The routine use of preoperative mechanical and antibiotic bowel preparation prior to radical cystectomy has recently been challenged (28,29). Concern over alteration of colonic bacterial flora with resulting higher rates of Clostridium difficile colitis due to preoperative oral antibiotic regimens remains controversial but has led many groups to discontinue them from routine use (30). While the majority of patients undergo urinary diversion utilizing small bowel and can probably do without an antibiotic preparation, any patient in whom a colonic diversion is contemplated may benefit from both mechanical and antibiotic bowel cleansing.

Patients are instructed to commence a clear liquid diet starting 1 or 2 days prior to surgery. Simple mechanical bowel preparation with one bottle of magnesium citrate the morning prior (and repeated additional times in the afternoon, if necessary) is well tolerated and obviates the need for enemas. For those that require preoperative antibiotic preparation (previous pelvic radiotherapy, possible urinary diversion with colon), a standard modified Nichols bowel preparation (31) may be used that is initiated the day prior to surgery: 1 g of neomycin orally at 10.00 a.m., 11.00 a.m., 12.00 p.m., 1.00 p.m., 4.00 p.m., 8.00 p.m., and 12.00 a.m., and 1 g of erythromycin base orally at 12.00 p.m., 4.00 p.m., 8.00 p.m., and 12.00 a.m. Patients are then admitted the morning of surgery.

Attention to fluid management is important, especially in the typical elderly bladder cancer patient, particularly on postoperative days 3 and 4 when mobilization of third-space fluid is highest, subsequently necessitating judicious use of diuretics. In addition, intravenous broad-spectrum antibiotics are administered en route to the operating room, providing adequate tissue and circulating levels at the time of incision. Perioperative antibiotics are continued for at least 24 hours.

The use of preoperative anticoagulation has been advocated by some. The possible benefit of anticoagulant use must be weighed against the risk of bleeding which may not be insignificant. We advocate the use of lower extremity intermittent compression devices prior to induction of anesthesia. Preoperative evaluation and counseling by an enterostomal therapy nurse is an absolutely critical component to the success of all patients undergoing cystectomy and urinary diversion. Patients determined to be appropriate candidates for orthotopic reconstruction are instructed as to the possibility that catheterization per urethra may be necessary postoperatively. All patients are site marked for a cutaneous stoma, instructed in the care of a cutaneous diversion (continent or incontinent form), and instructed in proper catheterization techniques should medical, technical, or oncologic factors preclude orthotopic reconstruction. The ideal cutaneous stoma site is determined only after the patient is examined in the supine, sitting, and standing position. Proper stoma site selection is important to patient acceptance should a cutaneous form of diversion be necessary. Incontinent stoma sites are best located higher on the abdominal wall, while stoma sites for continent diversions can be positioned lower on the abdomen (hidden below the belt line) since they do not require an external collecting device. The use of the umbilicus as the site for catheterization may be employed with excellent functional and cosmetic results (32).

The patient should be positioned in the hyperextended supine position with the top of the superior iliac crest located at the break of the operating table (Fig. 23A.1). The legs are slightly abducted to provide access to the rectum if digital examination during surgery is necessary. In the female patient several options for positioning may be used and include the modified frog-legged (Fig. 23A.2), lithotomy position or with the legs placed on spreader bars. All will allow access to the vagina. Care should be taken to ensure that all pressure points are well padded. The abdomen should then be placed parallel to the floor by placing the table in Reverse Trendelenburg position. This will help to keep the bowel in the epigastrium and out of the pelvis.
A nasogastric or orogastric tube is placed, and the patient is prepared from nipples to midthigh. In the female patient, the vagina is fully prepped. After the patient is draped, a catheter is placed in the bladder and drainage is left to gravity.

A midline incision is made extending from the pubic bone in a cephalad direction. For many patients, the incision may be limited to below or at the umbilicus. This incision allows for proper exposure when performing the extended lymphadenectomy. If extended above the umbilicus, it should be carried lateral on the contralateral side to that used for the cutaneous stoma. If the umbilicus is to be used as the site for a catheterizable stoma, the incision should be carried 2 to 3 cm lateral to the umbilicus. It is important to maintain this lateral location of the incision through the anterior rectus fascia as well to provide an adequate rim of fascia for closure of the abdomen once the procedure is completed. Once the peritoneum and posterior fascia are incised, the urachal remnant (median umbilical ligament) is identified, circumscribed, and removed en bloc with the cystectomy specimen (Fig. 23A.3). Early identification of the urachal remnant prevents early entry into a highriding bladder, and ensures complete removal of all perivesical tissues. In those patients who had previously undergone a cystotomy or a partial cystectomy, the cystotomy tract and prior full thickness of the incision scar should be circumscribed and excised with the specimen.

A careful, systematic intra-abdominal exploration is performed to determine the extent of disease and to evaluate for any hepatic metastases or gross retroperitoneal lymphadenopathy. The abdominal viscera should be inspected visually and manually to detect any unrelated disease. If no contraindication exists at this time, all adhesions should be incised and freed.

The bowel is mobilized by incising the peritoneal attachments of the right colon. The cecum and the ascending colon are reflected medially by incising the avascular/white line of Toldt. This incision is then carried up the root of the small bowel mesentery to allow full mobilization of the small bowl off its retroperitoneal attachments until the retroperitoneal portion of the duodenum is exposed. This mobilization will facilitate the urethra-neobladder anastomosis if orthotopic diversion is performed and importantly will provide excellent visualization of the cephalad extent of the extended lymph node dissection. The maneuvers of bowel mobilization are carried out along an inverted triangle whose base is formed by the third and the fourth portions of the duodenum, the right edge along lateral to the ascending colon, and left edge along the medial portion of the sigmoid and the descending colonic mesentery (Fig. 23A.4). This mobilization is critical in setting up the operative field for the LN dissection and the radical cystectomy.

The descending colon and the sigmoid mesentery are then mobilized to the region of the lower pole of the left kidney by incising the peritoneum attachments laterally along the avascular/white line of Toldt. The sigmoid mesentery is then elevated off the sacrum promontory and allows for exposure of the presacral region, iliac vessels, and distal aorta up to the origin of the inferior mesenteric artery (Fig. 23A.5). A wide mesenteric window is then opened within the sigmoid mesentery below
the inferior mesenteric vessels. This provides an angulation free-path for the left ureter to pass to the right side of the abdomen for a tension-free ureteroenteric anastomosis.

Following mobilization of the bowel, the bowel contents are carefully packed into the epigastrium. The descending and the sigmoid colon are not packed with the other intestinal segments as they require mobility to properly complete the PLND.

Successful packing of the intestinal contents is a critical step in the procedure as it allows for an unobstructed view of the operative field. Packing begins by sweeping the right colon and small bowel under the surgeon’s left hand along the right sidewall gutter. A moist open laparotomy pad is then swept with the right hand along the palm of the left hand, under the viscera along the retroperitoneum and sidewall gutter. In a similar fashion, the left sidewall gutter is packed ensuring not to incorporate the descending or the sigmoid colon. The central portion of the small bowel is packed with a third laparotomy pad. A moist rolled towel is then positioned horizontally below the laparotomy pads, but cephalad to the bifurcation of the aorta. Occasionally, prior to placement of the first moist pad, a mobile greater omental apron can be used to facilitate packing of the intestinal viscera in a similar fashion to the laparotomy pad. After the bowel has been packed, a wide retractor blade is placed with gentle traction on the previous packing to provide cephalad exposure. Care should be taken to avoid compression of the great vessels by the end of this retractor blade so as not to increase venous pressure within the pelvis and retroperitoneum.

The ureters are most easily identified in the retroperitoneum above their crossing over the common iliac vessels. The ureters are dissected into the deep pelvis (several centimeters beyond the iliac vessels) and divided between hemoclips. A segment of the distal ureter can be sent for frozen section analysis at this point to ensure the absence of carcinoma in situ or overt tumor based on surgeon preference. The role of frozen section analysis of the distal ureteral margin has recently been questioned (33,34,35,36,37,38); however, some continue to advocate the use of frozen section analysis of the distal ureteral margins to ensure a tumor-free ureteroenteric anastamosis (39,40,41).

The ureter is then minimally mobilized cephalad only enough to allow access to the common iliac vessels and lower aorta and vena cava. It is then packed to prevent inadvertent injury during the extended PLND. Frequently, an arterial branch originating from either the common iliac artery or the lower aorta will need to be secured to properly mobilize the ureter. Care must be taken to maintain the lateral blood supply to the ureter emanating from the gonadal vessels. The gonadal vessels provide an important collateral blood supply to the lower ureter that will be used in the ureteroenteric anastomosis. This is particularly important in patients that have received prior pelvic irradiated. In women, the infundibulopelvic ligaments are ligated and divided at the level of the common iliac vessels.

We advocate a meticulous, extended lymph node dissection when performing a radical cystectomy. The extent of the PLND may vary depending on patient factors and surgeon preference. When performing a salvage procedure following definitive radiation treatment (>5,000 rads), care should be taken as the lymphadenectomy may be difficult and associated with a significant risk of iliac vessel and obturator nerve injury (42).

The extended lymphadenectomy is initiated at the level of the inferior mesenteric artery. All lymphatic tissues are clipped proximally at this level out to the genitofemoral nerves, which represent the lateral limits of dissection. All lymphatic tissues are split and rolled off the lower aorta and vena cava. The common iliac vessels are then skeletonized, sweeping all tissues distally toward its bifurcation. At this point the distal limits of the dissection are delineated. The parietal peritoneum anterior to the external iliac vessels and medial to the gonadal vessels is divided, and in the process the vas deferens in the male (round ligament in the female) is clipped and divided. This provides exposure to the pelvic nodes (Fig. 23A.6). The medial aspect of the distal external iliac vein is identified. The lymphatics medial to the vein are traced to Cooper’s ligament where the node of Cloquet is identified, clipped, and divided. A row of clips is placed medial to lateral along the external iliac vein and artery at the level of the crossing of the circumflex iliac vein. Care is taken to avoid injury to the genitofemoral nerve fibers lateral to the artery. The lymphatics overlying the external iliac artery and vein are split and rolled to completely free both vessels circumferentially. Psoas branches from both vessels can be found on their lateral aspects near their midpoint, where they should be secured and divided. A small gauze sponge is then passed lateral to the external iliac artery and vein along the psoas muscle and gently pushed distally into the obturator fossa (Fig. 23A.7). This allows for complete clearance of all tissues in the groove between the external iliac artery and pelvic sidewall. The vein is then retracted anteriorly and the obturator nerve freed from all lymphatics (Fig. 23A.8). This allows for isolation of the obturator vessels that typically are secured and divided as they exit the obturator foramen (Fig. 23A.9). Occasionally, the obturator vessels lie in a more lateral location and may be preserved. The remaining lymphatics overlying the hypogastric vessels are then swept medially to completely free the hypogastric branches leading to the anterior pelvic organs. The remaining nodal region that requires clearance is the presacral chain, which can be freed from below the bifurcation of the common iliac vessels and swept off the sacral promontory. Removal of the presacral nodal packet and presciatic nodes can be completed after removal of the bladder specimen. Separate submissions for each nodal region are recommended to optimize the number of nodes reported.

After completion of the lymphadenectomy, the lateral vascular pedicles to the bladder are isolated and divided. Developing the correct plane for isolation of the lateral pedicles is a critical maneuver in performing a safe cystectomy. A finger is passed medial to the hypogastric artery, and posterior to the anterior visceral branches. The index finger is positioned parallel to the sweep of the sacrum. This maneuver defines the two major vascular pedicles that provide supply to the anterior pelvic organs: the lateral pedicle lies anterior to the index finger and the posterior vascular pedicle lies behind the index finger and is composed of the visceral branches located between the bladder and the rectum.

The lateral pedicle is entrapped between the left index and the middle fingers, and firm traction is applied vertically and caudally. This maneuver will facilitate the identification of the individual anterior vascular branches originating from the anterior portion of the hypogastric artery (Fig. 23A.10). The posterior branches of the hypogastric artery including the superior gluteal, iliolumbar, and lateral sacral arteries are preserved to avoid gluteal claudication. If necessary, the hypogastric artery may be ligated distal to the takeoff of the posterior divisions; however, this is only required for very large pelvic masses that otherwise obscure the view of the anterior hypogastric branches. The initial anterior branch to the bladder, the superior vesical artery, is identified and defines the cephalad border of the pedicle. Once isolated, the lateral pedicle is then divided between hemoclips or with a roticulating vascular stapler down to the endopelvic fascia, or as far as is technically possible. The inferior vesicle vein serves as an important landmark as it is located just cephalad to the endopelvic fascia. Incision of the endopelvic fascia just lateral to the prostate should be performed at this point to help identify the distal limit of the lateral pedicle.

Following division of the lateral pedicles, the bladder specimen is retracted anteriorly exposing the rectovesical cul-de-sac (pouch of Douglas). The surgeon elevates the bladder anteriorly while the assistant retracts on the peritoneum of the rectosigmoid colon in a posterior and cephalad direction. This will provide exposure to the deep regions of the cul-de-sac, and places the peritoneal reflection on traction, which will facilitate its division. The peritoneum lateral and then anteriorly to the rectum is incised across the cul-de-sac to join the incision on the contralateral side (Fig. 23A.11). The anterior and posterior peritoneal reflections converge to form Denonvilliers’ fascia within the cul-de-sac. Denovilliers provides an anatomic separation in the male as between the prostate and the seminal vesicles anteriorly from the rectum posteriorly. The prostate and seminal vesicles are separated from the rectum by incising through the posterior sheath of Denonvilliers’ (Denonvilliers’ space) allowing one to develop the plane between Denovilliers and the rectum. This separation should develop easily with blunt and sharp dissection. Incision of the peritoneal reflection in the cul-de-sac slightly on the rectal side rather than the bladder side will facilitate entering the proper plane of dissection (Fig. 23A.12). A posterior sweeping motion of the fingers with pressure maintained on the
back of the prostate and seminal vesicles will allow the rectum to be swept off the seminal vesicles, prostate, and bladder in men, and off the posterior vaginal wall in women. Sharp dissection may be required to dissect the anterior rectal fibers off the prostate and/or bladder and should be employed at any point during this dissection to prevent rectal injury.

Several situations may impede the proper development of the posterior plane, most commonly, when the incision in the cul-de-sac is made too far anteriorly and proper entry into Denonvilliers’ space is prevented. Additionally, posterior tumor infiltration, prior prostatectomy, prior low anterior resection of the colon, or previous high-dose pelvic irradiation can obliterate the plane making the posterior dissection difficult. To prevent injury to the rectum in these situations, only sharp dissection should be performed under direct vision to reduce the potential for rectal injury. If a rectotomy occurs, a two-layer or three-layer closure is recommended. A diverting proximal colostomy is not routinely required unless gross contamination occurs or if the patient has received previous pelvic radiation therapy. If orthotopic or vaginal reconstruction is planned, an omental interposition is recommended to prevent fistulization between suture lines.

Use of a roticulating vascular stapler may also be used to facilitate division of the posterior pedicles (Fig. 23A.13). The endopelvic fascia is incised lateral to the prostatic apex, medial to the levator ani muscles. Opening the endopelvic fascia facilitates the apical dissection of prostate and urethra. In the female patient, the posterior pedicles including the cardinal ligaments are divided 4 to 5 cm beyond the cervix. A sponge stick placed into the vagina will help identify the vaginal apex so that the posterior vaginal wall may be properly incised just distal to the cervix. The vagina is then circumscribed anteriorly with the cervix attached to the cystectomy specimen (Fig. 23A.14). If there is concern about an adequate surgical margin at the posterior or base of the bladder, then the anterior, and if necessary, lateral vaginal wall should be removed en bloc with the bladder specimen with subsequent vaginal reconstruction if sexual
function is desired. It is our preference to spare the anterior vaginal wall if orthotopic diversion is planned. This eliminates the need for vaginal reconstruction, which helps maintain the complex musculofascial support system and helps prevent injury to the pudendal innervation to the rhabdosphincter and proximal urethra, both important components to the continence mechanism in women. The anterior vaginal wall is then sharply dissected off the posterior bladder down to the region of the bladder neck (vesicourethral junction) which is identified by palpating the Foley catheter balloon (Fig. 23A.15). At this point, the specimen remains attached only at the apex in men and vesicourethral junction in women.

Once the cystectomy specimen is completely free and mobile posteriorly, attention may be directed to the anterior dissection. The opened endopelvic fascia lateral to the prostate is extended both anteriorly and posteriorly. The levator muscles are carefully swept off the lateral and apical portions of the prostate. Care should be taken in avoiding any extensive dissection in the region anterior to the urethra along the pelvic floor. Incision of the puboprostatic ligaments is limited to the portion that is directly attached to the apex of the prostate and should not be extended to the portion connected to the proximal urethra.

The dorsal venous complex is divided just distal to the prostatic apex and controlled with suture that reapposes the cut edges of the complex only. Excessive incorporation of levator muscle in this suture may inhibit proper function of the rhabdosphincter. The urethra is then incised 270° just beyond the apex of the prostate, and a series of 2-0 Monocryl sutures are placed in the anterior urethra, carefully incorporating only the mucosa and the submucosa of the striated urethral sphincter muscle anteriorly. Posteriorly, the rectourethralis muscle or the caudal extent of Denonvilliers’ fascia may be included in the urethral suture. After the posterior urethra is divided, the specimen is removed and a frozen section analysis of the distal urethral margin of the cystectomy specimen is performed to exclude tumor involvement.

If a cutaneous form of urinary diversion is planned, the urethra is stretched above the urogenital diaphragm, and divided at the level of the pelvic floor to excise all visible pelvic portions of the urethra.

When properly performed, a standard radical cystoprostatectomy should render men impotent in virtually all cases. Several modifications in the technique of radical cystectomy may allow preservation of the neurovascular bundles and hence spontaneous erectile function. Based on the neuroanatomic studies of Walsh and Donker (43), the relationship of the pelvic nerve plexus to the seminal vesicles and bladder was defined. In 1987, Schlegel and Walsh (44) modified the standard radical cystoprostatectomy to preserve the autonomic innervations arising from the pelvic plexus. Using the seminal vesicles as an intraoperative landmark, the pelvic plexus, which is located lateral and posterior to the seminal vesicles, may be spared during the ligation of the posterior pedicle.

Patient selection is absolutely critical when considering a nerve-sparing radical cystoprostatectomy (45). Patients who are impotent preoperatively and those in whom tumor involvement of the posterior pedicle of the bladder is suspected at the time of cystectomy are not candidates for a nerve-sparing approach. When performed correctly and applied to appropriately selected patients, oncologic efficacy does not appear to be compromised and erections adequate for intercourse postoperatively are achieved in approximately 45% of patients (46).

If anterior vaginal wall preservation is planned, the posterior vagina is incised at the apex just distal to the cervix (Fig. 23A.14). This incision is carried anteriorly along the lateral and anterior vaginal wall forming a circumferential incision. The anterior-lateral vaginal wall is then grasped with curved Kocher clamps to provide countertraction that facilitates the dissection between the anterior vaginal wall and the posterior bladder (Fig. 23A.15). Development of the posterior plane and vascular pedicle is best performed sharply and carried just distal to the vesicourethral junction which can be identified by palpation of the Foley catheter balloon (Fig. 23A.16).

The anterior vaginal wall should be removed en bloc with the cystectomy specimen when posterior invasive tumors are present to ensure that a complete resection is achieved. If the anterior wall is excised a 1-cm length of proximal urethra is mobilized while the remaining distal urethra is left intact with the anterior vaginal wall. Vaginal reconstruction may be accomplished by folding the posterior wall over or by closing the lateral walls side to side. Formal techniques for complete reconstruction of the vagina have been extensively described and can be fashioned from myocutaneous flaps (gracilus or rectus muscles), tabularized or detubularized bowel segments, or a peritoneal or an omental flap.

If an orthotopic reconstruction is planned in a female, it is critical that no dissection be performed anterior to the urethra along the pelvic floor. The endopelvic fascia should remain undisturbed to prevent injury to the innervation of the rhabdosphincter, which is critical in maintaining continence. Branches of the pudendal nerve course along the pelvic floor posterior to the levator muscles in women and supply the rhabdoid sphinteric complex (47,48,49). It is important to avoid
any dissection anterior to the urethra as this may injure the pudendal nerve branches and adversely affect continence.

Several anatomic factors in the female facilitate urethral suture placement including a more anteriorly positioned urethra and a wider pelvic inlet (Fig. 23A.17). An adequate number of sutures, usually totaling 8 to 12, should be used within the urethra depending on its size. Frozen section analysis is performed on the distal urethral margin of the cystectomy specimen to exclude tumor. Once the vaginal cuff is closed it is then anchored via a colposacralpexy using a strut of nonabsorbable material (Marlex, Prolene, or Goretex mesh) to the sacral promontory. Fixation of the vagina is performed with care to avoid angulation or undue tension. The peritoneal edge of the sigmoid mesentery can be used to cover the mesh anteriorly. A well-vascularized omental pedicle graft based on the left gastroepiploic arcade is placed between the reconstructed vagina and the neobladder prior to completion of the urethralneobladder anastomosis to separate the suture lines and prevent fistulization (Fig. 23A.18).

If a cutaneous diversion is planned in the female patient, the entire urethra is resected from either an abdominal or a vaginal approach.

The pathologic stage of the primary bladder tumor and the presence of lymph node metastases are perhaps the most important survival determinants in patients undergoing cystectomy for bladder cancer (17). Histopathologic assessment allows for risk stratification and provides important prognostic information that may direct subsequent treatment. The rate of organ-confined disease varies from 17% to 57% (Table 23A.1) (16,17,18,63,64,65). Organ-confined lesions include pTis, pTa, pT1, and pT2 lesions. Overall, the recurrence-free survival in this pathologic subgroup of organ-confined, lymph node-negative bladder tumors is as high as 85% at 5 years and 82% at 10 years (17). Noninvasive lesions, as expected, have a better recurrence-free survival compared to invasive disease (pT1-2) (26,66,67,68,69). There is evidence to suggest delays in treatment result in more advanced disease and poorer overall survival in patients with muscle-invasive bladder cancer (70,71,72,73). Furthermore, caution should be exercised when delaying definitive surgical therapy for patients with high-risk non-muscle-invasive tumors that do not readily respond to intravesical treatments (2,4).

For those with muscle-invasive node-negative disease, the prognostic significance of pathologic substaging based on the depth of invasion into the muscle layer remains controversial, with some reporting no difference in clinical outcomes between superficial (pT2a) and deep (pT2b) invasion into the muscularis propria (74,75,76) and others showing some prognostic stratification (77).

Non-organ-confined (extravesical), lymph node-negative tumors are reported in approximately 18% to 65% of patients undergoing cystectomy (Table 23A.1) (16,17,18,63,64,65). These include pT3a, pT3b, pT4a, and pT4b N0 lesions. There does not appear to be any difference in survival between those with microscopic and gross extravesical extension (pT3a vs. pT3b) (78,79,80). In an analysis of 129 patients with pathologic T3 disease managed with cystectomy alone, the 10-year recurrence-free survival was 54%. Lymph node metastasis was observed in 43 (33%) of the 129 patients and was associated with a lower 10-year recurrence-free survival rate (32%) than was lymph node-negative status in patients with extravesical extension (60%). The majority of recurrences occurred at a distant site, highlighting the excellent local control provided by radical cystectomy even in patients with locally extensive extravesical disease (78).

Direct tumor extension into contiguous organs such as the prostate and rectum in men and the vagina in women (pT4) portends a worse prognosis. In a large series, 5- and 10-year recurrence-free survival rates for surgically treated patients with pT4 disease were 50% and 45%, respectively (17). Many of these patients may benefit from neoadjuvant chemotherapy prior to radical cystectomy. Despite the overall poor prognosis for this group, a subset of patients with locally advanced disease
continues to benefit from radical cystectomy, and certain pathologic findings may permit prognostic stratification. Current staging designates pT4a disease as contiguous invasion into the prostate. Other forms of prostate involvement including CIS of the urethra or ducts and lamina propria invasion of the periurethral tissues may have an improved prognosis compared to invasive lesions into the prostatic stroma (81,82). Overall survival at 5 years for patients with carcinoma in situ of the prostatic urethra or ductal tumors is 71%, but only 36% for those with prostatic stromal invasion (83). Similarly, Pagano et al. (84) demonstrated that the 5-year overall survival for patients whose tumors invaded the prostate through the urethra was 46% compared with 7% for those whose tumor extended through the bladder wall to involve the prostate. Additionally, prostatic involvement has been noted to be an independent risk factor for subsequent urethral recurrences (85,86).

Radical cystectomy provides the best local (pelvic) control of the disease. An overall local pelvic recurrence rate of 7% was observed in the USC cystectomy series (17). Patients with organ-confined, lymph node-negative tumors demonstrated only a 6% local recurrence rate, compared to a 13% local recurrence rate in those with non-organ-confined, lymph node-negative tumors. Even those at highest risk of a local recurrence (lymph node-positive disease) had only a 13% local recurrence rate following cystectomy. Nearly all patients suffering a pelvic recurrence following cystectomy will die of their disease despite additional therapeutic efforts.

Anatomic studies performed around the turn of the 19th century, both in Europe and the United States, provided detailed descriptions of the lymphatic drainage of the urinary bladder. The classic descriptions of the lymphatic pathways within the bladder include a rich subepithelial lymphatic plexus that provides drainage through the detrusor musculature to the exterior of the bladder. Efferent lymphatic channels drain into perivesical lymph nodes that lie on the anterior, posterior, and lateral aspects of the bladder and the perivesical tissues. Larger lymphatic channels then pass to nodal basins located along the external iliac vessels, hypogastric vessels, and lateral sacral/sacral promontory. These basins form a ring of pelvic nodes and constitute the primary drainage sites. All primary landing sites subsequently drain into the more proximal common iliac node chain and then on to the lower retroperitoneal nodal groups.

Following the identification of the primary and secondary lymphatic drainage basins of the bladder within the pelvis, it remained to be determined whether a therapeutic advantage could be gained through their excision at the time of cystectomy. Experience with a similar radical surgical approach for cervical cancer, in which a therapeutic pelvic lymphadenectomy was combined with a total exenterative procedure, not only demonstrated that it could safely be executed but that it also potentially improved outcome (87). Early experience with radical cystectomy and regional lymphadenectomy for patients with nodal metastases suggested that despite a more aggressive surgical approach, most node-positive patients experienced an exceedingly poor long-term outcome. In 1956, Whitmore and Marshall (88) published their experience of 100 consecutive bladder cancer patients treated with radical cystectomy and a regional pelvic lymphadenectomy. In 32 patients with node-positive disease, only 22% were reported alive at 1 year. Three patients (9%), the majority of whom developed metastatic disease, survived 2 years and only one (3%) was alive at 3 years. While these data represented an overall improvement in outcome in node-positive patients, the poor rate of survival among patients with regionally advanced disease led many to question whether radical cystectomy was indicated in the presence of positive regional nodes.

Supported by reports of success with the Wertheim technique for cervical cancer, Leadbetter and Cooper (10) proposed inclusion of a thorough PLND that included the hypogastric, external iliac, presacral, and common iliac lymphatics at cystectomy. Subsequently Skinner (8) championed the benefits of a thorough lymph node dissection at the time of cystectomy and presented evidence supporting its therapeutic benefits. Contemporary experience with patients whose bladder cancer has involved the regional pelvic lymph nodes establishes a significantly more favorable outcome following radical cystectomy and pelvic lymphadenectomy compared with early published experiences. Current reports note that approximately 25% to 33% of all patients with invasive bladder cancer involving the regional lymph nodes will be rendered disease free following radical cystectomy and a thorough PLND (16,17,63,64,89,90). Stein et al. (91) reported the 5- and 10-year recurrence-free
survival to be 35% and 34%, respectively, for a series of 246 patients with node-positive bladder cancer following radical cystectomy and an extended pelvic lymphadenectomy that included all nodes from the aortic bifurcation distally to the inguinal ligament. The Memorial Sloan-Kettering experience with 193 contemporary node-positive bladder cancer patients that received radical cystectomy and a more limited PLND (proximal limit at the bifurcation of the common iliac vessels) reported a 31% disease-specific survival and 25% overall survival at 5 years (90). Ghoneim et al. found 188 node-positive patients out of 1,026 (18%) patients with bladder cancer. A 23.4% 5-year survival was reported in the 188 node-positive patients who received a lymphadenectomy at cystectomy that included the nodes up to the distal common iliac vessels (16). Mills et al. (89) reported a 29% overall survival in 83 node-positive patients who underwent cystectomy and a pelvic lymphadenectomy that included nodes from the bifurcation of the common iliac distally. Other groups have similarly documented that overall approximately one quarter to one third of all patients with regionally metastatic bladder cancer can be expected to survive 5 years following radical cystectomy and pelvic lymphadenectomy (63).

Subgroups of patients with regionally metastatic disease may actually demonstrate a significantly better prognosis and can be identified by having a lower volume of disease both at the level of the primary tumor and the regional lymph nodes. The survival for node-positive patients whose primary tumors are confined to the bladder is significantly better than that observed in node-positive patients with more invasive primary tumors. Vieweg et al. (64) reported a 51% 5-year overall survival in 44 P0-P3a, LN+ patients compared to a 17% 5-year survival in 149 patients with P3b-P4, N+ disease. The USC group reported a 46% versus 30% 5-year recurrence-free survival in node-positive patients with organ-confined versus non-organ-confined primary tumors, respectively (91). A similar stratification of outcome can be observed when considering the volume of disease at the level of the regional lymph nodes. If the number of positive nodes or size of the involved node(s) is considered as a measure of the extent of tumor involvement of the regional nodes, patients with fewer involved nodes or smaller tumor-bearing nodes have been found to have an improved outcome. Vieweg et al. reported the 5-year disease-specific survival of 193 node-positive patients that underwent cystectomy based on TNM node staging (1987 system) as 44%, 27%, and 0% for N1, N2, and N3 patients, respectively. The median survival for these three groups was 3.1, 1.9, and 0.9 years, respectively (p = 0.0006). The number of involved lymph nodes has been reported as an independent prognostic indicator following radical cystectomy (90). Lerner et al. (92) noted that patients with fewer than six involved lymph nodes exhibited a significantly improved 5-year survival compared to patients with greater than six positive nodes. An update of this series found that eight or fewer nodes involved was an optimized cutoff in that patients with eight or fewer involved nodes (n = 193) demonstrated a 41% 5-year recurrence-free survival and 37% overall 5-year survival compared with 10% and 4%, respectively, in patients with greater than eight involved nodes (n = 51, p < 0.001) (91). Other series have confirmed similar differences in outcome, as noted by Mills et al. (89) in which patients with fewer than five involved lymph nodes did significantly better compared with those with more involved nodes (5-year survival of approximately 50% vs. 10%).

Additional measurements of the burden of disease, such as the size of the involved nodes and the presence of extracapsular lymph node extension by tumor within the regional lymph nodes, may also prove to be prognostically important. Mills et al. reported that patients with involved nodes >0.5 cm or the presence of extracapsular extension within the involved nodes demonstrated a lower survival. Multivariate analysis of the relative importance of the number of positive nodes, size of involved nodes, or the presence of extracapsular lymph node perforation demonstrated that only the presence of extracapsular perforation remained independently predictive of outcome, with a hazard ratio of 2.6 (89).

While the number of lymph nodes involved with disease contains important prognostic information, simultaneous consideration of the extent of the lymphadenectomy performed appears to enhance the prognostic significance of such data. For example, consider two patients, both with four positive lymph nodes identified after cystectomy. Patient one had a total of 10 lymph nodes evaluated pathologically, while patient two had 40 lymph nodes analyzed. Should both patients be considered similarly staged? Would both patients have a similar anticipated outcome? Data from both University of Southern California and Memorial Sloan-Kettering Cancer Center (MSKCC) would suggest that evaluation of the number of involved nodes in the context of the total number of lymph nodes removed provides a more accurate means to identify higher risk node-positive patients (91,93). Using information obtained from the ratio of positive lymph nodes to the total number of lymph nodes evaluated (density of positive lymph nodes), both institutions have independently demonstrated an improved stratification of node-positive patients into differing risk groups. Using 20% as a cutoff for the percentage of involved lymph nodes, a 44% versus 17% 5-year recurrence-free survival was observed for patients with less than compared with >20% of their total nodes involved with disease, respectively. An even greater difference in outcome was reported using a ratio-based analysis in a series of 162 node-positive patients who had undergone cystectomy and lymphadenectomy. The 5-year disease-specific survival of patients with <20% of total nodes involved was approximately 65% compared with 5% for patients with >20% of evaluated lymph nodes involved with tumor. Positive-node ratio provided improved prognostic information compared to 1997 TNM nodal staging criteria or total number of involved lymph nodes.

Given the importance in staging information provided by the lymphadenectomy, the question of the extent of the dissection required for adequate staging or therapeutic value remained to be clarified. An established set of necessary anatomic boundaries and extent (number of lymph nodes) of the lymph node dissection that would provide optimal staging and therapeutic efficacy for bladder cancer is presently not available. A lack of prospectively validated studies has led to ongoing controversy regarding the necessary extent of dissection (94).

Wishnow et al. (95) attempted to determine the rate of involvement of the pelvic lymph nodes within the common iliac chain or more distally in the obturator, hypogastric, and external iliac nodes in bladder cancer patients undergoing radical cystectomy (95). In a series of 130 patients with grossly negative lymph nodes at the time of cystectomy, 88% of whom had common iliac nodes resected, 14% of patients were identified with microscopically involved nodes. Seventeen of the eighteen patients with microscopically involved lymph nodes had one or two positive nodes. None of the 17 patients with one or two microscopically involved nodes had common iliac or lateral external iliac lymph nodes involved. Based on these findings, the authors recommended limiting the proximal limit of the lymph node dissection to the bifurcation of the common iliac vessels for patients with no evidence of grossly positive lymph nodes.

More recently at MSKCC, a series of 144 bladder cancer patients undergoing radical cystectomy were prospectively evaluated to determine the site of regional lymph node involvement. Eighteen patients (14%) were found to have
microscopically involved regional lymph nodes including four with disease involving the common iliac nodes. In this series all but one patient with common iliac nodes had simultaneous involvement of a more distal lymph node region (hypogastric, obturator, or external iliac) suggesting that excision of the common iliac nodes would benefit a subset of microscopically LN-positive patients.

Additional information on the distribution of positive pelvic nodes is provided by a multicenter, prospective trial in which all patients underwent an extended PLND (proximal limits of dissection at or above the bifurcation of the aorta) (96). Of the 290 patients evaluated in this series, 81 (27.9%) demonstrated evidence of tumor involvement in 599 pelvic nodes. Involved nodes above the bifurcation of the common iliac vessels comprised 35% of all positive nodes. A total of 20 patients (6.9%) demonstrated involvement of the common iliac nodes without evidence of disease within the more distal nodal regions (obturator, hypogastric, or external iliac). While exact information on the nature of the nodes (gross or microscopically enlarged) was not available, in the 29 patients with only a single lymph node metastasis, 10% were located above the bifurcation of the common iliac vessels providing further strong support for the need to extend the dissection to minimally include the common iliac chain.

Data describing the relationship between the extent of the node dissection and outcome are provided by single-institution, retrospective reviews. Poulsen and Krarup (97) reported a comparative analysis of two consecutive series of bladder cancer patients undergoing either a limited (proximal limit at the bifurcation of the common iliac vessels) or an extended (including the nodes up to the level of the bifurcation of the aorta) lymphadenectomy performed at the time of radical cystectomy. Previously untreated bladder cancer patients were included in the study in which 126 underwent an extended (between 1993 and 1997) and 68 received a limited (between 1990 and 1993) LN dissection. The two groups were well matched demographically, with a slightly higher percentage of extravesical tumors in the extended lymphadenectomy group. As anticipated, the extended node dissection yielded a greater number of lymph nodes compared to the more limited excision (25 vs. 14, p < 0.001). Both groups had a similar proportion of node-positive patients, 27% in the extended and 24% in the limited LN groups. Despite the increased number of more advanced tumors in the extended dissection group, a similar 5-year recurrence-free survival, risk of local recurrence, and risk of distant metastasis was observed. In the subgroup of patients with organ-confined primary tumors, however, patients who underwent an extended dissection benefited with an improved 5-year recurrence-free survival (85% vs. 64%, p < 0.02).

Leissner et al. presented their analysis of 302 patients who received an extended lymph node dissection at the time of radical cystectomy for bladder cancer. They noted that patients with a greater number of lymph nodes reported in their pathology report had an improved disease-free and overall survival as well as improved local tumor control. At 5 years, 51% of patients with ≤15 lymph nodes removed were alive and disease free compared to 65% for those with ≥16 lymph nodes evaluated. The improvement in local control was also significant, with pelvic recurrences identified in 27% compared to 17% of patients with ≤15 and ≥16 lymph nodes evaluated, respectively (p < 0.01) (19).

Herr et al. reported a similar improved outcome in 322 patients, in which those that underwent a more extensive node dissection, as represented by a greater number of lymph nodes reported, exhibited an improved overall survival. For patients with node-positive disease, approximately 50% versus 20% of patients with greater or less than 11 lymph nodes removed were alive at 5 years, respectively (98). Additional supportive data obtained from a national cohort of patients are provided by a multivariate analysis of Surveillance, Epidemiology, and End Results (SEER) registry data on 1,923 radical cystectomy patients. This retrospective study found that the number of lymph nodes examined was positively associated with an improved survival, particularly in patients with higher stage disease. Patients with at least four lymph nodes evaluated demonstrated an improved outcome. However, patients with 10 to 14 nodes reported exhibited the greatest improvement in overall survival (99).

While suggestive of a therapeutic role, these data are retrospective in nature and do not include factors that may bias surgical selection in their analysis. Koppie et al. performed a sensitivity analysis that sought to take into consideration important factors that may have both biased selection for a particular extent of dissection and effected outcome (i.e., survival). Their cohort included 1,121 radical cystectomy patients in which age and comorbidity was considered when evaluating the potential therapeutic importance of the PLND. Their data demonstrated that older sicker patients were more likely to undergo a less aggressive PLND compared to younger, healthier patients. However despite correcting for age, comorbidity, tumor stage and other factors, they still observed an improved survival in patients undergoing a more extensive PLND (100).

While no prospective randomized data are currently available comparing a limited to more extended PLND, comparative data from different institutions performing either limited or more extensive dissections do exist. Dhar et al. reported a 13-year comparative study evaluating disease-specific outcomes in 336 patients who underwent radical cystectomy with a limited PLND to 322 patients that underwent a more extensive PLND at RC. Patients who underwent a more extensive PLND were more likely to be LN positive (improved staging) and demonstrated improved recurrence-free survival. Five-year disease-free survival for lymph node-positive patients undergoing an extended PLND was 35% versus only 7% in patients that received a limited PLND (101).

The ability to limit the dissection to the ipsilateral side of the pelvis in patients with tumors located on one side of the bladder has also been proposed. It is apparent, however, that despite a clear laterality in the location of the primary bladder tumor, contralateral nodal involvement will frequently be identified. Data from Leissner et al. (96) found that of 32 node-positive patients with primary tumors located specifically to one side of the bladder, the risk of contralateral lymph node involvement was only slightly less than that found for the ipsilateral nodes. Sentinal node studies have also confirmed that the initial node region involved with disease may be located in the contralateral side of the primary tumor (102).