Preventing Erectile Dysfunction after Radical Prostatectomy: Nerve-Sparing Techniques, Penile Rehabilitation, and Novel Regenerative Therapies Preventing Erectile Dysfunction after Radical Prostatectomy: Nerve-Sparing Techniques, Penile Rehabilitation, and Novel Regenerative Therapies

Erectile dysfunction is a known and much-dreaded functional consequence of radical prostatectomy. Dr. Patrick Walsh pioneered the nerve-sparing radical retropubic prostatectomy in the early 1980s, which has mitigated the morbidity of this surgery. Post- operative potency rates range widely from 20 to 80%, however, and depend on myriad factors including age, preoperative potency, and degree of nerve-sparing during surgery. Over the past four decades several developments have continued to offer hope to patients and clinicians alike, including refined understanding of cavernosal nerve neuroanatomy, beneficial modifications in surgical technique, as well as the advent of robotic surgery. Furthermore, multiple pre- and post-operative penile rehabilitation techniques using mechanotherapy and pharmaceuticals have also improved functional recovery. This paper examines erectile dysfunction as a consequence of radical prostatectomy, including the physiology of erections, the pathophysiology of post-operative erectile dysfunction, novel surgical techniques to enhance neurovascular bundle preservation, and penile reha- bilitation strategies involving hyperbaric oxygen, neuroprotective pharmaceuticals, dehydrated human amnion-chorion membrane allografts, and mesenchymal stem cell therapy.


Introduction
Erectile dysfunction is a known and much-dreaded functional consequence of surgery for prostate cancer. Although surgeons may cite oncologic control as the paramount component of the "Trifecta" (i.e. cancer cure, continence, and potency), many patients cite recovery of erectile function as the true measure of treatment success. In the early days of radical prostatectomy, post-operative potency rates were poor, and in fact largely non-existent. With the advent of nerve-sparing anatomic radical retropubic prostatectomy, a surgical approach pioneered by Dr. Patrick Walsh, the prospect of post-operative recovery of potency became not only a possibility but a reality for many men. The myriad factors that influence a patient's likelihood of sexual function recovery after both open and robotic radical prostatectomy have been examined and published in the literature. Also, there has been much investigation into the pathophysiology of iatrogenic erectile dysfunction (i.e. neurapraxia and nerve-traction injury) in the form of in vitro, pre-clinical animal studies and even translational studies with randomized human subjects. Given the pivotal importance of erectile function in a patient's perceived post-operative quality of life, there is much interest in the optimization of perioperative techniques to spare the integrity of the cavernous nerves and to develop efficacious mechanical and pharmacologic penile rehabilitation programs. Such programs employ an increasingly sophisticated arsenal of medical technologies such as pluripotent stem cell therapy, cytokine-rich human amnion-chorion membrane allograft, and even reappropriation of pharmacotherapies traditionally used for other disease states that have been found to have neuroprotective properties. This chapter will examine the evolution in the understanding of erectile dysfunction as a consequence of radical prostatectomy and examine novel strategies for prevention and amelioration of this condition.

Anatomy
The central role of smooth muscle dynamics in the corpora cavernosa in the development of erections was first elucidated in the 1980s [1]. Identification of nitric oxide (NO) as the principle neurotransmitter for tumescence and phosphodiesterases for detumescence were also major milestones with well-known pharmacologic ramifications. Anatomically, beneath the Buck's fascia, the corpora cavernosa are surrounded by the two-layered tunica albuginea, a reticulated network of collagen and elastin fibers that provides structural support during tumescence. The outer layer serves to compress the obliquely oriented emissary veins during tumescence that results in the "bottle-neck" effect of slower outflow than inflow that is essential for maintenance of an erection (Figure 1) [2]. The penile arterial supply arises from internal pudendal artery, which then gives rise to the common penile artery that branches into the dorsal, cavernous, and bulbourethral arteries. The cavernous arteries are responsible for the engorgement of the corpora cavernosa during tumescence. Accessory pudendal arteries may also be present in up to 4-25% of patients and these arise from the external iliac, obturator, vesical, and femoral arteries. Their preservation has been shown to be important for recovery of erectile function after radical prostatectomy [3,4].

Neuroanatomy
The vasomotor tone of the cavernous arteries is regulated by the autonomic cavernous nerves. They are the nerves that are being described during "nerve-sparing" techniques during radical prostatectomy. These nerves arise as an extension from the parasympathetic pelvic splanchnic nerves that originate from the pelvic plexus (S2-S4) located on either side of the rectum (Figure 2). These nerves innervate the endothelium of the cavernous sinuses and release acetylcholine which inhibits adrenergic neurons and stimulates NO release from endothelial cells [5]. NO increases intracellular production of cGMP with resultant decline in intracellular calcium and relaxation of the cavernous smooth muscle. Phosphodiesterase 5 is responsible for the degradation of cGMP and is the target of the well-known medications for erectile dysfunction.
Since Walsh and Donker pioneered the nerve-sparing radical prostatectomy in 1982, there has been much debate about the nature and trajectory of these nerves [6,7]. These nerves have been found to travel as a latticework of delicate fibers principally along the posterolateral and lateral aspects of the prostate, but with some fibers coursing ventrally as well. Invaluable work in human cadavers by several investigators has elucidated precise position and orientation of these nerves to optimize their preservation during radical prostatectomy.  [8,9]. Furthermore, Lunacek et al. showed that the cavernous nerves are displaced more anteriorly and splay along the lateral aspect of the prostate like a curtain [10]. These findings inspired the "curtain dissection" technique of high anterior release as well as the technique of preserving the lateral prostatic fascia within which the neurovascular bundle travels known as the "Veil of Aphrodite" technique to maximize the number of nerve fibers preserved [10,11]. Dr. Ashutosh Tewari has conceptualized the neuroanatomy as consisting of a tri-zonal neural architecture, comprised of the proximal neurovascular plate (PNP), predominant neurovascular bundle (PNB), and accessory neural pathways (ANP) [12]. The PNP is synonymous with the pelvic plexus and the PNB is the traditionally described NVB, which is enclosed within the layers of levator fascia and/or lateral pelvic fascia. The nerves are situated in a "hammock-like" distribution rather than a distinct, isolated structure (Figure 2).
Further anatomical studies have demonstrated that a significant proportion of the nerves are situated on the anterior surface of the prostate, 21.5-28.5% [13] and 19.9-22.8% (Figure 3) [14]. Structural configurations range from round and bundle-like to more widely distributed splaylike [15]. Ganzer et al. employed computerized planimetry software to analyze the topography of the nerves on whole-mount pathologic sections obtained during non-nerve-sparing radical prostatectomy [16]. Total nerve surface area was most concentrated dorsolaterally (74.5-84.1%), but up to 39.9% of nerve surface area was found ventrolaterally. These correspond to the ANPs described by Dr. Tewari's group, who found them in 41% of the cadavers [12]. It is possible that all of these nerves are not responsible for erections. Subsequent studies in electrophysiologic stimulation have shown cavernosal pressure responses with stimulation at all positions of the midprostate between the 1:00 and 5:00 positions for all patients, suggesting their role in potency [17]. The precision vs. degree of electrical spread of such electrical stimulation may represent a limitation of this testing. Conversely, Costello et al. reported that significantly parasympathetic nerve fibers only account for 4-6.8% of nerves on the anterolateral aspect of the prostate [18]. These findings were corroborated by Ganzer et al., who used immunohistochemistry to distinguish between parasympathetic and sympathetic nerves. They reported that only 14.6% of the parasympathetic nerves resided above a horizontal line drawn at the prostate base and only 1.5% above a horizontal line at the apex [19].
How these anatomical findings impact the functional outcomes of the various nerve-sparing approaches described below, some of which advocate for high anterior release of the prostatic fascia, is interesting to consider (see Section 5.2).

Pathophysiology of post-prostatectomy erectile dysfunction
There are several acute and chronic factors that contribute to decline in erectile function after surgery for prostate cancer. These factors may be classified as vasculogenic, neurogenic, and even psychological. The burden of cancer diagnosis, treatment, and need for long-term PSA monitoring, along with recovery from surgery, implications on self-image, awareness of mortality, and perceived or actual reduction in penile length represent a constellation of psychosocial factors that may contribute to ED.
The role of accessory pudendal arteries in vasculogenic erectile function has been described above. Chen et al. has also implicated the veins that travel longitudinally within the layers of the tunica albuginea. His group have reported that ligation of the DVC during prostatectomy results in dilatation of these veins which results in veno-occlusive dysfunction [20].
The vascular sequelae of radical prostatectomy were investigated by Mulhall et al., who reported cavernosometry and/or penile duplex ultrasonography [21]. They found arterial insufficiency in 59% and venous leakage in 26% of men after bilateral nerve-sparing RRP who had excellent pre-operative erectile function. Normal vascular status was found in 35% of men. Return to penetrative function was correlated with the vascular status, with 47% return if normal status, 31% in arteriogenic insufficiency, and only 9% with veno-occlusive dysfunction at 12 months. They also reported that the longer the duration of erectile dysfunction, the higher the risk of venous leakage. Zelefsky et al. reported venous leakage in 52%, arterial insufficiency in 32%, and neurogenic dysfunction in 12% [22]. Montorsi et al. found a higher proportion of patients with veno-occlusive disease in their randomized study of intracavernous alprostadil after open RP: 67% veno-occlusive dysfunction, 17% arterial insufficiency, 17% normal vascular dynamics [23].

Neurapraxia
The neurogenic basis for erectile dysfunction implicates the cavernous nerve architecture. Preservation of these architectural substrates may not be sufficient alone to engender recovery of post-operative function, as suggested by the well-documented latency period between surgery and functional recovery. This latency ranges from 6-24 months and has been suggested to be the result of nerve-traction injury from the physical manipulation/handling during surgery and resultant neurapraxia and axonotmesis [1,24,25].
Intraoperative manipulation and injury to the cavernosal nerves results in hemodynamic and histologic changes within the penis, which manifest clinically as erectile dysfunction. This injury may result from mechanical or thermal sources. Iatrogenic traction on the delicate neurovascular tissue can cause-stretch induced nerve injury and resulting dysfunction. Neuropathies may be classified into three histologic groups: neurapraxia, axonotmesis, and neurotmesis. Neurapraxia is the least severe and is characterized intact neural structural elements, but there may be ischemia and/or demyelination which leads to signal conduction block. Functional deficits in peripheral nerves manifest as motor, proprioceptive, and soft touch deficiencies, but these usually resolve in a few weeks (up to 12 weeks) [26]. The next level of injury is axonotmesis in which axons and their myelin sheaths over long segments of nerve are disrupted, while supporting structures such as the endoneurium are left intact [27]. There is consequent Wallerian degeneration distal to the level of injury and proximal axonal degeneration to the next node of Ranvier. Since the endoneurial tubes remain intact, recovery should be complete after a matter of several months but may not be complete. Frank transection of the nerve is termed neurotmesis, in which the endoneurial tubes and connective tissue components are disrupted. Intraneural fibrosis develops and impairs axonal regeneration and thus inhibits nerve functional recovery. Peripheral nerve regeneration is mediated by multiple factors including neurotrophic factors, extracellular matrix, and intact cellular components of the nervous system (i.e. endoneurium) [28]. Tissue trauma from surgery also generates an inflammatory response and oxidative stress around the degenerating axons with results in chromatolysis (degradation of the protein synthesizing infrastructure of the neuronal cell body) [29].
Nerve injury may have a vasculogenic etiology as well. Nerve ischemia may be a result of direct compression injury or stretch ("traction") injury, which produces a reduction in cross-sectional area and resultant compression of the vasculature [27]. Traction disrupts and occludes smallsized arteries traveling with the nerves (vasa nervorum) which supply pelvic tissues as well as the nerves themselves. Biochemically within the cavernosal smooth muscle cells, hypoxia induces production of superoxide which initiates oxidative reactions and attacks surrounding molecules to produce more free radicals. Oxygen free radicals in the setting of a nitric-oxide containing environment tends to combine to form peroxynitrite (O═NOO-) which is highly neurotoxic. Exposure of nerves to this compound leads to rapid excitation, excitotoxicity, and degeneration in the acute setting. Nitric oxide bioavailability is thereby reduced in this setting, which further impacts penile smooth muscle relaxation and exacerbates the hypoxic conditions.

Thermal injury
The importance of athermal dissection has been reinforced as classic teaching during nerve-sparing radical prostatectomy. The precise functional deficits induced by monopolar or bipolar cautery has been investigated in a canine model [30]. A total of 12 dogs were divided into four groups of neurovascular bundle dissection: Group 1, suture ligatures; Group 2, monopolar electrocautery; Group 3, bipolar electrocautery; Group 4, ultrasonic shears. Peak intracavernous pressures in response to distal cavernous nerve stimulation immediately post-op and at 2 weeks showed an attenuated response compared to controls (74-91% decrease and 93-96% decrease, respectively). In the dogs where electrocautery was employed, there was in fact almost no rise in the intracavernous pressure in response to stimulation. The findings in this study were presented in the context of having spared the contralateral neurovascular bundle from any dissection. Their results may have been even more dramatic if thermal energy had been used bilaterally. The follow-up was also admittedly short in this series at 2 weeks post-op only. Much longer durations at 6-, 12-, and even 24 months would be helpful to determine the long-term impact of thermal energy on recovery of potency (see Section 5.6).
This very objective was explored in humans after robotic prostatectomy, demonstrating delayed recovery after 12-18 months, but with 68% of bilateral nerve-sparing patient ultimately recovering function at 24 months [31] (see Section 5.3).

Chronic cavernosal tissue changes
In the chronic phase of injury, the persistent loss of nerve signal conduction results in loss of spontaneous nocturnal erections and relative cavernosal ischemia. The pathophysiologic consequence is cavernosal smooth muscle apoptosis, upregulation of TGF-beta and collagen deposition within the corpora [32][33][34][35]. Cavernous neurotomy studies in rats have demonstrated that corporal smooth muscle apoptosis begins 1 day after injury and peaks within the first week [34]. This fibrotic reaction impairs full expansion of the venous sinuses within the tunica and failure to adequately compress the emissary veins against the tunica. The result is "veno-occlusive dysfunction" in which the venous outflow occurs with the same velocity as arterial inflow. In this setting, tumescence is unable to be achieved or maintained. Furthermore, there is anatomical loss of penile length and girth as a result of the cavernosal smooth muscle fibrosis.

Sympathetic disinhibition
Erectile function is not only the result of parasympathetic input, but a dynamic interplay between enhanced parasympathetic and inhibited sympathetic function. The role of adrenergic, sympathetic signals in ejaculation and detumescence have been well-established [36]. The adrenergic system essentially inhibits tumescence via the release of presynaptic norepinephrine (NE) that binds to postsynaptic α1-and α2-adrenergic receptors that induce penile arterial and cavernosal smooth muscle contraction. Also, the activation of presynaptic adrenergic receptors on the parasympathetic nerves inhibits release of NO [37]. The α1 receptor subtype is predominant in human erectile tissue, and furthermore α1A and α1D are more common than α1B in humans [38,39]. Neurogenic contractile responses have been shown to be increased in the corpus cavernosum from rats after cavernous nerve injury and in cavernosal tissues from men with post-prostatectomy ED [40].
Recent animal studies have elucidated the role of adrenergic function on the contractile dynamics of cavernosal smooth muscle [41]. After bilateral crush injury to the cavernosal nerves rats were administered phentolamine (non-selective alpha-blocker), silodosin (α1Aselective alpha-blocker), or tap water only for 4 weeks and intracavernosal pressure (ICP) was monitored after (1) electrical stimulation of the cavernosal nerve proximal to the region of injury and (2) IV administration of tadalafil (phosphodiesterase 5 inhibitor). A significantly greater increase in ICP was observed for the silodosin group compared to the phentolamine or tap water groups after both electrical stimulation alone and co-administration with IV tadalafil. These findings suggest therapeutic benefit to alpha blockade for recovery of erectile function after RP. The authors translated their experiments to humans by obtaining strips of cavernosal smooth muscle at the time of inflatable penile prosthesis insertion. The response to electrical stimulation ex-vivo was enhanced by pretreatment of the muscle strips with both silodosin and tadalafil compared to tadalafil alone (Figure 4) [41].

Epidemiology of erectile dysfunction
The prevalence of erectile dysfunction in the general population has been reported by two large surveys: the Massachusetts Male Aging Study (MMAS) and the National Health and Social Life Survey (NHSLS). According to the MMAS the rate of complete, moderate, and mild ED for the study of n = 1709 community-dwelling men in their 40-70 seconds was 5.1-15, 17-34, and 17%, respectively [42]. The NHSLS examined n = 1410 community-dwelling men and women in 1992 and reported rates of ED by age: 7%, 18-29 years; 9% 30-39 years; 11% 40-49 years; 18%, 50-59 years [43]. The definition of ED in this study was not quantified with validated questionnaires. International studies have reported rates of 20-40% for men 60-69 years [44]. ED is therefore an age-dependent disease. Other established risk factors include diabetes mellitus, hypertension, hyperlipidemia, psychiatric/psychologic disorders, history of pelvic trauma, chronic disease states (i.e. hypogonadism, thyroid disease, chronic kidney disease), and socio-demographic status.
There are multiple validated questionnaires to accurately assess various aspects of a patient's baseline sexual function. These questionnaires are integral in the patient workup as both a quantifiable measure of their function and a method of realistic prognostication of their likelihood of meaningful recovery. These elements comprise an important aspect of patient preoperative counseling and informed consent. The International Index of Erectile Function (IIEF) consists of 15 items and five domains and was developed by an international panel of experts for uses in determining treatment efficacy in clinical trials [45,46]. Given its high sensitivity for detecting clinically significant treatment effects, it is regarded as the gold standard treatment outcome measure. Administration of this long questionnaire is somewhat cumbersome in a routine clinic setting, however. The National Institutes of Health's Consensus Panel on ED lead the development of an abridged five-item version of the IIEF, called the IIEF-5 or the Sexual Health Inventory for Men (SHIM). The SHIM is a powerful grading system and easily-administered patient reported tool [47]. It has high sensitivity and specificity and has been shown to be more reliable than a single item self-assessment of severity of ED [48]. Each question is rated on a Likert scale from 1 to 5, and consists of: "Over the past 6 months":

1.
How do you rate your confidence that you could get and keep an erection?

2.
When you had erections with sexual stimulation, how often were your erections hard enough for penetration (entering your partner)?
3. During sexual intercourse, how often were you able to maintain your erection after you had penetrated (entered) your partner?

4.
During sexual intercourse, how difficult was it to maintain your erection to completion of intercourse?

5.
When you attempted sexual intercourse, how often was it satisfactory for you?
Based on response to the questions, men may be categorized into one of five grades of ED severity: no ED (SHIM 22-25), mild (17)(18)(19)(20)(21), mild to moderate (12)(13)(14)(15)(16), moderate (8)(9)(10)(11), severe (1-7). Although there have been multiple other questionnaires published-the Quality of Erection Questionnaire, the Erectile Dysfunction Inventory of Treatment Satisfaction (EDITS), the Self-Esteem and Relationship questionnaire (SEAR); the Erection Hardness Score (EHS), the Sexual Experience Questionnaire-the IIEF and SHIM are the most widely employed, and have indeed been translated into over 30 languages [49,50]. Another popular health-related quality of life (HRQoL) questionnaire specific to prostatectomy patients is the Expanded Prostate Cancer Index Composite (EPIC), which examines urinary, bowel, sexual, and hormonal domains [51]. Some of these validated instruments have been used to measure outcomes in the randomized penile rehabilitation studies (see below).
Studies of ED prevalence among men prior to undergoing radical prostatectomy report potency rates ranging from 43 to 84%; the 43% value was obtained through interrogation with the IIEF [52][53][54]. The landmark prospective Prostate Cancer Outcomes Study (PCOS) included n = 3533 men from the Surveillance, Epidemiology and End Results (SEER) cancer registries diagnosed with prostate cancer in 1994-1995. A total of n = 1288 men who underwent radical prostatectomy had completed a baseline questionnaire were included. Baseline erections firm enough for intercourse was 81%, with 49% reporting a least "little/some" difficulty in maintaining erections [52,53]. A limitation of these figures, however, is the fact the FDA approval of sildenafil only occurred in 1998 and "baseline" function was assessed through post-hoc recall within 6 months after treatment in 90% of patients, thereby introducing recall bias.

Incidence of erectile dysfunction after radical prostatectomy
Recovery rates after bilateral nerve-sparing open RRP ranges from 31 to 86% of sexually active men with organ-confined disease [55]. The recovery rates for unilateral nerve-sparing range from 13 to 56%, and for non-nerve sparing 0-17% [56][57][58][59][60]. The CaPSURE database of community-based Urology practices has reported that only 20% of men return to preoperative baseline potency at 12 months after RP [61]. The metric for assessment has an impact on the incidence of ED, and the use of validated questionnaires such as the IIEF tends to expose higher incidence of ED [55]. The long-term outcomes of the PCOS study demonstrated [53] worse erectile function in men after RP compared to radiation therapy at the 2-and 5-year evaluation time points, odds ratio 3.46 and 1.96, respectively. There was no significant difference at the 15-year follow-up time point, however. Defining sexual function as "erections in sufficient for intercourse" produced absolute rates of ED at 2, 5, and 15 years of 78.8, 75.7, and 87%, respectively ( Figure 5). Of note, only 14.5% of men underwent bilateral nerve-sparing surgery in this series.. Among men who had bilateral nerve-sparing, 5-year erectile function firm enough for intercourse was reported in 40% of men vs. 23% and 23% for unilateral nervesparing and no nerve-sparing, respectively. Age was a significant predictor on multivariable analysis. Some limitations were the late approval of sildenafil in 1998 and the fact that RP was performed in an open manner, which limits applications to modern series. During the 3 years prior to the 5 year evaluation, only 43% of men had tried sildenafil.
Outcomes stratified by specific nerve-sparing approach are presented below (see Section 5.3 below).

Prognostic factors for recovery of erectile function post-prostatectomy
Factors for successful recovery of erectile function have been examined for patients after both open radical retropubic prostatectomy (RRP) and robotic prostatectomy in the modern era.
A systematic review post-RRP identified age <60 years, completeness of nerve-sparing, and pre-operative sexual function [55]. Indeed, in the PCOS study a significantly higher proportion of men <60 years reported satisfactory response to sildenafil compared to older men (p = 0.01) [52]. The systematic review of robotic prostatectomy by Ficarra et al. cited age, baseline potency status, comorbidities index, and extent of nerve-sparing as predictors of postoperative recovery of potency [63]. There may be a learning-curve effect with the robotic prostatectomy outcomes as well. As discussed below, since nerve-sparing may be performed in an incremental manner rather than an "all-or-none" phenomenon, grade of nerve sparing has shown an influence on recovery [12,65,66] (see Techniques of Nerve Sparing below).
The impact of surgeon experience/learning has been reported [67], but did not demonstrate statistical significance for potency rates between cases 1-300, 301-500, and 501-700 (61, 63, 65%, respectively) [64]. As mentioned previously, lack of electrocautery during neurovascular bundle dissection portends earlier recovery of erectile function [31]. While other factors have been cited, the previous predictors are the most consistent throughout the medical literature.
Post-operative potency rates may be influenced by the time point of assessment. Studies typically report 12 and 24 month outcomes as the longest follow-up. Although recovery typically occurs within the first 2 years after surgery, delayed recovery is also possible. A series of n = 1003 men who underwent either open or robotic RP between 2007 and 2013 reported on the achievement of "good erectile function" as defined by IIEF-6 score ≥ 22 [68]. Among men with poor function at 12 months, the probability of recovering erectile function at 24, 36, and 48 months was 22, 32, and 40% on Kaplan-Meier analysis. The 12-month functional score and patient age were the only significant predictors of delayed recovery on multivariable analysis. Also, very interestingly, the degree of nerve-sparing was not a predictor of delayed recovery; perhaps nerve-sparing only impacts early recovery at the 12-month time point. Surgical modality (open vs. robotic) was not explored. Similar findings of delayed recovery have been published in other reports [69,70]. Such findings may be tempered by the gradual decline in erectile function after year 5 observed in the Prostate Cancer Outcomes Study [52].

Preoperative planning with multi-parametric MRI
Multiparametric magnetic resonance imaging (mpMRI) has gained widespread use in the workup of elevated PSA and diagnosis of prostate cancer via MRI-transrectal ultrasound (TRUS) targeted fusion biopsy. This technique has been shown to increase the detection rate of high-grade (i.e. Gleason 4 + 3 = 7) prostate cancer by 30% and result in lower detection of low grade prostate cancers by 17% [71]. The recently published PROMIS trial evaluated the performance of mpMRI to the reference standard template prostate mapping (TPM) biopsy and reported superior sensitivity for mpMRI compared to TRUS biopsy (93% vs. 48%, p < 0.001) and negative predictive value (89% vs. 74%, p < 0.001), allowing 24% with negative MRI to safely avoid having to undergo biopsy [72].
The utility of MRI may be applied to the domain of pre-surgical planning as well. The aggressiveness of nerve-sparing is not solely based on surgeon experience, but also on the anatomical location of the tumor and the presence of locally advanced (i.e. pT3a-b) disease that may be invading the neurovascular bundle. In the setting of pT3 disease (i.e. extraprostatic extension (EPE) and seminal vesicle invasion (SVI)), aggressive nerve sparing may result in a positive surgical margin (PSM), which has been associated with increased rates of biochemical recurrence, systemic metastasis, and prostate cancer-specific mortality post-prostatectomy, especially for high grade disease (i.e. Gleason Score 8-10) [73]. Even with careful adherence to surgical technique and the use of intraoperative frozen section analysis, microscopic positive margins may be imperceptible during surgery. A meta-analysis of 75 studies comprising 9796 patients who underwent mpMRI between 2000 and 2015 demonstrated high specificity for the detection of EPE (90%) and SVI (95%) [74]. However, sensitivity for these endpoints in the best performing series was 71-73%. Therefore, mpMRI may be limited in detecting microscopic EPE and SVI, which surely impacts the safety of nerve-sparing, especially in high risk patients.
Despite these limitations, there is level one evidence published by a single center in Norway suggesting that preoperative mpMRI does indeed reduce the rate of positive surgical margins at robotic radical prostatectomy and influences the rate of nerve-sparing in patients who otherwise might not have been considered for a nerve-sparing approach [75]. Among the n = 438 men randomized in this study to preoperative MRI vs. no MRI, sensitivity and specificity for detection of pT3 disease was 73 and 65%, respectively. The mpMRI information altered surgical approach in 27% of patients. Bilateral nerve sparing was performed 6.7% less frequently in the mpMRI group. The positive margin rate was reduced in the mpMRI group for cT1c patients (16% vs. 27%, p = 0.035). Among the patients found to have pT3 disease, 89% of them had only unilateral nerve sparing or no nerve sparing. This group did not have improvement in positive margin rate with the mpMRI, however. Perhaps, even wider excision is required to render these patients free of positive margins, which has implications for post-operative potency. Further study is surely needed to clarify the role of mpMRI for pre-surgical planning.

Intraoperative nerve-sparing techniques
Refined understanding of the neuroanatomy of the cavernous nerves informs the surgical approach to and completeness of nerve-sparing. The endopelvic fascia is a multilayered sheath that encloses and buttresses the prostate and bladder and attaches these organs to the pubic bone via the puboprostatic ligaments. This fascia fuses as the arcus tendineus fascia pelvis just lateral to these organs. The fascial investments of the prostate may be further divided into the prostatic "capsule," periprostatic veins with their fascia, the lateral pelvic fascia (prostatic fascia), levator fascia, and levator ani muscles (Figure 6) [76]. Lepor and Walsh described nerve sparing in 1983, with the approach beginning at the prostate apex and proceeding in a retrograde fashion toward the prostatic vascular pedicle [6,7]. In the modern era of minimally invasive and robotic techniques, the nerve-sparing is usually performed in an antegrade manner after first controlling the prostatic vascular pedicle. Initial experiences with robotic prostatectomy employed monopolar or bipolar electrocautery to control the pedicle until the detrimental role of thermal injury was fully appreciated.
Classical approaches to nerve-sparing have been described as "interfascial" vs. "intrafascial" techniques, as well as the "extra-fascial" approach when nerve-sparing is not performed (Figure 7). Interfascial dissection follows the plane lateral to the prostatic fascia, which may render the NVB prone to partial resection. The intrafascial technique follows the plane directly on the prostatic capsule, medial to the prostatic fascia and anterior to the Denonvilliers' fascia, especially at the 5:00 and 7:00 positions. Dissection is typically performed with both blunt and sharp dissection in an athermal manner to reduce transmission of heat and electricity to the proximal NVB.
Refinements in the understanding of the neuroanatomy have resulted in more sophisticated classifications of nerve-sparing. An important concept is the ability to perform incremental nerve sparing, not just as an "all-or-none" phenomenon. It has been suggested that optical magnification and the extended degree of freedom afforded by robotic surgery facilitates  development of dissection planes within the NVB itself to perform partial nerve sparing in the setting of concern for pT3 disease [12,77]. This concept is illustrated in Figure 6. Dr. Tewari's series of n = 2317 patients was published in 2011 [76]. This approach relies on a risk-stratified approach to the "neural-hammock" as defined by the periprostatic veins and based on preoperative risk stratification based on Gleason score, PSA, digital rectal examination (DRE) findings, cancer volume, and mpMRI findings. The plane of dissection follows one of the grades illustrated in Figure 6. A similar grading system based on periprostatic arterial anatomy, rather than venous anatomy, has been proposed by Schatloff et al. (Figure 8). In their smaller series of n = 132 patients, they cite a landmark periprostatic artery to define grades as: 1, no nerve sparing; 2, <50% nerve-sparing; 3, 50% nerve sparing; 4, 75% nerve-sparing; 5, ≥95% nerve sparing [78]. The Menon et al. series reported on n = 2652 patients for whom nerve-sparing was initiated by incising the prostatic fascia anteriorly, termed "high anterior release" or the "Veil of Aphrodite" technique ( Figure 9) [79]. The authors originally reported the development of a plane between the prostatic capsule and prostatic fascia cranially at the base of the seminal vesicles. This plane is propagated deep to the periprostatic venous sinuses with careful blunt and sharp dissection. For patients with significant periprostatic fibrosis after biopsy that impairs development of this plane, the authors recommend initiation of the dissection at the 2:00 or 10:00 position [79].

Outcomes of nerve-sparing techniques
Comparison of potency rates after the various surgical techniques is not straightforward, as there may be differences in patient demographics such as age and baseline potency, as well as tumor characteristics such as stage and grade that may influence the ability to perform bilateral nerve-sparing. Furthermore, based on the aforementioned variations in neuroanatomy among different patients, perfect execution of a given surgical technique may not be enough to accommodate a particular patient's anatomy, resulting in the variability of erectile function recovery. Definitions of potency also vary, with some studies reporting percentage return at a given time point or "return to baseline." The most robust series employ the validated questionnaires (IIEF, SHIM, EPIC), but precise cut-offs for restoration of function may vary. To add the possibility of subjective satisfaction beyond the numbers of the questionnaires, some studies also define potency as "erections suitable for intercourse" that are "satisfactory." In the setting of these limitations, a recent systematic review and meta-analysis of six studies (only one randomized and three prospective; 4/6 minimally invasive approach) of n = 1663   Erectile function is not the only component of the Trifecta directly affected by nerve-sparing. Oncologic control with regard to the presence of positive surgical margins is also very important. Given the increased rates of biochemical recurrence and possibility of increased rates of metastasis and prostate-cancer specific mortality associated with the presence of a positive surgical margin, nerve-sparing outcomes must be understood within the context of margin status. Overall, the rates of PSM are around 15%, with rates ranging from 6 to 38% and influenced by pathologic stage, grade, and D'Amico risk category [80]. There have also been reports that bilateral vs. unilateral nerve-sparing in the setting of pT2 disease is associated with a higher rate of PSM, as demonstrated in a series of n = 9915 patients who underwent robotic prostatectomy at two institutions, with relative risk 1.52 [81]. Other studies have not corroborated an increased positive margin rate for intrafascial vs. interfascial nerve-sparing techniques (rate of 9% vs. 9.5%) [82]. The grades of nerve-sparing proposed by Dr. Tewari have been associated with excellent PSM rates, likely owing to preoperative risk stratification based on Gleason score, PSA, digital rectal examination (DRE) findings, cancer volume, and mpMRI findings. The rates of PSM for patients with nerve-sparing grades 1, 2, 3, and 4 were 9.9, 8.1, 7.2, and 8.7%, respectively (p = 0.636). The Schatloff series also reported their PSM rate, which was 9% overall. Although the rate of PSM for grade 1 NS was 0%, there was otherwise no consistent correlation between grade of NS and PSM rate (0, 5.7, 16.7, 7.5, 3.6% for grades 1, 2, 3, 4, and 5, respectively), perhaps reflecting good presurgical planning based on patient risk factors [78]. The Veil of Aphrodite technique by the Vattikuti Urology Institute reported a positive margin rate of 13%, which declined to 1.5% for pT2 disease after modification of approach to include en face oversewing of the DVC after apical transection [79].
Regarding the potency rates, it is necessary to standardize the definition of potency, which usually incorporates routine use of oral PDE 5 inhibitors. Furthermore, there is a distinction between restoration of penetrative sexual intercourse vs. return to baseline functioning. There are several contemporary series reporting erectile function outcomes after robotic prostatectomy. Tewari's risk-stratified approach to neural-hammock sparing in n = 2317 men resulted successful intercourse (score of ≥4 on question two of the SHIM and total SHIM >21) of 90.9, 81.4, 73.5, and 62% for nerve-sparing grades 1, 2, 3, and 4, respectively [76]. Regarding return to baseline function: grade 1, 81.7%; grade 2, 74.3%; grade 3, 66.1%; grade 4, 54.5%. Of note, this group also reported earlier return of continence associated with the higher grades of nerve-sparing, which has been controversial [83]. Incidentally, a recent systematic review and meta-analysis of 27 longitudinal cohort studies totaling n = 13,749 patients, however, reported that early urinary continence (at 6 months post-RP) was improved for patients undergoing nerve-sparing vs. The Schatloff series on grade stratification based on periprostatic arterial anatomy did not report functional outcomes [78]. The Vattikuti Institute series included n = 1142 patients with minimum follow-up of 12 months and among men with normal preoperative function (i.e. SHIM >21) potency rates were 68% in the standard bilateral nerve-sparing patients and 93% in the bilateral Veil nerve-sparing patients [79]. Return to baseline rates depended on preoperative function, and for those without preoperative dysfunction, return to baseline was 39% for standard nerve-sparing and 73% for the Veil technique. Despite these very favorable results, the authors disclosed that only 50% of these patients attained normal SHIM score without medication. Although these findings suggest that the Veil offers improvements in recovery of erectile function, there may be a bias of surgeon experience, as the Veil technique was employed later in the learning curve of this single-surgeon series. This series also demonstrated 84% total urinary control, and 95% social continence (one pad or liner per day) at 12 months follow-up. The role of nerve-sparing in earlier recovery of continence has been corroborated by many series [86][87][88].
Some limitations of these studies are that they are single institution and often single-surgeon series and there is no direct comparison among different techniques to be able to assess superiority. Furthermore, there may be shortcomings with translation of these techniques into the larger urologic community compared to the immensely high-volume centers where these techniques were invented.

Intraoperative penile oxygen monitoring
The real-time impact of neurovascular bundle tension on cavernosal ischemia was investigated by Tewari et al. in n = 64 patients [85]. During robotic prostatectomy, these patients underwent real-time penile oxygen monitoring with the Odissey Tissue Oximeter probe placed 2 cm from the base of the penis. Surgical dissection was altered whenever the O 2 alarm sounded until oxygenation was restored to 85%. Functional outcomes were compared to a propensity-matched historical control group of n = 192 patients (matched for age, preoperative prostate specific antigen (PSA), baseline erectile function, comorbidity status, and extent of nerve-sparing). Steps in the operation associated with significant decline in tissue oxygenation included opening the endopelvic fascia, all of the nerve-sparing, excessive traction on the Foley catheter, seminal vesicles, or prostate during the apical dissection, and control of the dorsal vein complex (DVC) prior to apical dissection. Of note, control of the DVC if done after apical transection was not associated with significant penile ischemia. At 6 weeks postoperatively, a larger proportion of patients in the O 2 monitoring group had no ED (24.5% vs. 10.4%, p < 0.05) and at 52 weeks this difference was persistent (84% vs. 68%, p < 0.05). Furthermore, using the Sexual Health Inventory for Men (SHIM) validated questionnaire at 1 year, a greater number proportion of patients reported minimal to no ED (94% vs. 78%, p < 0.05). These findings provide clinical evidence for the importance of minimizing neurovascular bundle manipulation during robotic prostatectomy as a means of preventing neurapraxia/axonotmesis.
Similar evidence comes from a well-designed prospective, randomized, single-blinded study of n = 61 with normal preoperative erectile function from 6 centers [89]. Patients underwent robotic prostatectomy with traditional bilateral nerve sparing compared to nerve-sparing using the Cavermap Surgical Aid. A 12 months post-op, the Cavermap group had mean 15.9 minutes of greater than 60% nocturnal tumescence vs. 2.1 minutes as measured by the RigiScan. The sexual function inventory questionnaire (SFIQ) scores at 12 months were not significantly different, however. Among those patients with intact response to nerve stimulation after nerve-sparing, 68% of those with bilateral response had recovery on SFIQ vs. 27% with unilateral response, and 0% with no response [89]. A subsequent multi-institutional study utilizing the Cavermap by five experienced surgeons demonstrated limited specificity to identify the precise location of the cavernous nerves, thus limiting its routine application during radical prostatectomy [90].

No countertraction technique
The detrimental impact of nerve traction injury may also be limited. A series emphasizing nerve-sparing with a lack of countertraction has been published (Figure 10). Kowalcyzk et al. reported statistical significantly different erectile function at 5 months post-robotic prostatectomy (24.9% vs. 18.4%, p = 0.004), which was confirmed in the multivariable regression model. These differences were no longer present at 12 months (34.7% vs. 33.5%, p = 0.849), independent of preoperative erectile function, laterality of nerve sparing, and inter-vs. intrafascial approach [91]. Therefore, "tractionless" surgery may accelerate functional return.

Minimizing use of electrocautery/thermal energy
In the early experience of robotic prostatectomy, many centers performed dissection of the neurovascular bundle with electrocautery. Ahlering et al. demonstrated slower return of erection function as measured by the IIEF-5 and two questions on the EPIC questionnaire. Of the n = 125 patients, only 36 met their inclusion criteria, (age < 66 and IIEF-5 score 22-25), with n = 3 having had monopolar electrocautery and n = 33 having had bipolar cautery. Of note men with Gleason 7-10 and high volume disease had ipsilateral wide excision of the NVB. Although there was no comparison group, erectile function recovery was modest especially when compared to modern historical series. Among those who had bilateral nerve-sparing, recovery at 15 months was 44.4% and at 24 months was 67.9% [31]. These findings help to support the clinical principle of avoidance of thermal energy during dissection of the neurovascular bundle.

Seminal vesicle preservation
Building on several observational studies [92,93], Gilbert et al. reported a randomized controlled, Phase II trial of n = 140 men who underwent radical prostatectomy [94]. Seminal vesicle-sparing approach was employed in 71 men. At 6 and 12 months post-operatively, sexual and urinary function scores were similar on the EPIC questionnaire (erections firm enough for intercourse in 67.7% vs. 56.3%, p = 1). In addition, positive surgical margin status and 12-month biochemical recurrence rates were similar. This approach has understandably not been widely adopted.

Hypothermic robotic radical prostatectomy
Finley et al. reported the potential benefit of regional hypothermia/cold dissection of the neurovascular bundle in a case-control study of n = 115 who underwent robotic prostatectomy [95,96]. The rationale for the endeavor mirrors developments in neurosurgery and cardiac surgery that have established significant benefits in randomized studies. Hypothermia was employed to minimize iatrogenic tissue inflammation and consequent cellular edema, lactic acidosis, nerve conduction blockade, free radical damage, and apoptosis that characterizes damage to muscle and nervous tissue. Cold intracorporeal irrigation was employed along with an endorectal cooling balloon cycled with saline at 4°C. Potency rates at 12-months were favorable. Patients subjected to hypothermia were compared to a historical cohort of n = 667 patients. Temperature probes monitored the endorectal and intracorporeal temperatures, which declined to mean 18.7 and 25.6°C, respectively. Potency was assessed during validated questionnaires and defined as "erections adequate for penetration" and "were the erections satisfactory." Potency at 3 months was similar, but at 15 months, the hypothermic group had significantly better function (83% vs. 66%, p = 0.045). There were no differences in oncologic outcomes and no complications related to the technology [97]. These results need further multi-institutional investigation.

Clipless antegrade nerve preservation
Rather than perform the nerve-sparing in the conventional antegrade (base to apex) fashion, some centers have described a medial to lateral approach. After division of the posterior bladder neck, the posterior plane along the prostate is developed toward the prostatic apex in the midline. The dissection proceeded laterally to release the vascular pedicles and neurovascular bundles in a medial to lateral direction, with sparing use of bipolar cautery.

Intraoperative frozen section analysis
There have been multiple reports of the utility of intraoperative frozen section to allow for more aggressive nerve-sparing in patients whose risk factors may have otherwise prompted a non-nerve sparing surgical approach. In an open series of n = 608 patients who underwent RP, 83 patients were found to have a palpable lesion close to the prostatic capsule [98]. A 4 cm wedge of tissue was excised in the suspicious area for intraoperative frozen section (IOFS). A total of 93% of these IOFS were positive for carcinoma, and 36% of these were pT3. Final positive margin rate overall was 16%. This real-time decision-making allowed for ipsilateral nerve-sparing in 52% of the cases without a negative impact on PSM. Of note, there was a false positive PSM rate of 17%. The applicability to the modern era of robotic prostatectomy may not be lost, but would be based on gross visual suspicion of tumor violation during nerve-sparing.
Another center has pioneered the Neurovascular Structure-Adjacent Frozen-section Examination ("NeuroSAFE") approach to nerve sparing [99]. In this technique, a bilateral nerve-sparing procedure is performed and then the prostate gland is promptly extracted from the surgical field and whole gland circumferential frozen section analysis is performed. The original series consisted of n = 11,069 who underwent open RRP from 2002 to 2011, n = 5392 of whom had the NeuroSAFE technique. When a margin was found to be positive, the ipsilateral neurovascular bundle was resected, including the rectolateral component of the Denonvilliers' fascia, prior to the vesicourethral anastomosis. The sensitivity and specificity of this approach was 93.5 and 98.8%, respectively, with accuracy of 97.3%. Of the 25% found to have PSM initially, 85% of these were converted to final negative margin. False negatives occurred in 2.5% and all of these margins were < 0.5 mm. There were significant reductions in PSM rates within each pathologic tumor stage (except for pT3b) and an increase in the rate of nerve-sparing for all stages. Of note processing time took about 35 minutes and there was no delay in surgery, as hemostasis, bladder closure, lymph node dissection, and posterior reconstruction could be performed during this time (Figure 11) [99].  [100]. These findings have contributed to the development of the "Safe-R score," a composite measure of margin status and laterality of nerve sparing [101].

The influence of surgical modality on nerve-sparing success
It has been postulated that the loss of haptic feedback renders traction-free nerve-sparing difficult during robotic prostatectomy. Conversely, the optical magnification and seven-degrees of freedom afforded by robotic surgery may allow for superior delineation of the neuroanatomy, precision of dissection, and even performance of partial or incremental nerve sparing for patients with concern for locally advanced disease. Such patients may have otherwise been subjected to a non-nerve-sparing technique. Tewari et al. reported earlier return of 50% erectile function (as reported on the EPIC questionnaire) after robotic prostatectomy vs. RRP (mean 180 days vs. 440 days, p < 0.05) and earlier return to intercourse (340 days vs. >700 days, p < 0.05) Interestingly, these findings were shown even in the setting of a greater number of patients post-RRP using sildenafil (65% vs. 42%) [102]. Such findings were also corroborated by a prospective, nonrandomized trial of robotic prostatectomy vs. RRP in n = 208 patients from 2006 to 2007. At 12 months follow-up, of the patients with bilateral nerve-sparing, those who underwent the robotic procedure had superior recovery of erectile function on the IIEF-5 questionnaire (81% vs. 49%, p < 0.001) [62]. Further contemporary evidence comes from a systematic review and meta-analysis of 31 studies published between 2008 and 2011 totaling n = 3491 patients [63].  [103]. Although the study randomized n = 326 men and plans to report on urinary and sexual function outcomes (via the EPIC and IIEF questionnaires) at 6 and 12 weeks, as well as 24 months, published results are immature at 12 weeks only. There were no significantly different urinary or sexual function scores at 6 and 12 weeks, nor were there differences in health reported quality of life.
A counterargument to the benefits afford by robotic surgery, however, states that more aggressive nerve sparing, although technically feasible, may not be oncologically safe, given the risk of positive margins. Indeed, the Preston et al. study revealed that positive margins were more likely in patients treated with robotic surgery (relative risk 1.76) compared to open surgery, while there was no difference between lap and robotic surgery [81]. These findings were not corroborated in a very large, albeit retrospective, multi-institutional, multi-national study of n = 22,393 patients, however, which found the lowest rate of positive margins in robotic prostatectomy ( One notable difference between the open and robotic approach is the use of a retrograde vs. antegrade approach to the dissection of the neurovascular bundle. There is a theoretically reduced risk of placing a clip across the neurovascular bundle with the retrograde approach, which releases the bundle from the prostate prior to obtaining vascular control [106]. These different approaches were compared in a propensity-matched series of n = 344 patients undergoing robotic prostatectomy. Using validated questionnaires at 3,6, and 9 months, the potency rate was significantly higher after the retrograde approach (92.9% vs. 72.1% at 9 months), and this finding was maintained with multivariable analysis. Of note, the PSM rate was similar between groups (11.6% vs. 7%, p > 0.05) [107]. Of note, this was a single-surgeon series and the approaches were performed in an interfascial manner, which may not reflect the most current understanding of the cavernous neuroanatomy. Also, the retrograde approach was generally performed more recently in the series with possible artifactual enhancement in outcome from being later in the learning curve. Furthermore, the seemingly excellent results from either approach in this series are likely a consequence of the permissive definition of erection function (erections firm enough for penetration in >50% of attempts).
Of note, although intriguing to consider if there are any technical advantages to nerve sparing with an extraperitoneal vs. transperitoneal approach to robotic prostatectomy, the extant literature has thus far focused exclusively on perioperative rather than functional outcomes [108,109].

Penile rehabilitation
Several options for penile rehabilitation have been proposed and investigated, including pharmacotherapy with oral phosphodiesterase inhibitors or penile intracavernosal injection, as well as use of a vacuum erection device and penile constrictive ring. Often a multimodality approach is advocated. Compliance with these recommendations depends on myriad factors, including physician counseling, patient motivation, and even socioeconomic status-as many of the therapies are not covered under health insurance. The PCOS trial reported that 43% of men tried sildenafil, 25% tried a vacuum erection device, and 17% tried intracavernous injections [52]. Admittedly, the denominator of men who were actually offered these treatments is not known.
The rationale for these treatments has been investigated in preclinical models. There is indeed precedent for treatment success in animal models of chronic low-dose tadalafil administration, with increased cavernosal smooth muscle, decreased fibrous tissue, and functional enhancement of erectile function [110][111][112]. Similar findings have been reported for sildenafil after bilateral cavernosal nerve damage in a rat model. Nerve damage resulted in elevation in several pro-inflammatory cytokines (interleukin-1β, transforming growth factor β) and markers of oxidative stress (nicotinamide adenine dinucleotide phosphate [NADPH] oxidase, myeloperoxidase, inducible nitric oxide synthase, tumor necrosis factor receptor superfamily member 5 [CD40]), which then normalized after administration of sildenafil in the drinking water. Levels were measured by polymerase chain reaction (PCR) and proteome expression of pelvic ganglia neurons [113].

Mechanotherapy
The vacuum erection device often in conjunction with penile constrictive ring are routinely recommended for use to assist with recovery of potency after radical prostatectomy. This therapy not only allows for tumescence and penetration, but also cavernosal sinus expansion, smooth muscle "stretching," and mitigation of hypoxia when used on a regular basis [114,115]. Compared to pharmacotherapies-which are often not covered by the patient's health insurance-VED may be more cost-effective, with a decreased side effect profile and the opportunity for the patient and his partner to take an active role in convalescence. In addition, the mechanism of action does not require intact cavernous nerves for success. There have been multiple retrospective studies examining the impact of the VED [115][116][117]. These suggest improvement in return of spontaneous erections. In particular, a retrospective study of n = 203 patients who underwent robotic radical prostatectomy between 2007 and 2011 investigated whether PD5I alone, VED alone, or a combination of the two yielded the highest improvement of the SHIM questionnaire, substratified into three groups of baseline EF (SHIM 8-16, SHIM 17-21, SHIM [22][23][24][25]. For each of the baseline EF groups the combination therapy resulted in the highest proportion of successful potency (erections suitable for penetration) and with the shortest latency period [118]. Compliance with the device was 80% with 55% partner satisfaction rate. After 9 months of treatment IIEF-5 score was significantly increased for both the NS and non-NS patients compared to the no treatment group. Furthermore, decreased penile length was reported in 63% of the control group vs. 23% among patients who responded to VED treatment [115].
Regarding timing of therapy, earlier initiation of 10 minutes daily VED therapy (1 month post-op vs. at 6 months) has been shown in a small randomized trial of n = 28 men to be superior in terms of 1. IIEF score at 3-and 6-months, and 2. preservation of stretched penile length (vs. 2 cm mean decrease observed in the delayed therapy group) [119].

Pharmacotherapy
The majority of penile rehabilitation studies incorporate early post-op therapy. Phosphodiesterase type 5 inhibitors have been demonstrated to be effective for the treatment of erectile dysfunction through their inhibitory effect on the enzyme that degrades cGMP. These medications augment the nitric oxide-mediation erectile response through increased relaxation of the cavernosal sinus smooth muscle [120]. The commercially available medications have different halflives of activity, with Tadalafil being the longest (T 1/2 sildenafil 3-5 mg; vardenafil 4-5 hours; tadalafil 17.5 hours) [121]. There is also experience with both intraurethral and intracavernosal prostaglandin E, which act via a cAMP-related mechanism to effect cavernosal smooth muscle dilatation [122]. Montorsi et al. published the first randomized, placebo controlled trial in 1997, involving intracavernosal injection of alprostadil three times per week × 12 weeks beginning 1 month after open RP [23]. Newer formulations of intracavernosal therapy include Trimix (prostaglandin, phentolamine-a non-selective alpha-blocker-and papaverine-a non-selective PDE5 inhibitor-and Bimix (papaverine and phentolamine). Given the invasive nature of ICI-as well as the putative higher complication rate regarding pain, hematoma, penile plaque formation, and priapism-this therapy is not universally accepted by patients. Therefore, there is much interest in the oral PDE5 inhibitors as a mechanism of increasing intracavernosal cGMP to promote smooth muscle relaxation and mitigate of the post-RP hypoxic state.

Oral PDE 5 inhibitors
There have been several prospective, randomized studies in this domain, which are summarized Table 1. Rationale for these trials was based on findings that daily sildenafil preserves intracorporeal smooth muscle after radical retropubic prostatectomy [123]. The studies tended to exclude men with high grade prostate cancer (i.e. Gleason 8) and those who required adjuvant radiation therapy. Latency of medication initiation is also variable, ranging from time of catheter removal to 4 weeks post op. Based on the cavernous neurotomy rat models in which post-injury cavernous smooth muscle apoptosis begins in 1 day and peaks at 1 week, earlier initiation of treatment has definite clinical rationale [34]. Furthermore, the study end-points are not identical. Some examine the impact on erections during an active treatment phase and others examine the rate of return of spontaneous erections (that is, without the need for pharmacotherapy). The proportion of open vs. robotic RP are also variable (open RP: 39.7-100%). The Jo et al. study out of South Korea is the only study comprised exclusively of patients who underwent robotic prostatectomy, and showed more complete return of erectile function at 12 months with immediate post-catheter removal initiation of 100 mg sildenafil twice weekly compared to waiting until 3 months after surgery [124]. The concomitant use of VED in these studies is not clear, which adds another limitation to their interpretation. Also of interest is a follow-up study to the study by Pavlovich et al. [172] suggesting poorer EPIC scores with nightly sildenafil 50 mg compared to PRN dosing, largely in the urinary irritative and bother subscales of the questionnaire [125]. Further study is required to clarify the implications of this report.   Table 1. Randomized controlled trials of post-prostatectomy pharmacologic penile rehabilitation.

Prostatectomy
A majority of these trials for ED after radical prostatectomy examine daily vs. on-demand dosing schedules. Of note, a randomized, double-blind trial (RESTORE study) in men with mild to moderate ED (IIEF-EF 15-20; not post-prostatectomy) of 10 mg vardenafil daily vs. PRN showed no differences in increase of IIEF-EF score [126]. Similar findings appear to be borne out in the post-RP series as well (see below). Side effect profiles of the treatment groups were largely favorable and usually without the need for drug discontinuation. Furthermore, daily vs. on demand schedules did not have significantly different adverse effects in a recent metaanalysis, and absolute rates of side effects were not substantially different between treatment and placebo groups overall (59.6% vs. 48.4%) [127]. Side effects generally reported included headache, flushing, dyspepsia, and rhinitis, none of which are severe.

Intraurethral prostaglandin
Intraurethral alprostadil (Medicated Urethral System for Erection, MUSE™) has also been developed for use in erectile dysfunction. Experience with this drug for non-post-RP erectile dysfunction has been limited by lack of efficacy and penile pain/dysuria in a large number of patients. Early experience with MUSE after 6 months of post-RP ED showed 55% of the n = 54 patients able to achieve erections suitable for penetration, although only 48% continued longterm therapy (57% of men discontinued therapy for inefficacy) [128]. The author's experience with application of the therapy at 3 weeks post-op for a duration of 6 months showed erections firm enough for penetration in 74% of patients compliant with treatment. All patients reported mild penile pain or urethral burning and 32% of patients discontinued treatment [129].

Pelvic floor PT
The role of pelvic floor physical therapy (PFMT; "Kegel" exercises) on return of urinary continence has been well understood, as evidenced by meta-analyses demonstrating earlier return of continence with preoperative PMFT and biofeedback [130,131]. The impact of such therapy on recovery of sexual function has also been explored. Physical therapy efforts focus on the bulbocavernosus (bulbospongiosus) and ischiocavernosus muscles. Contraction of the ischiocavernosus muscle may compress the proximal aspects of the corpora cavernosa and increase intracavernosal pressure during erection [132]. Also, muscle contraction has been shown to increase levels of brain-derived neurotrophic factor in muscle cells, which may have a role in promoting neuronal growth after nerve injury [133]. A trial of n = 55 men with ED for ≥6 months randomized one group to PFMT with biofeedback + lifestyle changes vs. lifestyle changes alone [134]. Lifestyle changes consisted of alcohol intake reduction, smoking cessation, increasing exercise, and avoiding bicycle riding. Outcomes were assessed with IIEF and anal pressure measurements. At 3 months follow-up, the PFMT group had greater improvement in IIEF than the lifestyle changes alone group (p < 0.001). Other concomitant therapies for ED (i.e. PDE5 inhibitors or vacuum erection device) were not disclosed by the authors, limiting the strength of the results. Also, there were no post-prostatectomy patients in this cohort.
With regard to prostatectomy, the timing is such physical therapy is also interesting to consider and there is no established consensus of whether pre-and postoperative physical therapy is superior to postoperative therapy alone. Perez et al. examined the use of biofeedback preoperatively to strengthen the levator muscles prior to radical prostatectomy [135]. Their methods employed a device that provided visual feedback based on intra-anal pressures.
In this prospective cohort study, a total of n = 20 patients completed 10 sessions pre-RP and n = 32 patients proceeded directly to surgery. Potency outcomes were assessed using the IIEF-5, although follow-up time was not clear. The erectile dysfunction rate in the physical therapy group was 5% vs. 48.6% in the control group < p < 0.001). There are likely variations in physical therapy technique and not all methods are standardized.
The impact of PFMT on erectile function after radical prostatectomy has also been examined. Prota et al. studied n = 52 patients who underwent open RP and randomized them to PFMT + biofeedback weekly × 3 months (beginning at time of catheter removal) vs. verbal instructions only [136]. Nerve-sparing was performed in a similar proportion of patients in each group (64.7% vs. 68.8%). There was earlier recovery (IIEF >20) in the treatment group which persisted at 12 months post-op (47.1% vs. 12.5%). The authors report an absolute risk reduction of 34.6%, with number needed to treat (NNT) of 3. These results are even more encouraging given that oral PDE5 inhibitor therapy was withheld during the study. This study provides level 1b evidence supporting the use of post-operative PFMT for erectile function recovery [136]. Optimal schedule and intensity has yet to be determined.
Additional randomized controlled trials are needed to further assess the utility of PFMT in the restoration of potency.

Hyperbaric oxygen
The role of hyperbaric oxygen (HBO) in Urology for treatment of radiation-induced hemorrhagic cystitis is well known. Application of this technology to post-RP erectile dysfunction is controversial. There have been preclinical studies in rats with cavernous nerve injury undergoing HBO that demonstrate higher intracavernosal pressure/mean arterial pressure ratio, increased levels of penile nerve growth factor (NGF) and endothelial nitric oxide synthase (eNOS) compared to the controls. Tissue studies on smooth muscle to collagen ratio, however, did not show a significant difference after HBO [137]. Exposure to 100% oxygen at 2 atm pressure induces stem cell differentiation and neovascularization, as well as vasoconstriction that attenuates tissue edema [138]. These findings underlie the rationale for translational studies. Chiles et al. published the first randomized, double-blind controlled trial of HBO vs. air for men after RP [139]. Although the authors were unable to demonstrate significant difference in IIEF score at 18 months, uncertainty about the proper regimen for therapy (total number and frequency of treatments) and lack of a true "sham" group may have limited their ability to detect a clinically relevant difference in outcome.

Neuroprotective agents
There is a growing body of literature on drug therapy to preempt or mitigate the postcavernosal nerve injury pro-inflammatory environment. The immunophilin ligand FK506 (tacrolimus)-traditionally thought of as an immunosuppressive agent and widely used in for solid organ transplant-has been implicated in a neuroprotective role when administered as early as 1 day following partial nerve-crush injury in a rat model [140]. Further work has demonstrated improved intracavernous pressure/mean arterial pressure ratio, restoration of inducible nitric oxide synthase (iNOS) staining, reduced apoptosis, preservation of cavernosal architecture, and upregulation of glutathione peroxidase (GPX) with resultant decrease in oxidative-stress-induced tissue damage [141,142].
Pioglitazone, a thiazolidinedione used for the treatment of diabetes mellitus, may enhance neuronal survival and regeneration and decrease inflammation, and has been shown to be neuroprotective in models of sciatic nerve ischemia and optic nerve crush injury, as well as BCNI [143][144][145]. Furthermore, a small randomized controlled trial has shown efficacy of this agent for erectile dysfunction refractory to sildenafil [146]. A study by Katz et al. investigated the impact of pioglitazone on pelvic ganglion neurons after bilateral cavernosal nerve injury (BCNI) in Sprague-Dawley rats [147]. Four groups were examined: sham surgery, BCNI, BCNI + post-operative pioglitazone, BCNI + pre-and post-operative therapy. Gene expression profiles of neuronal nitric oxide synthase, neurturin, glial cell line-derived neurotrophic factor family receptor alpha-2 (GFRα2), and β-III tubulin were upregulated in the pre-operative therapy group [147]. Further work is necessary to fully explore the utility of this therapy.

Amnion-chorion membrane
Dehydrated human amnion-chorion membrane allograft (dHACM) is a source of implantable neurotrophic factors and cytokines which promote neural survival and facilitate axonal regeneration. Its application has been examined in a rat model of axonal regeneration after spinal cord injury [148]. Clinically, it has been applied in the treatment of burns, corneal injuries, chronic venous ulcers, and chronic wounds [149]. There has been preliminary work in the placement of this membrane after bilateral nerve-sparing robotic RP to accelerate erectile functional recovery after surgery [150]. A single-surgeon propensity-matched analysis of preoperatively potent (SHIM >19) and continent (American Urological Association Symptom Score < 10) patients (n = 58) demonstrated earlier return of continence (1.2 months vs. 1.8 months, p = 0.033) and potency (1.34 months vs. 3.39 months, p = 0.007) with the wrap. (AmnioFix; MiMedx Group, Marietta, GA, USA) Some limitations of this series include the small size, lack of randomization, and short mean follow-up of 4 months, which limits conclusions about oncologic safety vis-à-vis risk of biochemical recurrence (BCR). Furthermore, functional recovery rates appear much higher than other series in the literature, and may be a consequence of the recall bias used to assess level of function. These preliminary results are certainly encouraging.
The dHACM allograft wrap was recently examined in a series of n = 940 patients (preoperative SHIM > 20) who underwent robotic RP with bilateral nerve-sparing [151]. A total of n = 235 had bilateral dHACM placement and these were propensity matched in a 1:3 proportion to non-dHACM patients (n = 705). Potency recovery rates were higher in the dHACM group at all time points except 12 months. Time to potency was significantly shorter in the dHACM group after bilateral NS (2.2 months vs. 2.8 months, p = 0.029) and partial NS (3 months vs. 3.9 months). After 12 months follow-up, erections sufficient for penetration were similar. Of note, the recovery rates in this study were higher than what is generally reported in the literature. Also, although the rate of biochemical recurrence at 12 months was similar between groups, longer follow-up is certainly needed to demonstrate the oncologic safety of this application of technology.
A recent pre-clinical study of the role of hemostatic tissue sealing sheets has also investigated the impact on erectile function recovery in a series of 21 Sprague-Dawley rats [152]. The TachoSil (CSL Behring, Tokyo, Japan) is a collagen sponge coated on one side with fibrinogen and thrombin and is approved for achieving hemostasis during surgery. Contact between the sheet and blood or serosanguinous fluid results in deposition of a fibrin clot. Compared to sham surgery, the rats who underwent cavernous nerve dissection and had the TachoSil placed demonstrated similar intracavernous pressure/mean arterial pressure ratios at 4 weeks post-op. Furthermore, PCR-measured expression of inflammatory and oxidative markers (internleukin-6, tumor growth factor beta1, and heme-oxygenase-1) in the major pelvic ganglion was significantly reduced in the sheet group (p < 0.05).

Stem cell therapy
Stem cells may be harvested from growing embryos (embryonic stem cells) or as allografts from bone marrow or adipose tissue (mesenchymal stem cells). This technology has been applied to rat models of bilateral cavernosal nerve crush injury and shown considerable promise. Bochinski et al. [153] conducted a study of embryonic stem cells induced along the neuronal cell line with brain-derived neurotrophic factor [153]. These stem cells were then injected into the major pelvic ganglion (MPG) (group 3) and into the corpora cavernosa (group 4) of rats after BCNI. The study was well controlled with sham surgery (group 1) as well as a BCNI group with injection of culture media only (group 2). Volume of stem cells injected was 500 μL of a 10,000 cells/ mL solution. Erectile response was assessed by electrostimulation of the cavernosal nerve at 3 months. Immunohistochemistry was also performed of the penile tissue to assess levels of nitricoxide synthase-containing fibers and neurofilament concentration. Intracavernosal pressure in response to electrostimulation was greatest for sham surgery and lowest for group 2, which was also significantly lower than groups 3 and 4. The neurofilament stain of tissue taken from the MPG and the corpus cavernosum was also greater in groups 3 and 4 compared to group 2 ( Figure 12). Such neurofilaments are involved in establishing tensile strength and putatively intracellular transport guidance to axons and dendrites [154]. The authors conclude that preservation of the neuronal architecture may promote/facilitate nerve regeneration after nerve injury.
There have been several more studies investigating the role of stem cells in BCNI in rats, either alone [155] or in combination with PDE5 inhibitors [156,157]. Furthermore, Lin et al. reported that adipose-derived stem cells (ADSCs) injected into the corpora cavernosa migrated within days to the bone marrow and then to the MPG [158]. Subsequent work has been reexamined in multiple systematic reviews and meta-analyses, reiterating the efficacy of these methods in 12 studies, n = 319 rats and 20 studies, n = 248 rats, respectively [159,160]. The combination of stem cells + PDE5 inhibitor therapy appears to have the greatest effect [159]. Consistent outcomes among the studies were increase in the ICP/MAP ratio, levels of neuronal nitric oxide synthase, cavernous smooth muscle content, ratio of cavernous smooth muscle to collage, and cGMP levels [160]. Furthermore, stem cells modified by growth or neurotrophic factors prior to implantation appeared to exert the greatest benefit [160]. Although there are many processing steps to refine and standardize, this approach remains a riveting endeavor in the arena of penile rehabilitation.

Penile prosthesis implantation
Despite these advances in the restoration of endogenous erectile function after radical prostatectomy, some men will have dysfunction refractory to the above treatments. The final line treatment option for such patients is insertion of a penile prosthesis, which may in fact be conceived as a type of mechanotherapy. This procedure eliminates the possibility of any Preventing Erectile Dysfunction after Radical Prostatectomy: Nerve-Sparing Techniques, Penile… http://dx.doi.org/10.5772/intechopen.79398 spontaneous erectile function in the future and has many potential complications, including pain, erosion, extrusion, infection, mechanical failure, need for revision, and altered penile sensation. There have been a number of surgical improvements that have reduced the infection rate, including the "no touch" technique, standardized protocols for perioperative antibiotics, and the development of antibiotic-impregnated and hydrophilic devices [161][162][163][164]. Although the technique of 3-piece penile prosthesis insertion is beyond the scope of this chapter, there are a number of considerations with regard to radical prostatectomy. The cavernosal smooth muscle fibrosis and penile shortening can have an impact on the size of the cylinders able to be accommodated by the corpora, leading to actual or perceived reduction in penile length. Furthermore, the lack of glans tumescence with cylinder expansion may exacerbate the perceived loss of size. After appropriate dilation of the corporal bodies with the dilators, fibrosis should not preclude placement of a three-piece inflatable prosthesis, which has been reported to have the highest satisfaction rates and lowest rate of mechanical failure [165]. Some models of the 3-piece prosthesis (i.e. American Medical Systems, AMS 700 CX) allow for cylinder axial expansion to allow girth. This is important for men with penile plaques/Peyronie's disease and penile curvature that must be corrected during placement of the prosthesis [164].
Although not previously discussed, Peyronie's disease is another "sexual" complication of radical prostatectomy, thought to be from repetitive "buckling" injury to the phallus as a result of intercourse in the setting of an incomplete erection.
Reservoir placement during the three-piece prosthesis surgery also deserves consideration, given the previous dissection of the space of Retzius during the prostatectomy. Given the increased risk of bladder perforation, bowel and vascular injury in this setting, some authors favor ectopic reservoir placement just posterior to the rectus muscle and anterior to the transversalis fascia [166,167]. Single-institution series have reported excellent outcomes with this approach [166,167].
Satisfaction with this treatment appears favorable, as exemplified in a recent study of n = 71 patients who underwent penile prosthesis implantation after radical prostatectomy [168]. Pillay et al. employed the EDITS and SEAR questionnaires, as well as the Prostate Cancer-Related Quality of Life Scale and Patient Health Questionnaire-9 (PHQ-9) for both patients and their partners. They reported good sexual function (EPID score > 60) in 77% of men and treatment satisfaction in 94% (EDITS score > 50). Other studies have actually reported improved satisfaction for men undergoing early penile prosthesis insertion compared to those receiving sildenafil or intracavernosal injection therapy 6 months after radical prostatectomy [169,170]. Such studies did not employ early aggressive pharmacologic penile rehabilitation programs, however. Certainly, for the appropriately selected patient, penile prosthesis implantation has a very high level of success and satisfaction for patients and their partners alike.

Conclusions
There have been many advances in the understanding of erectile dysfunction as a result of radical prostatectomy since the initial pioneering work by Walsh & Donker almost 40 years ago. Refined understanding of neuroanatomy, beneficial modifications in surgical technique, the advent of robotic surgery, and the exploration of pre-and post-operative rehabilitation techniques using mechanotherapy and pharmaceuticals have improved the prognosis for potency recovery after this once morbid surgery. Further developments in the realm of local and systemic therapies for cavernous nerve neuroprotection and regeneration may mitigate the cascade of cavernosal smooth muscle apoptosis, fibrosis, and veno-occlusive dysfunction that jeopardizes further erectile function recovery after the period of post-operative neurapraxia. Achievement of the Trifecta is extremely important to patients and clinicians alike, and will surely inspire the future waves of investigations that continue to elucidate this field.