Upper urinary tract TCC (UTTCC) is an uncommon disease. Because the incidence and mortality of renal pelvic TCC continues to be grouped with RCC specific epidemiological data are not available . Nevertheless, assuming the historical estimate that renal pelvic TCC constitutes ≈ 10% of renal malignancies, 6000–8000 new cases of UTTCC were estimated to occur in the USA in 2003, accounting for ≈ 0.5% of all new cancer diagnoses and ≈ 1.7% of all new genitourinary malignancies [1,2]. An analysis of the National Cancer Institute Surveillance, Epidemiology, and End Results database from 1973 to 1996 showed a slight increase in the rate of ureteric TCC during 1985–96, which was speculated to result from increasing ureteroscopic diagnoses . There was a slight overall improvement in 5-year disease-specific survival in that study, with reduced survival rates among African-Americans and women . Few publications have reported the long-term results of therapy in substantially many patients . Tumour stage and grade are the major prognosticators of recurrence and survival . The multifocal nature of these tumours is increasingly recognized by mapping studies and on careful systematic pathological evaluation and, similar to the bladder, the upper urinary tract may have patchy involvement by carcinoma in situ (CIS) [2,5]. Munoz and Ellison  detected an increase in the rate of CIS from 7.2% to 23.1% during their 23-year analysis of the database. Tobacco, environmental and occupational exposures have been linked to the risk of developing TCC in general . Patients with a history of bladder TCC are at added risk, as 4.5–21% develop UTTCC during the long-term follow-up [6,7]. This risk appears to be equivalent over 5, 10 and 15 years after the diagnosis of bladder TCC . Conversely, 25–75% of patients first presenting with UTTCC subsequently develop de novo bladder TCC [2,8]. Patients with a history of TCC are also recognized to have a high risk of subsequent recurrence in a retained ureteric stump . Thus, patients with a history of TCC require surveillance of all remaining urothelium using appropriate imaging and endoscopic studies.
The standard treatment of UTTCC has traditionally consisted of radical nephroureterectomy (RNU), i.e. extrafascial dissection of the kidney with the entire length of ureter and adjacent segment of bladder cuff (Fig. 1) . This procedure generally requires two major incisions or one long midline incision. Many alternatives have emerged in an effort to minimize the morbidity of this procedure. Advances in endoscopy and laparoscopy have dominated these efforts . An endourological approach to the distal ureter was first proposed in 1952 . Since then, many other minimally invasive approaches to the distal ureter and bladder cuff have been described . Clayman et al. were the first to report on radical laparoscopic nephroureterectomy (LNU) in 1991. While these initial procedures were marked by long operative times, patients had less morbidity afterward than with open RNU . Over 20 clinical series have since been reported at the time this review was compiled. Here I evaluate contemporary perioperative and oncological outcomes after various minimally invasive techniques for excising the kidney, ureter and bladder cuff.
A pre-Medline and Medline search using the keywords ‘nephroureterectomy’, ‘laparoscopy’, and ‘transitional cell carcinoma’ was used to identify publications dating 1996 to mid-2003. The primary source was examined for single-series publications detailing histopathological and follow-up information of > 10 patients. Publications published before 1996 and other relevant reports were selectively reviewed.
Since the first report of LNU ≈ 400 patients have been reported to undergo the procedure via conventional transperitoneal [14–18], conventional retroperitoneal [19–22], hand-assisted transperitoneal [23–26] and hand-assisted retroperitoneal [27,28] approaches, with a variety of laparoscopic, endoscopic or open methods for control of the distal ureter and bladder cuff . Most reports of LNU are predominantly small, single-institution series, with few reports involving ≥ 25 patients [15–17,19,24,28].
Contemporary single-institution series representing the most significant experience with LNU, stratified by type of laparoscopic approach, are listed in Table 1. While there are some major differences among these single-centre series, e.g. the length of hospitalization and complication rates, these differences probably reflect heterogeneous practice patterns, variable definitions of outcome and the influence of experience, rather than real differences in technique. No consistent pattern of significant, clinically relevant differences is evident between the different laparoscopic approaches used for this portion of LNU.
Table 1. Perioperative results from larger international series of LNUX, according to laparoscopic approach
To date, no prospective randomized studies comparing conventional and hand-assisted LNU have been reported. In a multi-institutional retrospective study, Landman et al. evaluated hand-assisted and conventional transperitoneal LNU in 16 and 11 patients, respectively. Operations for the conventional approach were longer than for the hand-assisted approach, with equivalent blood loss, narcotic use and hospitalization. However, patients undergoing the hand-assisted technique had a 1–3 week longer convalescence than those undergoing conventional transperitoneal LNU. In a retrospective single-institution study Baldwin et al. compared the results of conventional laparoscopic, hand-assisted laparoscopic, and open nephrectomy (13, eight and 26 patients, respectively). The conventional laparoscopic approach in that study was associated with fewer complications than the hand-assisted and open approaches. Also, conventional laparoscopic nephrectomy was associated with 21% lower costs than the other two approaches. For nephroureterectomy, some consider a hand-assisted approach cumbersome, as the incision can be either too high for adequate bladder dissection or too low for comfortable assistance of renal dissection, causing significant discomfort to the surgeon's arm and hand .
For transperitoneal vs retroperitoneal approaches, various rationales have been used to support either technique . Concerns about a longer postoperative ileus associated with transperitoneal procedures do not appear to be supported by the clinical experience (Table 1), possibly because of the minimal bowel manipulation that occurs with a laparoscopic approach, even with the transperitoneal route. A randomized prospective trial of laparoscopic transperitoneal vs retroperitoneal laparoscopic nephrectomy conducted at the Cleveland Clinic identified only quicker surgery and more rapid control of the renal vessels with the retroperitoneal approach, with no difference in the rates of postoperative ileus between the transperitoneal and retroperitoneal approaches . The hand-assisted retroperitoneal approach is a relatively novel development that would initially appear to combine the advantages of retroperitoneal and hand-assisted techniques. However, the limited working space afforded by pneumoretroperitoneum is unlikely to allow this approach to become widely used, and experience with this operation remains limited to a few institutions in Asia (Table 1). Regarding costs, limited data suggest that, overall, LNU is 6% more efficient than open RNU at centres where laparoscopy is used routinely . With the continued decline in operative times and reduced reliance on disposable equipment, the two major cost determinants of laparoscopic surgery, it is anticipated that the cost of LNU will continue to decrease.
In summary, results from centres using routine laparoscopic surgery show that LNU is feasible, with acceptable morbidity and overall favourable perioperative outcomes, regardless of the type of laparoscopic approach used (Table 1). The retroperitoneal approach avoids entry into the peritoneal cavity, affords quicker access to the renal vessels, and is associated with a shorter operation. Limited data suggest that hand-assisted LNU is learned more easily and is faster but at the potential cost of a longer convalescence, higher operative expenses and reduced ergonomic comfort for the surgeon. Additional data from a prospective randomized study are still required to clarify cogent differences between the conventional and hand-assisted methods.
TECHNIQUE OF DISTAL URETERECTOMY AND BLADDER CUFF RESECTION
More than five different methods of controlling the distal ureter and bladder cuff have been described during conventional and hand-assisted LNU (Table 1). In 1952, McDonald et al. reported the first endoscopic method of distal ureter and bladder cuff control, first performed in 1948, in a patient presenting with extensive tumour in the residual ureter after a nephrectomy. The ureter was disconnected from the bladder by transurethral resection (TUR) of the perimeatal bladder and intramural ureter . The open distal end of the ureter was sealed with fulguration, and the remaining detached ureter was then ‘plucked’ through the open incision by traction. This technique, with various modifications thereof, has been described in many contemporary series of LNU [14,15,34,35]. TUR of the distal ureter and bladder cuff has been associated with local recurrences in several publications (see below under Port Site and Ureteric Stump Recurrences). Other disadvantages of this approach include urine and tumour cell spillage from the upper tract, despite attempted sealing of the distal ureter with fulguration, and the risk of leaving a discontinuous, isolated segment of ureter in the extraperitoneal space. The presence of distal ureteric tumours or CIS is a contraindication to this approach. This approach may be advantageous for patients who have no bladder or distal ureteric tumours and have had previous pelvic surgery, whereby an additional lower abdominal incision may be avoided.
In 1953, McDonald  described another endoscopic technique of distal ureterectomy, whereby the ligated ureter is stripped, or intussuscepted, through the urethra with the aid of a ureteric catheter, after endoscopic incision of the peri-meatal bladder wall. Clayman et al. reported 18 patients undergoing this technique in 1983, with 14 followed for 5 years with no any local recurrences. This method is still reported occasionally in contemporary series . The presence of any ureteric tumour is a contraindication to this approach.
Laparoscopic stapling of the distal ureter and associated bladder cuff is another method used during LNU, and is often combined with cystoscopic unroofing and scarification of the ureteric orifice [16,17]. This method has the unique advantage of keeping the urinary tract closed, which may be advantageous for patients with a history of radiation therapy or other risk factor for poor bladder healing, and those in whom urine or irrigant extravasation is to be avoided . Stone encrustation along the staple line has not been a problem in the few patients on long-term follow-up, although the staples are occasionally visible intraluminally [16,38]. The small segment of tissue within the incorporated staple line on the specimen side becomes unavailable for pathological evaluation of the margins, whereas at the other end the same tissue has been found to remain viable  and a potential source of extravesical recurrence. Patients with distal, intramural or peri-meatal tumours are not candidates for this approach because of the concerns outlined here.
A novel method of en-bloc transurethral detachment and ligation (TUDL) was described by Gill et al. in 1999 . This method mimics the classic open approach of distal ureter and bladder-cuff resection. A ureteric catheter is initially placed through an endoloop tie and into the ureter. The bladder cuff is circumferentially incised, and the intramural ureter dissected using a Collins knife aided by two transvesically placed 5-mm laparoscopic ports. The endoloop is then used to ligate the ureter against the catheter, preventing spillage from the upper tract into the extravesical space during the subsequent laparoscopic approach. After 11.1 months of follow-up in 35 patients, there were no differences in bladder, local or metastatic recurrences between patients undergoing laparoscopic and open RNU . An updated analysis of 59 patients with 23-month follow-up by the same investigators showed this technique to be associated with significantly fewer positive margins and local recurrences than extravesical stapling . Stifelman et al. used a variation of the described TUDL technique in 11 patients undergoing hand-assisted LNU. The presence of active bladder disease, intramural ureteric tumour and previous pelvic radiation therapy are contraindications to this approach.
Open dissection remains one of the most common methods of controlling the distal ureter and bladder cuff (Table 1). This is the method of choice at most centres, including ours. Among the advantages of this approach are that the patient does not need to be repositioned, and it is rapid. The same incision is used for simultaneous intact extraction of the en-bloc specimen while minimizing the risk of tumour spillage, as the ureter is clipped laparoscopically after vascular control has been established and urinary production is terminated. An open approach may be used transvesically or extravesically, usually via a low Gibson, modified Pfannenstiel, or lower midline incision. Some surgeons place a ureteric catheter at the beginning of the procedure to facilitate laparoscopic dissection, but we do not find this manoeuvre necessary. Care must be taken to avoid injury to the contralateral ureteric orifice. Open RNU, when performed through a single incision, has been associated with a 50% residual ureteric stump rate , with recurrences in the stump occurring in 30–64% of cases [5,44]. Thus, an open approach does not guarantee complete ureteric resection or negative margins, especially when extravesical and ‘blind’ (Fig. 2) or in the presence of unrecognized invasive disease of the distal ureter. Because contemporary pathological practice does not identify nor confirm the presence of bladder mucosa at the distal margin of resection, it is the surgeon's responsibility to visually confirm the presence of this tissue during open, laparoscopic or endoscopic surgery (Fig. 1B). Thus, for an open extravesical approach, our routine practice is to dissect the entire lateral pedicle of the bladder, which allows direct visualization of the ureteric hiatus. The bladder cuff and subsequent bladder repair can thus be completed under direct vision. Such dissection may be difficult in patients with previous pelvic surgery or radiation.
MORBIDITY OF LNU
Minor complications after LNU have been reported in 7–40% of patients and major complications in 4.8–8% (Table 1). In comparison, open RNU has been associated with a 29% complication rate [16,19]. It has been suggested that patients undergoing LNU have fewer pulmonary complications than those undergoing open RNU, possibly as a result of reduced splinting, improved ambulation and reduced narcotic usage afterward [14,16]. Perioperative morbidity and mortality after LNU or RNU is probably higher than that encountered for other forms of elective surgery, and probably attributable to the interrelated causative risk factors associated with TCC, e.g. heavy tobacco use, advanced age and the corresponding high frequency of associated cardiopulmonary comorbidity in these patients [2,14,15].
Table 2 lists the histopathological results of LNU from the larger single-institution series, representing in all 198 patients from the USA, UK and Japan. Low-stage (pTis-pT1) lesions were identified in 50–68% of patients, and high-stage (pT2–pT4) in 28–43%. Interestingly, most lesions (52–94%) were high-grade. Table 3 reviews the results of six studies comparing follow-up data in patients undergoing open RNU (156) and LNU (154). Positive margin rates were not consistently reported by all institutions but appear equivalent for the few studies reporting the results of both approaches. No port-site recurrences were reported in these laparoscopic cohorts or in other reports [15–17,19,21,22,25,27–29,42]. Bladder recurrences for open (24–64%) and laparoscopic (9–54%) groups are within the expected ranges. Rates of local and distant recurrence are likewise comparable between open (0–24% and 0–23%, respectively) and laparoscopic (0–15% and 0–23%, respectively) groups. Importantly, all patients undergoing open RNU had a longer follow-up, and thus the open cohorts appear to have higher rates of bladder, local and distant recurrence in these reports (Table 3). The higher recurrence rates may also represent selection bias, as the laparoscopic cases correspond to the initial series of LNU from each respective centre, whereby patients with more favourable disease were probably selected. Reports on open RNU show that most patients with low-stage, low-grade disease had favourable long-term outcomes after surgery, whereas those with high-stage, high-grade disease had poorer survival rates [3,4]. In one of the largest published series of open RNU, Hall et al. reported 5-year disease-specific survival rates of 100%, 92%, 73% and 41% in 252 patients with stage Ta/Tis, T1, T2 and T3 disease, respectively. The median survival for those with T4 disease was 6 months . The limited stage-dependent survival data from reports of LNU seem to parallel this trend. Whereas most recurrences for this disease tend to occur within the first 2 years , the follow-up for LNU remains limited and the availability of long-term data will be a critically important determinant of oncological efficacy.
Table 2. Histopathology results from eight contemporary international single-institution series of LNU for UTTCC
Port-site metastasis is a rare event that in most cases is related to biological and pathophysiological factors , although the importance of maintaining sound principles of surgical oncology remain critical for diseases such as TCC that have a propensity for local implantation . One case of port-site recurrence after laparoscopic biopsy of bladder TCC has been reported . There have been two reports of port-site recurrences after LNU [47,48]. In one, TUR of the ureter was initially performed  and in another TCC was unsuspected . Notwithstanding these three case reports, all contemporary single-institution series of LNU to date have reported no port-site recurrences [15–17,19,21,22,25,27–29,42].
Much more recognized and ominous are local recurrences after endoscopic control of the distal ureter and bladder cuff, which pre-date the advent of the laparoscopic approach [49–52]. Recurrences after TUR or ‘pluck’ procedures are the most frequently reported. This procedure is associated with irrigant extravasation, urine spillage from the upper urinary tract into the extraperitoneal space despite attempted ‘sealing’ by fulguration, and discontinuous segments of distal ureter remaining in situ. All these events increase the risk of extravesical tumour recurrence, resulting in iatrogenic T3 or T4 disease that may not be seen cystoscopically or on CT . Patients whose upper urinary tract has not been completely evaluated are at particular risk, because they may have occult distal ureteric tumours in addition to an obvious renal pelvic tumour. CIS may also be present, which may not be visible endoscopically nor evident radiographically. It is generally accepted that patients with distal or intramural ureteric tumours are not candidates for any form of endoscopic resection of the distal ureter [14,15]. Evaluation of 60 patients undergoing LNU for curative intent of UTTCC at the Cleveland Clinic also gave compelling data about the method of distal ureter and bladder cuff control. While only 2.8% of patients undergoing the novel TUDL method had a recurrence at the ipsilateral ureteric orifice scar, 25% of patients undergoing laparoscopic stapling had a local recurrence . Accordingly, caution should be exercised when routinely using TUR or stapling for excising the distal ureter. However, in selected surgically challenging patients the variety of minimally invasive options provides the surgeon with additional operative flexibility for an anatomical area that has historically defied a standardized approach.
DISCUSSION AND CONCLUSIONS
LNU results in less blood loss, less postoperative pain, quicker oral intake, shorter hospitalization and a more rapid recovery than with open RNU [15,16,19,21,28,42,53]. LNU is most commonly performed by the transperitoneal approach, although the retroperitoneal approach is gaining popularity. In general, flexibility with both transperitoneal and retroperitoneal approaches is advantageous for the surgeon and the patient. No studies have shown a clear advantage of one approach over another. Preliminary evidence suggests a financial disadvantage with the use of hand-assist devices, but this may be offset by a quicker operation; additional data are required in this regard. The distal ureter and bladder cuff are controlled by various endoscopic, laparoscopic and open methods. No studies have prospectively evaluated one approach against another, although a recent retrospective evaluation raises concern over the routine use of laparoscopic stapling . Increasing awareness of the high prevalence of multifocality and CIS in patients with UTTCC raises concern over any method compromising the principles of surgical oncology [2,3,41,54]. Caution is warranted with TUR or ‘pluck’ procedures, as these are the most commonly reported to be associated with local recurrences. Even with the open approach, complete excision is not guaranteed if done ‘blindly’. Despite the significant improvement in perioperative outcomes after LNU a long-term oncological follow-up is not yet available, and intermediate-term follow-up for a significant number of patients remains limited. The 1- and 2-year follow-up data thus far show cancer-specific survival rates of 63% to 97% in 154 patients after LNU, compared to 53% to 87% after open RNU for 156 patients followed for 1.5–4 years (Table 3). Limited data suggest survival rates for patients with low-stage, low-grade disease undergoing LNU to be quite favourable, while those with high-stage, high-grade disease continue to have markedly diminished survival, a finding consistent with results after open surgery. Overall, the published international data support the continued selective application of LNU for treating localized UTTCC.
The assistance of H. Barton Grossman MD, who critically reviewed this manuscript, and Vickie J. Williams, who provided editorial support, is gratefully appreciated.