Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Pelvic exenteration (PE) is a very complex procedure, even currently. In addition to unique surgical erudition and interdisciplinary surgical team involvement, these operations require special expertise for the right candidate selection, postoperative care, and handling of complications. Because of the narrow spectrum of indications, PE is a rare procedure, and so the only solution for achieving good results is establishing a system of candidate centralization. Such a system requires an excellent interplay among all experts and an agreement among gynecologic oncologic institutions for the referral of candidates to a few centers, which can therefore accumulate sufficient experience.
Achieving this ideal network can be limited by an array of obstacles, such as reluctance of an institution to refer its patients to a different institution; late diagnosis of local progression in patients after primary radiotherapy; consideration of PE after too long an interval after detection of recurrence, often as late as after palliative chemotherapy has been given; or even the unavailability of these procedures in many regions in the world. Results of a survey evaluating the patterns of care in 61 institutions in Germany and the United States were published in 2009. In these two countries, which both have a high quality of care in gynecologic oncology, only four centers carried out more than 10 PEs a year. In the Czech Republic, based on data from the National Oncology Register (NOR) and the National Register of Hospitalized Patients (NRHOSP), PE is performed in only approximately 20% of the assumed number of suitable candidates. In most patients PE is not considered at all.
After more than half a century, tribute still must be paid to the seminal work of Alexander Brunschwig. It was an outstanding achievement to perform such extensive operations in the 1940s, with inaccurate preoperative imaging diagnostics, limited possibilities to control bleeding, a much lower level of intensive care, and inadequate options for urinary tract reconstruction. The first procedures were carried out with palliative intent with an almost 25% intraoperative mortality rate (5 of 22 patients), which should be appreciated as a unique accomplishment at that time.
In fewer than 20 years (1947–1965), Brunschwig compiled a large cohort of 925 patients with PE. Taking into account the lack of preoperative imaging at the time, it is not surprising that many of the patients had lymph node (LN) involvement or that no tumor was found in the final specimen from 24 patients. He routinely applied gauze packs in the pelvis and kept them in place for 72 hours. The intestinal fistula formation rate reached 15%, and mortality in the whole series reached 8%. He performed even extremely extensive procedures, which to this day are rarely used and in a few selected institutions, such as external or common iliac artery resection in 11 patients and pelvic bone resection up to the extent of hemipelvectomy in five patients with cervical cancer. The overall survival reached 20%, even in such a heterogeneous and risky cohort.
A major weakness of Brunschwig’s operation was frequent recurrent pyelonephritis and renal failure due to the drainage of ureters to the active part of the colon. The use of the excluded part of the bowel for urinary diversion, which was described by Bricker, introduced a significant improvement. Both ureters were anastomosed into the loop of the distal ileum, which was closed on the cranial end and brought as a urostomy on the opposite end. In patients with previous pelvic radiation, including the ileocecal bowel, the author later described the use of transverse colon.
During its more than 60-year history, PE has undergone many improvements, which, above all, dramatically decreased mortality. New techniques of bleeding control, especially advanced bipolar coagulation, the use of antibiotics, a significant improvement in intraoperative and postoperative intensive care, and new developments in the reconstructive phase of the procedure, contributed to the improvement of outcome. Techniques in pelvic floor and vaginal reconstruction notably decreased the rate of complications caused by empty pelvis syndrome. Other significant enhancements resulted from increased accuracy of preoperative imaging, which facilitates selection of good candidates, and the ability to tailor the extent of the procedure. It should be noted that the basic principles of PE, the surgical technique, and the classification are still described in the same way they were more than 50 years ago by Alexander Brunschwig.
The most frequent indication for PE is recurrent cervical cancer after primary radiotherapy. Radical surgical treatment is the only modality in this indication with a potentially curative outcome. Vulvar, vaginal, and endometrial cancer represent about one-third of the cases in large series of PE performed for mixed types of malignant gynecologic tumors.
Whereas in rectal cancer PE is performed more often in primary treatment, in gynecologic tumors radiotherapy or chemoradiotherapy is mostly preferred. Nevertheless, there are cohorts of patients in whom PE is the treatment of choice in primary treatment, such as patients with pelvic radiotherapy in their history (i.e., lymphoma treatment), patients with locally advanced vulvar or vaginal cancer, or those with a locally advanced tumor and the presence of fistula. The oncologic outcome in cohorts after PE conducted in primary or secondary treatment does not differ.
Because of high morbidity, most PEs are performed as therapeutic procedures with the goal of achieving disease cure. In some reports, patients in whom complete resection (R0) was not performed are enrolled into the category of palliative procedures, which is not correct because the decisive factor is the primary intention. The actual use of PE in palliative care with the aim of increasing quality of life is rather rare. In a small group of 18 patients with cervical, vulvar, or endometrial cancer, the mean overall survival after palliative PE was less than 1 year. Even though the majority of patients were satisfied with the procedure, half of them had major surgical complications. In a different group of mixed pelvic tumor types, there were 14 women referred for palliative PE; four of the procedures were aborted, and the outcome was evaluated as having zero benefit in two patients and excellent or good in eight patients. Although data are limited, it is obvious that the morbidity of the procedure is high even in palliative treatment, so the palliative indications require careful candidate selection and extensive counseling. The main reason for palliative PE is usually uncontrollable symptoms such as intractable pain or severe bleeding.
One of the main messages of this chapter should emphasize that the extent of PE must always be meticulously individualized, which is crucial on one hand for achieving the best oncologic outcome and on the other hand for minimizing patient morbidity. The majority of complications, particularly severe ones, are caused by a large volume of resected tissue and formation of dead space that cannot be replaced by tissue or material. Close attention should be paid to the preservation of tissue that does not necessarily have to be resected.
The extent of the procedure can be tailored on three planes. It is not only a choice of one of the three basic PE types in a sagittal pelvic plane among anterior, posterior, and total. Equally relevant is the craniocaudal extent, with the possibility of preserving the vulva, vagina, and urogenital diaphragm. Moreover, the majority of lateral tumors are not located centrally, which enables significant adjustments to the radicality on each side of the pelvis in the frontal plane.
The basic classification distinguishes in the sagittal plane an anterior (including cystectomy and urethrectomy), posterior (including rectosigmoid resection), and total PE ( Fig. 15.1 ). The consequence of an anterior or posterior exenteration is one permanent stoma, and two stomas in the case of total PE. This tentative classification can be modified in many cases. The resection of the sigmoid colon and/or rectum followed by colorectal anastomosis can be sufficient for the achievement of R0 resection. In anterior or total PE, only cystectomy can be performed and the intact urethra used for continent urinary diversion.
In planning for a type of PE, not only the localization of the recurrent tumor should be taken into account but also the extent of the disease before primary treatment (i.e., infiltration to other pelvic structures) and the radiation dose received by the adjacent organs. Previous radiation worsens the quality of the rectal or urinary bladder wall. The dissection of fibrotic rectovaginal or vesicovaginal spaces is associated with a higher risk of fistula formation in the postoperative period. This is a potentially severe complication in patients who have undergone such a demanding procedure. In many cases the final decision on whether the rectum or urinary bladder can be preserved can be made based only on the intraoperative findings. This implies that the patient must be informed about more and less radical alternatives before the operation.
Each of these types of PEs can be performed in different craniocaudal extents in the coronal plane as a supralevator type I (preserving endopelvic fascia and pelvic diaphragm) or infralevator (including resection of levator ani muscle) procedure. Infralevator PE is further classified as type II (including vaginectomy) or type III (including vulvectomy) ( Fig. 15.2 ). Type II PE can be performed with a resection of the levator ani muscle but with preservation of the urogenital diaphragm and distal vagina (type IIA), or with total vaginectomy and resection of both pelvic and urogenital diaphragms (type IIB). Type I and IIA posterior exenteration usually enables the surgeon to make a colorectal anastomosis.
All of the aforementioned types of exenterations can be performed as laterally extended procedures when the tumor is fixed to the pelvic side wall. The planned procedure should always be described in great detail to thoroughly indicate its extent (i.e., laterally extended [left side] anterior infralevator type IIA PE), with a further description of the intended reconstructions ( Fig. 15.3 ).
It is necessary to emphasize the three different goals of the preoperative workup: diagnosis of the recurrence, exclusion of distant metastases, and assessment of the local extent of the disease.
Early diagnosis of local pelvic recurrence before it is too advanced or spread is key for the prognosis after PE. Current guidelines for surveillance after primary treatment are generally the same for all patients, irrespective of the risk of recurrence. Imaging studies are usually indicated only in the presence of symptoms or a suspicious finding at physical examination. Positron emission tomography–computed tomography (PET-CT) in routine surveillance is generally considered to be too expensive and not cost-effective. Cost-effectiveness, however, has been evaluated mostly in large groups of patients with cervical cancer regardless of their primary disease stage. PET-CT in surveillance can be of great importance in patients with locally advanced disease who have a higher risk of local recurrence with the option of curative salvage treatment. A total of 105 patients with locally advanced cervical cancer who underwent primary chemoradiation were observed in an Australian prospective cohort study. Single PET-CT was performed after the primary treatment was terminated. The 5-year progression-free survival rate reached 86% and 0% in patients with and without complete metabolic response, respectively. Even more important, of 16 patients who underwent salvage therapy, the disease was detected in 12 of them with a surveillance PET scan.
The key step before any PE for curative intent is performed is the exclusion of metastatic lesions in distant locations. The gold standard is currently considered to be PET-CT, which has a higher sensitivity than CT. Experiences with whole-body diffusion-weighted magnetic resonance imaging (MRI) and PET-MRI show comparable or even higher accuracy and better diagnostic confidence with these modalities than with PET-CT; the results are, however, based on small cohorts. Given the high demands of PE for the patient, and also the costs of the procedure and perioperative management, the most accurate available method should always be used to exclude a distant tumor spread before the operation, such as PET-CT or PET-MRI.
A distinct issue is the assessment of the local tumor extent in the pelvis, such as the infiltration of adjacent organs and key structures in the pelvis. Precise knowledge of local tumor spread is crucial for determining the extent of the procedure and providing adequate patient counseling before the operation and also for guiding the surgeon during the procedure. Pictures from preoperative imaging studies (ultrasound, MRI, CT, PET-CT) from two patients with locally advanced cervical cancer are shown in Fig. 15.4 .
MRI is considered the gold standard, due to its excellent soft tissue contrast. In a cohort of 50 patients from Memorial Sloan Kettering Cancer Center (MSKCC), preoperative MRI was highly accurate not only in the assessment of bladder and rectum invasion (sensitivity, 81%–87%; specificity, 93%–97%), but also in the identification of pelvic side wall invasion (sensitivity, 88%; specificity, 97%). It is important to note, though, that its sensitivity is not 100%, and the extent of surgery could be altered according to any intraoperative findings. Imaging reliability may be lower after radiotherapy or chemotherapy, in particular when the interval from the previous treatment is short. Unfortunately, disease extent tends to be larger than predicted with preoperative imaging.
The use of quantitative metrics of pretreatment imaging offers a new, potentially useful tool. Metabolic tumor volume and total lesion glycolysis of FDG (18F-fluorodeoxyglucose) uptake derived from PET scans and the mean apparent diffusion coefficient derived from diffusion-weighted MRI were shown to be associated with the prognosis. Functional metrics could be used in the future to triage patients with higher risk of a recurrence so that they are surveilled more intensely after primary treatment. It might also contribute to the decisions about PE execution as a prognostic marker.
The 5-year cumulative survival rate after PE varies greatly in the literature, from as low as 20% to up to 70% ( Table 15.1 ). Such a span cannot be caused by a volatile quality of surgical performance. Even though the ability to reach free margins can be higher in institutions in which surgeons have expertise with combined procedures, these cases always represent a limited part of a cohort. The decisive factor for the oncologic outcome in the whole cohort are the selection criteria. They correspond to the main prognostic parameters, such as positive pelvic and paraaortic LNs, lateral pelvic attachment, histologic types, the disease-free interval from the primary treatment, or the presence of symptoms at the time of diagnosis. The overall outcome can, however, be influenced by other criteria, such as age, obesity, or performance status. Even though they are not typically listed as prognostic factors, they can significantly increase morbidity and also diminish the ability to reach R0.
Year | No. of Patients | Morbidity (%) | Mortality (%) | 5-Year Survival (%) | |
---|---|---|---|---|---|
Series Published Before Year 2000 (N ≥ 50) | |||||
Kiselow et al. | 1967 | 207 | 42 | 8 | 35 |
Symmonds et al. | 1975 | 198 | 78 | 8 | 32 |
Rutledge et al. | 1977 | 296 | NA | 14 | 42 |
Averette et al. | 1984 | 92 | NA | 10 (29) a | 22 (36) a |
Jakowatz et al. | 1985 | 104 | 49 | 3 | 27 |
Kraybill et al. | 1988 | 99 | 71 | 14 | 45 |
Anthopoulos et al. | 1989 | 20 | 74 | 5 | 58 |
Hatch et al. | 1990 | 31 | 32 | 0 | 68 |
Lawhead et al. | 1989 | 65 | NA | 9 | 23 |
Shingleton et al. | 1989 | 143 | NA | 6 | 50 |
Soper et al. | 1989 | 69 | 84 | 7 | 48 |
Morley et al. | 1989 | 100 | 49 | 2 | 61 |
Ketcham et al. | 1970 | 162 | NA | 17 | 38 |
Hafner et al. | 1992 | 75 | 75 | 5 | 43 |
Lopez et al. | 1993 | 248 | 60 | 12 | 40 |
Series Published After Year 2000 (N ≥ 50) | |||||
Berek et al. | 2005 | 75 | 86 a | 0 | 62 |
Yoo et al. | 2012 | 61 | 44 | 0 | 56 |
Nguyen et al. | 2005 | 76 | 28 | 0 | 53 b |
Goldberg et al. | 2006 | 103 | 71 a | 1 | 47 |
Marnitz et al. | 2006 | 55 | 57 | 6 | 37 |
Maggioni et al. | 2009 | 106 | 66 | 0 | 52 |
Fotopoulou et al. | 2010 | 47 | 70 | 9 | — |
De Wever | 2011 | 106 | 60 | 2 | 40 |
Baiocchi et al. | 2012 | 107 | 53 | 0 | 27 |
Schmidt et al. | 2012 | 282 | 51 | 5 | 41 |
Chiantera et al. | 2014 | 167 | 35 | 0 | 38 |
Fleisch et al. | 2007 | 203 | 41 | 1 | 21 |
Urh et al. | 2013 | 133 | NA | NA | 57 |
Jurado et al. | 2010 | 48 | 71 | 0 | 21 c |
The decision to perform PE is always a complex and complicated process, one that must be made individually after careful consideration of all potential prognostic parameters. Patient preferences must always play a significant role in the entire process. More stringent selection criteria improve the institution’s results; at the same time, they inevitably exclude patients who would potentially benefit from PE. If the institution does not offer PE to patients with pelvic LN involvement, with pelvic side wall fixation, with the presence of symptoms corresponding to pelvic large nerve compression, or with a short interval from primary treatment, the proportion of cured patients will be higher. At the same time, a chance is then taken away from the 20% to 30% of patients who can be cured by an adequately performed salvage procedure.
In planning and carrying out a procedure, it is important to bear in mind that the main prognostic factor for PE is achieving free surgical margins. The long-term survival of patients with remaining tumor residuum after PE reaches less than 10% irrespective of further adjuvant treatment and other prognostic factors. If there is any doubt about tumor involvement during the operation, a biopsy specimen should always be taken from the suspicious location and sent for intraoperative histologic review. To sum up, the goal of every procedure with a curative intent must be complete tumor removal.
The importance of histologic type differs according to tumor type. In cervical cancer, the data in the literature are inconsistent, showing the same or worse survival rates in adenocarcinoma in comparison with squamous cancer. In the largest published cohort (282 patients) with cervical cancer, the 5-year survival rate was 31% in patients with adenocarcinoma, but 41% in those with squamous cancers. It is not surprising that in uterine cancer a much better prognosis is associated with endometrioid histotype, whereas in serous cancer, mixed cancer, or carcinosarcoma the 5-year survival oscillates below 20%.
LN involvement is not a rare finding in patients referred for PE. The risk oscillates between 10% and 30%, depending on tumor types, previous treatment, and the method used in the preoperative workup. In considering the benefits of the procedure and counseling patients, the right assessment of prognostic significance is of great importance. In mixed cohorts of gynecologic tumors, a positive LN is mostly a considerably negative prognostic factor. These data are, however, biased to a certain extent, because a systematic lymphadenectomy is not typically a routine part of PE, so LNs are removed selectively only if they are enlarged and patients with a recurrent tumor have often undergone a lymphadenectomy already. In addition, the localization of positive LNs is not distinguished in many papers. The authors from two institutions in Germany and Switzerland published their results on PE in a large group of women with primary (25%) or recurrent (75%) cervical cancer (N = 282). In 21% of cases, patients had positive LNs only in the pelvis and 28% in the paraaortic region. The involvement of pelvic LNs did not decrease the 5-year survival compared with the group without positive LNs (45% vs. 50%), whereas the positivity of paraaortic LNs was associated with a dismal 5-year survival (17%). Available data show that pelvic LN involvement can worsen the prognosis, but it should not be taken as an absolute contraindication; positive paraaortic LNs must be considered to be distant metastases associated with a very low probability of long-term survival. Limited data are available regarding the prognostic significance of involved inguinal LNs in patients undergoing PE, even though it is not a rare finding.
In the majority of tumors, the interval from primary treatment to recurrence reflects the tumor’s biological behavior. Although some smaller studies did not find prognostic significance of a progression-free interval from the primary treatment in patients after PE for a recurrent tumor, the majority of authors reported a substantially better prognosis with a long interval from the initial treatment. Marnitz reported 5-year survival rates in patients with a recurrence up to 2 years, between 2 and 5 years, and more than 5 years after the initial treatment of 17%, 28%, and 83%, respectively. It should be emphasized that patients with a very early tumor progression, up to 1 year after the primary treatment, represent an extremely unfavorable prognostic group for PE procedures.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here