Resecting Beyond the Central Pelvic Compartment in Pelvic Exenterations for Gynaecological Malignancies: Outcomes From a Tertiary Unit

ABSTRACT

Background

Pelvic exenterations (PE) in solid organ gynaecological malignancies have historically been confined to the central pelvic compartment but an expansion of resection extent to include the lateral compartment is now technically feasible. There remains uncertainty about its influence on outcomes.

Methods

Retrospective study of consecutive patients undergoing PE for non‐ovarian gynaecological malignancies in a tertiary referral unit. Two groups were assessed and compared: standard resections including but not beyond the pelvic sidewall (SR); and extended resections beyond the pelvic side wall (lateral compartment excisions; internal iliac vessels and/or obturator internus/piriformis resection and/or bone resection) (ER).

Results

Thirty‐one patients underwent 32 PE: of these, 21 patients underwent SR (65.6%) with 11 who underwent ER (34.4%). Clear microscopic resection margins (R0) were achieved in 72.7% for ER versus 90.5% for SR. The overall 30‐day major complication rate (≥Clavien‐Dindo III) was 45.5% for the ER group versus 19.9% for SR. Median length‐of‐stay for ER was longer than for SR (38 vs. 18.5 days). The 1‐year DFS for ER was 66.7% versus 77.8% for SR. The 2‐year DFS for ER group was 60% versus 57.1% for SR.

Conclusions

In this study of contemporary patients, the short‐term patient outcomes of ER were consistent with other similar cohorts in the literature and comparable to SR in this cohort. There is an urgent need for larger multicentre studies to clarify the influence of PE resectional extent on resection margin status, patient outcomes and survival.

1. Introduction

Pelvic exenteration (PE) was first described as a palliative surgical option for recurrent and advanced gynaecological malignancy but has now evolved into a gold‐standard therapeutic option with curative intent [1].

A negative resection margin has consistently been shown to be the most important determinant of recurrence free survival following pelvic exenteration across different surgical cohorts [2]. Improved surgical technique, neoadjuvant and adjuvant multimodal systemic anticancer therapeutic options have helped contribute to increasing R0 resection rates and survival in this patient group [1, 3, 4].

Specifically, there has been an expansion of resection extent in PE; while previously, malignancies with lateral pelvic sidewall involvement were considered irresectable, PE extending into the lateral pelvic compartment can now be offered to achieve clear margins [1, 5]. While there has been an evolution towards more radical PE, reliable reports of clinical outcomes remain scarce.

The aim of this study was to determine the influence of lateral resectional extent on outcomes of PE for gynaecological malignancy.

2. Methods

A retrospective cohort study was performed of consecutive patients undergoing PE for primary and recurrent gynaecological malignancies from 2018 to 2022 in a single tertiary centre covering the West of Scotland geographical region (Glasgow Royal Infirmary). PE was defined as resection of organs in 2 or more contiguous pelvic compartments performed by a multidisciplinary team of pelvic exenterative surgeons. All consecutive patients undergoing PE with curative intent were included in this cohort. Exclusion criteria were procedures for ovarian malignancy, and benign disease on final histopathology and procedures performed for palliative non‐curative intent.

Procedure, disease characteristics and clinical outcomes were collected from retrospective review of electronic patient records.

In our centre, surgical margin status was defined as R0 if all resection margins were microscopically clear of tumour by > 1 mm.

Post‐operative complications were classified using the Clavien‐Dindo (CD) grading system [6]. The 30‐day all complications rate and major complications rate was defined as the proportion of patients suffering, at least one complication of any grade and at least one complication graded CD3 or above, respectively, in the first 30 post‐operative days. The specific complication rates were defined as the number of patients suffering from the specified complication in the first 30 post‐operative days. Within individual patients, where multiple treatments are administered for a specific complication, the highest corresponding CD grade is assigned.

Disease‐free‐survival was defined as the length of time in months from date of surgery to date of death, or local or distance recurrence, or in the absence of these events, date of last follow‐up.

Disease‐free‐survival (DFS) rates at 1 and 2 years were defined by the proportion of patients alive and free of local or distant recurrent disease at those time‐points from the date of surgery.

Overall survival was defined as the length of time from date of surgery to date of death or if patient is alive, date of last follow‐up.

2.1. Pre‐Operative Staging of Patients

All patients were routinely staged with magnetic resonance imaging (MRI) of the pelvis and contrast‐assisted tomography (CT) of the chest, abdomen and pelvis. Positive emission tomography (PET) was used selectively in patients where regional or distant metastases were suspected but not fully elucidated by CT or MRI.

All patients are discussed in a multidisciplinary team (tumour board) setting to guide management. In the local setting, patients with widespread and/or irresectable metastatic disease would in general be considered a contraindication to proceeding to PE with curative intent.

2.2. Surgical Technique and Extent of Resection

All patients had PE tailored to the specific degree of disease involvement across the anatomical pelvic compartments with the aim of achieving a negative surgical margin with curative intent.

All procedures were performed via a midline laparotomy. All patients within this cohort underwent routine lateral pelvic side wall dissection, including the pelvic side wall fascia, obturator lymph nodes and distal internal iliac branches. This reflects our practice in pelvic exenteration for gynaecological disease to ensure the plane is entered wide on the central portion of the exenteration envelope around the cervix to reduce the risk of a positive microscopic resection margin. In the smaller subset of patients with genuine lateral pelvic wall involvement, lateral compartment dissection envelope is extended to the deeper side wall structures involving liberal resection of internal iliac vessels, incorporating obturator internus and piriformis with the specimen, with or without bone resection (or a combination of the aforementioned components) to try and achieve an R0 resection based on pre‐operative imaging and intra‐operative findings.

Patients undergoing resection with the lateral margins confined to the side wall fascia, obturator nodes and distal internal iliac branches were defined as having undergone a standard PE (SR). Any patients with a resection extending laterally beyond this, incorporating components of the lateral pelvic compartment, as detailed above, were defined as having undergone an extended PE (ER).

An ileal conduit is the urinary reconstructive option of choice in total and anterior pelvic exenterations. Depending on the extent of the perineal resection required, either a permanent end colostomy is matured or a primary stapled colonic anastomosis is performed. Similarly, the perineum is closed primarily, or a flap reconstruction is performed, tailored to the size of the perineal defect.

2.3. Statistical Analyses

Continuous data were presented as numerical ranges with means or medians. Categorical data were reported as frequencies and/or proportions. Comparisons between key pre‐operative categorical variables were analysed using Chi‐square tests or Fisher exact tests with or without Monte‐Carlo approximation for categorical data and Mann–Whitney tests were used for continuous data. Kaplan–Meier curves were used to analyse survival. In survival analysis, survival for the included patient (n = 1) who underwent 2 PE within the study period is in reference to the first procedure. All statistical analysis was performed using SPSS (Version 28, IBM Inc., Armonk, NY).

2.4. Ethics

As this study involved the use of anonymised, retrospective data, formal ethical review was not required. Permission to collect, and maintain this database was granted by the local institutional review body governing the use of clinical data (NHS Greater Glasgow and Clyde Caldicott Guardian). This report is part of a larger study registered at clinicaltrials.gov (NCT05696002). All stored and analysed data were anonymised.

3. Results

3.1. Patient and Procedure Characteristics

There were 37 patients identified who underwent 38 PE for gynaecological malignancy (Figure 1). A total of 6 patients were excluded from further analysis as 5 patients had malignancies of ovarian origin and 1 patient had benign pathology on final operative specimen histopathology. The final cohort therefore comprised of 31 patients undergoing 32 PE. No patients had occult metastatic disease found intraoperatively which precluded proceeding with the planned PE with a curative intent.

FIGURE 1.

Patient inclusion flowchart.

A single patient had two PE within the study period 40 months apart. Both were SR procedures (initially an anterior pelvic exenteration followed by a completion posterior pelvic exenteration) for recurrences of endometrioid adenocarcinoma. Histopathology of both tumour specimens were morphologically similar.

The whole cohort comprised 21 SR patients (65.6%) and 11 ER patients (34.4%). Median follow‐up of the cohort was 18 months.

There were no statistical differences in any of the clinicopathological and patient characteristics between the two groups (Table 1).

TABLE 1.

Disease and procedure characteristics.

	Standard resections, n = 21 (%)	Extended resections, n = 11 (%)	p
Age	53 (30–76)	42 (36–77)	0.721
ASA grade ≥ 3	8 (38.1)	6 (54.5)	0.465
Recurrent/progressive/persistent malignancies after previous treatment	17 (81)	11 (100)	0.272
Previous surgical resection of tumour	9 (42.9)	4 (36.4)	1.0
Previous pelvic radiotherapy	16 (76.2)	11 (100)	0.138
Previous chemotherapy	17 (81)	10 (90.9)	0.637
Malignancy site of origin
Vulval	3 (14.3)	1 (9.1)	0.303
Vaginal	6 (28.6)	2 (18.2)
Cervical	6 (28.6)	6 (54.5)
Uterine	6 (28.6)	2 (18.2)
Cancer histological subtype
Squamous cell carcinoma	9 (42.9)	9 (81.8)	0.089
Adenocarcinomas (including endometrioid carcinomas)	7 (33.3)	2 (18.2)
Other (signet ring, leiomyosarcoma)	5 (23.8)	0 (0)
Type of exenteration
Total	13 (61.9)	9 (81.8)	0.628
Anterior	5 (23.8)	1 (9.1)
Posterior	3 (14.3)	1 (9.1)
Bone resection	0 (0)	2 (18.2)	0.111
Urinary diversion and reconstruction	18 (85.7)	10 (90.9)	1.0
Perineal closure with flap	12 (57.1)	9 (81.8)	0.248
Type of flap
Gracilis	6 (50)	4 (44.4)	0.544
ALT thigh	2 (16.7)	4 (44.4)
Gluteal/buttock	3 (25)	1 (11.1)
VRAM	1 (11.1)	0 (0)

Abbreviations: ALT, antero‐lateral thigh; VRAM, vertical rectus abdominus muscle.

The majority of patients had PE for recurrent, progressive or persistent disease after initial treatment (surgery and/or systemic anticancer therapy) (ER 100% vs. SR 81%). The rate of previous surgical resection of their tumour before their index PE was 36.4% for ER and 42.9% for SR. Most patients had previous pelvic radiotherapy (ER 100% vs. SR 76.2%). Similarly, most patients had previous chemotherapy for their disease prior to the index PE (ER 90.9% vs. SR 81%).

For ER, the most common site of disease origin was the cervix (54.5%) followed by the vagina and uterus (both 18.2%). For SR, the most common sites of disease origins were vaginal, cervical and uterine (each 28.6%).

The most common pathological subtype of cancers for both ER and SR were squamous cell carcinomas (ER 81.8% and SR 42.9%) followed by adenocarcinomas (ER 18.2% vs. SR 33.3%).

The majority of resections were total pelvic exenterations (ER 81.8% vs. SR 61.9%). Urinary diversion/reconstructions were performed in the majority of cases (ER 90.9% vs. SR 85.7%).

Similarly, the majority of cases had perineal flap reconstructions (ER 81.8% vs. SR 57.1%). For ER, the most common type of flap reconstruction were gracilis and ALT thigh flaps (both 44.4%). For SR, gracilis flaps were the most common reconstruction method (50%). A total of 2 bone resections were performed in the cohort, both of which were ER patients.

3.2. Resection Margins

R0 was achieved in 84.4% of the cohort, corresponding to an R0 rates of 72.7% and 90.5% for ER and SR, respectively. The median distance from tumour to closest surgical margin in the whole R0 resection group was 6 mm (range, 2–18 mm). Of the 5 R1 cases, 1 case had a tumour which had 0.5 mm clearance from the closest surgical margin. The other R1 tumours were present at the resection margin. In the single patient undergoing 2 PE, R0 was achieved in both procedures.

3.3. Post‐Operative Complications and Survival

The overall 30‐day major complication rate (Clavien‐Dindo IIIa, IIIb and IV) was 45.5% for ER versus 19.9% for SR (Table 2). In terms of 30‐day all complication rates, 90.9% for ER had developed complications versus 85.7% for SR.

TABLE 2.

Comparative complication outcomes by type of pelvic exenteration.

	Standard resections, n = 21 (%)	Extended resections, n = 11 (%)
30‐day all complications	18 (85.7)	10 (90.9)
30‐day major complications	4 (19.0)	5 (45.5)
30‐day re‐operation	2 (9.5)	2 (18.2)
30‐day readmission to Intensive care	0 (0)	0 (0)
30‐day re‐admission to hospital after discharge	1 (4.8)	2 (18.2)
Length of stay, median days (range)	18.5 (9–308)	38 (16–50)

In terms of specific complications (Table S1), a low haemoglobin level requiring post‐operative packed red cell transfusions was the most common minor complication in ER whereas surgical site infections were the most common specific minor complication for SR. Urinary leaks and myocutaneous flap issues were the most common major complications for ER.

For ER, 5 patients experienced major complications in the first 30 post‐operative days comprising 2 patients developing pelvic collections managed by percutaneous drainage including one patient developing their pelvic collection due to a urinary anastamotic leak, 2 patients developing flap issues requiring wound refashioning, and one patient developing a urinary leak which was managed non‐operatively with antibiotics and single‐organ support. For SR, four patients developed major complications. One patient required a percutaneous drainage of a pelvic collection while another patient requiring a flap refashioning under local anaesthesia. The third patient developed a urinary leak requiring a laparotomy and subsequently multi‐organ support in an intensive care setting. This patient then unfortunately developed an anastomotic leak from the conduit small bowel anastomosis and had a second re‐laparotomy. The last patient in this group developed a rectal stump blow‐out necessitating a transrectal pelvic drain insertion under local anaesthesia. This patient also subsequently developed a distal obstruction in her urinary conduit necessitating bilateral nephrostomies.

The 30‐day re‐operation rate was 18.2% versus 9.5% for the ER and SR groups respectively. For ER, 2 patients had reoperations for flap related issues. For SR, one patient had a reoperation for a flap related issue and one patient had a relaparotomy for urine leak.

There were no readmissions within the first 30 post‐operative days to intensive care. There were two readmissions back into hospital within 30 days of discharge for ER versus one readmission for SR.

Median length of stay for ER was longer than SR (median 38 vs. 18.5 days).

For the whole cohort, 1‐ and 2‐year DFS were 73.9% and 57.9% respectively (Table 3).

TABLE 3.

Survival and recurrence outcomes by resection extent and resection margin status, n = 31 ^a .

	All patients, n = 31 (%)	Standard resections, n = 20 (%)	Extended resections, n = 11 (%)
30‐day mortality	0 (0)	0 (0)	0 (0)
90‐day mortality	0 (0)	0 (0)	0 (0)
1‐year disease free survival (n = 23) (SR n = 18) (ER n = 6)	17 (73.9)	14 (77.8)	4 (66.7)
2‐year disease free survival (n = 19) (SR n = 14) (ER n = 5)	11 (57.9)	8 (57.1)	3 (60.0)
Margin status	R0, N = 26 (%)	R1, N = 5 (%)
1‐year disease free survival (R0 n = 19) (R1 n = 5)	15 (78.9)	3 (60.0)
2‐year disease free survival (R0 n = 15) (R1 n = 4)	8 (53.3)	2 (50.0)

^a In survival analysis, survival for the included patient (n = 1) who underwent 2 PE within the study period is in reference to the first procedure.

The 1‐year DFS for ER was 66.7% versus 77.8% for the SR group. The 2‐year DFS for the ER group was 60% versus 57.1% for the SR group.

Stratifying by surgical margin or status, R0 resections were associated with a 1‐ and 2‐year DFS of 78.9% and 53.3% respectively, while R1 resections were associated with a 1‐ and 2‐year DFS of 60% and 50%, respectively.

Survival curves relating to overall and disease‐free survival in both groups are demonstrated in Figure 2.

FIGURE 2.

Disease free and overall survival in SR and ER groups*. *In survival analysis, survival for the included patient (n = 1) who underwent 2 PE within the study period is in reference to the first procedure.

4. Discussion

Our results suggest that expanding the extent of resection in PE to encompass the lateral compartment yields acceptable R0 rates in the context of lateral pelvic recurrence. Both procedures had comparable intermediate term survival and largely similar complication rates; the modest sample size precludes interpretation of formal statistical testing of differences between these groups.

The cohort's patients are often young patients with advanced disease or those who have failed first line therapy. These patients often have limited curative options and thus there is an increasing interest in strategies which optimise an R0 resection.

The R0 rates of 72.7% and 90.5% for ER and SR (84.4% overall), respectively, in our cohort are in line with other undifferentiated cohorts in the literature. The PelvEx Collaborative multicentre study reporting 299 patients from 22 centres from 2006 to 2017 with cervical, vaginal or endometrial cancer (primary and recurrent disease) reported an R0 rate of 73%–87% depending on cancer site [2]. Similarly in a multicentre German and Italian study involving 167 patients with primary or recurrent cervical cancer, the R0 resection rate was 72.5% [7].

Lago et al. in a more contemporary analysis of a single centre study of pelvic exenterations for gynaecological malignancies reported an R0 rate of 91% but defined R0 resections as resections with no microscopic tumour at the margin [8]. The lower R0 resection rate for ER (72.7%) likely reflects more advanced disease and therefore, more technically challenging surgery in this group [7]. This rate however remains compatible with other undifferentiated cohorts undergoing PE in the literature.

Comparative evidence for the effect of extended lateral compartments resections versus standard resections are limited by relatively modest cohort sizes, differing definitions of resectional extent, inclusion and exclusion criteria, treatment selection bias and R0 definitions.

For example, Hockel et al. report a cohort of 102 pelvic exenterations and describe lateral compartment excision with a 97% R0 rate; however only 76 patients in the cohort had disease fixed to the pelvic side wall [9]. In another case series of relapsed pelvic malignancies, Hockel reports an R0 rate of 98.8% but this cohort was a mixed cohort of both patients with and without lateral compartment resections [10]. Similarly, Park et al. describe a cohort of lateral compartment excisions with a 92.9% R0 rate but the study was limited to recurrent cervical carcinoma and had a less conservative R0 definition; the absence of microscopic tumour at the resection margin [11]. A similarly contemporary cohort by Vizzieli et al. of patients with pelvic side wall involvement of gynaecological malignancies was associated with an R0 rate of 85.7% [12]. In Kanao's series of pelvic exenterations (n = 17) and laterally extended endopelvic resections (LEER) (n = 11), all performed in a laparoscopic approach, for recurrent post‐irradiated recurrent cervical, meanwhile, a pathological complete resection was achieved in 100% and 73% respectively [13].

Definitions of R0 resection are important when interpreting and comparing these studies. Park et al., Lago et al. and Vizzieli et al. all defined R0 resections as the absence of microscopic tumour at the margin regardless of distance in contrast to our more conservative definition [8, 11, 12].

Our definition is similar to the one used by the PelvEx collaborative study and is extrapolated from primary rectal cancer in which there is evidence that microscopic tumour within 1 mm are associated with worse prognosis and local recurrence rates than margins > 1 mm [2, 14]. In gynaecological malignancies, different thresholds and small sample sizes are likely explain the varying degrees of association between distance of tumour to closest resection margin and patient outcomes [15, 16, 17].

Our results show that both our groups have post‐operative complication rates compatible with rates reported in the literature; SR 19.0% versus ER 45.5%, respectively. This is compatible with the major complication rates in the PelvEx collaborative ranging from 15%–60% depending on cancer site [2]. Hockel's single centre cohort of 102 cancers treated with pelvic lateral compartment resections in contrast had a moderate‐to‐severe complication rate of about 70% [9]. In a single centre case series by Hockel of patients undergoing resections for relapsed pelvic malignancies the overall rate of both moderate to severe early and late complications was 57% although this cohort includes a proportion of patients without a lateral compartment resection [10]. In the context of the wider literature, Ubinha et al.'s narrative review of pelvic exenterative surgery in cervical cancer, they reported a general and major complication rate ranging from 25% to 94% and 23% to 70%, respectively [18].

There was a longer length of stay for ER versus SR. While further interpretation is limited due to the small sample sizes, these trends reflect the recognised morbidity of major surgery in PE and perhaps reflect the more radical nature of ER. These is in keeping and consistent with reports of other cohorts and consistent with Ubinha et al.'s inference that the significant increase in length of inpatient hospital stay observed over the last few years can be explained by an increase radicality in procedures [18]. Kanao et al.'s cohort reported a median stay of 34 days for pelvic exenteration and 41 days for LEER respectively [13].

The short and intermediate term mortality outcomes were also compatible with other cohorts in the literature. The absence of 30‐ and 90‐day mortalities in our cohort is consistent with the 0.7%–3% 30‐day mortality in other cohorts [2, 19].

Similarly, our 2‐year survival rates are compatible with the 3‐year‐overall‐survival rates in the PelvEx collaborative 12.5%–48% depending on site and resection margin status [2].

The contemporaneous 4‐year period of this cohort is a strength compared with some reported cohorts which span sometimes longer periods of time with changes in operative and non‐operative factors across time potentially confounding outcomes. These rapid changes towards more complex, aggressive surgery in more patients in recent years has been eloquently elucidated by Waters et al. [20]. In Solomon et al.'s single‐centre cohort, the effects of an evolution in surgical techniques, including adoption of a more lateral plane of excision, are seen when examining a large cohort over a 20‐year interval, with improved R0 resection rates from 21% to 66.5% [21].

As mentioned, the small cohort size and limited longer‐term follow‐up are limitations of this study, as is selection bias due to its retrospective nature. This study was not statistically powered for patient outcome or survival analysis. This is a limitation of many PE studies and indeed the cohort sizes are similar to other single centres such as the MSKCCC cohort which comprised 71 R0 patients over a 8‐year period [22]. Similarly, Vizzieli et al.'s cohort included 63 patients over a 10‐year period from 2007 to 2017 [12].

Similar to the heterogenous definitions of R0, another key issue are the heterogenous technical definitions. The more lateral dissection described is broadly but not entirely synonymous with a LEER and this reflects the inherent challenges of studying these patients with a tailored resectional extent to their disease in a clinical study [9].

There is an urgent need to fully eludicate the relationship between the radicality of resection, resection margin status, complications and survival from interrogating larger clinical cohorts to clarify the risks and benefits of such resectional strategies.

There also remains an unmet research need for patient reported outcomes in PE which are magnified in the context of considering the utility of a more radical resectional strategy [23]. Recent published data by Steffens et al. for example while shows sustained quality‐of‐life improvements over the course of the patients operative journey from baseline pre‐operative assessments in patients undergoing pelvic exenteration for advanced and recurrent pelvic malignancies [24]. The authors however noted a higher proportion of partial exenterations in the long term survivor cohort and postulate that this may partially explain the favourable quality of life in the long‐term survivor cohort, highlighting again, as previously mentioned, the relationship between surgical radicality and quality of life as an urgent clinical and research question [24].

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Number of specific 30‐day complication rates by severity.

Kong C. Y., Galbraith N. J., McCormick A., et al., “Resecting Beyond the Central Pelvic Compartment in Pelvic Exenterations for Gynaecological Malignancies: Outcomes From a Tertiary Unit,” Australian and New Zealand Journal of Obstetrics and Gynaecology 66, no. 1 (2026): e70079, 10.1111/ajo.70079.

Data Availability Statement

Data not available—participant consent.