Principles and Practice of Surgical Treatment for Cervical Cancer: It's Time for a Change

Despite a decreasing incidence within developed countries, along with optimism regarding prevention of future disease through vaccination, carcinoma of the uterine cervix is still the second most frequent cancer among women, with 490,000 new cases and 270,000 deaths annually worldwide [1,2]. The number of deaths attributed to this malignant disease is particularly high since it hits women at a relatively young age – 20% of the patients are aged 40 years or younger, resulting in a mean loss of 26 years of life per patient succumbing to cervical cancer [3]. Approximately 40% of cervical cancer patients are diagnosed with early macroscopic disease (i.e., International Federation of Obstetricians and Gynaecologists [FIGO] stages IB–IIA) and are candidates for surgical treatment with radical hysterectomy, an operation that, conceptually, is over 100 years old [4,5]. Although various modifications have been applied since the introduction of abdominal radical hysterectomy by Wertheim in the early 20th Century, such as the vaginal approach, systematic pelvic and periaortic lymph node dissection, extension of the (paracervical) dissection by including the ‘vesicouterine ligament’, tailoring the radicality of the pericervical dissection, autonomic nerve sparing, fertility preserving, minimally invasive techniques and, recently, robotic assistance, the principles on which these operations are based have remained essentially unchanged [6,7]. These are: surgical anatomy deduced a posteriori from dissection and imaging of the mature organism; and a model of local tumor spread that assumes undirected linear intra- and trans-cervical permeation with microscopic and occult disease preceding the macroscopic tumor front. Consequently, for radical treatment the tumor is resected with a ubiquitous metrically defined margin of intra- and para-cervical tumor-free tissue [8,9]. Close margins are regarded as an indication for adjuvant radiation [10]. Moreover, it is anticipated that surgical treatment is not sufficient to control discontinuous local and metastatic pelvic disease [11,12].

However, a closer look at the current practice and its results, both with regard to standard treatment and innovative surgical techniques, questions the validity of the underlying principles; the fact that adjuvant radiation is added to radical hysterectomy in approximately 50% of the cases undoubtedly contradicts the claim that the surgical treatment is ‘radical’.

Locoregional failure rates of 15% and higher and a 5-year overall survival probability rate of 80% associated with a rate of major (i.e., grades 2 and 3) treatment-related morbidity of up to 30% that are accepted as a treatment standard appear to be unfavorable for early-stage (FIGO IB–IIA) disease [13]. Following the literature on radical hysterectomy, inconsistencies in surgical anatomy and clinical results cannot be ignored. Surgical anatomy applied with radical hysterectomy is traditionally uterocentric and focused on ligaments. However, it has been demonstrated that most of the so-called ligaments connecting the uterovaginal tract to the pelvic wall and to the adjacent organs, such as the broad ligament, cardinal ligament and vesicouterine ligament, have no significant suspensory function – some of them are mere dissection artifacts [14–16]. Recently, the uterocentric view has been refuted by the demonstration that the infraureteral parametrium, which is a major subject of the Okabayashi modification of radical hysterectomy, does not exist [17]. Several retrospective studies have demonstrated that the extension of paracervical radicality and lymph node dissection did not improve the overall outcome in early cervical cancer [18,19]. The methodological flaws of these studies might be considered as an explanation for these unexpected results; however, a randomized, prospective study of tailored paracervical radicality (type II vs type III radical hysterectomy) supplemented with adjuvant radiation confirmed the lack of superiority of the more extended resection with regard to pelvic control. Type III radical hysterectomy only served to increase morbidity [20].

I argue that the relatively poor overall outcome, the necessity of adding adjuvant radiation to an operation termed ‘radical’ in almost every second case of stage IB–IIA disease and the cited inconsistencies of many aspects of radical hysterectomy cannot be explained by suboptimal performance of the procedures, inadequate diagnostics or inappropriate trials but, instead, falsify the principles on which the current surgical treatment of cervical cancer is based. Indeed, following the anatomy of the female genital tract from the anlage in the embryo through fetal differentiation to maturation in the adult and tracing the local permeation of cervical cancer with high resolution MRI clearly shows that the surgical anatomy of radical hysterectomy and the appreciation of tumor spread do not match. I have suggested revisiting the surgical anatomy for cancer treatment in general a priori from embryonic development and have established the compartment theory of local tumor spread [21,22]. The theory states that malignant solid tumors are confined for a relatively long phase during their natural course to a permissive compartment that can be deduced from embryonic development as the mesenchymal differentiation product of the corresponding anlage. Compartment borders are tumor suppressive. For transgression into adjacent compartments of different embryonic origin, phenotypical changes that evolve relatively late during malignant progression are necessary. Local relapses may arise from remnants of the compartment remaining in situ after treatment harboring or recruiting residual tumor cells/tumor stem cells.

The compartment theory sets up the new principles of radicality for surgical tumor treatment, namely the resection of the tumor-bearing compartment with intact borders. Nonlymphatic adjacent tissues of embryologically different compartments can safely be retained despite their close proximity to the tumor front. On one hand, compartment resection should result in maximum local tumor control without adjuvant radiation, and on the other hand, compartment resection should result in minimal treatment-related morbidity. A metrically defined tumor-free resection margin has to be achieved only at the site of intracompartmental resection, which may be indicated to preserve functional aspects of the morphgenetic unit.

We have identified the uterovaginal compartment in the adult female by following the differentiation and maturation of the Müllerian anlage mesenchyme (i.e., the paramesonephric–mesonephric complex connected to the deep urogenital sinus) and demonstrated that carcinoma of the uterine cervix inflates, distorts and destroys, but does not transgress, this compartment for long phases during its natural course [21,22].

The later event of tumor transgression into primarily nonpermissive adjacent compartments, such as the ureter, bladder, rectum and peritoneum, is focal and accompanied by inflammation hampering the exposition of intact compartment borders. The extirpation of the uterovaginal compartment with intact borders (except its distal part to retain a functional vaginal vault) by total mesometrial resection (TMMR) in 212 consecutive patients with FIGO stage IB, IIA and selected IIB cervical cancers achieved recurrence-free and overall-survival probabilities of 94 and 96%, respectively, at a median follow-up of 41 months (range: 5–110 months). Adjuvant radiation was not administered, although 63% of the patients had histopathological risk factors regarded as an indication for adjuvant radiotherapy in case of conventional radical hysterectomy. Treatment-related grade 2 morbidity was detected in 9% of the patients. Resection of the uterovaginal compartment led to excellent (98%) pelvic control irrespective of the metric extension of the resection margins, thus proving the underlying principles [22].

Total mesometrial resection is distinct from conventional radical hysterectomy in several aspects. Most importantly, radical hysterectomy resects anterior and deep lateral paracervical tissue that does not belong to the morphogenetic unit of the uterovaginal compartment, and is therefore retained with TMMR. Conversely, perirectal tissue, which is part of the Müllerian compartment, is ignored by conventional radical hysterectomy, but is subjected to complete sharp resection with TMMR. Moreover, differences in lymph node dissection with TMMR result from an altered exposition of draining lymph node basins after developing the pelvic retroperitoneum with regard to the embryologically deduced visceroparietal compartments.

The pattern analysis from pelvic relapse landscapes obtained by coregistration of individual MRI scans of patients failing after conventional radical hysterectomy clearly demonstrates that recurrent disease mainly arises and grows within remnants of the Müllerian compartment. Much of the treatment-related morbidity after conventional radical hysterectomy originates from the resection of the paracervical tissue, which is not part of the uterovaginal compartment, and from adjuvant radiation. The insufficient resection of the tissue of potential locoregional tumor spread by conventional radical hysterectomy explains both the necessity of adjuvant radiation in early-stage disease with high-risk factors and the fact that the extension of paracervical resection could not improve pelvic tumor control.

How are the new developments in surgical treatment of cervical cancer to be interpreted in the light of the compartment theory? Autonomic nerve preservation has long been rejected by Wertheim surgeons. It was argued that sparing these structures would only be possible at the cost of reduced pelvic tumor control. From the collective evidence of all published results, it is now clear that autonomic nerve preservation does not result in more pelvic relapses [23]. These results are fully consistent with the compartment theory as the hypogastric nerves and the inferior hypogastric plexus are not part of the uterovaginal compartment. However, with the nerve-sparing techniques, other non-uterovaginal tissues, such as the bladder vessels, are still sacrificed for the sake of ‘radical’ tumor resection, whereas perirectal parts of the Müllerian compartment are not completely removed.

The sophisticated, minimally invasive techniques established and evaluated for the surgical treatment of cervical cancer, including the most recent robotic assistance, are all based on the conventional views of surgical anatomy and tumor permeation and apply the principle of wide tumor excision. With regard to the compartment theory and resection, all arguments given for the open procedures are also valid here. Consequently, the oncologic outcome and most of the treatment-related morbidity, except that of the laparotomy, could not be improved.