- Top of page
- Patients and Methods
- Conflict of Interest
The role of pelvic lymph node dissection (PLND) in patients with prostate carcinoma (PCa) has been a highly debated topic. While in bladder cancer surgery there is strong evidence for a definite therapeutic benefit of extended PLND, such evidence is lacking for prostate cancer, and no consensus exists . Thus, practices with regard to PLND in PCa are inconsistent between surgeons and institutions. Some investigators argue for extended PLND based on data showing that a significant number of positive lymph nodes (LNs) are missed with a limited dissection [2-4], while others highlight the low rates of LN metastasis with localized PCa and lack of a prospective randomized controlled trial investigating survival in patients with extended dissection . Although high risk patients are most vulnerable to LN invasion, this cohort has rarely been examined independently in PLND studies. It would be beneficial to study the extent of PLND in this group and to determine a minimum LN yield (LNY) necessary for detecting nodal metastasis for accurate staging.
Additionally, the effects of radical prostatectomy on potency and continence have been well documented in the literature; however, there have been few, if any, studies investigating functional outcomes in relation to PLND in PCa. Colleagues interested in these outcomes after lymphadenectomy during surgery for rectal cancer have presented some initial compelling data relating erectile dysfunction to the extent of lymphadenectomy [6, 7]. Because dissection templates are similar for rectal and prostate cancer surgery, it would seem reasonable to assess functional outcomes in relation to LN dissection during radical prostatectomy.
We aimed to investigate the extent of LN dissection necessary in a high risk cohort and to assess the functional outcomes and complications of PLND during radical prostatectomy.
- Top of page
- Patients and Methods
- Conflict of Interest
With respect to the total cohort stratified by D'Amico low, intermediate and high risk classification, there was a statistically significant difference in median LNY between the three groups (low risk 5, intermediate risk 7, high risk 13; P < 0.001). LN positivity also varied significantly between the risk groups (low risk 0%, intermediate risk 3 [0.8%], high risk 11 [13%]; P < 0.001). Console time increased significantly with each risk group (P < 0.001) (Table 1).
Table 1. Baseline demographics, preoperative, intraoperative and postoperative variables, variables for the study cohort stratified by D'Amico risk category.
|Variable||Low risk (n = 309)||Intermediate risk (n = 369)||High risk (n = 82)||P|
|Median age (IQR)||58 (53, 63)||61 (56, 66)||62 (56, 67)||<0.001|
|Median BMI (IQR)||27 (25, 29)||26.5 (25, 29)||27.5 (24, 30)||0.81|
|Median PSA (IQR)||4.6 (3.4, 5.8)||5.1 (4.1, 7)||7.4 (5, 12.2)||<0.001|
|Gleason biopsy grade (%)|| || || ||<0.001*|
|≤6||301 (99.7)||36 (10)||3 (3.6)|| |
|7||0||325 (89.7)||4 (4.8)|| |
|≥8||0||0||76 (91.6)|| |
|Clinical stage (%)|| || || ||<0.001*|
|≤T1c||256 (81)||262 (69.3)||34 (40.5)|| |
|≥T2a||60 (19)||116 (30.7)||50 (59.5)|| |
|Gleason pathology grade (%)|| || || ||<0.001*|
|≤6||118 (38.3)||36 (10)||3 (3.6)|| |
|7||187 (60.7)||310 (83.1)||39 (48)|| |
|≥8||3 (1)||27 (7.2)||39 (48)|| |
|Lymph node yield (IQR)||5 (2, 10)||7 (3, 12)||13 (6, 20)||<0.001|
|Lymph node positivity (%)||0||3 (0.8)||11 (13.4)||<0.001*|
|Positive surgical margins (%)||18 (5.7)||28 (7.4)||10 (12)||0.144|
|Pathology prostate volume (IQR)||46.3 (37.9, 58.5)||46.9 (37.6, 57.4)||50.1 (24, 64)||0.057|
|Percentage cancer (IQR)||3 (1, 8)||5 (3, 10)||10 (5, 15)||<0.001|
|Estimated blood loss||150||150||150||0.143|
|Console time (IQR)||85 (67, 110)||86 (68, 110)||111 (85, 135)||<0.001|
With respect to the relationship of LNY and LN involvement in high risk patients, there was no significant difference in baseline characteristics between the two study groups (LNY < 13 and LNY ≥ 13). The median LNY in the LNY < 13 group was 6 and the median LNY in the LNY ≥ 13 was 19 (P < 0.001). In all, 5.1% (n = 2) of the patients in the LNY < 13 group and 20.9% (n = 9) of patients in the LNY ≥ 13 group had positive LNs (P = 0.036). In other words, 82% of the patients with positive LNs had an LNY ≥ 13. There was also a statistically significant difference between these two groups with respect to median console time with the LNY < 13 and LNY ≥ 13 groups having a console time of 100 and 120 min, respectively (P = 0.04) (Table 2).
Table 2. Baseline demographics and postoperative variables, pathological and intraoperative variables for D'Amico high risk patients stratified by lymph node yield.
|Variable||LNY < 13 (n = 37)||LNY ≥ 13 (n = 43)||P|
|Median age (IQR)||63 (52, 67)||62 (56, 67)||0.904|
|Median BMI (IQR)||27 (24, 29)||28 (25, 32)||0.589|
|Median PSA (IQR)||7.7 (5.5, 15)||6.8 (4.6, 10.3)||0.163|
|Median lymph node (IQR)||6 (3, 9)||19 (14, 24)||<0.001|
|Lymph node positivity (%)||2 (5.1)||9 (20.9)||0.036*|
|Positive surgical margins (%)||4 (10)||6 (14)||0.609*|
|Pathology prostate volume (IQR)||51 (43, 62)||50 (41, 68)||0.747|
|Percentage cancer (IQR)||10 (3, 16)||10 (5, 15)||0.696|
|Estimated blood loss (IQR)||150 (150)||150 (150)||0.291|
|Console time (IQR)||100 (85, 120)||120 (95, 137)||0.044|
When comparing high risk patients based on LN positivity, there was no difference between the LN positive and negative groups based on age, body mass index, Gleason biopsy grade, clinical stage, surgical margins and estimated blood loss. All patients (100%) with LN positivity were Gleason biopsy ≥8 and 73% were clinically staged ≥T2, whereas 91% of the LN negative group were Gleason biopsy ≥8 and 58% were clinically staged ≥T2 (P = 0.60 and P = 0.35 respectively). Patients with LN involvement had a median PSA of 22 while patients without LN involvement had a median PSA of 7 (P = 0.02). Patients with LN positivity had a median LNY of 20 while patients who had negative LNs had a median LNY of 11 (P = 0.05). There were statistically significant differences between the two groups with respect to pathological prostate volume (P = 0.01), percentage cancer (P = 0.01) and console time (P = 0.04) with the LN positive group having higher values in all categories (Table 3). Overall, complications of extended PLND in the high risk cohort included a 19.5% rate of lymphorrhoea and a 2.4% rate of CSLs.
Table 3. Baseline demographics and postoperative, pathological and intraoperative variables for D'Amico high risk patients stratified by lymph node invasion.
|Variable||LN negative (n = 71)||LN positive (n = 11)||P|
|Median age (IQR)||62 (56, 67)||59 (55, 65)||0.234|
|Median BMI (IQR)||27 (24, 29)||30 (25, 34)||0.252|
|Median PSA (IQR)||7.1 (5, 11)||22 (9.1, 24)||0.02|
|Gleason biopsy grade (%)|| || ||0.601*|
|≤6||2 (3)||0|| |
|7||4 (5.7)||0|| |
|≥8||64 (91.4)||11 (100)|| |
|Clinical stage (%)|| || ||0.346*|
|≤T1c||30 (42.3)||3 (27.3)|| |
|≥T2a||41 (57.7)||8 (72.7)|| |
|Gleason pathology grade (%)|| || ||0.06*|
|≤6||2 (3)||0|| |
|7||37 (54)||2 (18)|| |
|≥8||30 (44)||9 (82)|| |
|Lymph node yield (IQR)||11 (5, 18)||20 (13, 22)||0.050|
|Pathology prostate volume (IQR)||50 (40, 61)||75 (50, 134)||0.013|
|Percentage cancer (IQR)||8 (4, 15)||15 (10, 25)||0.01|
|Estimated blood loss (IQR)||150 (150)||150 (150, 200)||0.670|
|Console time (IQR)||108 (85, 125)||137 (96, 147)||0.041|
We found no effect on postoperative continence as a result of LNY, with rates of 90% and 90.2% in patients with LNY < 13 and LNY ≥ 13, respectively (P = 0.845) (Table 4). Additionally, there was no effect on postoperative continence when patients with LNY < 20 and LNY ≥ 20 were compared, with rates of 92.8% and 86.6%, respectively (P = 0.529) (Table 5). We also found no difference in postoperative potency in patients with LNY < 13 compared with LNY ≥ 13, with rates of 73.9% vs 77.8% (P = 0.764) (Table 4). However, patients with LNY < 20 had a statistically significant higher rate of potency than those patients with LNY ≥ 20, with rates of 70.0% vs 55.2% (P = 0.020) (Table 5).
Table 4. Functional outcomes with LNY < 13 and LNY ≥ 13 (bilateral nerve sparing).
|Variable||LNY < 13 (n = 360)||LNY ≥ 13 (n = 99)||P|
|No. of patients continent at 26 weeks (%)||272 (90)||74 (90.2)||0.845a|
|No. of patients with erection sufficient for intercourse at 26 weeks (%)||266 (73.9)||70 (77.8)||0.764a|
Table 5. Functional outcomes with LNY < 20 and LNY ≥ 20 (bilateral nerve sparing).
|Variable||LNY < 20 (n = 430)||LNY ≥ 20 (n = 29)||P|
|No. of patients continent at 26 weeks (%)||336 (92.8)||19 (86.6)||0.529a|
|No. of patients with erection sufficient for intercourse at 26 weeks (%)||301 (70.0)||16 (55.2)||0.020a|
- Top of page
- Patients and Methods
- Conflict of Interest
PLND is the most effective method for detecting LN invasion in prostate cancer as imaging modalities such as CT and MRI have poor accuracy in this regard. Predictive models using preoperative criteria have been developed and validated to more precisely determine which patients would benefit from PLND . There has been a decline in the rate of PLND since the 1990s, perhaps due to the stage migration of PCa since the advent of PSA screening . Therefore many urologists do not believe the risks associated with PLND are worth the low probability of detecting LN invasion. Minimally invasive techniques have also been blamed for the decrease in PLND, as a sampling of Medicare patients from 2003 to 2005 showed that patients who underwent open radical prostatectomy were four times more likely to undergo PLND vs patients who underwent any minimally invasive technique . Initially, there was a paucity of literature describing PLND in RARP series; however, more recent reports demonstrate that PLND during RARP is technically feasible with good LN yields and comparable complication rates to open series [14, 15].
There have been variations in the definitions used for PLND, but most authors agree that a limited (or standard) template includes the obturator and external iliac packets and an extended template adds the hypogastric nodes (and often the common iliacs and triangle of Marcille as in our own series). The European Association of Urology states that men with intermediate to high risk PCa should receive an extended PLND; on the other hand, the AUA guidelines state that PLND should generally be reserved for patients with higher risk of nodal involvement, but the extent is not defined . Some believe that a limited PLND provides an adequate sampling for staging, as Weingartner et al.  demonstrated that the predominant region for LN metastasis in their study was to the external iliac nodes. However, there is a legitimate concern for inaccurate staging with a limited dissection, as Bader et al.  demonstrated in 2002 by finding LN positivity exclusively along the hypogastric artery in 19% of his patients. Heidenreich et al.  compared an extended technique to a limited dissection and found twice as many positive nodes using the former (26% vs 12%; P < 0.03). Their extended dissection also included the presacral and common iliac regions, but only added 3.1% of patients with metastasis to these regions. Additionally, a subgroup analysis found a low risk of 2% for LN disease in patients with serum PSA < 10.5 ng/mL and biopsy Gleason < 7. Thus, Heidenreich et al.  concluded that an extended dissection should include the obturator, external iliac and hypogastric LNs and could be omitted in low risk patients.
To better understand the extent of PLND necessary, Matthei et al.  mapped out the primary lymphatic drainage of the prostate in 34 patients receiving radical prostatectomy. They used ultrasound guided intraprostatic injection of technetium (Tc-99m) nanocolloid with a single-photon emission CT 1 h later. The single-photon emission CT was fused with a preoperative CT or MRI to construct a three-dimensional lymphatic map. The locations of the Tc-99m LNs were confirmed using an intraoperative gamma probe. Matthei et al. found that the highest proportion of positive LNs (38%) was located along the external iliac and obturator packets. However, they also identified 25% of primary lymphatic drainage to the hypogastric nodes and 16% to the common iliacs. They recommended that an extended dissection to include the common iliac LNs up to the ureteric crossing would be ideal and would cover about 75% of the prostate's primary lymphatic drainage .
Briganti et al.  investigated whether the number of LNs removed correlated with the probability of LN invasion. The probability of correctly predicting the rate of LN invasion was close to zero when <10 LNs were removed, rose sharply with 10–20 LNs removed and plateaued at ≥28 LNs. These data support the rationale for a LNY of at least 10 LNs and demonstrate the extremely low probability of accurate staging with lower LN yields; interestingly, such yields are often reported with limited PLND [18, 19].
The complications of PLND have been well documented and include most commonly lymphocoeles, prolonged lymphatic drainage (lymphorrhoea), vessel injury, obturator nerve injury and deep vein thrombosis . Although the authors in the previous study  promoted very extensive PLNDs, it was clear that they began to note complications with their dissections, as demonstrated by their 2011 study: using radical retropubic prostatectomy, Capitanio et al.  found that the risk of developing a CSL (defined as the presence of a symptomatic lymphocoele requiring any type of invasive treatment) increased by 5% for every LN removed. There was also a significant positive relationship with LNY and lymphorrhoea (volume of lymphatic drainage). Twenty LNs or more removed was a significant predictor in multivariable analysis of CSL development .
The main objective of our study was to establish a minimal LNY necessary for accurate staging in a high risk cohort. Unfortunately, the high risk PCa group is not frequently isolated in studies due to low sample size, even though this group is most likely to experience LN invasion. Briganti et al.'s  study which found the probability of detecting LN invasion to be nearly zero with <10 LNs removed did not examine the high risk cohort (n = 70) independently. In our study, high risk patients (n = 82) underwent an extended PLND which included the external iliac, obturator and hypogastric nodes up to the common iliac bifurcation, as well as the triangle of Marcille. Overall median LNY in this cohort was 13, which is comparable to open radical prostatectomy series . Median LNYs (IQR) among the high risk LN positive and negative patients were 20 (13–22) and 11 (5–18) (P = 0.05), indicating the greater likelihood of detecting invasion with greater LNs retrieved. As expected, median console time was significantly higher in the LN positive group (137 min vs 108 min, P = 0.04), as well as the preoperative PSA group (22 vs 7, P = 0.02). The high risk group (n = 82) was subdivided into patients with ≥13 LNY vs <13 LNY, as 13 was the overall high risk median LNY, and the incidence of LN invasion was compared between these groups. In the LNY ≥ 13 group, 21% of patients had positive LNs vs 5.1% of the patients in the LNY < 13 group (P = 0.04). In other words, 82% of the patients with positive LNs had a LNY ≥ 13. These data suggest that, in order to accurately stage a high risk cohort, LNYs of ≥13 LN must be achieved through an extended PLND.
Overall, the high risk cohort demonstrated a low rate of complications, with only 2.4% of patients developing CSLs requiring intervention. This rate is certainly on the lower end of the spectrum compared with other studies which have found 1–12% of patients developing CSLs . Undoubtedly, more lymphocoeles could have been diagnosed had they been routinely looked for but this would have been inconsequential given their asymptomatic nature. The 19.5% rate of lymphorrhoea in our patients was expected given the extent of dissection, and highlights the expectations on postoperative recovery that must be made clear to these patients.
There has been minimal discussion of whether the extent of PLND negatively impacts the results of nerve-sparing radical prostatectomy with regard to potency and continence. In a 2007 review, Heidenreich et al.  postulated this but were unable to make any conclusions, as PLND studies normally do not report on functional outcomes. To our knowledge, there have not been any studies evaluating the effect of the extent of PLND on functional outcomes in the PCa literature. However, in the rectal cancer literature, authors have begun to evaluate the effects of ‘lateral LN dissection’ in Japan where this type of dissection is routinely practised during total mesorectal excision for cancer; the template includes (i) internal iliac (proximal to the superior vesical artery), (ii) common iliac, (iii) external iliac, (iv) obturator, (v) presacral and (vi) internal pudendal. Nishizawa et al.  examined International Index of Erectile Function scores preoperatively and at 3 and 12 months postoperatively. Of those receiving a lateral LN dissection 86% had erectile dysfunction at 12 months vs 50% of those who did not receive a lateral LN dissection (P < 0.01). The lateral node dissection turned out to be the only statistically significant risk factor for postoperative sexual dysfunction, even when examining patients with and without pelvic plexus or hypogastric nerve preservation. Akasu et al.  had similar results in his total mesorectal excision data set, where 95% of patients with pelvic plexus preservation and no lateral LN dissection were able to maintain sexual intercourse 1 year postoperatively vs 56% with pelvic plexus preservation and lateral LN dissection.
Variations in surgical technique and depth of dissection make it difficult to generalize these potency outcomes to PLND during radical prostatectomy; however, it is worth further examining this hypothesis. Our study looked at 760 patients who underwent RARP by a single surgeon. All risk patients were included who were preoperatively potent (SHIM ≥ 17), continent, and who received bilateral nerve sparing. Among patients who fitted the inclusion criteria for functional outcomes (n = 459), 55.2% (16 of 29) with ≥20 LNs removed had erection sufficient for intercourse at a median follow-up of 6 months postoperatively vs 70.0% of patients with <20 LNs (301 of 430) (P = 0.020). There was no significant difference in erection sufficient for intercourse in patients with ≥13 LNs removed vs <13 LNs (73.9% vs 77.8%, P = 0.764). When examining continence, there were no differences between groups (either with 20 LNs or 13 LNs as a cut-off).
Since we can assume the ≥20 LNY group had a more extensive PLND, we can hypothesize that the worse potency outcomes may have been a result of the deeper dissection in this cohort. In fact, our group has recently shown that limiting damage to the pelvic plexus and its downstream fibres, located in proximity to the hypogastric vessels, aids in the preservation of orgasmic function postoperatively ; it is not unreasonable to suggest that an extended PLND involving the hypogastric LNs may contribute to autonomic nerve injury and subsequent reduced potency. However, the exact mechanism of neural damage remains unknown and will certainly be investigated in the future. These data become relevant in low and intermediate risk patients who are more likely to receive bilateral nerve sparing than high risk patients. With stage upgrading occurring rarely in low risk patients, an extended PLND may be counterproductive to the aims of nerve sparing in this cohort. In order to further confirm our results and understand potential mechanisms, decreased potency should be addressed as a potential complication in future studies on extended PLND.
Our study is not without limitations; a small sample size of 82 high risk patients precludes definite conclusions, although this cohort is larger than in most RARP series focusing on PLND. We also lacked information on the locations of the dissected LNs which may have provided us with greater regional understanding of the LN positive patients. Regarding the functional outcomes data, different methods were used to assess preoperative potency (full SHIM score ≥ 17) and postoperative potency (erection sufficient for intercourse by using questions 2 and 3). Additionally, the reduced potency data only added greater strength to the hypothesis since a mechanism has not yet been elucidated. Lastly, the retrospective nature of our study produces inherent limitations.
Thus, robotic-assisted extended PLND is feasible with satisfactory LNY comparable to open series. In order to accurately stage a high risk cohort, an LNY of ≥13 LNs must be achieved through an extended PLND. Additionally, worse potency outcomes with extended PLND involving ≥20 LNs should be weighed against the potential benefits of nerve sparing in low risk patients. PLND studies should begin to incorporate potency data and further investigation is needed to elucidate a mechanism for these reduced outcomes.