Lymph node dissection during nephroureterectomy: Establishing the existing evidence based on a review of the literature

ABSTRACT Abstract Objective: To determine the role of lymph node dissection (LND) in the treatment of upper tract transitional cell carcinoma (UTTCC), as the role of LND along with nephroureterectomy in treating UTTCC is unclear and several retrospective studies have been published on this topic with conflicting results. Methods: The Medical Literature Analysis and Retrieval System Online (MEDLINE), the Excerpta Medica dataBASE (EMBASE), Cochrane Central Register of Controlled Trials database (CENTRAL), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Clinicaltrials.gov, Google Scholar, and individual urological journals, were searched for all studies investigating the role of LND in the treatment of UTTCC. Of the studies identified, those that met inclusion criteria were included in this review. Results: In all, 27 studies were included in this review, with 9303 patients who underwent LND. No randomised controlled trials (RCTs) were identified. Tumours were located in the renal pelvis in 62% of patients, in the ureter in 35.5%, and multifocal in 2.3%. In total: 77.1% were LN-negative and 22.9% had LN metastasis. For all patients undergoing LND, the 5-year recurrence-free survival (RFS) and cancer-specific survival (CSS) rates were 27–65.4% and 32.3–95%, respectively. For patients who underwent a LND in accordance with a standardised anatomical template, the 5-year RFS and CSS rates were 84.3–93% and 83.5–94%, respectively. Conclusion: LND may provide a survival benefit in patients undergoing nephroureterectomy for UTTCC, particularly if following a standardised anatomical template and in those patients with muscle-invasive disease; however, a prospective RCT is required to confirm this. Abbreviations: CSS: cancer-specific survival; LN(D): lymph node (dissection); MeSH: Medical Subject Headings; OS: overall survival; pT: pathological T stage; RCT: randomised controlled trial; RFS: recurrence-free survival; UTTCC: upper tract TCC


Introduction
Urothelial carcinomas are the fifth commonest malignancy [1], of which 5-10% occur in the upper urinary tract [2,3]. They occur twice as commonly in the pelvicalyceal system as they do in the ureter [4]. In contrast with TCC of the bladder, where the majority of patients are diagnosed with non-muscle-invasive disease, about two-thirds of patients have muscleinvasive disease at diagnosis [5].
Current guidelines by the European Association of Urology (EAU) advocate open or laparoscopic radical nephroureterectomy with bladder cuff excision as the standard treatment for high-risk upper tract urothelial cancer, regardless of tumour location [4]. Alternative treatments are generally reserved for patients with low-risk disease or those with significant renal impairment or solitary kidney.
The role of LN dissection (LND) for upper tract TCC (UTTCC) is not known and currently there are no guidelines regarding its role [4]. Given its rare nature, it is a difficult topic to establish an evidence base. In TCC of the bladder, there is increasing evidence that more extensive LND improves prognosis after radical cystectomy [16,17]. It may be the case that this same benefit may exist in treating UTTCC; however, the current evidence is based on small retrospective studies. Results from these studies has been conflicting, with some reporting a survival benefit of LND [9], whilst others report the only benefit being that of accurate staging for prognostication [18][19][20].
The aim of the present review was to systematically review the literature to establish the role of LND in patients undergoing nephroureterectomy for UTTCC.

Methods
A review of the literature was conducted using Cochrane and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [21,22].

Study selection
All languages were included if data were extractable, also references of searched papers were evaluated for further studies for potential inclusion. Authors were contacted wherever the data were not available or not clear, to be able to adequately assess inclusion of their study. If data were not extractable, provided or clarified, the study was excluded. Inclusion criteria were: papers publishing outcome data for patients undergoing LND with nephroureterectomy for UTTCC, papers publishing original data (i.e., not review papers), and English language. Exclusion criteria were: abstracts published from conference proceedings with no full manuscript available, papers not providing outcome data specifically for patients who underwent LND at the time of nephroureterectomy, and papers publishing data not specifically for TCC.

Data extraction
All types of publications were included. Studies were excluded if based on children or LN excision in other conditions than for UTTCC.
The following variables were extracted from each study: patient and cancer demographics, operation, LN yield, operative outcomes, and survival outcomes [recurrence-free survival (RFS), cancer-specific survival (CSS), and overall survival (OS)].

Statistical analysis
We used the Review manager (RevMan) version 5.2 program (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) to conduct the analysis. For continuous data, a Mantel-Haenszel chisquared test was used and expressed as the mean difference with 95% CI and for dichotomous data an inverse variance was used and expressed as risk ratio (RR) with 95% CI. A P < 0.05 was considered statistically significant [21,22].
Heterogeneity was analysed using a chi-squared test on N-1 degrees of freedom, with an α of 0.05 used for statistical significance and with the I 2 test. I 2 values of 0-40%, 30-60%, 50-90%, and 75-100% correspond to 'heterogeneity may not be important', 'may indicate moderate heterogeneity', 'may indicate substantial heterogeneity', and 'may indicate considerable heterogeneity', respectively [21,22]. A fixed-effect model was used unless statistically significantly high heterogeneity (I 2 > 75% was considered as significantly high heterogeneity) existed between studies. A random-effects model was used if heterogeneity existed [22,23].

Patient demographics
The 27 studies included a total of 25 969 patients, 9303 of whom underwent LND along with nephroureterectomy for UTTCC; patient characteristics are summarised in Table 1. Of those undergoing LND, 60.0% were male and ages ranged from 25 to 99 years.
There was no significant difference in the 5-year RFS rate between those who underwent LND compared to those who did not (57.8% vs 64%; P = 0.11; Figure 2).

CSS
The 5-year CSS ranged from 32.3% to 66%. One paper reported 10-year CSS at 32% [33]. Patients with pN0 disease had 5-year CSS of 56-84.5%, whilst patients with LN-positive disease had 5-year CSS ranging from 0% to 47%. Three studies published 5-year CSS for patients undergoing specific-template LND. The 5-year CSS was much better in these patients and ranged from 83.5% to 94% [31,36,39]. There was no significant difference in CSS between those who did and did not undergo LND, with a 5-year CSS of 74.0% and 80.5%, respectively (P = 0.1; Figure 3).

OS
The 5-year OS for patients with pN0 disease ranged from 61% to 68% and was 22.3-66% for those with LN-positive disease. Kondo et al. [37] reported 3-year OS of 86.1% for patients with pT2-4 disease undergoing LND. The 5-year OS was 60.7% and 69.6% (P < 0.001) for those who did and did not undergo LND, respectively ( Figure 4).

Impact on number of LNs removed or template
Six studies investigated the impact on number of LNs removed at LND [23,24,27,36,42,48]. One paper suggested there was no benefit in recurrence rates with increased LNs removed [26], whereas two papers   reported reduced recurrence rates associated with increased number of LNs removed [29,44]. The number of LNs removed had a lesser impact on survival, with four papers reporting no difference in survival according to the number of LNs removed [25,26,38,44], whilst only one paper reported a small survival benefit with increased number of LNs removed [50]. Bolenz et al. [26] reported that increased number of LNs removed had no impact on either recurrence or survival, but did investigate proportion of positive LNs as a prognostic factor. Patients with >30% positive LNs had a higher rate of recurrence at 5 years compared with patients with <30% positive LNs (38% vs 25%, P = 0.021). Similarly, those with >30% positive LNs had a higher 5-year mortality rate (48% vs 30%, P = 0.032).
Five studies investigated whether a more standardised anatomical template for LND had an impact on outcomes [31,33,36,38,39]. Patients undergoing LND following a standardised template generally had better outcomes than patients undergoing LND not according to an anatomical template. Only one paper reported no difference according to type of LND [33]. Four papers [31,36,38,39] reported improved RFS after a standardised template and three papers reported improved survival when a standardised template was used for LND [31,36,38].
pN0 vs pNx Five studies compared outcomes on patients who had undergone LND and been staged as pN0 with patients who were Nx [24,28,32,40,47]. One paper found patients with pN0 to have favourable RFS compared with patients who were Nx [24]. Two papers reported no difference in RFS between the two cohorts [40,47]. Burger et al. [28] reported that, overall, there was no difference in RFS between patients with pN0 and Nx staging, but when comparing patients with locally advanced disease, pN0 had improved RFS compared with those with Nx. Ikeda et al. [32] also reported better RFS in pN0 patients when only including patients with locally advanced disease. Four papers compared survival between pN0 and Nx patients [28,32,45,47]. Similarly to RFS, when all patients were included in analysis, there was no statistically significant difference noted between the two cohorts; however, when focusing on patients with locally advanced disease, two papers reported improved survival in patients who were pN0 [28,32].

Discussion
There is a lack of high-quality evidence on the role of LND along with nephroureterectomy in treating patients with UTTCC. Furthermore, a vast disparity between countries and centres exists. It is not clear whether or not LND reduces recurrence or increases survival, or which patients may benefit most. However, patients who potentially benefit from LND are those with advanced disease and those who undergo LND according to a standardised anatomical template.

Survival
Meta-analysis revealed no statistically significant difference in RFS and CSS between patients who did and did not undergo LND, but patients who underwent LND had poorer OS.

LND
Another factor making it difficult to draw firm conclusions on the benefit of LND is the variation in what was included as LND in the different studies. Most of the studies simply stated that the extent of LND was at the discretion of the operating surgeon. In some cases, this resulted in as few as one LN being resected and in other cases as many as 65. Some studies performed LND in accordance with predetermined anatomical templates depending on the site of the primary tumour; these patients had much more favourable outcomes with 5-year CSS up to 94%.

Chemotherapy
Comparison of outcomes between studies was further complicated by some patients undergoing neoadjuvant and/or adjuvant therapies in conjunction with nephroureterectomy. A large degree of variation was seen between the chemotherapy regimens described.

Strengths and limitations of the review
The majority of published data are evidence based on retrospective studies with large degrees of heterogeneity between studies, and there were no RCTs. Nonetheless, the present review was conducted in a methodological protocol-driven method based on Cochrane and PRISMA guidelines.
The results presented represent the published existing data. Albeit, high risk of biases exist due to the heterogeneity between studies; however, this should emphasise the need for a multicentre RCT.

Implications for clinical practice
Currently, with the existing studies from the literature, routine LND should not be advocated. If LND is required, a pre-determined templated technique should be used.

Implications for research
It is clear that an RCT comparing between LND and no LND should be carried out to establish the evidence.

Conclusion
There is currently insufficient evidence to support the role of LND along with nephroureterectomy in patients being treated for UTTCC. Some studies have suggested that patients with higher pT stage may be more likely to benefit from this and that a standardised anatomical template of LND results in better outcomes. A prospective, RCT is required to determine if there is a survival benefit in LND along with nephroureterectomy in patients being treated for UTTCC and in which group of patients, if at all, this is most appropriate.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
No funding was received for this work.