancer-related deaths (56.2%). While mortality rates were higher in the PLND group (41.2%) compared to the PPLND group (29.9%), the difference was not statistically significant (p = 0.151). Likewise, the proportion of cancer-related deaths did not differ significantly between groups (p = 0.681).

Disease-free survival and overall survival analyses between groups

According to the Kaplan–Meier analysis, there was no significant difference in DFS or OS between patients who underwent PLND alone and those who received combined PPLND. The hazard ratio (HR) for DFS was 0.88 (95% CI: 0.47–1.64, p = 0.690), indicating no statistically significant difference in the risk of recurrence between the two groups. Similarly, the HR for OS was 0.98 (95% CI: 0.60–1.62, p = 0.941), showing no significant impact of para-aortic lymphadenectomy on overall survival. These findings suggest that the addition of para-aortic lymphadenectomy to pelvic lymphadenectomy does not provide a survival benefit in terms of either DFS or OS (Fig. 2).

Fig. 2.

Kaplan-Meier curves for disease-free survival and overall survival in patients who underwent PLND alone versus those who underwent combined PPLND

Univariate and multivariate analysis of risk factors for survival

Table 2 presents the univariate and multivariate analyses of OS and DFS. In univariable Cox regression analysis, age ≥ 60 years was associated with significantly worse OS (HR = 3.08; 95% CI: 1.59–5.99; p < 0.001) and demonstrated a strong effect also in the multivariable model (HR = 2.83; 95% CI: 1.45–5.52; p < 0.001). For DFS, age showed a trend toward poorer outcomes (HR = 2.10; 95% CI: 0.99–4.46; p = 0.06). Histologic subtype emerged as an independent prognostic factor for both DFS and OS. Compared to patients with grade 3 endometrioid tumors, those with non-endometrioid histologies had significantly worse DFS (multivariable HR = 4.08; 95% CI: 1.55–10.73; p < 0.001) and OS (multivariable HR = 1.71; 95% CI: 1.02–2.86; p = 0.04). Other clinicopathological parameters—including number of lymph node harvested, FIGO stage, depth of myometrial invasion, presence of LVSI, cytology results, surgical approach, and extent of lymphadenectomy—were not significantly associated with either DFS or OS in multivariable analysis (all p > 0.05). Similarly, type of adjuvant treatment did not show a statistically significant impact on survival outcomes.

Table 2.

The univariate and multivariate analyses of OS and DFS

Variables	DFS				OS
	Univariable		Multivariable		Univariable		Multivariable
	HR (95% CI)	p value	HR (95% CI)	p value	HR (95% CI)	p value	HR (95% CI)	p value
Age
<60 years	1.00	0.06			1.00	< 0.001	1.00	< 0.001
≥60 years	2.10 (0.99–4.46)				3.08 (1.59–5.99)		2.83 (1.45–5.52)
BMI	1.01 (0.98–1.05)	0.51			1.01 (0.98–1.04)	0.56
Pre-operative CA-125	0.99 (0.98–1.00)	0.16			1.00 (1.00–1.00)	0.80
Tumor size	1.00 (0.87–1.14)	0.94			1.03 (0.93–1.15)	0.59
No. of Lymph node harvested
PLND	1.00	0.78			1.00	0.66
PPLND	1.64 (0.85–5.57)				1.27 (0.93–3.22)
Histological types
Grade 3 endometrioid	1.00	< 0.001	1.00	< 0.001	1.00	0.02	1.00	0.04
Non-endometrioid	3.67 (1.76–7.69)		4.08 (1.55–10.73)		1.84 (1.10–3.07)		1.71 (1.02–2.86)
FIGO stage
IA	1.00	0.74			1.00	0.63
IB	1.21 (0.56–2.23)				1.14 (0.66–1.96)
II	1.53 (0.65–3.59)				1.32 (0.62–2.77)
Depth of myometrial invasion
None	1.00	0.64			1.00	0.79
< 50	1.02 (0.52–2.00)				1.08 (0.61–1.93)
≥ 50	1.13 (0.68–1.89)				1.06 (0.68–1.64)
LVSI
Absent	1.00	0.41			1.00	0.54
Present	0.74 (0.36–1.51)				1.18 (0.70–1.99)
Peritoneal washing cytology
Negative	1.00	0.36			1.00	0.78
Positive	1.62 (0.58–4.57)				0.87 (0.31–2.41)
Methods of surgery
Laparoscopy	1.00	0.22			1.00	0.58
Laparotomy	0.58 (0.24–1.39)				0.78 (0.34–1.83)
Extent of lymphadenectomy
PLND	1.00	0.69			1.00	0.94
PPLND	0.88 (0.47–1.64)				0.98 (0.60–1.62)
Adjuvant treatment
None	1.00	0.98			1.00	0.76
Radiotherapy only	0.58 (0.19–1.74)				1.03 (0.36–2.93)
Chemotherapy only	0.98 (0.24–3.95)				1.15 (0.32–4.12)
Chemoradiotherapy	0.90 (0.30–2.71)				0.94 (0.32–2.79)

PLND pelvic lymphadenectomy, PPLND pelvic and para-aortic lymphadenectomy, BMI body mass index, FIGO International Federation of Gynecology and Obstetrics, CA-125 cancer antigen 125, LVSI lymph-vascular space invasion

Subgroup analyses for disease-free survival and overall survival

Because age and histologic subtypes were identified as independent risk factors for survival in the multivariate analysis, a subgroup analysis was conducted. Subgroup analyses were performed between the PLND and PPLND groups in terms of DFS and OS according to age and histological types (Figs. 3, 4, 5 and 6). In patients under 60 years of age, a slight trend in favor of survival was observed in the PPLND group compared to the PLND group, although this difference was not statistically significant. The hazard ratio (HR) for DFS was 0.48 (95% CI: 0.10–2.40, p = 0.373), and the HR for OS was 0.87 (95% CI: 0.21–3.54, p = 0.847) (Fig. 3).

Fig. 3.

Kaplan–Meier curves for disease-free survival and overall survival in patients aged < 60 years according to extent of lymphadenectomy (PLND vs. PPLND)

Fig. 4.

Kaplan–Meier curves for disease-free survival and overall survival in patients aged ≥ 60 years according to extent of lymphadenectomy (PLND vs. PPLND). Similarly, in patients aged 60 and older, a slight trend toward a DFS and OS advantage was observed in the PPLND group compared to PLND, but these differences did not reach statistical significance. The HR for DFS was 0.90, (95% CI: 0.45–1.80, p = 0.773) and the HR for OS was 0.84, (95% CI: 0.49–1.45, p = 0.535), (Fig. 4)

Fig. 5.

Kaplan-Meier curves for disease-free survival and overall survival in patients with endometrioid type grade 3 according to the extent of lymphadenectomy (PLND and PPLND)

Fig. 6.

Kaplan-Meier curves for disease-free survival and overall survival in patients with non-endometrioid type according to the extent of lymphadenectomy (PLND and PPLND)

In patients with grade 3 endometrioid endometrial carcinoma, a comparison between those who underwent PLND alone and those who received combined PPLND revealed a statistically significant difference in OS. Patients in the PLND group demonstrated superior OS (HR: 2.41; 95% CI: 1.01–5.74; p = 0.046). Although DFS was also higher in the PLND group, the difference did not reach statistical significance (HR: 3.93; 95% CI: 0.79–19.54; p = 0.094). These results indicate that incorporating para-aortic lymphadenectomy does not improve survival outcomes in patients with grade 3 endometrioid endometrial cancer (Fig. 5).

In a survival analysis of patients with non-endometrioid endometrial carcinoma, no significant difference was observed between those who underwent PLND and those who underwent PPLND. Although the risk of recurrence was lower in the PPLND group in terms of DFS, it was not significant (HR: 0.51; 95% CI: 0.25–1.03; p = 0.059). Similarly, OS appeared to favor PPLND, but was not significant (HR: 0.54; 95% CI: 0.29–1.01; p = 0.054) (Fig. 6).

Discussion

Main findings

In this retrospective study investigating the impact of lymphadenectomy extent on survival outcomes in patients with early-stage grade 3 endometrioid and non-endometrioid endometrial carcinoma, we found no significant survival benefit associated with the addition of para-aortic lymphadenectomy to pelvic lymphadenectomy. Although para-aortic dissection was more frequently performed in patients with larger tumors and lower BMI, and was associated with a higher rate of combined chemoradiotherapy, the survival outcomes were not significantly improved compared to pelvic lymphadenectomy alone.

Evaluation of retroperitoneal lymph nodes provides prognostic information and plays a critical role in adjuvant treatment planning by identifying the presence of lymph node metastases in patients with endometrial cancer clinically confined to the uterus. Lymphadenectomy is generally limited to dissection encompassing the iliac and obturator chains in the pelvic region; however, in some cases, especially in high-risk patients, extended dissection to the para-aortic region is also performed. Para-aortic nodal dissection is preferred, particularly for tumors with poor prognostic features such as deep myometrial invasion, lymphovascular invasion, or high-grade histology (e.g., grade 3 endometrioid, non-endometrioid types). The therapeutic potential of lymphadenectomy, beyond staging, has been debated for years. According to the updated FIGO 2023 staging system and the molecular classification framework (POLE-mutated, p53-abnormal, MMR-deficient, and NSMP subtypes), risk stratification in endometrial cancer has evolved toward a more biologically driven approach. These molecular profiles complement traditional clinicopathologic factors in predicting prognosis and guiding adjuvant treatment decisions [3].

Two important randomized controlled trials evaluating the effect of PLND on survival demonstrated that systematic pelvic lymphadenectomy statistically significantly improved surgical staging but not DFS or OS [16–19]. The ASTEC trial has faced considerable criticism due to methodological limitations, particularly the insufficient extent of lymph node dissection [17]. Notably, in approximately 35% of patients, fewer than 10 lymph nodes were removed, raising concerns about the adequacy of nodal assessment. Furthermore, the study did not employ a standardized approach to systematic PPLND, which may have limited the ability to detect any potential therapeutic benefit of the procedure. Similarly, the randomized trial conducted by Benedetti Panici et al. also failed to demonstrate a survival advantage associated with lymphadenectomy [18]. Although a higher number of lymph nodes were resected in that study compared to ASTEC, the lack of a standardized para-aortic dissection protocol was a major limitation and was a frequent point of critique. Another important limitation of both studies is the inclusion of a large proportion of low-risk patients—specifically those with early-stage, low-grade, and superficially invasive tumors—who are less likely to benefit from extensive surgical staging. However, these data have focused largely on pelvic dissection, and no randomized controlled trials have directly and controlledly examined the impact of para-aortic lymphadenectomy on survival. This leaves the therapeutic contribution of para-aortic dissection questionable and leads to the current practice being largely based on observational studies.

The results of our study are consistent with previous findings suggesting limited therapeutic benefit of para-aortic lymphadenectomy in node-negative high-risk endometrial cancers. For example, Lai et al. demonstrated in a multicenter study from Korea and Taiwan that para-aortic dissection did not significantly improve recurrence-free or overall survival in early-stage grade 3 and non-endometrioid cancers, even in patients at high risk of occult metastasis [9]. The SEPAL study reported that combined pelvic and para-aortic lymphadenectomy improved survival in intermediate- and high-risk cases. However, the intermediate- and high-risk group in this study included patients with FIGO stages I–II, grade 3 endometrioid, and non-endometrioid endometrial cancer, as well as patients with stage 3 and 4 cancer. It is possible that the inclusion of patients with enlarged pelvic or para-aortic lymph nodes who underwent removal in the SEPAL study led to optimal cytoreduction and consequently resulted in a significant improvement in survival. Furthermore, adjuvant therapy was used more frequently in the pelvic and para-aortic dissection group in the SEPAL study, directly impacting survival outcomes [8]. In our study, no difference in survival was found because the lymph node-negative group was included and adjuvant therapy was administered equally between the two groups.

The therapeutic and prognostic value of para-aortic lymphadenectomy in patients with high-grade endometrial carcinoma remains insufficiently studied, with evidence limited to a small number of publications. Vatansever et al. demonstrated that in patients with FIGO stage I–II type II endometrial carcinoma and carcinosarcoma, those who underwent systematic PPLND had significantly longer overall survival and recurrence-free survival compared to those who did not undergo systematic lymphadenectomy [19]. Venigalla et al. reported that para-aortic lymph node dissection performed in addition to pelvic lymphadenectomy significantly improved overall survival in patients with clinically diagnosed stage I serous endometrial carcinoma [16]. However, this survival advantage was not observed for patients diagnosed with clear cell carcinoma or uterine carcinosarcoma. This finding suggests that the therapeutic effect of extended lymphadenectomy may not apply to all histological subtypes and may be particularly pronounced in certain histological types, such as serous carcinoma. In their study, approximately 20% of patients who underwent PPLND developed lymph node metastases, which may have contributed to the improved survival due to the debulking effect. In contrast, our study focused only on patients with pathologically confirmed FIGO stage I–II grade 3 endometrioid and non-endometrioid endometrial carcinoma without lymph node metastases. The exclusion of lymph node-positive patients prevented the study from assessing the potential therapeutic effect of lymphadenectomy in patients with nodal disease and allowed for a more accurate assessment of the procedure’s contribution to early-stage, lymph node-negative patients. This difference may explain the lack of a survival advantage observed after para-aortic lymphadenectomy in our study.

Subgroup analyses in our cohort revealed that para-aortic lymphadenectomy did not confer a statistically significant survival advantage in patients under or over 60 years of age. Interestingly, among patients with grade 3 endometrioid carcinoma, overall survival was significantly better in the pelvic-only group. This finding is counterintuitive, as one might expect more extensive staging to be associated with improved outcomes. However, the result may be influenced by unmeasured confounders such as surgical morbidity, tumor biology, or differences in adjuvant treatment compliance. In contrast, a non-significant trend toward improved DFS and OS was observed in non-endometrioid histologies undergoing PPLND, which could be due to the higher propensity for para-aortic nodal spread in this group [15, 16].

Importantly, multivariable Cox regression identified histologic subtype as an independent prognostic factor for both DFS and OS, with non-endometrioid histologies associated with significantly worse outcomes compared to grade 3 endometrioid tumors. Age ≥ 60 years was also a strong independent predictor of poorer OS, consistent with the established adverse prognostic role of older age in endometrial cancer. Other variables, including LVSI, tumor size, FIGO stage, and type of adjuvant therapy, were not independently associated with survival in our multivariate models. These findings highlight the dominant prognostic influence of histologic subtype and patient age over surgical variables such as the extent of lymphadenectomy.

Consistent with the findings of several randomized controlled trials that have evaluated the role of adjuvant therapy in patients with intermediate- and high-risk endometrial cancer, our study also found no statistically significant association between the type of adjuvant treatment administered and survival outcomes in patients with FIGO stage I–II grade 3 endometrioid and non-endometrioid tumors. These results suggest that the choice of adjuvant modality—whether radiotherapy, chemotherapy, or a combination of both—may have a limited impact on disease-free or overall survival in this subset of patients, particularly when surgical staging is comprehensive and lymph node-negative status is confirmed [20, 21].

Extensive nodal dissections such as PPLND may lead to significant perioperative and long-term morbidities (e.g., lymphocele, lower extremity lymphedema, chylous ascites, vascular or nerve injury) and may delay the initiation of adjuvant therapy [22]. Conversely, in high-risk EC, SLN mapping has been increasingly recognized as a staging strategy due to its potential to reduce nodal morbidity while preserving oncologic accuracy in appropriately selected patients [23].

Strength and limitations

The strength of our study lies in the strict inclusion of pathologically node-negative patients, which eliminates the potential confounding effect of nodal metastasis and allows for a more accurate evaluation of the therapeutic value of para-aortic dissection. Furthermore, the analysis spans three decades of data and includes a well-characterized cohort treated in a single high-volume center, ensuring consistency in surgical and pathological evaluation.

Nevertheless, the study is not without limitations. Its retrospective nature inherently introduces the possibility of selection bias. Decisions regarding the performance of para-aortic lymphadenectomy were not randomized but based on physician discretion, potentially influenced by intraoperative findings or patient characteristics. Additionally, although we adjusted for known prognostic factors in multivariate analysis, residual confounding from unmeasured variables cannot be excluded. An additional significant limitation is the absence of molecular tumor classification—specifically markers like POLE mutations, p53 alterations, and mismatch repair deficiency—which are now recognized as essential components in the personalized risk assessment and management of endometrial cancer. Finally, despite the relatively large sample size for a single-institution study, subgroup analyses—particularly when stratified by histology and lymphadenectomy type—may be underpowered to detect subtle but clinically relevant differences. Because some information regarding postoperative complications was missing due to the retrospective design of the study, these data are not presented.

Conclusion

Our findings suggest that the addition of para-aortic lymphadenectomy to pelvic lymphadenectomy does not offer a survival benefit in patients with early-stage grade 3 endometrioid and non-endometrioid endometrial cancer who are pathologically node-negative. However, in the grade 3 EC group that underwent PPLND, overall survival was poorer, although not statistically significant, and no difference was observed in the non-endometrioid group. Histologic subtype and age were the most significant independent predictors of survival, underscoring the importance of individualized risk assessment in surgical decision-making. Given the increased surgical complexity and potential morbidity associated with para-aortic dissection, its routine application in this patient population should be reconsidered. Prospective studies or large-scale multicenter analyses are warranted to further clarify which subgroups, if any, might benefit from extended lymphadenectomy.

Abbreviations

BMI

Body mass index

CA-125

Cancer antigen 125

CT

Computed tomography

DFS

Disease-free survival

EC

Endometrial cancer

FIGO

International Federation of Gynecology and Obstetrics

LVSI

Lymphovascular invasion

OS

Overall survival

PET/CT

Positron emission tomography combined with CT

PLND

Pelvic lymphadenectomy

PPLND

Pelvic and Para-aortic lymphadenectomy

SLN

Sentinel Lymph Node

RECIST

Response Evaluation Criteria In Solid Tumors

Authors’ contributions

Yakup Yalcin: Design, data collection, analysis, drafting the manuscript, manuscript writing.Bahadir Kosan, Mehmet Sakinmaz: Data collection and analysis.Serenat Yalcin, Kemal Ozerkan: Study conception, review of study design and review of draft manuscript and supervision.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and approved by the Bursa Uludağ University Clinical Research Ethics Committee (Protocol No. 2025 − 800/14–19). Due to the retrospective nature of the study, the ethics committee exempted the requirement for informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.