(Table 2).[20,21,23]
Similarly, the use of combination chemotherapy regimens as neoadjuvant treatment initially appeared attractive, but most later follow-up studies did not indicate any apparent long-term benefit.[24] By contrast, the Nordic Cooperative Bladder Cancer Study Group has reported that two cycles of neoadjuvant cisplatin and doxorubicin confer a reduced death rate in patients with T3 and T4 bladder cancer who are treated by cystectomy, albeit with a relatively weak statistical power.[25]
A recent meta-analysis[26] of all known randomized trials reviewed 479 cases and compared local treatment vs neoadjuvant chemotherapy followed by local treatment. The use of neoadjuvant chemotherapy was associated with an overall hazard ratio of 1.02 (favoring local treatment) and a 2% increase in relative risk of death. However, this study was dominated by single-agent trials. Similar findings were provided by the International Intergroup (MRC/EORTC) Trial[22] in which nearly 1,000 patients were randomly allocated to receive either neoadjuvant cisplatin, methotrexate, and vinblastine (CMV) plus local treatment or local treatment alone.
The first report of this study failed to reveal a significant survival difference between the two arms.
Although an important United States intergroup trial has not yet been reported, little current evidence supports the routine use of neoadjuvant chemotherapy. It is our belief that such treatment programs should be conducted only in the context of a randomized clinical trial.

Adjuvant Chemotherapy
The beneficial use of adjuvant systemic chemotherapy, in which cytotoxics are delivered after the control of local tumor (by complete resection or by radical radiotherapy), has its paradigms in the management of breast cancer and colonic cancer. In the context of bladder cancer, several early nonrandomized phase II trials have been reported but without useful conclusions. More recently, Skinner and colleagues[27] conducted a randomized trial for patients with deeply invasive bladder cancer that was staged at cystectomy. Half of the patients enrolled on this trial received adjuvant treatment with cyclophosphamide, doxorubicin, and cisplatin. Although the interpretation of this study has been controversial, survival benefit achieved by this strategy is small, as the survival curves had already crossed at four years. The results of a randomized trial by Stockle et al[28] may hold more interest for future strategic planning. In this study, patients with stages pT3b or pT4 bladder cancer (defined at cystectomy and lymph node dissection) were randomly allocated to receive adjuvant M-VAC or equivalent chemotherapy vs observation. A disease-free survival benefit was reported (73% vs 19%, respectively). However, the statistical power of the study was diminished by limited patient enrollment due to early closure when it was recognized that the trial design required patients on the observation arm to forego salvage chemotherapy at the time of relapse. Thus, this study assessed only the impact of chemotherapy at some time after cystectomy and was not a true test of adjuvant treatment.
Of greater relevance to this hypothesis is a randomized trial[29] in which cystectomy plus adjuvant CMV was compared with cystectomy alone (but with CMV offered at the time of relapse). While this study also was limited by small patient numbers, it demonstrated a statistically significant increase in disease-free survival and a trend toward improved overall survival from the use of adjuvant CMV chemotherapy. However, caution is indicated by the report of a nonrandomized experience involving 56 patients with adjuvant CMV chemotherapy for stages pT2 through pT4 and node-positive disease.[30] The median disease-free survival was 15.5 months, but the three-year disease-free survival probability was only 28%. Again, a definitive statement regarding the role of adjuvant chemotherapy requires the results of a statistically valid randomized trial, although the current weight of evidence suggests a survival benefit from adjuvant M-VAC or CMV chemotherapy.

Perioperative Chemotherapy
Another variant of the combination of systemic chemotherapy with definitive locoregional treatment is the use of cytotoxic agents administered before and after the local therapy. The first major trial[20] of such an approach predominantly assessed the impact of neoadjuvant methotrexate but also included a component of adjuvant therapy after completion of definitive treatment. No survival gain was noted when this approach was compared with standard treatment.
In a pilot study conducted by the Eastern Cooperative Oncology Group,[31] two cycles of M-VAC were administered before cystectomy, followed by two additional cycles. Seventeen patients had T3 disease and one had a stage T2 tumor, and nearly half the cases showed downstaging in response to M-VAC. However, at a median follow-up of 23 months, 50% had died. Logothetis and colleagues[32] tested a similar hypothesis in a randomized trial in which 100 patients were randomized to receive either two cycles of M-VAC followed by cystectomy and three adjuvant cycles of M-VAC or initial cystectomy followed by five cycles of adjuvant M-VAC. Survival was equivalent in both arms, despite the significant level of downstaging after neoadjuvant chemotherapy. Both trials revealed an unexpectedly high rate of intercurrent deaths from vascular complications, which further demonstrates the importance of randomized trials in the assessment of these novel strategies of management.

Concurrent Chemoradiation
Doxorubicin, fluorouracil, mitomycin C, carboplatin, and cisplatin exhibit the characteristic of radiosensitization, ie, altering the responsiveness of both tumor and normal tissues to the cytotoxic effects of irradiation. Predicated on this feature, clinical trials have been initiated in which these cytotoxic agents are given during the period of radiotherapy to increase the level of tumor kill, with the hope that a sufficient discriminant in toxicity between normal and malignant tissues will protect the patient from excessive side effects.

Several phase II trials have assessed the impact of single agents (fluorouracil, carboplatin, doxorubicin, and cisplatin) and combination regimens (fluorouracil and cisplatin) in radiosensitizing protocols[33-36] and have demonstrated significant tumor reduction (Table 3). In most of these studies, chemotherapy has been combined with radiotherapy as the definitive local treatment, although one study assessed the use of doxorubicin-induced radiosensitization as adjuvant therapy after surgery. To date, it is unclear if these approaches confer a survival benefit, notwithstanding their efficacy in achieving local tumor control.[40] In a report of a randomized trial conducted by the National Cancer Institute of Canada, Coppin et al[37] showed that a statistically significant increase in local tumor control could be achieved by the concurrent use of cisplatin and radiotherapy (compared with radiotherapy alone), but a survival benefit from the combination regimen was not seen. No other randomized trials have assessed the potential survival benefit of this approach. Given the changes in tumor classification, staging, available cytotoxics, and support therapy, it is our view that at least one convincing randomized trial is required to establish the role of chemoradiation as a standard of therapy. However, this view is not universally held. For example, Shipley has expressed the view that the Canadian NCIC study has established the role of chemoradiation in achieving local control and that this should be incorporated into other randomized trials to assess innovations in neoadjuvant or adjuvant systemic therapy (personal communication, 1996).

Conclusions
After more than a decade of investigation on the combination of systemic chemotherapy and definitive local treatment, it appears that this regimen may be beneficial in the management of locally advanced bladder cancer. However, the optimal schedules and the extent of true survival benefit have not yet been demonstrated. Current data indicate that the most promising future strategy will be the incorporation of adjuvant systemic chemotherapy in some way into treatment programs. However, the design and implementation of well-structured, randomized clinical trials are needed to resolve these issues. In this era of managed care and contained medical costs, it may be difficult to conduct these studies unless the academic leaders of the medical community take a stand on these issues and secure cooperation from community practitioners to enroll eligible patients in these trials.

Bladder Replacement and Urinary Diversion After Radical Cystectomy

Julio M. Pow-Sang, MD, Evangelos Spyropoulos, MD, PhD, Mohammed Helal, MD, and Jorge Lockhart, MD

Advancements in bladder replacment construction and continent urinary diversion have reduced treatment mobidity for patients facing cystectomy.

Background: The optimal mode of urinary tract reconstruction following cystectomy continues to challenge the urologic surgeon. Disadvantages with bowel conduits have prompted the search for better techniques to improve patient outcomes.
Methods: The development of urinary tract reconstruction is reviewed, and results from several forms of continent urinary diversion and bladder replacement construction are presented. The authors report on their experience in creating continent reservoirs or neobladders in over 400 patients.
Results: Several surgical approaches are now available for continent urinary diversion. Metabolic and nutritional abnormalities, stone formation, infection, and cancer formation are potential complications.
Conclusions: Advances in surgical techniques, an understanding of the physiology of isolated bowel segments, and improvements in pre- and post-operative care have altered the field of urinary reconstruction after cystectomy for bladder cancer. Most patients can expect minimal morbidity and mortality.

Introduction
Determining the optimal mode of urinary tract reconstruction following cystectomy is a challenge for the urologic surgeon. Until recently, bowel conduits were considered the gold standard and represented the most popular form of urinary tract restoration.[1-5] However, the negative body image associated with an external ostomy appliance, as well as the risk of renal damage, led to the development of continent urinary diversion and bladder replacement reconstruction to improve outcomes for patients who undergo cystectomy.[6-13] A better understanding of isolated bowel segment physiology, improvements in surgical technique, and the acceptance of intermittent catheterization have promoted the widespread popularity of these forms of reconstruction.[7] Cancer eradication, preservation of renal function, and optimal quality of life are the ultimate goals of surgery for bladder cancer.[14]

Historical Review
The quest for an ideal technique for urinary tract reconstruction following cystectomy dates back to 1852 when Simon[15] first reported diversion of urine to a segment of bowel by creating fistulas between the ureters and the rectum in a patient with bladder exstrophy. Initially, efforts were aimed at either bringing the ureters to the skin or diverting the urine to the sigmoid colon to benefit from continence provided by the anal sphincter.[16,17] Prior to the 1950s, the use of the anal sphincter for continence established ureterosigmoidostomy as the urinary diversion of choice. During this era, techniques of nonrefluxing ureteral anastomoses were improved.[18-22] However, the risk of long-term complications with ureterosigmoidostomy was significant (hydronephrosis: 32%; pyelonephritis: 57%; metabolic derangements: 47%).[23] In 1950, Bricker[2] popularized the use of the ileum as a urinary conduit, which constituted the gold standard for patients who underwent urinary diversion until the 1980s. The need for improvements in the quality of life of patients led to the era of continent urinary diversion and bladder replacement. By applying the concepts of a cutaneous catheterizable ileocecal reservoir developed in 1950, several investigators reported encouraging initial results with colonic reservoirs in the mid 1980s,[7,8,12] and Kock et al[6] concurrently developed a catheterizable ileal pouch. Camey and LeDuc[24] reintroduced the concept of the neobladder in 1979, and other investigators improved the technique by applying the experiences of the early continent urinary diversion.[11]

Selection of Type of Urinary Diversion
The goal of surgery in the management of infiltrating bladder cancer is either curative or palliative. If the intent is palliative, then the simplest and most expeditious type of urinary diversion is best. If the goal is curative, then the patient is apprised of reconstruction options for the urinary tract and undergoes preoperative evaluation. Although the psychologic impact of surgery and diversion is significant, any type of urinary reconstruction should be acceptable with good preoperative assessment and education. Enterostomal therapists and urology nurses play pivotal roles in improving patients' coping abilities, both preoperatively and postoperatively.[25,26]
Factors that affect the choice of urinary diversion include patient age, manual dexterity, body habitus, physical and mental status, renal function, prognosis of the primary disease, existing bowel pathology, prior radiation or chemotherapy, the presence of urethral disease, the expectations, preferences, and fears of the patient, the experience and preference of the surgeon, and cost.[27] Since there is no unanimous choice for the best method of urinary diversion, all options should be considered.[28]
Indications for an external collecting device diversion (bowel conduit) are either absolute or relative. Absolute indications include impaired renal function, impaired physical ability to perform self-catheterization, and inability to understand the significance and possible complications of a continent diversion. Relative indications include advanced age, need for postoperative chemotherapy, previous pelvic irradiation, bowel disease (Crohn's disease, colitis, cancer), body habitus, diseased urethra, and impaired functional status.[29,30] Patient choice also is a key factor in selection.

Options in Continent Urinary Diversion

Continent Cutaneous Urinary Reservoir
In 1982, Kock et al[6] described a technique for construction of an internal ileal reservoir that consists of a 80-cm segment of terminal ileum isolated on its mesentery at approximately 50 cm proximal to the ileocecal valve. Proximal and distal 17-cm sections are used to construct the afferent and efferent limbs to the pouch, and two medial 23-cm segments are detubularized, approximated, and remodeled to form the reservoir. Afferent and efferent continent nipple valves are then created by intussuscepting sections of bowel 5- to 6-cm in length with strips of Marlex or polyglycolic acid mesh around the bases of the intussusceptions. The ureters are anastomosed to the afferent limb using a mucosa-to-mucosa anastomosis, the pouch is closed, and the efferent limb is brought through the abdominal wall and fixed to the rectus fascia using a Dexon collar to form a stoma through which urine can pass.
In 1992, Fisch et al[31] described a form of continent urinary diversion termed the Mainz pouch, which utilized cecum and ileum. To create the reservoir, 10 to 15 cm of cecum and ascending colon, as well as terminal ileal segments of equal length, are isolated and detubularized. The posterior wall of the pouch is completed by anastomosis of the ascending colon with the ileal loop, starting at the inferior aspect. The latter is then anastomosed with the next proximal ileal segment. The ureters are implanted in an antirefluxing manner in an open-end technique through a submucosal tunnel of 4 cm to 5 cm in length. To create the continence mechanism, an additional 8 to 12 cm of ileum is isolated to form an ileal intussuscepted valve by invaginating and fixing 6 cm of this latter segment with metal staples. Alternatively, continence can be achieved by submucosal embedding of the appendix.
In 1985, Rowland et al[32] described the cecoileal continent urinary reservoir, in which approximately 8 to 10 cm of terminal ileum and 25 to 30 cm of cecum and ascending colon are isolated.[33] The colonic segment is detubularized either by incising along its antimesenteric surface with scissors or cautery or by placing a 60- to 75-mm gastrointestinal anastomosis (GIA) stapler between the two more lateral tenia. The continence mechanism is then created by tapering the efferent limb (terminal ileum) over a 12F red rubber catheter resting against the antimesenteric surface of the ileum. A 60-mm GIA metal staple is applied to excise the redundant antimesenteric portion of the ileum and to create a smooth tube for catheterization using 16F to 18F catheters. The ureters are tunneled into the tenia of the colonic segment through an inverted "T" incision. A mucosal incision is then made for the orifice, the ureter is cut either obliquely or spatulated, and a ureter-to-mucosa anastomosis is performed over a 5F to 8F stent using interrupted 5-0 absorbable, synthetic, monofilament sutures. The cephalad end of the pouch is folded to the caudal end, and the reservoir is closed with a single layer of running 3-0 braided synthetic absorbable suture.
In 1986, Light et al[34] described Le Bag, in which 20 cm of cecum and ascending colon are isolated with a corresponding length of terminal ileum. Following detubularization, the free ileal and colonic borders are sutured together, and the pouch is folded as described in the Kock procedure. The ureters are reimplanted on the colonic portion of the pouch according to the preference of the surgeon. After tapering and reinforcing the ileocecal valve, the ileal tail is brought through the abdominal wall as the continent segment.
In 1986, we described a different form of continent urinary diversion utilizing an extended colon segment.[35] Creation of the reservoir begins with wide mobilization of the right colon and terminal ileum. The mid-transverse colon and distal ileum are transected using automatic staplers. The last 10 cm to 15 cm of ileum is preserved, depending on the abdominal wall thickness. Using standard stapling techniques, an ileocolonic anastomosis is performed to restore bowel continuity. The colonic segment is turned into itself in the form of a "U" and is detubularized, either by opening the bowel along its antimesenteric border or by using an absorbable automated surgical stapler. The medial edge is then closed with running 3-0 absorbable suture, and the ureters are brought through the posterior wall of the colon where they are anastomosed directly, mucosa to mucosa. To create the continence mechanism, the redundant antimesenteric portion of the distal ileum is excised with surgical staplers. The reservoir is then closed with a running, locking 3-0 absorbable suture (Figs 1-5).[36]

Orthotopic Bladder
Camey and LeDuc,[24] Hautmann et al,[11] and Studer and Turner[37] described the creation of a bladder from different bowel segments as an alternative for handling continuity of the urinary tract after cystectomy.

Complications Relating to Techniques
Complications from earlier techniques affect 2% of patients with continent urinary diversion and 4.5% of patients with neoplasms. Complications include infection, wound dehiscence, urinary fistulas, prolonged ileus (longer than seven days), small bowel obstruction, respiratory distress (atelectasis, pneumonia, pulmonary embolus), myocardial infarction, deep venous thrombosis, and bleeding.

Metabolic and Nutritional Effects
Possible metabolic and nutritional consequences associated with small and large intestinal segments for continent diversion of the urinary tract include disturbances of electrolyte metabolism, abnormal drug metabolism, calculus formation, altered hepatic metabolism, nutritional disturbances, osteomalacia, impaired sensorium, growth retardation, infection, and cancer development.[38,39]

Electrolyte Abnormalities
Hyperchloremic metabolic acidosis develops as a result of sodium secretion (in exchange for hydrogen) and bicarbonate (in exchange of chloride), as well as reabsorption of ammonia, ammonium, hydrogen ions, and chloride when these segments are exposed to urine. The mechanism that appears to be most responsible for hyperchloremic metabolic acidosis is excess absorption of chloride and ammonia, which maintains a chronic endogenous acid load.[40] Since chloride seems to be more readily absorbed from colonic than from ileal reservoirs and since electrolytic derangements predominate when longer colonic segments are used for reservoir construction, the use of an ileal segment may be preferable in patients with impaired renal function.[41]
Hypokalemia and total body depletion of potassium may occur in patients with urinary intestinal diversion. Potassium depletion is probably the result of renal potassium wasting as a consequence of renal damage, osmotic diuresis, and gut loss through intestinal secretion.[40]
Hypocalcemia is a consequence of depleted body calcium stores and excessive renal wasting.[40] The chronic acidosis is buffered by carbonate from the bone with subsequent release of calcium into the circulation, which is then cleared by the kidney and results in a gradual decrease in body calcium stores. An impairment of renal tubule calcium reabsorption also occurs. Normal bone mineral metabolism requires the interaction of calcium, magnesium, and phosphate, which are influenced by parathormone, calcitonin, and vitamin D. Osteomalacia in adults and rickets in children -- essentially the same condition -- are characterized by chronic loss of bone buffers and calcium and lead to hypercalciuria and bone demineralization. Mineral losses are eventually replaced by osteoid with a resultant decrease in bone strength. Alterations in bone mineral content occur in most patients who have had a urinary intestinal diversion for extended periods of time.[42]

Calculus Formation
The incidence of renal stone formation increases in patients with intestinal urinary reconstruction. The increases range between 16.7% and 26.5% with the Kock pouch, 5.4% with the Indiana pouch, and 9.8% with the Mainz pouch.[43,44] In our series, at a mean follow-up of 6.3 years, we found a 15% incidence of stone formation.[45] With a shorter follow-up, the incidence of urinary calculi in neobladders ranges between 2.1% and 2.7% (hemi-Kock neobladder and Hautmann ileal neobladder, respectively).[46] Generally, the stones are comprised of struvite, calcium oxalate, calcium phosphate, or uric acid, and mixtures of these minerals often are present in the same stone. Most stones reported to be infectious are comprised of struvite and/or carbonate apatite and are related to foreign materials and infection. A small but significant portion of stones are metabolic and consist of calcium phosphate and/or calcium oxalate secondary to hyperchloremic metabolic acidosis.[47] Common risk factors for urolithiasis are chronic colonization of the reservoir with bacteria secondary to urine alkalinity,[45] renal infection with urease-producing bacteria, the presence of foreign materials (eg, sutures, metallic staples, nonabsorbable collars) in the reservoir, retained intestinal mucous, and increased urinary excretion of phosphate, sulfate, and magnesium, and hypocitraturia.[47]

Nutritional Disturbances
The liver synthesizes and conjugates bile salts that are necessary for proper fat digestion and for the uptake of vitamins A and D. After fat stimulates their release into the duodenum, bile salts are actively reabsorbed by the distal ileum and returned to the liver by the enterohepatic circulation to be used again. After ileal resection, length-dependent alterations in bile metabolism can lead to a multitude of intestinal events that may result in diarrhea. Even though considerable amounts of bile salts are lost in the colon, the liver can synthesize and maintain the salt pool after resection of up to 100 cm of ileum. If ileal resection is greater than 100 cm, hepatic bile salt synthesis cannot match the losses. In this case, micelle formation in the jejunum decreases, and fat malabsorption leads to steatorrhea (fecal fat of more than 20 g per day) and diarrhea. Hydroxylated fatty acids directly decrease colonic absorptive capacity, cause active secretion of electrolytes and water, and form soaps, which are cathartic.[48]
Vitamin B12 is excreted exclusively into the bile. It is highly conserved by active uptake at the terminal ileum and is returned to the liver by the enterohepatic circulation. Body stores of vitamin B12 may last three to six years in complete malabsorption and six to 30 years in partial malabsorption.[48] Loss of the distal ileum can impair vitamin B12 absorption. A loss of 50 cm of terminal ileum appears to be the critical margin for sufficient vitamin B12 absorption. Substitution of vitamin B12 should be prescribed to patients who lose more than 50 cm of terminal ileum beginning several years after surgery.
Following removal of the ileocecal valve, the absorptive processes in some patients may be affected due to the development of high concentrations of bacteria in the ileum. Severe diarrhea may occur as a result of fat malabsorption or irritation of unreabsorbed bile salts on the colonic mucosa.[38] Diarrhea also may occur when major portions of the large bowel are removed. In this case, a significant amount bicarbonate can be found in the fecal fluid, since alkaline ileal contents drain into a shortened large bowel segment, which may result in acidosis and dehydration.[38]

Infection
Approximately 80% of patients with continent intestinal diversion are bacteriuric with diverse bacterial flora. In the first year of reconstruction, the incidence of septic episodes varies from 5% to 20%. The frequency of bacteriuria, pyelonephritis, and sepsis is higher in patients with continent intestinal diversion than in those with an intact bladder that is subjected to daily instrumentation by intermittent catheterization.[40]

Carcinogenesis
The incidence of malignancy in intestinal segments used for urinary reconstruction is currently unknown. If cancer develops, the most common site is the ureterointestinal anastomosis. The most common types of tumor are adenocarcinoma (85%) and transitional cell carcinoma (10%), with the remaining 5% consisting of signet ring cell carcinoma, adenomatous polyps, sarcoma, and undifferentiated carcinoma.[49] A possible mechanism is an increase in exposure to carcinogens such as N-nitroso compounds, which are highly mutagenic and induce tumors in many animal species. Nitrate is normally excreted by the kidney into the urine, and many species of Gram-negative bacteria (Escherichia coli, Proteus, Klebsiella, Pseudomonas) can reduce nitrate and catalyze the conversion of nitrite and secondary amines present in the urine into N-nitroso compounds. Fecal bacteria are presumably responsible for the formation of these substances, although the admixture of urine and feces is not considered an absolute requirement for this production. Long-term surveillance is mandatory for patients who have undergone urinary reconstruction with intestinal segments.

Complications Related to the Reconstructed System

Obstruction
Ureterointestinal anastomosis obstruction is a serious complication, and surgical intervention is usually required to preserve the upper urinary tract. Common factors predisposing to anastomotic structure formation are inadequate ureteral length, poor vascular supply, poor surgical technique with ureteral twisting, and possibly an increased angulation with chronic reservoir distension.[45] The mean incidence for this complication is 7.5% with continent reservoirs; with neobladders, the incidence is higher.[46] When the ureters are reimplanted, the incidence of obstruction is even higher (28%). Ureterointestinal anastomosis obstruction may be managed either by balloon dilatation and stenting or by an open surgical procedure through a transreservoir approach.
The incidence of acute pyelonephritis ranges up to 5.8% with continent diversions and up to 8.0% with neobladders. In most cases, its onset is related to obstruction of the ureterointestinal anastomosis.[45]

Reflux
The estimated incidence of intestinoureteral reflux is 2.6% with continent reservoirs and 0.4% with neobladders.[45,46] Despite the controversy regarding the optimal type of ureterointestinal reimplantation (tunneled vs nontunneled), the incidence of reflux is low regardless of which reimplantation technique is used.

Reservoir Complications
Hypertonicity of the bowel reservoir with associated episodes of urine leakage has been noted in 5.6% of pouches and in 4.2% of neobladders.[46] Whether the bowel is detubularized or left in its original tubular form, bowel motility resumes in some segments across anastomotic lines. Pressure spikes may be noticed in both detubularized and tubularized segments of bowel.[40]
Spontaneous perforation of the urinary reservoir is a rare complication. The incidence with continent reservoirs is 4.8%, and no cases have been reported with neobladders.

Efferent Limb Complications
Dysfunction of the continence segment occurs in 6% of patients with continent reservoirs, and dysfunction of the intestinourethral anastomosis with neobladders occurs in 2.75% of patients.[46] Dysfunction of the continence segment (ileocecal valve) may be due to intrinsic factors (eg, a dysfunctional plicated bowel limb) or extrinsic factors (eg, a parastomal hernia).[50] Multiple abdominal wall scars, weight gain, and a chronic increase in intra-abdominal pressure due to constipation or chronic obstructive pulmonary disease may favor hernia development.[45] Difficulty with emptying the reservoir is encountered in 7% of patients with continent cutaneous reservoirs and in 12% of those with neobladders.[46] In the former, the difficulty may be related to a long and tortuous efferent limb, the creation of a false passage, or the development of a stricture along the efferent limb. For patients with neobladders, the main causes of difficulty are intestinourethral strictures (6.26%) and urethral cancer recurrence (3% to 18%).[48,51]
Protrusion of a ventral hernia through the incision line developed in one (1.7%) of our 60 patients. Other series report an incidence rate of ventral hernia that ranges from 4.4% to 14%.[27,30] Meticulous closure of the abdominal wall with appropriate suture materials is the cornerstone in preventing this complication.

Conclusions
Significant advances in surgical techniques, a better understanding of isolated bowel segment physiology, and improvements in preoperative and postoperative care have revolutionized the field of urinary reconstruction after cystectomy for bladder cancer. The majority of patients who undergo this procedure can expect minimal morbidity and mortality and an enhanced quality of life. The stride still continues to refine the surgical techniques for urinary tract reconstruction

Hormonal and Chemotherapeutic Systemic Therapy for Metastatic Prostate Cancer

Ravat Panvichian, MD, and Kenneth J. Pienta, MD

Promising new approaches in the treatment of prostate cancer include gene therapy, tumor-biology-based therapies, and the development of new agents and combination chemotherapy.

Background: Prostate cancer is the most frequently diagnosed cancer and the second leading cause of cancer death in men in the United States. It is estimated that over 300,000 men will have been diagnosed with prostate cancer in 1996, and more than 40,000 will have died of this disease.
Methods: The authors combined their experience with a review of the literature on management of this disease to examine the effectiveness of treatments for both localized and metastatic prostate cancer.
Results: Surgery and radiation therapy are potentially curative modalities for cancer still limited to the gland. Androgen ablation therapy results in stabilization or regression of metastatic disease in most instances but is not curative. Some new approaches are described for patients with hormone-refractory prostate cancer.
Conclusions: Newer tumor-biology-based combinations are promising in the treatment of hormone-refractory prostate cancer, but their effect on patient survival needs to be evaluated in larger clinical trials.

Introduction
Prostate cancer is the most frequently diagnosed cancer and the second leading cause of cancer death in men in the United States. In 1996, prostate cancer will have been diagnosed in approximately 317,000 men and will have caused more than 41,000 deaths.[1] While radiation therapy and surgery are potentially curative treatment modalities for cancer that is still limited to the gland, treatment of metastatic disease remains palliative. In those symptomatic patients with newly diagnosed metastatic prostate cancer, androgen deprivation is the mainstay of treatment. Androgen ablation therapy results in stabilization or regression of the disease in approximately 80% of patients but often fails to prevent progressive disease. Men with progressive prostate cancer in the presence of total androgen blockade are defined as having hormone-refractory prostate cancer (HRPC). In the past, no effective "standard" chemotherapy was available for patients with HRPC, which has a median survival of six to nine months.[2] However, recent developments in basic and clinical research have shown promise in improving the morbidity and mortality rates of patients with HRPC.

Androgen Dependence
The prostate gland is sensitive to a variety of hormones, but androgens are the key components in prostate cellular proliferation and growth throughout a man's life. The importance of androgens in prostate cancer became apparent in the seminal works of investigators in the 1940s[3-5] when they established the concept of androgen dependence of prostate cancer and demonstrated that surgical castration (orchiectomy) or medical castration (estrogen therapy) produced a reduction in cancer mass and a clinical remission in 80% of patients with advanced metastatic disease. As a consequence, the disruption of the hypothalamic-pituitary-gonadal axis by surgical or medical castration has been the mainstay of therapy for metastatic prostate cancer (stage D1 or D2). In 1960, the five-year survival rates in patients so treated were 20% compared with 0% in placebo or untreated patients.[6]

Role of Androgens in Prostate Biology
Androgens are derived from the testis and the adrenal cortex. Testicular androgens are released by the Leydig cells after stimulation by luteinizing hormone (LH), which in turn is controlled by a pulsatile release of the hypothalamic gonadotropin-releasing hormone (GnRH=LHRH). Testosterone is converted into dihydrotestosterone (DHT) in the peripheral tissues and in the prostate by the activity of the 5-alpha-reductase enzymes. Adrenal androgens are released by the adrenal cortex after stimulation by adrenocorticotropin. These adrenal androgen, mainly composed of dehydroepiandrosterone, its sulfate, and androstenedione, undergo a multistep conversion to testosterone and DHT in the peripheral tissues and in the prostate itself.[7] The binding of androgens, mainly DHT, to nuclear androgen receptors allows the receptor to dimerize and then associate with androgen-responsive elements on androgen-regulated genes. This modulates the transcription of specific genes and the regulation of particular biologic responses, eg, production of growth factors and programmed cell death.

Primary Hormonal Treatment Options
Four major methods of androgen deprivation can be used in the palliative treatment of metastatic prostate cancer: (1) surgical castration by orchiectomy to remove the primary androgen-producing organs, (2) medical castration by estrogen therapy or LHRH therapy to reduce LH production, (3) antiandrogen therapy directed primarily at the target organs (ie, prostate and metastatic sites), and (4) combined androgen blockade.

Bilateral Orchiectomy
Bilateral orchiectomy, the standard by which other forms of hormonal therapy are measured, removes approximately 90% of circulating testosterone and decreases bone pain almost immediately. Thus, surgical castration may be an appropriate choice for patients with aggravated symptoms and impending complications of paralysis by metastatic disease. Complications are minimal, and the side effects are associated with androgen withdrawal (ie, loss of libido, impotence, and hot flashes). Removal of the testis is a major psychologic aspect associated with orchiectomy.

Estrogen and LHRH Therapy
In the treatment of prostate cancer, estrogens exert their effect primarily by a negative feedback at the hypothalamic-pituitary level, which results in reduced LH secretion and testicular testosterone synthesis (Fig 1). Castrate levels of testosterone (less than 50 ng/dL) in patients receiving an oral dose of 3 mg of diethylstilbestrol (DES) daily were achieved in a range of 21 to 60 days.[8] Generally, this dosage of DES is necessary to obtain adequate castrate levels.[9] A dose level of 1 mg daily of DES will lower testosterone to castrate levels in 80% of patients, but its effectiveness varies among individual patients. Doses of less than 1 mg daily of DES do not result in an appreciable decrease in testosterone levels. A regimen containing a dosage level of 5 mg of DES has been associated with thromboembolic toxicity and thus is no longer administered.[10] Adequate dosage of oral estrogen treatment is as effective as orchiectomy or an LHRH analog in the treatment of metastatic prostate cancer; however, changes in the blood (eg, increases in platelet aggregation, low-density lipoprotein, and certain clotting factors) that may result in thromboembolic sequelae have tempered enthusiasm for the use of estrogens.

Acute administration of the LHRH analogs stimulates the secretion of LH in the pituitary gland. Conversely, chronic administration suppresses LH secretion by the down-regulation of the receptors in the pituitary, leading to a decrease of plasma testosterone and resulting in a complete and reversible medical castration. The different LHRH formulations share the same properties, and the depot preparations (eg, leuprolide, goserelin acetate) appear to be equally efficacious in controlling prostate cancer when compared with bilateral orchiectomy or additive estrogen therapy. A transient worsening of signs and symptoms (tumor flare) during the first week of therapy is a side effect of LHRH analog treatment. This effect is the result of the transitory LH and testosterone surge. The peak increase in serum testosterone occurs within 72 hours, and achievement of castrate levels occurs within four weeks from the initial dose. Antiandrogen administration is recommended before or simultaneously with the first LHRH analog depot injection to prevent tumor flare and its clinical complications in patients with overwhelming metastatic disease.

Antiandrogen Therapy
An antiandrogen is any compound that blocks the interaction between androgens and their receptors in the presence of normal or even increased target tissue levels of DHT. Currently, the primary role of antiandrogens is in combination treatment. Antiandrogens consist of two types: nonsteroidal agents and steroidal antiandrogens with progestational activity. The two progestational antiandrogens -- cyproterone acetate and megestrol acetate -- block androgen receptor function and inhibit the release of LH by their progestational action. As monotherapy, neither compound alone suppresses androgen production completely, and a rise toward normal in plasma testosterone concentration occurs after several months. In general, progestational antiandrogens are not used as initial hormonal monotherapy for patients with metastatic disease.
Nonsteroidal antiandrogens in combination therapy currently receive more attention, particularly in their role in combination therapy to achieve maximal androgen withdrawal. The current pure antiandrogens are flutamide, nilutamide, and bicalutamide, which are all structurally related. They act directly on the prostate cancer cells by competitively blocking the binding of DHT to the nuclear androgen receptor.

Combined Androgen Blockade

While plasma testosterone dramatically decreased by 90% or more following surgical or medical castration, prostate DHT, the major stimulus for prostate epithelial growth, remained at approximately 25% of precastration levels due to the intraprostatic synthesis from the adrenal androgens (Fig 2).[11] Maximal or combined androgen blockade (CAB) is achieved by the central inhibition of androgen production through medical or surgical castration combined with peripheral blockade of circulating androgens by the use of an antiandrogen (eg, flutamide or casodex). A study in support of CAB was conducted by the European Organization for Research of the Treatment of Cancer (EORTC 30853)[12] in which the combination of goserelin and flutamide was compared to orchiectomy. Another study from the NCI and the Southwest Oncology Group (SWOG)[13] compared leuprolide with leuprolide plus flutamide. In both studies, patients receiving the CAB showed a survival advantage of approximately seven months. However, several other studies have not shown a measurable difference in outcome between monotherapy and CAB, and a recent meta-analysis of all randomized clinical trials of monotherapy vs CAB reported little difference between the two groups.[14] Given the conflicting data, the benefit of CAB remains unclear. Ongoing studies should clarify this issue.

Early or Delayed Endocrine Therapy
Determining the most effective timing of endocrine treatment of prostate cancer remains controversial. Endocrine therapy is the mainstay of treatment for all symptomatic patients with metastatic prostate cancer, while asymptomatic patients should be evaluated according to their general health and concomitant diseases. A review of data from the Veterans Administration Cooperative Urological Research Group (VACURG) using a covariate analysis suggests that some patients may benefit from early treatment.[15] A survival benefit was seen in younger patients with more aggressive stage D prostate cancers (Gleason score 7-10) when hormonal treatment began at the time of diagnosis, but a follow-up period of at least three years was needed to show differences between the early-treatment group and the deferred-treatment group.

Intermittent or Continual Therapy
In experimental studies with the androgen-dependent Shionogi mouse mammary carcinoma, Akakura et al[16] showed that postcastration progression of tumors to an androgen-independent state was linked to the cessation of androgen-induced differentiation of stem cells resulting from deprivation of androgen. The androgen-induced differentiation of stem cells with recovery of apoptotic potential is the basis for the concept of intermittent androgen-deprivation therapy of androgen-dependent tumors. Several cooperative groups are investigating this theory to determine if temporary interruption of androgen deprivation can delay or prevent the development of androgen independence.

Prostate-Specific Antigen Level as a Predictor of Treatment Outcome
Prostate-specific antigen (PSA) is a 34 kDa glycoprotein found in prostatic tissue and seminal plasma. Serum levels of PSA correlate well with extent of disease. Serum PSA has been used to evaluate response to treatment in both hormone-dependent and hormone-refractory disease. PSA can rapidly assess tumor response because of its short half-life of two to four days. After the initiation of hormonal therapy, PSA levels decrease over a period of three to four months. The PSA nadir, representing quiescent disease, lasts approximately 18 to 24 months in the average patient. The PSA level is often the earliest sign of treatment failure or relapse in patients with metastatic prostate cancer who undergo primary androgen ablation therapy. The rise in PSA portends clinical progression by approximately six months. Decreases in baseline PSA with treatment have been shown to correlate with improved prognosis in both hormone-dependent disease and hormone-refractory disease.[17,18] In trials of HRPC, a posttherapy decrease in PSA of 50% or more that was documented on multiple determinations and maintained over time (more than six months) was the most significant prognostic factor in predicting prolonged survival.[18] The majority of men with metastatic prostate cancer who are treated with androgen ablation respond initially, thus demonstrating that at least a proportion of their cancer cells are androgen responsive. However, most of these patients eventually relapse to a state that is unresponsive to further antiandrogen treatment, regardless of the aggressiveness of their secondary antiandrogen manipulations. Androgen ablation therapy fails to cure the disease because the prostate cancer in each patient is heterogenously composed of clones of both androgen-dependent and independent prostate cancer cells at the time of first presentation.
A significant proportion of primary and metastatic prostate adenocarcinomas contains a subpopulation of neuroendocrine cells.[19] These prostatic neuroendocrine cells produce a variety of biogenic amines and neuropeptides that can impact on tumor growth and metastatic behavior in a paracrine or autocrine manner.[19,20] Prostatic neuroendocrine cells may contribute to the androgen-independent progression of prostate cancer by clonal expansion or through the paracrine stimulation of adjacent adenocarcinoma. The development of future therapies that use neuroendocrine pathways for therapeutic benefit is promising.

Management of HRPC

Continued Androgen Suppression
In the past, most patients with metastatic prostate cancer were treated with orchiectomy either as first-line treatment or at the time of failure of primary medical androgen suppression. Thus, these patients who enrolled in trials for HRPC were androgen-deprived. With the availability of medical forms of reversible androgen blockade involving LHRH analogs and antiandrogens, the role of continued androgen suppression has become important. A recent retrospective review[21] from the Southwest Oncology Group (SWOG) concluded that continued androgen suppression was not a significant factor in patient survival. Taylor and colleagues,[22] however, retrospectively analyzed 341 patients from four clinical trials and demonstrated a modest survival advantage for patients on continued androgen suppression. Currently, SWOG policy recommends that patients continue on androgen suppression during chemotherapy trials.

Flutamide or Antiandrogen Withdrawal
Flutamide withdrawal syndrome refers to a significant decline in PSA levels following withdrawal of antiandrogen therapy due to evidence of disease progression. This phenomenon was documented in 10 (29%) of 35 patients in whom disease had progressed following combined androgen blockade.[23] The duration of the PSA decline seen with discontinuation of flutamide was short (median = five months), but the decline was associated with symptomatic improvement. The response to flutamide withdrawal may be explained by the presence of functionally altered androgen receptors that recognize flutamide as an androgen agonist or by the unmasking of the agonistic property of flutamide. A similar phenomenon has been seen in case reports of casodex. When disease progression occurs following antiandrogen therapy, observation is recommended to determine the effects of flutamide withdrawal before evaluating subsequent interventions.[23]

Therapies Based on Tumor Biology
Currently, no single or combination cytotoxic chemotherapy regimen for HRPC has consistently shown objective tumor regression or prolonged patient survival. However, new therapies for HRPC are being developed that are based on new approaches to cancer therapy rather than the traditional chemotherapeutic targets such as DNA and RNA. These approaches rely on an understanding of dynamic structure and function of cancer cells (ie, cell shape, movement, and signaling). The dynamic structure of the cell is composed of a tissue matrix system that interacts to organize and process spatial and temporal information to coordinate genetic information and cell function.[24] Virtually every part of the tissue matrix system is altered in cancer cell therapy, thus providing the potential for the development of cancer-specific targets. Four chemotherapeutic regimens (suramin, estramustine plus vinblastine, estramustine plus etoposide, and estramustine plus paclitaxel) based on inhibiting dynamic cell structure are being studied in clinical trials of HRPC and appear promising based on results of preliminary clinical studies.

Suramin
Suramin, a polysulphonated naphthylurea, is the first of a new class of growth factor antagonists with a 55-day serum half-life(T1\2-Beta). Suramin affects not only interactions between growth factors and their respective receptors, but also several cellular functions. It possesses striking antiangiogenesis activity based on its ability to competitively block the binding of fibroblast growth factors (and other growth factors) to their respective transmembrane receptors.
Suramin was evaluated in prostatic cancer on the basis of its inhibitory effects on growth factor-induced proliferation and antitumor activity against human-derived prostatic cancer cell lines, both in vitro and in vivo. The initial trial[25] of suramin in HRPC demonstrated a 35% objective response rate in patients with measurable soft-tissue disease and a 35% response rate based on a decrease in PSA of 75% or more in patients with bone-only disease. Subsequent reports have confirmed the activity of suramin in HRPC (Table 1).[25-29]

The overall response to suramin in these trials is 55% using PSA declines of 50% or more and 30% when considering only patients with measurable disease. Its principal side-effects are neurotoxicity and fatigue.
Because suramin therapy can cause adrenal insufficiency, hydrocortisone replacement is needed with suramin therapy. The percentage of the observed response rate to suramin that is due to hydrocortisone coadministration or flutamide withdrawal has recently been questioned. In addition to its beneficial effect on pain palliation from bone metastasis, corticosteroids also cause decreases in PSA levels. Hydrocortisone in physiologic replacement doses can result in a decline in PSA levels to 50% or more in 20% to 50% of patients.[30] These issues are being addressed in ongoing prospective, randomized trials, including a phase III trial comparing suramin plus hydrocortisone with placebo plus hydrocortisone.

Estramustine and Vinblastine
Estramustine phosphate consists of an estradiol molecule attached to a nornitrogen mustard through a carbamate ester linkage. Estramustine cytotoxicity is predominantly attributed to its ability to bind microtubule-associated proteins, which are essential to the stability of microtubules.[31] Estramustine causes microtubules to disassemble and prevents their de novo formation, resulting in mitotic arrest during metaphase and the disruption of many vital cellular functions that lead to cell death. Vinblastine is a vinca alkaloid whose cytotoxicity is also attributable to microtubule inhibition, but it acts by binding to the beta subunit of the tubulin monomer, a distinctly separate microtubular target.

Estramustine and vinblastine were combined based on in vitro evidence of additive antimitotic activity and nonadditive toxicity. The major side effect of estramustine is nausea, and the major toxicity of vinblastine is myelosuppression. Three trials were conducted to evaluate the combination of estramustine and vinblastine in the treatment of HRPC (Table 2).[32-34] Based on a decrease in PSA levels of 50% or more, responses in these trials were similar at approximately 50%, and the overall response rate was 30%.

Estramustine and Etoposide
The nuclear matrix, the RNA-protein network of the nucleus, plays an important role in DNA replication and gene expression. The DNA matrix attachment site is part of a "replication complex" that includes the topoisomerase II enzyme. Etoposide is a topoisomerase II inhibitor that selectively inhibits DNA replication at the level of the nuclear matrix.[35] In addition to acting as a microtubule-associated protein inhibitor, estramustine also binds to the nuclear matrix.[35]
Etoposide and estramustine interact in a synergistic fashion at the level of the nuclear matrix in both in vitro and in vivo animal models.[35] A clinical trial based on these data was conducted for patients with HRPC in which 50 mg/m2 of oral etoposide daily was combined with 15 mg/kg of estramustine daily for 21 days with cycles repeated at 28-day intervals.[35] In this phase II clinical trial of 42 patients, 18 had measurable soft-tissue disease. Using standard response criteria, nine (50%) of these patients responded (three complete and six partial). Of 24 patients with bone-only disease, 14 (58%) showed a decrease in PSA levels of more than 50%, and bone scans improved in six (25%) patients. The overall response rate was 15 (36%) of the total 42 patients, and 22 (52%) had a decrease in PSA levels of more than 50% (Table 3). A concurrent trial of oral etoposide alone failed to show significant activity.[36]

Estramustine and Paclitaxel
Paclitaxel inhibits microtubule disassembly and freezes cells in mitosis. Although neither paclitaxel nor estramustine appeared to have significant single-agent activity in HRPC, Speicher et al[37] evaluated their combined activity in human prostatic carcinoma cell lines based on the potential complementary mechanisms of action of estramustine and paclitaxel. These investigators demonstrated in vitro the synergistic cytotoxic effect of the estramustine and paclitaxel combination in human prostate cancer cells. Hudes et al[38] used this information to initiate a phase II trial of estramustine and paclitaxel in 17 patients with HRPC. Ten (58%) of 17 patients demonstrated a decrease of more than 50% in their baseline PSA, and three (50%) of six patients achieved a partial response, suggesting that the combination of estramustine and paclitaxel is an active regimen in the management of HRPC (Table 4).

Conclusions
Metastatic prostate cancer remains an incurable illness with multiple factors influencing patient survival. A clearer understanding of the process by which a tumor develops and recurs following treatment has fostered the development of new approaches in the prevention and treatment of prostate cancer (Fig 3, Table 5). New combination therapies based on tumor biology are promising in the treatment of HRPC, although their effects on patient survival must be evaluated in larger clinical trials. Studies are ongoing to investigate strategies such as chemotherapy and radiation therapy or gene therapy to treat the cancer before it metastasizes.
Many of the current controversial issues associated with management of metastatic prostate cancer should be resolved in the near future. The next decade should provide us with new weapons by which to attack and defeat prostate cancer.

This study was supported by grant #1 P50 CA69568 from the Specialized Program of Research Excellence of the National Cancer Institute, Bethesda, Md.
From the Departments of Internal Medicine (RP) and Surgery (KP) at the University of Michigan Comprehensive Cancer Center, Ann Arbor, Mich

Role of Retroperitoneal Lymph Node Dissection in the Management of Testicular Cancer

Randall G. Rowland, MD, PhD

Retroperitoneal lymph node dissection for management of testicular cancer at various stages is reviewed.

Background: Retroperitoneal lymph node dissection is an important component of staging and management of nonseminomatous germ-cell carcinoma of the testis. Ejaculatory impotence has been a dominant aspect of operative morbidity.
Methods: The author has led the investigation of a series of modifications of operative techniques with the aim of reducing morbidity while retaining the prognostic and therapeutic benefits for retroperitoneal lymph node dissection.
Results: The advances in surgical techniques have reduced the incidence of ejaculatory impotence to less then 5%. Guidelines for the type of retroperitoneal lymph node dissection for different clinical stages of disease are presented.
Conclusions: The advances in surgical techniques for retroperitoneal node dissection have minimized morbidity. The procedure plays a role in many clinical stages of testicular cancer.

Introduction
The lymphatic drainage of the testis was studied by Jamieson and Dobson[1] in 1910. Based on the knowledge of the lymphatic distribution, Cooper et al[2] performed retroperitoneal lymph node dissections (RPLNDs) on patients with testicular cancer in the 1950s, and this technique was then studied and expanded by Donohue.[3] In early experience, the extent and thoroughness of the dissection increased based on observations of occasional involvement of suprahilar left periaortic and interaortocaval nodes in patients with low-volume disease. The suprahilar zones were routinely resected as well as all of the node-bearing tissue between the ureters from the renal vessels to the bifurcation of the common iliac arteries. Donohue described a "split and roll" technique to ensure clearance of the nodal tissue completely around the great vessels. This technique of extended RPLND was used for patients with low-stage clinical disease as primary treatment and for patients who had presented with advanced stage and/or bulky retroperitoneal metastases who did not achieve a complete remission with chemotherapy.
Approximately 60% of patients with low-stage clinical disease who had pathologically positive nodes were rendered disease-free by RPLND alone.[4,5] In patients with pathologically negative nodes, only 5% to 10% relapsed at a later time. Relapses in the retroperitoneum were rare and were detected early by elevation of serum markers or by chest radiographs. The subsequent salvage rate with chemotherapy was virtually 100%.[4]

Evolution of Surgical Technique

By the early to mid 1980s, many investigators were attempting to modify the RPLND technique to decrease or eliminate ejaculatory impotence, a major side effect in this population of young patients. Ejaculatory impotence, which causes infertility, is a result of the resection of the lumbar sympathetic chains and/or the postganglionic branches of the lumbar sympathetic nerves. Fig 1 shows the course of the lumbar sympathetic chains and their postganglionic branches. Initial attempts at sparing the lumbar sympathetic chains involved modifying the template of the dissection based on whether the primary tumor is left- or right-sided.
A study performed at Indiana University Medical Center[6] described the distribution of positive lymph nodes by the size and total number of nodes involved and by the location of the tumor. These data showed that in clinical stage I patients, nodal involvement was highly predictable as to the location of positive nodes. A rationale for altering the templates of dissections was based on this study. Figs 2A and B show the margins of dissection for the modified templates for right- and left-sided tumors.