Whether extended pelvic lymph node dissection should be performed in prostate cancer: The present evidence from a systematic review and meta‐analysis

To compare non‐extended pelvic lymph node dissection (nePLND) with extended pelvic lymph node dissection (ePLND) in outcomes and complications of patients with prostate cancer (PCa).


| INTRODUCTION
Prostate cancer (PCa) disseminates initially to regional lymph nodes (LNs). 1 Pelvic lymph node dissection (PLND) is the most accurate procedure to identify the histological status of LNs. Besides, PLND con- recommend a risk of nodal metastases >5% was an indication to perform an extended pelvic lymph node dissection (ePLND). [3][4][5][6] In addition, the individual risk of lymph node invasion (LNI) can be estimated using models based on preoperative characteristics such as the Briganti and Memorial Sloan Kettering Cancer Center nomograms. 4 Historically, the lack of standardized definitions and terminologies of PLND has led to the difficulty of comparing various PLND approaches. It was reported that a wider extent of PLND led to longer operative duration 7 and more hospitalization days. 8 Conversely, some researchers held opposite opinions. 9,10 The incidences of perioperative and postoperative complications are emphatically considered factors of urologists to decide the extent of PLND. Numerous studies revealed that ePLND had significantly more overall complications than non-extended PLND (nePLND). 11,12 What is more, a randomized prospective self-control study demonstrated that an increased risk of complications attributable to the lymphadenectomy occurred to an extended dissection. 13 However, the assertion that expending of extensive dissection leads to more overall complications has not always been confirmed. Yuh reported that there was no significant difference in overall complications in patients from nePLND group and ePLND group. 14 Furthermore, it is difficult to draw a definite conclusion of specific complications due to no relatively consistent records on the type of specific complications to existing researches.
In this study, we sought to obtain more definitive results of the comparison between nePLND and ePLND, and explore the association with varying extent of PLND and complications by performing meta-analysis.

| MATERIALS AND METHODS
The Preferred Reporting Items for Systematic Reviews and Meta-analyses 15 guideline was conformed when our study was conducted and reported. Comparative studies between radical prostatectomy (RP) with any extent PLND in PCa patients were enrolled in our present study.
In our study, we defined ePLND as types 3 and 4; whereas defined nePLND as types 1 and 2.

| Search strategy
A systematic literature searched on PubMed, EMBASE, and Web of Science was performed to identify all published potentially appropriate studies (till 20 April 2017). The key words were "extended pelvic lymph node dissection," "pelvic lymph node dissection," "PLND," "prostate cancer," "PCa," "radical prostatectomy," or "RP." Additional publications were identified when we searched the reference list of original articles manually. A flow diagram of the selection process is presented in Figure 1.

| Inclusion and exclusion criteria
Articles met the following criteria were included: (a) randomized con- radiotherapy, adjuvant or neoadjuvant chemotherapy, or hormone therapy before PLND was excluded. In addition, references without available data were disregarded.

| Risk of bias assessment
Two investigators independently evaluated the quality of each reference using the Cochrane Risk of Bias Assessment Tool for Non-Randomized Study tool. Cochran's Q test (chi square) and the I 2 method were performed to assess heterogeneity between articles. 17 Begg's funnel plots were compiled to estimate the presence of publication bias; P value <.05 and asymmetric plot suggest a potential publication bias.

| Statistical analysis
We used odds ratio (OR) with corresponding 95% confidence interval (CI) to evaluate the difference between nePLND and ePLND in metaanalysis. Z test was performed to determine the statistical significance of the OR, and a P value of <.05 was considered statistically significant. Statistical analyses were performed using STATA 12.0 software (Stata Corp, College Station, TX). All results were reported with 95% CIs. The modified Newcastle-Ottawa scale was applied to assess the level of evidence of all controlled studies. 18 We did not analyze the blinding method since it was not applicable for surgery clinical trials.
Lindberg's 12 and Eden's 7 researches focused on the difference between limited PLND and extended PLND, super-extended PLND and extended PLND, respectively.
Extender PLND led to more blood loss (650 mL, range 200-1950 vs 590 mL, range 150-2100 mL) and longer operating room time (179 minutes, range 140-235 vs 125 minutes, range 85-150 minutes) in patients with ORP. However, no significant difference in blood loss was observed in patients with LRP (P = .13) or RALP (P = .088, P = .322, and P = .7). A longer operating room time was recorded in LRP and RALP patients (P < .001, P = .027, P < .03, and P < .001) except in Liss's study (P = .211). In addition, no statistics discrepancy between postoperative length of stay was shown in two individual articles (P = .77 and P = .998) apart from Liss's study (P = .002). What is more, there were no significant difference was observed in transfusion (P = .27 and P = .436) and prostate weight (P = .09, P = .125, and P = .582).

| Complications
We extracted information on overall complications from seven references with 2051 patients, and the results are shown in Table 3. 7,8,11,12,14,22 Besides, specific complication statistics were revealed in Table S1. Data demonstrated significantly more complications from ePLND group compared with nePLND group  F I G U R E 3 Forest plots and meta-analysis of overall complications subgroup divided in surgical approaches in extended pelvic lymph node dissection vs non-extended pelvic lymph node dissection

| Evaluation of publication bias
Publication bias evaluated by Begg's test was found for all analyses (P = .035). The presence of publication bias is mainly due to the lack of enrolled studies. However, given the lack of current relevant researches and limited references, our study was still performed.

| DISCUSSION
As the most common malignancy that affects the male genitourinary system, 24 PCa disseminated initially to regional LNs. 1 Conventionally, LNs available for disseminating including the obturator LNs, external iliac LNs, internal iliac LNs, common iliac nodes, and presacral LNs. 16,25 Accepted widely, accurate LNs staging was the most important way to identify localized PCa patients with a higher risk of adverse oncologic outcomes and decide their optimal clinical treatment strategy. [26][27][28][29] The risk of LNI assessment provides the most important refer- Compared with nePLND, ePLND may have a therapeutic role in removing more micrometastases. 31 An inverse correlation was reported on the number of normal LNs removed and the risk of biochemical recurrence. 32 In addition, a retrospective study shown that a more extensive PLND might increase biochemical recurrence free survival in node-negative PCa patients. 31 Although one study enrolled in our study revealed that ePLND might have the trend to improve overall survival in PCa patients, the difference was not statistically significant (P = .07). 19 Our findings revealed that compared with nePLND, more LNY and metastasized LNs could be dissected by ePLND. However, the study of prostate lymph reflow may change our existing view. It was reported that the lymphatic channel draining from the prostate was detected using a fluorescence navigation system after injecting indocyanine green, and the result shown that the major lymphatic pathway involved in the spreading of PCa is internal iliac LNs, which would mean that the dissection of external iliac LNs and obturator LNs covered in standard PLND might not keep the cancer from spreading. 33 To date, the strategy of ePLND application is still controversial because of the lack of solid evidence regarding its oncological benefit and the adverse events associated with it. Given to the improved LNs staging, it is advisable to replace nePLND with ePLND. 16 and NCT01555086) could not be included in this study.
Limitations also inevitably existed in our study. As mentioned before, significant heterogeneity which could not be adjusted by sensitivity analysis was identified among enrolled studies. What is more, limited original studies led to insufficient data available for further analysis.

| CONCLUSION
The results of this systematic review and meta-analysis indicated that extender PLND led to more LNY and more metastasized LNs. More harm would be brought by ePLND in ORP, whereas not in LRP and RALP. In addition, ePLND may lead to more overall compilations than nePLND. Finally, RCTs and high-quality prospective comparative studies are needed to complete this study.