Correspondence: Atsushi Takenaka M.D., Ph.D., Division of Urology, Department of Surgery, Faculty of Medicine, Tottori University, 86 Nishi-cho, Yonago 683-8503, Japan. Email: email@example.com
The procedure of radical prostatectomy has changed rapidly since the introduction of laparoscopic and robotic surgery, which allows for clear, highly magnified visualization. The da Vinci surgical system enables three-dimensional visualization and makes use of instruments with seven degrees of freedom. Since its introduction, the use of robot-assisted radical prostatectomy has spread rapidly worldwide. However, adaptation to changes in surgical techniques using knowledge of classical pelvic anatomy has been difficult. In this report, we reviewed the progress in knowledge of pelvic anatomy, particularly regarding the cavernous nerves, prostatic fascia, Denonvilliers' fascia, endopelvic fascia, bladder neck and apex of the prostate, with regard to radical prostatectomy.
Laparoscopic procedure was first introduced to radical prostatectomy in 1997. Although LRP had a steep learning curve, it was less invasive and enabled more precise observation of architecture. The da Vinci surgical system was developed in 1999. Afterward, RALP was initially reported in 2001. Since the operative procedure was established by Menon et al. the next year, RALP has been widely expanded. LRP and RALP provide a precise, highly magnified view and less venous oozing associated with pneumoperitoneum, which has led to a better anatomical understanding of fine membranous structures. It has become difficult to understand the progression and changes in function-preserving radical prostatectomy procedures with only knowledge of classical pelvic anatomy. Great progress has been made in the knowledge of urological surgical pelvic anatomy during the past several years.
An improved operative view during LRP and RALP increases the possibility of cancer control and functional preservation. To maintain function, a precise understanding of autonomic nerves is indispensable. In the early 1980s, Walsh et al. and Lepor et al. proved histologically that both nerves and blood vessels are densely present in NVB. Since then, the concept that “all routes of NVB correspond to those of cavernous nerves” has prevailed. Recent research, however, has shown that nerve fibers involved in nerve-sparing radical prostatectomy were wider than previously thought.
In 2004, it was verified that the main route of the cavernous nerve branches from around the root of the pelvic splanchnic nerves and joins in a spray-shaped distribution to the central area of the NVB, grazing along the distal side of the pelvic plexus. Furthermore, it was found that cavernous nerves exist in a plate-like distribution in the anterolateral or posterior region of the prostate in some cases. Costello et al. also reported that in the NVB, both cavernous nerves and nerve fibers that stretch toward the levator ani muscle, prostate gland and rectum are present. Since then, many reports have shown clinical data indicating preservation of postoperative sexual function and early return of postoperative continence using operative procedures that preserve the nerves anterolateral to the prostate.[9-13]
In terms of the scientific basis for these operative procedures, the distribution and function of nerve fibers around the prostate have been discussed. Eichelberg et al., Lee et al., Ganzer et al. and Hisasue et al. examined the nerve distribution surrounding the prostate from a histological viewpoint using non-nerve-spared radical prostatectomy specimens. Hisasue et al. examined 23 non-nerve-spared prostatectomy specimens and found that of all nerve fibers, the number of anterolateral parasympathetic nerve fibers accounted for 27% in the apex, 13% in the middle and 21% in the base of the prostate. Ganzer et al. pointed out that in some cases, the number of anterolateral parasympathetic nerve fibers accounted for 39.9% in the anterolateral and 45.5% in the posterior prostate, indicating that parasympathetic nerve fibers in some patients were not concentrated in the posterolateral region. Savera et al. counted the number of nerve fibers anterolateral to the prostate, which had been dissected using either a standard nerve-ST or VT, and reported that the mean number of fibers with the ST and VT was 10 and two, respectively. Data of the innervation in the anterolateral, posterolateral and posterior regions are shown in Table 1. It is certain that nerve fibers are most densely present in the posterolateral region. It has also become clear that approximately 20% of them are present in the anterolateral region.
Table 1. Data of the innervation in the anterolateral, posterolateral and posterior regions
In a study using donated cadavers, Alsaid et al. succeeded in rebuilding a 3-D model of nerve fibers using six fetus specimens. They discovered that nerve fibers in the anterolateral region of the prostate were present as far as the 2- to 10-o'clock positions, as previously reported.[13, 14, 16, 18, 20, 21] They also analyzed sympathetic, parasympathetic and sensory nerves using immunohistochemical staining. All three types of nerves were present as far as the anterior portion, although the numbers of nerves were not counted.
In addition to these histological examinations, electrophysiological studies have been pursued. The tissues around a NVB were electrically stimulated during radical prostatectomy, and the pressure of both the corpus cavernosum and urethral sphincter was simultaneously monitored to examine the distribution of nerves that caused erectile function and urinary continence. Of all nerves around the prostate, considerable differences were found among the routes of the cavernous and continence nerves. Kaiho et al. electrically stimulated the 5-, 4-, 3-, 2- and 1-o'clock positions of the lateral side of the prostate and monitored the intracavernosal pressure. The intracavernosal pressure was increased by the stimulations at every location, and as the stimulations approached the 12-o'clock position, the pressure increase was reduced. However, a report stating that the cavernous nerve is limited to the posterolateral aspect of the prostate[4, 8] has not been totally refuted, and we have not reached a conclusion in terms of the route of the parasympathetic nerves that innervate the corpora cavernosa.[14, 24, 25]
Costello et al. recently examined immunohistochemical staining of the nerves around the prostate using four donated cadavers. They stained both sympathetic and parasympathetic nerves around the prostate. Although 27.8% of nerve fibers were present in the anterolateral area, as many as 52.3% to 55.7% were sympathetic, and nerve-specific nNOS-positive nerves in the anterolateral area accounted for just 9–15%. In a recent report, Ganzer et al. affirmed these findings using 38 prostates from patients who underwent non-nerve-sparing radical prostatectomies. They deduced that although these sympathetic nerves in the anterolateral area of the prostate might be branches to the urethral sphincter muscle, a small number of nerves in the anterolateral area innervate the prostatic stroma, and most parasympathetic nerves that control erection are present in the classical NVB. They concluded that an anterolateral nerve-preservation technique, such as the VT, could avoid the traction injury and thermal injury associated with the classical NVB, thus improving the potency rate.
The aforementioned reports vary among one another, which can be explained by individual differences in periprostatic tissues and cavernous nerves. We examined the distribution of cavernous nerves through electrical stimulation during the operation. Approximately 30% of cavernous nerves had bundle formation and 70% had plate formation. Clear differences were present among the cases (Fig. 1). Kiyoshima histologically investigated the cavernous nerves using radical prostatectomy specimens. They reported that just 48% of specimens had bundle formation in the posterolateral region. The difference among the reports might have been caused by the small number of donated bodies (Alsaid, n = 6; Costello, n = 4). Anterolateral nerve preservation must have contributed to the successful results shown in the clinical data. Whether the preserved nerves around the prostate actually maintained their function, however, has not been determined.
Regarding the relationship between nerve preservation and postoperative continence, Kundu et al. studied 3477 cases of radical prostatectomy and reported that there was no relationship between nerve preservation and postoperative continence. Gralnek et al. reported that neither urinary function nor urinary bother had an association with nerve preservation using the University of California Los Angeles Prostate Cancer Index. In contrast, Kaiho et al. reported that compared with unilateral nerve-sparing or non-nerve-sparing, bilateral nerve-sparing resulted in much better urinary function 3 months after the operation. They carried out an identical examination for each type of operation based on the operator judgments and reported that there was no significant difference. The results of LRP showed that the nerve-sparing procedure was associated with a significantly superior continence rate compared with that of the non-nerve-sparing procedure 3 months after the operation. With respect to RALP, bilateral nerve-sparing significantly improved continence after 24 months without affecting the margin-positive rate.[34, 35] Abel et al. showed that nerve-sparing prostatectomy was associated with improvement in urinary function and bother. The continence mechanism is a multifactorial phenomenon, and there are methodological differences in terms of the definition of continence, questionnaires used for evaluation and time for estimation of continence return. Thus, whether nerve-sparing prostatectomy will contribute to postoperative continence remains controversial. One theory suggests that nNOS-positive nerves partially innervate the sphincter. However, from an anatomical point of view, continence nerves are present around the prostate, and delicate athermal and atraumatic operative procedures can indirectly achieve maximal preservation of the sphincter muscle; thus, we should be clinically ready to preserve nerves under the condition that oncological safety is guaranteed.
Fascias around the prostate
Both the connecting fibers swathing the surface of muscular tissues, and the tissue that covers organs, glands, and vessels are anatomically referred to as “fascia.” The former comprises histologically clear collagenous fibers, so there is no objection to the term “fascia” in a narrow sense. In contrast, the latter is sometimes difficult to identify anatomically, although it is often regarded as an important landmark in operations. In surgical dissection, the concept of the fascia has seemed to prevail without an adequate basis. Prostatic fascia, Denonvilliers' fascia and endopelvic fascia are present around the prostate. The number of fascial layers has been discussed. Higher magnification and clearer vision during LRP or RALP have shown that multiple layers exist. As a result, a new concept about these fascias is now required in surgical anatomy.
The anatomy of the fascia around the prostate is strongly associated with the various aforementioned methods of cavernous nerve sparing. This fascia is, however, referred to by different names in different references: lateral pelvic fascia,[8, 37, 38] periprostatic fascia,[16, 39-41] parapelvic fascia[42, 43] and prostatic fascia.[9, 44] In the present review, it is called “prostatic fascia” and lies between the levator ani fascia and the prostatic “capsule”, although the prostate does not have a true capsule, but only an outer fibromuscular band so called “capsule” (Fig. 2b).
Although it was once believed to be divided into two layers (inner and outer), new discoveries have shown that from a histological point of view, the prostatic fascia is a multilayered connective tissue that comprises collagenous fibers, fat tissues (Fig. 2a,b), nerves and blood vessels.[11, 18, 20, 46, 47] Its thickness varies. Kiyoshima et al. investigated 79 non-nerve-sparing radical prostatectomy specimens and reported that the levator fascia was adhered to the prostatic capsule, and that no areolar tissues were found in 48% of the specimens. Recently, Hirata et al. histologically examined the structure of the parietal pelvic fascia (levator fascia) using 11 fetus specimens (5 males) and 27 adult specimens (10 males). They reported that the levator fascia in a male's pelvis is a multilayered structure that comprises elastic fibers and smooth muscles that course in random directions.
Intraoperative visual recognition of multilayered fascial structures depends on the magnification power. One of the merits of LRP and RALP in contrast to open surgery is the ability to confirm the structure of the prostatic fascia. This enables the surgeon to select the best layer among interfascial, intrafascial or extrafascial when a nerve-sparing procedure is carried out (Fig. 3a). Based on a study showing that nerve fibers exist posterior to the classical NVB, nerves can be partially spared if thermal and traction injuries are minimized, even in extrafascial dissection. Thus, we used the nerve-sparing grade called “partial nerve-sparing” and three other nerve-sparing grades (Fig. 3b). Five or more nerve-sparing grades have also been reported in RALP. Historically, only nerve-sparing and non-nerve-sparing procedures were available. Now, however, it is possible to carry out more fine-tuned procedures using multigrade nerve-sparing systems. Furthermore, it is possible to convert the layers during dissection in RALP.
Still, we must pay attention to the following three facts: first, it is difficult to recognize peripheral autonomic nerves during surgery; second, nerve courses differ from one person to another, and it is difficult to distinguish these differences during surgery; finally, the more aggressively we try to spare nerves, the smaller the prostatic capsule safety margins.
In 1836, CP Denonvilliers discovered a firm membranous structure between the rectum and the prostate or bladder. It was subsequently named Denonvilliers' fascia and is sometimes called the posterior prostatic fascia/seminal vesicle fascia. The understanding of the anatomy of DF is still developing, and a full consensus has not yet been obtained. For instance, although the most widely accepted theory is that DF is derived from the cul-de-sac of the peritoneum posterior to the embryonal bladder, another theory is that the cranial origin of DF exists anterior to the caudal limit of the cul-de-sac. DF histologically comprises collagen, elastic and muscle fibers. Some of these fibers are reportedly thin and permeable, and others are thick enough to be considered a single layer. DF has also been described as a sac-like structure from the pouch of Douglas to the pelvic floor.
In LRP or RALP, the surgical step from the bladder neck transection to the seminal vesicle exposure should be very carefully carried out, as mentioned later in the bladder neck section. When this step is carried out, the fibromuscular layer that covers the seminal vesicles should be identified and cut to reach the seminal vesicle. The fibromuscular layer reportedly corresponds to the posterior longitudinal fascia of the detrusor muscle, which is externally upholstered by the bladder adventitia. A firm membranous structure that originates from the pouch of Douglas is found between the rectum and the prostate or bladder, and adheres tightly to the prostate near the base of the seminal vesicle. Posterior to this, somewhat thick loose connective tissue is present (Fig. 4). This connective tissue peripherally converges on the rhabdosphincter or rectourethralis muscle. It radiates and disappears at the posterolateral aspect of the prostate. Which area should be called DF is a terminological issue. Regardless, no sac-like structure comprising both anterior and posterior layers is present. DF reportedly conglutinates to the prostatic capsule with considerable frequency.[45, 58] Thus, to avoid entering the prostatic capsule, the firm membranous tissue should first be cut at the base of the seminal vesicle, and then the loose connective tissue should be separated. It has become clear that cavernous nerves lie not only at the 5- to 7-o'clock positions, but also in the posterior area in the apex. The dissection plane should be as close to the prostate as possible when nerve-sparing is intended. Thus, it seems important to gradually transfer the dissection layer anteriorly as the surgeon advances peripherally; that is, the surgeon should not advance in the preconceived manner between the anterior layer and posterior layer, but should devise and create the dissection plane, transferring the layers.
Although incision of the EPF used to be an indispensable step in open radical prostatectomy, it is not always necessary in an antegrade approach for LRP or RALP. Pelvic organs are covered with fascia, and the pelvic fascia is divided into visceral and parietal fascia.[37, 38, 51, 60] The parietal component is the levator ani fascia and is sometimes called the EPF.[61, 62] In some reports, the parietal and visceral components are collectively called the EPF.[40, 63] The parietal EPF and visceral EPF are fused on the lateral side of the prostate and bladder. In this fusion, the levator fascia approaches the prostate and then turns back laterally. This fusion site is sometimes recognized as a whitish line and is called the “arcus tendineus fascia pelvis”. The ATFP runs from the puboprostatic ligament to the ischial spine. The levator fascia was found to be able to attach to either the visceral or parietal side depending on the method of separation. When the EPF is incised inside the ATFP and advances toward the side of the prostate, the prostate is covered only with prostatic fascia and the levator fascia is spared.[18, 37] Furthermore, if the levator fascia is simultaneously incised, the levator fascia attaches to the prostate. The inner thickened area of the levator fascia is the ATFP, and the area that adheres to the pubic bone is the puboprostatic ligament.
The ATFP- and puboprostatic ligament-sparing procedure carried out by avoiding incision of the EPF is expected to result in improved postoperative early continence recovery and erectile function.[64, 65] However, incision of the EPF is necessary to confirm prerectal fat posterolateral to the prostate in extrafascial dissection or non-nerve-sparing procedures.
An antegrade approach in LRP or RALP requires early transection of the bladder neck before apical dissection, so suitable traction cannot be applied during bladder neck transection. In addition, few useful anatomical landmarks are present when the bladder neck is transected. For these reasons, bladder neck transection, especially posterior bladder neck transection, is one of the most difficult surgical steps in RALP. If the transection too closely approaches the trigone, unnecessary removal of the detrusor or detrusor nerves results. On the contrary, caudal deviation too closely approaches the glandular prostate.
Anterior bladder neck
To approach the bladder neck, it is necessary to cut a part of the anterior wall of the bladder detrusor called the DA (Fig. 5), which extends over the bladder neck and hangs over the anterior prostate.[46, 66] Because the DA contains a wide range of structures, “incision of the DA” describes many types of procedures. McNeal stated that the AFS “consists of a thick sheet of tissue covering the entire anterior surface of the prostate and hides the urethra and glandular portions of the prostate from view anteriorly”. Thus, the AFS and DA share some aspects of this concept. How, then, should the DA be handled during bladder neck transection? It is reported that the AFS might play both sphincteric and detrusor roles. McNeal's anatomy as aforementioned states that the AFS does not contain glandular tissues, so it theoretically cannot be an origin of adenocarcinoma. However, glandular tissues are reportedly found in the AFS in some cases. Furthermore, it is necessary to extensively resect the DA and reduce the possibility of a positive surgical margin as much as possible when the cancer is present at the base of the prostate. In contrast, over-resection of the DA can cause anastomotic insufficiency and should be avoided by all means. It is thus necessary to determine the resection volume in each case depending on the preoperative risk as determined by preoperative imaging diagnosis. It is also essential to operate with a direct view after removing the adipose tissue in front of the bladder and prostate, because the size and form of the prostate and the thickness of the DA differ from case to case. Middle-bunching sutures are not recommended for three reasons: first, they make it difficult to visually recognize the bladder neck by artificially straining the bladder neck; second, adjustment of the resection volume of the tissues becomes difficult; finally, consecutive nerve sparing becomes more difficult to carry out after the prostatic fascia is forcibly lifted anteriorly and made thinner.
Posterior bladder neck
The bladder wall is historically described to contain three smooth muscle layers; the middle circular layer is sandwiched between the inner and outer longitudinal layers. In the bladder neck, however, the inner longitudinal layer is atrophic or defective in some cases. Continuation of the ureteral smooth muscle to the verumontanum is termed the verumontanum muscle. However, it is difficult to identify each muscle layer during the operation. The morphological relationship between these muscle layers and the glandular prostate differs among cases. In some cases, the glandular prostate and these muscles fuse to each other, and no clear boundary exists.
An important anatomical landmark during posterior bladder neck dissection is a structure termed the “longitudinal fascia” by Secin et al. or “retrotrigonal layer” by Tewari et al. Although the terminologies have not yet been coordinated, either name refers to downward-running muscle fiber bundles that are encountered before the vas deferens and seminal vesicles are exposed during posterior bladder neck dissection (Fig. 6). Secin et al. proved that these bundles comprise two different structures. The inner layer comprises muscle fiber bundles that originate from the outer longitudinal detrusor muscle. These muscle fiber bundles run longitudinally and fuse to the base of the prostate, and it is reported that this layer might be the counterpart of the anterior DA. Anatomically, it is termed the musculus vesicoprostaticus.[51, 69] Much of the outer layer near the seminal vesicle comprises fibroadipose tissue that is continuous with the bladder adventitia.
With respect to posterior bladder neck transection procedures, a lateral approach in which the bladder neck is incised after the seminal vesicles are exposed has been reported; however, tissues can be unexpectedly torn during blunt dissection if the dissected site has no surgical plane or the tissues are fused. Inadvertent incision into the glandular prostate is a particular problem that cannot be ignored, even if it can be compensated for afterward. Thus, sharp dissection under a direct view is recommended. Although bladder neck preservation contributes to early recovery of continence because the circular smooth muscle is spared,[71, 72] its usefulness is still controversial. We do not use bladder neck preservation as a standard procedure because of the possibility of incomplete extirpation of the glandular prostate and positive surgical margins.
Apex of the prostate
Apical dissection is the most important step in radical prostatectomy, because it affects not only cancer control, but also postoperative continence and sexual function. Positive surgical margins occur frequently in apex lesions,[73, 74] and a positive surgical margin is an important factor in control of prostate cancer. In contrast, preservation of urethral sphincter length is one of the most important factors for postoperative continence recovery.[76, 77] Because the apical surgical margin rate is reportedly affected by institutional experience, apical dissection is a difficult step in radical prostatectomy.
The urethral sphincter complex comprises three layers: the rhabdosphincter, circular smooth muscles and longitudinal smooth muscles.[79, 80] These three layers can be clearly recognized with high magnification (Fig. 7). To balance oncological and continence outcomes, the urethra at the apex should be clearly distinguished and meticulously dissected. The anatomical morphology of the male rhabdosphincter reportedly shows interindividual variation. In aged males, the posterior rhabdosphincter has thinned or is absent in most cases. The rhabdosphincter sometimes invades the pseudocapsule of the prostate or between the glandular tissues at the apex. Thus, blunt dissection in this area can cut into the apex of the prostate. In addition, some anatomical points to be noted while dissecting the apex include lack of a capsule-like structure in the apex,[45, 83] variation in the form of the prostatic apex and variation in the structure of the sphincter complex in the apex.
The DVC should be ligated to prevent intraoperative blood loss in open prostatectomy. However, if the DVC is bunched and ligated before transection, a form of the rhabdosphincter deviates and apical dissection seems to become complicated. A method in which sharp transection of the DVC is carried out before ligation was recently reported for RALP. This method is associated with a decreased apical margin-positive rate and earlier return of postoperative continence.[85, 86]
Problems to be solved in the future
The appearance of LRP and RALP allowed for precise recognition of the structures related to radical prostatectomy. Furthermore, the anatomical understanding of nerves, fascias and muscles around the prostate has recently made great progress. Membranes that were previously believed to be single or double-layered were proven to be multilayered structures. However, even with highly magnified vision, it is still impossible to confirm autonomic nerves in terminal branches during the operation. Thus, more detailed anatomical knowledge is required. The routes of nerves involved in erection and micturition have not been fully elucidated, and the roles of nerves running around the prostate must be clarified from an anatomical viewpoint. Further discussion on appropriate case selection and operative methods is necessary. In addition, studies involving intraoperative visualization of nerves or cancer sites should also be desired. Tewari et al. reported preclinical application of multiphoton microscopy as a tool for intrasurgical identification of periprostatic structures. Further physiological or intraoperative imaging study should also be encouraged.