Urothelial carcinomas are a type of transitional epithelium cancer. The majority of urothelial carcinomas affect the lower urinary tract and only approximately 5% to 10% are upper tract neoplasms involving the renal pelvis and ureters.1 The current gold standard treatment for patients with ipsilateral upper tract urothelial malignancy and a normal contralateral kidney is an open radical nephroureterectomy with excision of the bladder cuff.2 The removal of the bladder cuff is of prime importance, as research has shown a high rate of recurrence in the ureteral stump if it is not removed completely.3,4
Management of upper tract urothelial carcinoma (UTUC) has advanced significantly since Le Dentu and Albarran performed the first open nephroureterectomy for UTUC in 1898.5 Surgery has moved toward less invasive procedures that are beneficial for myriad reasons, including faster recovery time and a decreased likelihood of certain intraoperative complications. In 1991 Clayman and colleagues6 performed the first laparoscopic nephroureterectomy; from there, progression to the use of robotic surgery occurred. Use of robotic surgery in urology then led to the first robotic nephrectomy, which took place in 2001 using the Zeus robotic system (Computer Motion, Sunnyvale, CA).7 Following this, the da Vinci® Surgical System (Intuitive Surgical, Sunnyvale, CA) was introduced for use in other urologic procedures and, in 2004, the first bilateral heminephroureterectomy with robotic assistance took place.8
The aim of this study is to systematically review all relevant literature (randomized and observational studies) on the following surgical modalities of nephroureterectomy: open, laparoscopic, and robot assisted. We performed a comprehensive comparison to assess both perioperative and oncologic outcomes of these types of surgery in order to examine whether one surgical modality is advantageous over another.
This study was carried out using guidelines from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.9
A sensitive search strategy identified any potentially relevant articles examining the surgical management of UTUC. This search contained words including transitional cell carcinoma and urothelial carcinoma combined with the management options including laparoscop*, open surgery, and robot*. MEDLINE (1946-November 2014), EMBASE (1974-November 2014), Cochrane, and PubMed databases were searched. A list of potentially relevant titles and abstracts were imported to bibliographic software. Duplicates, review articles, non–English-language articles, short surveys, and abstracts from conferences were deleted. Reference lists of articles were checked to identify any missing relevant articles.
To identify relevant articles, titles and abstracts were read and reviewed. If more detail of a study was required to make a decision of inclusion, then the full article was investigated.
Inclusion criteria were that the article must have examined open, laparoscopic, or robot-assisted nephroureterectomy (RANU) surgery. With regard to laparoscopic surgery, all approaches were included: hand-assisted laparoscopy, and transperitoneal or retroperitoneal approaches to surgery. Both the conventional multiport approach and laparoendoscopic single-site approaches to RANU were included in this review. Furthermore, there are a number of different ways in which the distal ureter is handled, all of which were accounted for in this review. For articles to be considered, a minimum sample size of nine was required; otherwise, the statistical power of the study did not reach an acceptable threshold. Endoscopic and laser ablation management were excluded in this review, as these methods are not considered standard treatment of UTUC.
In this review, the perioperative outcomes of interest were length of time in surgery, estimated blood loss (EBL), and length of hospital stay. The oncologic outcomes included were cancer-specific survival, overall survival, incidence of malignancy recurrence, and metastases.
Overall, 7194 potentially relevant articles were identified, of which 1 was a randomized controlled trial (RCT) and the remainder were observational studies. After removal of duplicates, 5141 articles remained, after which exclusion criteria were applied, leaving 350 articles after abstract review. A full-text review was carried out on these articles; ultimately, 50 were identified for inclusion in the review (Figure 1).
Following the literature search, one RCT was identified that compared perioperative outcomes and oncologic outcomes of open nephroureterectomy (ONU) versus laparoscopic nephroureterectomy (LNU) procedures (Table 1).10 The study included 80 patients, of whom 40 were randomly assigned to receive ONU and 40 to receive LNU. The surgery was performed at a single institution by one experienced surgeon. Operative times were very similar, with a difference of only 4 minutes. However, the differences between EBL and length of hospital stay for ONU and LNU were statistically significant (P < .001). This P value showed that LNU was superior to ONU when comparing both surgical outcomes.
In all, 38 observational studies were identified, all of which were retrospective in nature (Table 2). Comparing the average length of time in surgery, ONU averages ranged from 156.211 to 324.012 minutes, LNU averages ranged from 180.013 to 498.014 minutes, and RANU averages ranged from 184.015 to 326.016 minutes. There were a number of studies that directly compared ONU and LNU by designating ONU recipients as the control group and LNU recipients as the case group. In 1995 McDougall and colleagues14 conducted a study in which 13 patients underwent ONU and 10 patients underwent LNU. The outcomes were that the average time for ONU was 234.0 minutes versus 498.0 minutes for LNU; therefore, ONU was the significantly shorter procedure (P = .01). Furthermore, another study carried out by Seifman and associates17 in 2001 showed that the ONU procedure was significantly shorter than LNU (ONU, 199.0 min vs. LNU, 320.0 min; P < .001).
However, not all studies draw the same conclusions. In 2000 Gill and associates18 found that the LNU procedure was significantly shorter than ONU (ONU, 282 min vs. LNU, 222 min; P = .003). Other studies have also demonstrated findings in which LNU is of shorter duration when compared with ONU, including studies by Aguilera and colleagues,19 Blackmur and colleagues,20 Hattori and associates,12 and Kawauchi and associates.21 However, not all of these studies show LNU to be superior to ONU to a suitable degree of significance.
No studies were found that directly compared the operating duration of RANU with another method of nephroureterectomy surgery. The largest RANU study that included length of time in surgery as an outcome was by Pugh and coworkers22 from 2007 to 2011. The study took place in three UTUC centers and the average operating time was 247.0 minutes, which falls within the average time bracket for ONU.
EBL was reported in 27 observational studies. The average range of EBL for ONU was 29623 to 696 mL,18 for LNU the average range was 13019 to 479 mL,23 and for RANU the average range was 5024 to 284 mL.25 Although average ranges of blood loss for each surgical method show overlap, RANU results were superior to other surgical methods. Yang and coworkers24 showed the lowest average EBL of all the studies, at a total EBL of 50 mL. Although data regarding very few robotic cases have been published, the results of EBL demonstrate the potential benefits of RANU with regard to this perioperative measure.
Finally, the last perioperative outcome analyzed was length of stay in hospital. This bears great importance, as it is related to the patient’s general well-being following surgery, and it has implications on the overall total cost of the procedure. The length of stay average for ONU was from 5.217 to 21.121 days, for LNU the average ranged from 2.318 to 13.526 days, and for RANU the average ranged from 2.327 to 6.724 days. The results show that LNU is associated with a shorter hospital stay than ONU. In all of the studies directly comparing the two surgical modalities, the length of stay for patients undergoing LNU was shorter than that for those undergoing ONU. Raman and coworkers28 showed that patients undergoing LNU had a significantly shorter length of hospital stay compared with ONU patients (LNU, 4.6 d vs. ONU, 7.1; P < .01).
The RANU study that had the shortest average length of stay for patients postoperatively was by Hu and colleagues.27 In this study, the procedure that the patients underwent was a laparoscopic nephrectomy followed by robotic surgery to excise the distal ureter and bladder cuff. The average length of stay following the procedure was 2.3 days; however, the small study size of nine must be taken into account when considering the statistical power of the study.
The oncologic outcome results of the one RCT in this category regarding ONU versus LNU were as follows: 5-year cancer-specific survival (CSS), 89.9% versus 79.8%; 5-year metastasis-free survival (MFS), 77.4% versus 72.5%; and 5-year bladder tumor-free survival (BTFS), 77.4% versus 72.5% (Table 3). None of these results were statistically significant; however, the difference in results was most prominent when looking at 5-year MFS with 6 patients having subsequent metastasis after ONU and 11 after LNU (P = .124). The other results that showed some difference were 5-year CSS rates; 12 cancer-related deaths occurred within the groups, 4 in the ONU group and 8 in the LNU group. This difference in mortality associated with cancer led to a P value of 0.2 for 5-year CSS, which had no statistical significance.
Overall, 30 observational studies were found that looked at oncologic outcomes (Table 4); 26 of the studies looked at oncologic outcomes in ONU and LNU procedures, and 4 studies looked at these outcomes in RANU surgery, of which 1 had a reasonable sample size.
It is difficult to draw conclusions from overall results, as data were heterogeneous, and many of the studies looked at slightly different outcomes measures across a variety of timescales. However, what was clear from the consensus of the studies was that ONU and LNU were of equal oncologic safety. Few studies were found to address oncologic outcomes with regard to RANU; therefore, it is difficult to draw an overall conclusion, but it can be said that the oncologic outcome results compare well with those of ONU and LNU currently.
One outcome measure that was commonly reported across studies was 5-year CSS. The range for 5-year CSS in the ONU category was 73.1% to 92.6%,11,29 for LNU it was 75.2% to 95.2%,11,30 and for RANU just one study looked at this outcome, with a result of 75.8%.31 When looking at metastases rates, the range for ONU was 6.9% to 21.1%,32,33 for LNU the range was 0% to 17.2%34,35 and for RANU the range was 0.0% to 18.2%.15,16 Finally, when looking at bladder cancer recurrence, the range for ONU was 11.1% to 42.0%11,36 and for LNU it was 6.0% to 36.0%34,37; the only RANU study that looked specifically at bladder recurrence was by Hu and coworkers,27 with a result of 33.3%.
The study with the largest cohort was carried out by Capitanio and associates29 in 2009, and included 1249 participants. It was a multi-institutional study with a median follow-up time of 49 months. When data remained unadjusted and were analyzed, ONU compared with LNU was associated with higher recurrence rates and higher cancer-specific mortality (P < .001). However, when results were adjusted for tumor stage there was no statistically significant difference found among the data regarding these outcomes (P = .1). This demonstrates clearly how selection bias can influence results when comparing different institutions if potential confounding variables are not adjusted for.
Bariol and colleagues36 performed a single-center retrospective study looking at ONU versus LNU, with a follow-up duration of 101 and 96 months, respectively. In this study, the raw data demonstrated LNU was favorable to ONU; bladder tumor recurrence rates were 28.0% versus 42.0%, and 5-year MFS rates were 72.0% versus 82.1%, respectively. However, no statistical significance was shown (P = .2 for bladder recurrence and P = .26 for 5-year MFS).
Finally, the largest robotic study addressing these oncologic outcomes, with 38 participants, was by Lim and associates31; it had a median follow-up of 45.5 months. The overall outcomes in this study were promising; 5-year overall survival was 60.9% and 5-year CSS was 75.8%. The other RANU studies, despite their small sample sizes, also show encouraging results with outcomes that appear to be on par with alternative surgical management options.
This systematic review did confirm that there is a paucity of good-quality evidence regarding ONU, LNU, and RANU procedures, with just one relevant RCT identified. Even though an RCT is considered the highest level of evidence, the one identified must be viewed with caution as it was carried out in a single center by a single surgeon. If this individual surgeon had greater skill and experience in either the ONU or LNU procedure, the results would not be generalizable beyond the single surgeon’s expertise.
When examining the perioperative outcome measurement of length of time in surgery, it is expected that, because LNU is a newer technique, it will take longer to complete. However, this value is likely to decrease as a surgeon becomes more skilled and experienced. Results of the other perioperative outcomes of EBL and length of hospital stay showed that LNU was superior to ONU. When interpreting data regarding EBL it is important to consider that often volume is defined at the discretion of the surgeon. Therefore, a more objective way of quantifying EBL is to compare preoperative and postoperative hemoglobin levels, producing a more valid result. Length of stay is also not easily compared across centers, as it is heavily influenced by patient discharge pathways that are not standardized. Length of stay is also not easy to compare across centers, as it is heavily influenced by patient discharge pathways that are not standardized. Uniform criteria would minimize bias and help in comparing this measure.
Some well-accepted advantages of laparoscopic over open surgery for these types of procedures were not examined and must be taken into account, such as postoperative pain. Results were not shown for this outcome in the RCT directly comparing ONU and LNU, but it is expected that the ONU procedure, which is carried out through a flank incision combined with a lower quadrant incision, causes more pain postoperatively than the LNU technique, which uses the four-trocar technique.10 The cosmetic impact of the incisions should also be considered, despite it being a qualitative measure.
When exploring oncologic outcomes there were many factors that may have affected result synthesis, including age, pathologic tumor stage or grade, lymphovascular node invasion, and previous malignancies of the bladder. Fortunately, most studies, when processing the results, used multivariable models that adjusted for these potential confounding variables. In the study by Capitanio and associates,29 patients who were at favorable risk (malignancy being at a nonmuscle invasive stage and an absence of metastases) were assigned to the LNU group rather than the ONU group, as LNU was the more novel technique.
One of the most important factors that may have influenced results, particularly concerning malignancy recurrence, is the management of the distal ureter in nephroureterectomy surgery. An abundance of studies examined the controversies of this matter and there is no consensus as to the best available technique; what is clear is that the surgeon operating must adopt a safe technique with good margins.39 Another area of discord surrounds the risk of port-site metastasis in LNU and RANU; however, in a recent review, it was found that the actual risk of this event occurring is low and can be further minimized by using an organ bag when retrieving the specimen.40
It is important to consider the overall cost efficiency of the different surgical procedures, as decisions made regarding the surgical modality for a nephroureterectomy must be economically sustainable. In the study carried out by Trudeau and colleagues41 overall costs of LNU and RANU were compared, taking into account costs of complications, blood transfusions, and length of stay. It was found that a significantly higher cost was incurred during hospitalization if RANU was the chosen procedure. This poses the question of whether RANU is a procedure that should be offered universally from an economic perspective, no matter how good its outcomes are.
When examining perioperative outcomes, the laparoscopic procedures (LNU and RANU) had superior outcomes to those of ONU and, when looking at longer-term oncologic outcomes, results were comparable. However, there are limited studies analyzing the use of RANU, which led to inconclusive results. Studies were all retrospective in nature, with the exception of one; consequently, results should be viewed with caution. RCTs comparing RANU directly with LNU and ONU will be particularly useful to draw further conclusions regarding the procedure with the best outcome. Ultimately, the procedure with the best outcome results will be the most beneficial for patients. There is, however, a significant gap in data regarding the nephroureterectomy procedure; in order to confirm and further review findings, larger, well-designed RCTs are needed.
