Secondary Logo

Share this article on:

Prognostic Factors for Curative Pelvic Exenterations in Patients With Recurrent Uterine Cervical or Vaginal Cancer

Sardain, Hugo MD*†; Lavoué, Vincent MD, PhD*†‡; Laviolle, Bruno MD, PhD‡§∥¶; Henno, Sébastien MD#; Foucher, Fabrice MD*; Levêque, Jean MD, PhD*†‡

International Journal of Gynecological Cancer: November 2014 - Volume 24 - Issue 9 - p 1679–1685
doi: 10.1097/IGC.0000000000000288
Surgeon’s Corner

Objectives The aim of this study was to assess the prognostic factors after curative pelvic exenterations performed for recurrent uterine cervical or vaginal cancers in the era of concomitant chemoradiotherapy.

Methods We retrospectively enrolled 16 patients with recurrent uterine cervical or vaginal cancer and tumor-free resection margins on pelvic exenteration pathological analysis between October 1997 and April 2014.

Results Pelvic exenterations were performed for 13 recurrent cervical cancers and 3 recurrent vaginal cancers. All of the patients had received pelvic irradiation (external radiotherapy for 14 patients and brachytherapy for 2 patients). The median age at the recurrence was 59.5 years (49–77 years), and the median tumor size was 4.35 cm (2–9 cm). There were no intraoperative or postoperative deaths. The 5-year disease-free survival and overall survival were 30% and 34.1%, respectively. The following 3 factors affected the disease-free survival: tumor size greater than 5 cm (P = 0.05), mesorectal lymph node involvement (P = 0.02), and vascular emboli (P = 0.0093).

Conclusions The presence of vascular emboli is a new prognostic factor in cases of recurrent cervical or vaginal cancer. Assessing the presence of vascular emboli on pretherapeutic biopsies could facilitate the selection of patients eligible for curative pelvic exenterations.

*Gynecology Department, Tertiary Surgery Center, Teaching Hospital of Rennes; †Faculty of Medicine, Rennes University; ‡Team Death Receptors and Tumor Escape, Inserm U1085; §Inserm, CIC-P0203 Clinical Investigation Centre; ∥Experimental and Clinical Pharmacology Laboratory, University of Rennes 1; and ¶Department of Clinical Pharmacology and #Pathology Department, Teaching Hospital of Rennes, Rennes, France.

Address correspondence and reprint requests to Vincent Lavoué, MD, PhD, Gynecology Department, Tertiary Surgery Center, Teaching Hospital of Rennes, Hôpital Sud, 16, Bd de Bulgarie, 35000 Rennes, France. E-mail:

The authors declare no conflicts of interest.

Received July 31, 2014

Received in revised form August 26, 2014

Accepted August 31, 2014

Concomitant chemoradiotherapy (CCR) increases survival in locally advanced cervical carcinoma (≥IB2 classification of the International Federation of Gynecology and Obstetrics [FIGO] 2009).1,2 It is estimated that 70% of patients with uterine cervical cancers were treated with CCR in developed countries.3 Despite these advances in treatment, a recurrence in the initially irradiated area occurs in 20% to 30% of patients.4 Most of these recurrences occur within 18 to 24 months after the initial treatment. The rate of recurrence is 10% for stage IB, 17% for stage IIA, 23% for stage IIB, 42% for stage III, and 74% for stage IV.4 There are fewer data regarding squamous cell cancers of the vagina because of their relative rarity, but their recurrence rate also increases with the initial stage.5 In cases of isolated pelvic recurrence, pelvic exenteration (PE) is currently the only therapeutic option.4 Although the procedure is mutilating, this radical surgery has improved outcomes since the first cases reported by Dr Brunschwig in 1948.6 Because of better preoperative patient selection and improved surgical techniques, the postoperative morbidity and mortality have become acceptable with survival rates at 5 years reaching up to 64%.4 However, studies in the literature are difficult to compare. There is heterogeneity in the inclusion criteria, including the inclusion of several types of cancers (eg, cancers of the cervix, endometrium, vulva, and ovaries).7–9 This heterogeneity most likely leads to bias in the analysis of patient survival.

The purpose of this study was to evaluate the prognostic factors for disease-free survival (DFS) in a monocentric consecutive series of patients treated for recurrent uterine cervical or vaginal cancer by curative PE, that is, with tumor-free resection margins, in the era of CCR.

Back to Top | Article Outline


We retrospectively studied a cohort of patients who underwent PE between October 1997 and April 2014 in the gynecological and surgical oncology center of the University Hospital of Rennes, France, which is a tertiary care center. The study was approved by the institutional review board of the French College of Obstetricians and Gynecologists (CEROG 2012-GYN-08-05). The inclusion criteria were centropelvic recurrence of uterine cervical or vaginal cancer, that is, limited to the vagina, bladder, and/or rectum, and treatment by curative PE. Curative exenteration was defined by tumor-free resection margins in the final pathological examination (macroscopic and microscopic). The exclusion criteria were malignancy other than uterine cervical or vaginal cancer (thus, recurrent ovarian or endometrial cancers were excluded) and tumor metastatic disease in the lymph nodes or visceral organs diagnosed during the staging before the PE. Finally, if macroscopic or microscopic tumor invasion was diagnosed in the resection margins during the final pathological examination, the patient was excluded.

Data were collected from the clinical records of patients and were classified into 4 groups. The first group of data focused on the diagnosis and initial treatment of the disease, including the patient age at the time of the initial diagnosis of uterine cervical or vaginal cancer, histological characteristics (type, squamous cell carcinoma or adenocarcinoma; the presence of vascular emboli; tumor location, endocervical or exocervical), FIGO stage, lymph node dissection (pelvic and/or para-aortic), and the type of treatment (concomitant radiochemotherapy or radiosurgical association). The second group of data focused on the diagnosis and assessment of tumor recurrence, including the time of recurrence with respect to the time of initial treatment, the histological characteristics of recurrences, and preoperative imaging (magnetic resonance imaging [MRI], computed tomography [CT], or positron emission tomography-CT [PET-CT] scans). The third group of data focused on the PE procedure, including the age of the patient at the time of exenteration, whether surgical biopsies were performed before exenteration (para-aortic lymphadenectomy, mediastinal biopsies) to remove metastatic disease, and the type and level of exenterations. An exenteration was total when the entire pelvic area was removed (uterus, bladder, and rectum). An anterior exenteration included removal of the bladder and uterus, and a posterior exenteration included the removal of the rectum and uterus. The resection level was defined following the work of Magrina et al,10 who described 3 resection levels corresponding to the levator anus muscles: type I is supralevator, type II is infralevator, and type III is infralevator with vulvectomy. The type of urinary, digestive, and pelvic floor reconstruction; the operative time; and the final histological characteristics (size, bladder, rectal or ureteral involvement, presence of vascular emboli, lymph node involvement) were identified. The length of hospital stay and information regarding various postoperative complications were collected. The complications were classified as early (within 30 days after surgery) and late (more than 30 days after surgery) according to the international classification of Clavien and Dindo.11 Grade I and II complications according to the Clavien-Dindo classification were considered minor, and grade III and IV complications were considered major. The fourth group of data focused on the survival and recurrences after PEs. Recurrences were clinically suspected and confirmed either through histological specimens or by imaging (CT). When recurrences were confirmed, the following information was collected: delay to PE and recurrence location. Recurrences were considered local if they were perineal and metastatic if the recurrence was nodal or visceral. In case of death, the time between the PE and death was recorded.

The primary outcome measure was the recurrence rate of oncological disease after PE. The secondary outcome was postoperative complications (early and late), with a focus on severe complications, defined as stages III and IV according to the Clavien-Dindo classification.

The statistical analyses were performed using SAS version 9.3 (Statistical Analysis System, Inc, Cary, NC).

The discontinuous data are presented as percentages (%), and the continuous data are presented as median values along with the minimum and maximum values.

The DFS and overall survival (OS) were calculated according to the Kaplan-Meier method. We studied the effects of clinical and histological data on recurrence using the log-rank test and a Cox model to define a hazard ratio (HR) for univariate analysis. Significant variables in univariate analysis were considered for inclusion in the multivariate analysis using a Cox model with an adjusted HR.

Back to Top | Article Outline


Between October 1997 and April 2014, 17 patients underwent surgical procedure at the teaching hospital of Rennes for recurrent uterine cervical or vaginal cancer with curative intent to treat. One patient was excluded because of positive margins on the final pathological examination. The study included 16 patients with tumor-free resection margins.

Back to Top | Article Outline

Epidemiological and Clinical Characteristics of Patients at Initial Diagnosis and Initial Treatment

At the time of initial diagnosis, 13 patients had uterine cervical cancer (81.25%), and 3 patients had vaginal cancer (18.75%). The median age was 54.5 years (34–76 years), and the median tumor size was 5 cm (2–8 cm). Of the 16 patients, 93.75% (15/16) had squamous cell carcinoma, and 6.25% (1/16) had adenocarcinoma.

The initial FIGO stage, the types of lymph node explorations, and the types of initial treatments are summarized in Table 1. Eleven patients (68.75%) underwent pelvic or lombo-aortic lymph node dissections: 6 patients before CCR and 5 patients with uterine removal surgical procedure. In 3 patients (27.3%), lymph node metastases were found. The laparoscopic approach was always used during lymph node staging. Eight patients (50%) were treated with CCR, and 2 of them underwent closure surgery. Four patients were treated exclusively with radiotherapy (25%). Finally, 2 patients (12,5%) were treated with radiotherapy and surgery combination (12.5%), and 2 patients (12.5%) were treated with surgery and brachytherapy. Radiotherapy regiment was 50 to 60 Gy.



Back to Top | Article Outline

Clinical Characteristics of Tumor Recurrence and Preoperative Assessment

The median time of recurrence after the initial treatment was 8 months (1–516 months).

Preoperatively, 81.25% (13/16) of the patients underwent MRI, and 68.75% (11/16) underwent PET-CT. The patients who did not undergo MRI underwent CT. At the end of this preoperative evaluation, metastatic disease was suspected in 5 patients. To formally eliminate these suspicions, 3 lombo-aortic lymphadenectomies, 1 biopsy of the cardia, and 1 mediastinal biopsy were performed. Two of the lombo-aortic lymphadenectomies were performed at the beginning of the PE. Pathological results were tumor free.

On the imaging (MRI), the median tumor size was 4 cm (1.5–8 cm), and all of the recurrences were centropelvic.

Back to Top | Article Outline

PE and Final Histology

At the time of the PE, the median age was 59.5 years (49–77 years).

The PE was total in 68.75% (11/16), anterior in 12.5% (2/16), and posterior in 18.75% (3/16) of the patients. The PE was supralevator in 62.5% (10/16) and infralevator in 37.5% (6/16) of the patients. When the surgery was infralevator, vulvectomy was performed (83.3%, 5/6). When infralevator exenteration was performed, pelvic reconstruction was performed with a gracilis muscle flap (100% of cases).

In 92.3% (12/13) of the cases, noncontinent urinary diversion (Bricker procedure; ileal conduit) was performed, and in 1 case, a transcutaneous ureterostomy was performed. Digestive diversion was performed with terminal colostomy in 92.8% (13/14) of the cases, and a colorectal anastomosis was performed in 1 case.

The median operative time was 6 hours (3.5–7 hours).

The median tumor recurrence size at the final pathological examination was 4.35 cm (2–9 cm). Pathological analysis showed squamous cell carcinoma in 15 cases (93.75%) and high-grade urothelial carcinoma in 1 case (6.25%). There were vascular emboli in 6 cases (37.5%). There were bladder involvement in 10 cases (76.9%), rectal involvement in 12 cases (85.7%), and mesorectal nodal involvement in 3 cases (18.75%).

Back to Top | Article Outline

Postoperative Complications

The median hospitalization time was 26 days (16–52 days).

Early complications occurred in 12 patients (75%), although only 4 of these patients (25%) had severe complications. The minor complications (grades 1 and 2 according to the Clavien-Dindo classification) were pyelonephritis and subocclusive syndrome, which were medically treated. The patients with major complications (grades 3 and 4) had ileo-ileal anastomosis failure, colostomy obstruction in parietal transition, pelvic peritonitis, or seroma of the lower limb. Three of these complications required surgical reintervention. All of these severe complications occurred in women with supralevator exenteration.

A total of 68.75% (11/16) of the patients had late complications, and these complications were severe in 6 patients (37.5%). Tumor recurrence was the cause of these severe complications in 66% of the patients (4/6). Not including these recurrences, the rate of severe complications was 12.5% (2/16). The complications are summarized in Table 2.



Back to Top | Article Outline

Postoperative Follow-Up and Clinical Characteristics of Recurrences

The median follow-up was 11.5 months (4–163 months).

Recurrences occurred in 11 patients (68.75%); 10 of these patients died, and one was still alive. Five patients were alive without recurrence (31.25%).

The median time of recurrence after PE was 4.5 months (2–15 months). Recurrences were local in 36.3% (4/11), metastatic in 36.3% (4/11), and synchronously local and metastatic in 27.4% (3/11) of the cases. In 75% of the local recurrence cases (3/4), patients were treated with supralevator exenteration.

When a recurrence occurred, palliative treatment was established, including palliative surgery, palliative chemotherapy, or palliative follow-up with appropriate treatment of pain.

The 5-year DFS was 30%, and the 5-year OS was 34.1%.

Back to Top | Article Outline

Clinical and Histological Prognostic Factors

The prognosis of recurrence was not significantly related to the characteristics of the original tumor (FIGO stage, lymph node involvement, type of initial treatment).

The characteristics of tumor recurrence in relation to patient prognosis are summarized in Table 3. In the univariate analysis, 3 factors significantly influenced DFS; these factors were tumor recurrence larger than 5 cm (P = 0.05), mesorectum lymph node involvement (P = 0.02), and the presence of vascular emboli based on the histological analysis (P < 0.01) (Fig. 1). The other criteria studied were not significant, including the time to recurrence after initial treatment and the level of surgical resection (infralevator or supralevator). In the multivariate analysis, none of these factors were statistically significant.





Back to Top | Article Outline


The PE is the only curative therapeutic option when a pelvic recurrence of uterine cervical or vaginal cancer is diagnosed, and many prognostic factors influence patient survival. Our study shows that the presence of vascular emboli is a new and major prognostic factor after curative PE (ie, pathological tumor-free resection margins), and vascular emboli should be considered in addition to other prognosis factors before proceeding with PE.

Since the Shingleton et al12 study in 1989, it has been assumed that the main prognostic factor for DFS and OS is the presence of tumor-free resection margins in the final surgical analysis.13,14 Other known prognostic factors include a recurrence time of less than 18 months or 2 years,9,15 lymph node involvement at the time of initial treatment,16 and perinervous invasion.7 We also determined that 2 previously known factors are prognostic indicators, namely, tumor recurrence size greater than 5 cm11 and mesorectal lymph node involvement.17 Our study identified a new DFS prognostic factor after PE, namely, vascular emboli at the pathological examination, with a 5-year DFS of 0 for these patients (P = 0.0093). This factor was previously known to be poor at the time of initial management of uterine cervical cancer because of the metastatic potential of these emboli.18 In the future, recurrence diagnostic biopsies should be evaluated for vascular emboli. Indeed, vascular emboli were currently assessed by pretherapeutic biopsy or cone biopsy in primary uterine early cervical cancer.19

Because of improved patient selection, the 5-year survival rates after PE range from 24% to 64% in the literature.4 Our DFS and OS rates were 30% and 34.1%, respectively, which are in the lower part of this range, although our patients were treated with curative PE. To select appropriate candidates, clinical characteristics and modern preoperative imaging are the most important diagnostic tools for recurrence assessments. The preoperative imaging, including MRI and PET-CT, has to show that a recurrence is local without lymph node involvement or distant metastases and without pelvic wall involvement. In our study, 81.25% of the patients underwent MRI, and 68.75% underwent PET-CT, which are comparable with the rates in the literature (22%–100% of the patients underwent PET-CT).4 In the present study, if any suspicious metastases were observed on preoperative imaging, further explorations were performed as biopsies or lombo-aortic lymph node dissections before the PE, and all of the specimens were histologically negative. Despite this active approach, when a recurrence was diagnosed after PE, it was metastatic in 7 of the 11 cases (63.7%), which most likely explains our survival rates. In the literature, the metastatic recurrence rates vary from 20% to 40%.12 These findings reflect the need to determine selection criteria, such as vascular emboli, before treatment before confirming the surgical indications for PE. Another factor could also explain our survival rates. Baiocchi et al determined that squamous cell carcinoma was a poor prognostic factor compared with adenocarcinoma (P = 0.03) because of its higher perineural invasion.4 In our study, squamous cell carcinoma was diagnosed at the time of PE in 93.75% (15/16) of the cases. Furthermore, in the present study, the median time of recurrence after initial treatment is low (8 months) compared with literature (18–24 months).4 This point is probably due to selection bias because our tertiary cancer center is a reference center with referred serious cases. This low median time of recurrence after initial treatment probably induced the low 5-year survival rates in the present study despite curative PE.

In the literature, early complication rates vary from 20% to 80%; however, some of the series did not use the international classification system,7,8 which makes comparisons between studies difficult. We found an acceptable rate of severe complications of 25% with the Clavien-Dindo classification. All of the patients with severe complications had undergone supralevator exenteration, which is in accordance with data from studies that determined that the surgical techniques of pelvic reconstruction used for type 3 exenterations decreased pelvic abscess, fistula, and occlusion rates.20 Surgical reintervention was required in 18.75% (3/11) of cases, which is low compared with the literature data; with the exception of McLean et al,21 who found 4% of rate of reintervention, the reintervention rates vary from 23% to 38%.7,9 When severe late complications were diagnosed, the etiology was tumor recurrence in 66% of the cases. The major late complications were predominantly digestive fistula and obstruction of 22 the urinary tract, which is comparable with literature data.4

Our study has some limitations. First, we collected the data retrospectively; however, all of the studies on PE are retrospective. Second, this was a short study with 16 patients, which prevents significant multivariate statistical analysis, and the results should be interpreted with caution. However, our patients were carefully selected to undergo curative PE, that is, with tumor-free resection margins, for recurrent uterine cervical or vaginal cancer after pelvic irradiation. To our knowledge, no study investigated the prognostic factors after curative PE.

In cases of recurrences after the pelvic irradiation of uterine cervical or vaginal cancer, patient selection for curative PE is essential. This surgery is a mutilating procedure, although we found an acceptable morbidity rate (25% of rate of severe early complications) and a poor prognosis (5-year DFS of 30%) despite curative surgery (ie, tumor-free resection margins at the final pathological examination). Before confirming the surgical indication for PE, several prognostic factors must be evaluated, including recurrence tumor size and mesorectum lymph node involvement on modern imaging and vascular emboli on pretherapeutic biopsies. The presence of vascular emboli is a new poor prognostic factor that should be considered before performing PE in patients with recurrent uterine cervical or vaginal cancer.

Back to Top | Article Outline


1. Pecorelli S, Zigliani L, Odicino F. Revised FIGO staging for carcinoma of the cervix. Int J Gynaecol Obstet. 2009; 105: 107–108.
2. Reducing uncertainties about the effects of chemoradiotherapy for cervical cancer: a systematic review and meta-analysis of individual patient data from 18 randomized trials. J Clin Oncol. 2008; 26: 5802–5812.
3. Gadducci A, Tana R, Cosio S, et al. Treatment options in recurrent cervical cancer (review). Oncol Lett. 2010; 1: 3–11.
4. Peiretti M, Zapardiel I, Zanagnolo V, et al. Management of recurrent cervical cancer: a review of the literature. Surg Oncol. 2012; 21: e59–e66.
5. Beller U, Benedet JL, Creasman WT, et al. Carcinoma of the vagina. FIGO 26th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet. 2006; 95: S29–S42.
6. Brunschwig A. The surgical treatment of advanced carcinoma of the cervix. N Y State J Med. 1948; 48: 1733.
7. Baiocchi G, Guimaraes GC, Rosa Oliveira RA, et al. Prognostic factors in pelvic exenteration for gynecological malignancies. Eur J Surg Oncol. 2012; 38: 948–954.
8. Kaur M, Joniau S, D’Hoore A, et al. Pelvic exenterations for gynecological malignancies: a study of 36 cases. Int J Gynecol Cancer. 2012; 22: 889–896.
9. Maggioni A, Roviglione G, Landoni F, et al. Pelvic exenteration: ten-year experience at the European Institute of Oncology in Milan. Gynecol Oncol. 2009; 114: 64–68.
10. Magrina JF, Stanhope CR, Weaver AL. Pelvic exenterations: supralevator, infralevator, and with vulvectomy. Gynecol Oncol. 1997; 64: 130–135.
11. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004; 240: 205–213.
12. Shingleton HM, Soong SJ, Gelder MS, et al. Clinical and histopathologic factors predicting recurrence and survival after pelvic exenteration for cancer of the cervix. Obstet Gynecol. 1989; 73: 1027–1034.
13. Marnitz S, Kohler C, Muller M, et al. Indications for primary and secondary exenterations in patients with cervical cancer. Gynecol Oncol. 2006; 103: 1023–1030.
14. Berek JS, Howe C, Lagasse LD, et al. Pelvic exenteration for recurrent gynecologic malignancy: survival and morbidity analysis of the 45-year experience at UCLA. Gynecol Oncol. 2005; 99: 153–159.
15. Chiantera V, Rossi M, De Iaco P, et al. Survival after curative pelvic exenteration for primary or recurrent cervical cancer: a retrospective multicentric study of 167 patients. Int J Gynecol Cancer. 2014.
16. Moutardier V, Houvenaeghel G, Martino M, et al. Surgical resection of locally recurrent cervical cancer: a single institutional 70 patient series. Int J Gynecol Cancer. 2004; 14: 846–851.
17. Mourton SM, Chi DS, Sonoda Y, et al. Mesorectal lymph node involvement and prognostic implications at total pelvic exenteration for gynecologic malignancies. Gynecol Oncol. 2006; 100: 533–536.
18. Quinn MA, Benedet JL, Odicino F, et al. Carcinoma of the cervix uteri. FIGO 26th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet. 2006; 95: S43–S103.
19. Kodama J, Kusumoto T, Nakamura K, et al. Factors associated with parametrial involvement in stage IB1 cervical cancer and identification of patients suitable for less radical surgery. Gynecol Oncol. 2011; 122: 491–494.
20. Ferron G, Martel P, Querleu D. Vaginal reconstruction after pelvic exenteration: when and which techniques? Bull Cancer. 2003; 90: 435–440.
21. McLean KA, Zhang W, Dunsmoor-Su RF, et al. Pelvic exenteration in the age of modern chemoradiation. Gynecol Oncol. 2011; 121: 131–134.

Pelvic exenteration; Recurrent cervical or vaginal cancer; Vascular emboli; Prognostic factors

© 2014 by the International Gynecologic Cancer Society and the European Society of Gynaecological Oncology.