Platinum Priority – Review – Prostate Cancer Editorial by Judd W. Moul on pp. 1266–1268 of this issue Pelvic Lymph Node Dissection in Prostate Cancer Alberto Briganti a, *, Michael L. Blute b , James H. Eastham c , Markus Graefen d , Axel Heidenreich e , Jeffrey R. Karnes b , Francesco Montorsi a , Urs E. Studer f a Department of Urology, Vita-Salute University, Milan, Italy b Department of Urology, Mayo Clinic, Rochester, MN, USA c Urology Service, Department of Surgery, Memorial Sloan-Kettering Cancer Centre, New York, NY, USA d Martini Clinic, Prostate Cancer Centre, Hamburg, Germany e Department of Urology, University of Aachen, Aachen, Germany f University Hospital of Bern, Department of Urology, Bern, Switzerland european urology 55 (2009) 1251–1265 available at www.sciencedirect.com journal homepage: www.europeanurology.com Article info Article history: Accepted March 3, 2009 Published online ahead of print on March 10, 2009 Keywords: Prostate cancer Pelvic lymph node dissection Radical prostatectomy Imaging Complications Please visit www.eu-acme.org/ europeanurology to read and answer questions on-line. The EU-ACME credits will then be attributed automatically. Abstract Context: Pelvic lymph node dissection (PLND) is considered the most reliable procedure for the detection of lymph node metastases in pros- tate cancer (PCa); however, the therapeutic benefit of PLND in PCa management is currently under debate. Objective: To systematically review the available literature concerning the role of PLND and its extent in PCa staging and outcome. All of the existing recommendations and staging tools determining the need for PLND were also assessed. Moreover, a systematic review was performed of the long-term outcome of node-positive patients stratified according to the extent of nodal invasion. Evidence acquisition: A Medline search was conducted to identify original and review articles as well as editorials addressing the significance of PLND in PCa. Keywords included prostate cancer, pelvic lymph node dissec- tion, radical prostatectomy, imaging, and complications. Data from the selected studies focussing on the role of PLND in PCa staging and out- come were reviewed and discussed by all of the contributing authors. Evidence synthesis: Despite recent advances in imaging techniques, PLND remains the most accurate staging procedure for the detection of lymph node invasion (LNI) in PCa. The rate of LNI increases with the extent of PLND. Extended PLND (ePLND; ie, removal of obturator, exter- nal iliac, hypogastric with or without presacral and common iliac nodes) significantly improves the detection of lymph node metastases com- pared with limited PLND (lPLND; ie, removal of obturator with or without external iliac nodes), which is associated with poor staging accuracy. * Corresponding author. Department of Urology, Vita-Salute University, San Raffaele Hospital, Via Olgettina, 60, 20132, Milan, Italy. Tel. +39 02 26437286; Fax: +39 02 26437298. E-mail address: [email protected](A. Briganti). 0302-2838/$ – see back matter # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2009.03.012
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Platinum Priority – Review – Prostate CancerEditorial by Judd W. Moul on pp. 1266–1268 of this issue
Pelvic Lymph Node Dissection in Prostate Cancer
Alberto Briganti a,*, Michael L. Blute b, James H. Eastham c, Markus Graefen d,Axel Heidenreich e, Jeffrey R. Karnes b, Francesco Montorsi a, Urs E. Studer f
aDepartment of Urology, Vita-Salute University, Milan, ItalybDepartment of Urology, Mayo Clinic, Rochester, MN, USAcUrology Service, Department of Surgery, Memorial Sloan-Kettering Cancer Centre, New York, NY, USAdMartini Clinic, Prostate Cancer Centre, Hamburg, GermanyeDepartment of Urology, University of Aachen, Aachen, GermanyfUniversity Hospital of Bern, Department of Urology, Bern, Switzerland
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 5
avai lab le at www.sciencedi rect .com
journa l homepage: www.europeanurology.com
Article info
Article history:Accepted March 3, 2009Published online ahead ofprint on March 10, 2009
Please visitwww.eu-acme.org/europeanurology to read andanswer questions on-line.The EU-ACME credits willthen be attributedautomatically.
Abstract
Context: Pelvic lymph node dissection (PLND) is considered the mostreliable procedure for the detection of lymph node metastases in pros-tate cancer (PCa); however, the therapeutic benefit of PLND in PCamanagement is currently under debate.Objective: To systematically review the available literature concerningthe role of PLND and its extent in PCa staging and outcome. All of theexisting recommendations and staging tools determining the need forPLND were also assessed. Moreover, a systematic review was performedof the long-term outcome of node-positive patients stratified accordingto the extent of nodal invasion.Evidence acquisition: A Medline search was conducted to identify originaland review articles as well as editorials addressing the significance ofPLND in PCa. Keywords included prostate cancer, pelvic lymph node dissec-tion, radical prostatectomy, imaging, and complications. Data from theselected studies focussing on the role of PLND in PCa staging and out-come were reviewed and discussed by all of the contributing authors.Evidence synthesis: Despite recent advances in imaging techniques,PLND remains the most accurate staging procedure for the detectionof lymph node invasion (LNI) in PCa. The rate of LNI increases with theextent of PLND. Extended PLND (ePLND; ie, removal of obturator, exter-nal iliac, hypogastric with or without presacral and common iliac nodes)significantly improves the detection of lymph node metastases com-pared with limited PLND (lPLND; ie, removal of obturator with or withoutexternal iliac nodes), which is associated with poor staging accuracy.
* Corresponding author. Department of Urology, Vita-Salute University, San RaffaeleHospital, Via Olgettina, 60, 20132, Milan, Italy. Tel. +39 02 26437286; Fax: +39 02 26437298.E-mail address: [email protected] (A. Briganti).
0302-2838/$ – see back matter # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2009.03.012
Because not all patients with PCa are at the same risk of harbouring nodalmetastases, several nomograms and tables have been developed andvalidated to identify candidates for PLND. These tools, however, arebased mostly on findings derived from lPLND dissections performed inolder patient series. According to these prediction models, a stagingPLND might be omitted in low-risk PCa patients because of the low rateof lymph node metastases found, even after extended dissections (<8%).The outcome for patients with positive nodes is not necessarily poor.Indeed, patients with low-volume nodal metastases experience excellentsurvival rates, regardless of adjuvant treatment. But despite few retro-spective studies reporting an association between PLND and PCa pro-gression and survival, the exact impact of PLND on patient outcomes hasnot yet been clearly proven because of the lack of prospective random-ised trials.Conclusions: On the basis of current data, we suggest that if a PLND isindicated, then it should be extended. Conversely, in view of the low rateof LNI among patients with low-risk PCa, a staging ePLND might bespared in this patient category. Whether this approach is also safe fromoncologic perspectives is still unknown. Patients with low-volume nodalmetastases have a good long-term prognosis; to what extent this prog-nosis is the result of a positive impact of PLND on PCa outcomes is still tobe determined.# 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved.
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 51252
1. Introduction
Pelvic lymph node dissection (PLND) representsthe most accurate and reliable staging procedurefor the detection of lymph node invasion (LNI) inprostate cancer (PCa) [1]. Unfortunately, imagingprocedures such as computed tomography (CT)and standard magnetic resonance imaging (MRI)have very limited ability to predict LNI [2–4]. Otherinteresting imaging techniques such as [11C]cho-line positron emission tomography/CT or MRI withlymphotropic superparamagnetic nanoparticlesare currently under investigation [5–9]. The lattertechnique is not yet available on the market, andthe use of these sophisticated imaging techniquesis limited by significant costs. Thus, for the timebeing, PLND remains the gold standard for nodalassessment. Which candidates to select for thisprocedure and the optimal extent of PLND (limitedvs extended) are still points of discussion. Debatecentres on three issues. First, not all patients are atthe same risk of harbouring PCa nodal metastases[10–29]. Second, a meticulous PLND is a time-consuming and challenging procedure whichrequires skilled surgeons [30–32]. Third, the impactof PLND on PCa outcome is currently unknown.Indeed, no prospective randomised clinical trialhas ever tested the impact of PLND on PCaoutcomes, not even in the high-risk patient group.
Although one may object that several surgicalprocedures which are currently considered asstandard treatment were never proved efficaciousin randomised clinical trials, this requisiteundoubtedly remains important. Some authorsbase their decision on the need for PLND onpreoperative nomograms which are mainly basedon routinely available preoperative variables[10–29]. Such nomograms allow them to identifythose patients for whom a routine staging PLNDmight be omitted; however, whether this reason-able approach is also safe from an oncologicperspective is still unknown because of the lackof prospective randomised clinical trials. Conver-sely, other authors favour performing PLND in allpatients for whom a radical prostatectomy (RP) istruly indicated [33]. This approach is clearlyassociated with higher staging accuracy, especiallyif an extended PLND (ePLND) is performed [34–40].The aim of this paper is to systematically reviewthe available literature concerning the role of PLNDand its extent in PCa staging and outcome. Thepotential benefits as well as the side-effects ofPLND in PCa are presented. All of the existingrecommendations and staging tools determiningthe need for PLND are also critically evaluated.Moreover, the long-term outcome of node-positivepatients is reviewed and stratified according to theextent of nodal invasion.
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 5 1253
2. Evidence acquisition
A Medline search was conducted to identify originalarticles, review articles, and editorials addressingthe role of PLND in PCa. Keywords included prostatecancer, pelvic lymph node dissection, radical prostatec-tomy, imaging, and complications. All of the keywordsare within the Medical Subject Headings (MeSH)database, which represents the controlled vocabu-lary used for indexing articles for Medline andPubMed. The articles with the highest level ofevidence were identified with the consensus of allof the collaborative authors and were criticallyreviewed.
3. Evidence synthesis
3.1. Improving the detection of lymph node metastases in
prostate cancer: critical assessment of currently available
imaging techniques
Currently, none of the standard radiologic techni-ques predicts the presence of LNI accurately whencompared with ePLND. Some innovative techniques,however, might overcome this clinically significantstaging problem in the near future. Reported CTsensitivity for the detection of lymph node metas-tases is typically in the range of about 35% [2]. Thislow sensitivity can be attributed to the fact that alymph node size >1 cm in diameter is required forthe identification of lymph node metastases [2].Similarly, standard MRI, dynamic enhanced MRI,and even magnetic resonance spectroscopic ima-ging (MRSI) have shown no advantage over CT inpredicting the presence of LNI [3–4]. Conversely, theuse of lymphotropic paramagnetic iron oxide nano-particles with a size of 30–50 nm as a contrast agentat MRI (ie, lymphotropic nanoparticle–enhancedMRI [LNMRI]) might improve the detection of nodaldisease [5–7]. Initial results in a group of 30 patientswith genitourinary malignancies demonstrated asignificantly improved sensitivity and specificity of100% and 80%, respectively, for accurately detectingpelvic lymph node metastases [6]. In a more recenttrial in 80 men with clinically localised PCa, LNMRIwas shown to increase the sensitivity for detectinglymph node metastases from 35% when using MRIalone to 90% [5]. Specificity also increased from 90%to 98%, making LNMRI a potentially useful imagingtechnique for preoperative staging of the smallpelvis. Similarly, the sensitivity and negative pre-dictive value (NPV) of magnetic resonance lym-phoangiography (MRL) using ferumoxtran-10 as acontrast agent were as high as 82% and 96%,
respectively, in 375 patients with intermediate- tohigh-risk PCa [7]. These studies, however, havesome limitations which have to be addressed inthe near future before LNMRI will become a routinestaging method for PCa. Patients enrolled in thesetrials underwent a limited PLND (lPLND). An ePLNDwas performed in a few cases only in the presence ofsuspicious lymph nodes outside the boundaries oflPLND. Therefore, the high reported sensitivityand NPV of LNMRI might have been falsely inflatedbecause of the significant understaging associatedwith lPLND [34–41]. Moreover, the conventionalLNMRI has its own limitations. First, in the presenceof fibrosis or lipomatosis within the lymph node, it isdifficult to discriminate benign tissue from cancer.In such cases, there also might be a lack of contrastagent uptake. Second, the reading time required forthis technique is long (several hours per patient),and high interobserver variability can be found.Third, small nodal micrometastases can be missed.To solve these issues, a novel approach consisting ofMRI enhanced with ultrasmall superparamagneticparticles of iron oxide (USPIO) combined withdiffusion-weighted MRI (DW-MRI) has been pro-posed. This approach has been shown to be a fastand accurate method for detecting pelvic lymphnode metastases in patients with prostate and/orbladder cancer, even in normal-sized nodes [9].Similarly, [11C]choline positron emission tomogra-phy (PET)/CT has also been tested recently in thedetection of PCa nodal metastases [8]. Interestingly,this imaging technique showed high accuracy indetecting LNI in intermediate- and high-risk PCapatients treated with ePLND. The sensitivity, speci-ficity, NPV, and number of correctly recognisedcases at PET/CT were 60.0%, 97.6%, 87.2%, and 87.7%,respectively [8].
Sentinel lymphoscintigraphy (SLN) has beendescribed as an imaging staging tool for planningthe necessity and the extent of PLND in patientsundergoing RP. Planar films are taken preopera-tively, and intraoperatively, the use of gamma probefacilitates dissection of all lymph nodes storingthe technecium (99mTc) nanocolloid. This has led tothe concept of laparoscopic or open sentinel lymphnode dissection in PCa, which would eventuallydecrease the rate of unnecessary ePLNDs [42–48].Interestingly, the sensitivity of the radioguidedsentinel lymph node dissection for detectingpatients with positive nodes is extremely high(96%) [42]. This approach, however, has somesignificant limitations. First, in about 5% of patients,no marker is taken up on one pelvic sidewall, andePLND has to be performed [43]. Second, SLN is notable to identify all metastatic lymph nodes either
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 51254
due to the presence of micrometastases with adiameter below the resolution of SLN or due tomacrometastases blocking the lymphatic drainageof 99mTc-nanocolloid into the lymph nodes [47].Indeed, 32% of positive nodes were falsely negative[48]. Third, technecium-containing nodes can onlybe found intraoperatively with the collimator if it isin direct contact with the lymph node.
Single photon emission CT (SPECT) fused with CTor MRI has been shown to improve spatial resolutionand orientation, thus allowing for a more preciselocalisation of 99mTc-containing lymph nodes [49].The procedure, however, is time consuming anddepends on the skills and endurance of the reader.Moreover, experience with this tool is limited, and itcannot overcome the problem of false-negativenodes.
3.2. Importance of the extent of pelvic lymph node
dissection in prostate cancer staging
Several studies have shown that the rate of LNI inPCa patients almost linearly increases with theextent of PLND [34–41]. Indeed, ePLNDs might benecessary to detect occult lymph node metastasesthat would not otherwise be detected by lPLNDs, asPCa nodal metastases do not follow a predefinedpathway of spread [50]; however, what does repre-sent an ePLND in PCa is still a matter of debate. Someauthors consider ePLND to be the removal ofobturator, external iliac, and hypogastric nodes[14,37,39]. Others include the removal of presacralnodes [36,51], which are part of the hypogastricpackage in some series [33,38]. Golimbu et al showedthat the deep presacral–presciatic nodes wereinvolved almost as often as the more superficialexternal iliac-obturator group, which demonstratesthat ePLNDs excluding the presacral region still havea substantial likelihood of overseeing positive nodes[51]. Finally, other authors advocate the additionalremoval of common iliac nodes, at least up to theureteric crossing, on the basis of imaging studies[38,49]. Yet, even in the presence of such extensivenodal dissections, approximately 25% of lymphnodes potentially harbouring PCa nodal metastaseswould not be removed [49]. Regardless of thedefinition used, general agreement has beenreached on the fact that an extended nodal dissec-tion should always include removal of lymph nodesalong the hypogastric artery. Indeed, several studieshave demonstrated that up to 50% of lymph nodemetastases are located in this landing site [38,40,49–52]. Therefore, removal of lymph nodes located inthe obturator fossa alone or in conjunction with thelymphatic tissue along the external iliac vessels
might significantly underestimate the true inci-dence of nodal metastases in PCa. Heidenreich et al[36] as well as Bader et al [38] pioneered a systematicassessment of the concept of PLND extent and LNIrate. Heidenreich et al [36] found twice as manypositive nodes using the extended versus limitedtechnique in a historical control group (26% vs 12%; p< 0.03). Similarly, ePLND with a mean count of 13.1lymph nodes was associated with a 2.8-fold higherLNI rate versus lPLND (mean: 10.1 removed lymphnodes; 11.4% vs 4.1%; p = 0.009) in another recentretrospective laparoscopic series [39]. Interestingly,the rate of false-negative findings associated withlPLND (restricted to external iliac area and obturatorfossa) would have been 19% and 16% in Bader et al’s[38] and Heidenreich et al’s [36] series, respectively;this rate increases up to 60% if only patients withlymph node metastases are considered [38]. Otherinvestigators confirmed these findings [49–52]. Therelationship between PLND extent and the rate ofLNI was also examined by Briganti et al [34,35].These authors showed that the ability correctly topredict the likelihood of LNI increases when thenumber of removed nodes is increased [34]. Inter-estingly, the probability of correctly predicting therate of LNI was close to zero when <10 nodes wereremoved. Conversely, a virtually perfect ability wasreported when �30 lymph nodes were removed.These results seem indirectly to confirm the resultsof an autopsy study which found that an average of20 dissected pelvic lymph nodes can be considered arepresentative sampling that enables exact loco-regional staging of PCa [53]. Taken together, thesedata show that lPLND is associated with a dismalstaging accuracy that is falsely biased towards lowrates of LNI due to inadequate nodal sampling. Theonly prospective randomised study assessing therate of LNI in 123 patients randomly assigned toeither lPLND or ePLND did not find a significantdifference in the rate of LNI between the two surgicalapproaches (3.2% vs 4%, respectively; p = 0.1) [31].This study, however, is flawed by several limita-tions. First, the vast majority of patients includedhad low-risk PCa, which is associated with a low rateof LNI, even in patients treated with ePLND. Second,ePLND was performed on only one side. Third, thefield of ePLND was not defined, and no data aregiven regarding the number of lymph nodesremoved in each group or the pathologic assessmentperformed in detecting lymph node metastases.Fourth, the study was seriously underpowered toallow for a conclusion of noninferiority. Takentogether, these limitations strongly restrict thevalidity of this trial. Therefore, available data seemto support the statement that if PLND is planned in
Table 1 – Available preoperative staging tools predicting the presence of lymph node metastases in prostate cancer
Study No. of patients Predictors Extent ofPLND
Prevalenceof LNI, %
Predictiveaccuracy, %
Cagiannos et al [11] 7014 PSA, clinical stage, biopsy Gleason score Limited 3.7 76
Kattan et al [12] 697 PSA, clinical stage, biopsy Gleason score Limited 8 76.8
Makarov et al [13] 5730 PSA, clinical stage, biopsy Gleason score Limited 1 88
Briganti et al [14] 602 PSA, clinical stage, biopsy Gleason score Extended 11 76
Briganti et al [15] 278 PSA, clinical stage, biopsy Gleason score,
percentage of positive cores
Extended 10.4 83
Bluestein et al [16] 1632 PSA, clinical stage, biopsy Gleason score Limited NA NA
Bishoff et al [17] 481 PSA, clinical stage, biopsy Gleason score Limited 7.7 NA
Narayan et al [18] 932 PSA, biopsy Gleason score Limited 11 NA
Conrad et al [22] 344 No. of positive biopsies, no. of biopsies
containing any Gleason grade 4 or 5 cancer
Limited 8.1 NA
Roach et al [23] 212 PSA, biopsy Gleason score Limited 17 NA
Crawford et al [24] 4133 PSA, clinical stage, biopsy Gleason score Limited NA NA
Batuello et al [25] 6135 PSA, clinical stage, biopsy Gleason score Limited 4.6 81
Han et al [26] 5744 PSA, clinical stage, biopsy Gleason score, age Limited 5 88
Poulakis et al [27] 201 PSA, clinical biopsy Gleason score, and
pelvic coil MRI findings
Limited 10 91
Karam et al [28] 425 PSA, clinical stage, biopsy Gleason score,
preoperative plasma endoglin
Limited 3.3 97.8
Wang et al [29] 411 PSA, clinical biopsy Gleason score, and
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 5 1255
patients with PCa, it should be extended. Thisapproach significantly increases the nodal stagingaccuracy by decreasing the rate of false-negativefindings associated with lPLNDs.
3.3. Critical evaluation of predictive models used to assess
the need for pelvic lymph node dissection in prostate cancer
Several nomograms and predicting tables [10–29]have been developed to predict LNI and to assess theneed for PLND (Table 1). Most of these tools werebased on routinely available variables such aspreoperative prostate-specific antigen (PSA) level,clinical stage, and biopsy Gleason sum. These toolscan identify patients at low risk of LNI and have
Table 2 – Currently available guidelines regarding the need forcancer
contributed to a steep and unrelenting decrease inthe utilisation of routine PLND at RP [54]. All of thesetools, however, except for two [14,15] were devel-oped and validated in patients treated with lPLND.Therefore, despite their apparently high accuracy(range: 76–97.8%; Table 1), they may significantlyunderestimate the true prevalence of LNI due to thelimited nodal sampling. Makarov et al [13] publishedan update of the Partin tables developed to predictpathologic stage (including LNI) using preoperativePSA, clinical stage, and biopsy Gleason score. In thisstudy, LNI rate and predictive accuracy were 1% and88%, respectively. Conversely, lower accuracy wasreported when LNI predictions from the Partintables were validated in a population-based cohort
and the extent of pelvic lymph node dissection in prostate
on for PLND Extent of PLND
T2a, PSA 10–20 ng/ml, biopsy
risk (>cT2b, PSA >20 ng/ml,
te cancer
Extended
for patients with higher
t
Not indicated
patients with <7%
ymph node metastases
some patients with nodal
d. An extended
PLND is performed.
Extended
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 51256
and in European patients, in which accuracies of76% were found [20,21]. Cagiannos et al also reporteda preoperative nomogram aimed at identifyingpatients at low risk of LNI based on PSA, clinicalstage, and biopsy Gleason sum [11]. The bootstrap-corrected accuracy of this model was 76%. None ofthese studies, however, provided the number ofremoved lymph nodes. Moreover, all mainly reliedon lPLNDs, which limits their validity and applic-ability in cohorts treated with ePLND. To circumventthis limitation, Briganti et al developed a nomogrampredicting the rate of LNI in patients who underwentan ePLND at a single high-volume centre [14]. Theirnomogram was 76% accurate and relied on PSA,clinical stage, and biopsy Gleason sum. This nomo-gram represents the first tool based on ePLNDpatients, but it still awaits prospective externalvalidation. Even higher LNI predictive accuracycan be reached if data on tumour volume, suchas percentage of positive cores, are included inmultivariable models and applied to ePLND-treatedpatients [15]. All of these findings were recentlyreviewed and included in the currently available PCaguidelines (Table 2) [1,55,56].
3.4. Is there a need for pelvic lymph node dissection in
low-risk prostate cancer patients?
Several trials have assessed the rate of LNI in low-risk PCa patients treated with either lPLND or ePLND[13,57–62]. Despite a lack of uniformity in definingthe low-risk PCa group, the rate of LNI in lPLNDseries is invariably low, ranging between 0.5 and0.7% [13,56–59]. In the largest low-risk PCa seriesfocussing on patients with cT1 PCa and PSA �6 ng/ml, the rate of LNI was as low as 0.7% [60].These results have been confirmed by the mostrecently updated Partin tables, where the rate of LNIwas <1% in patients with favourable cancer char-acteristics (PSA <10 ng/ml, T1c PCa, and biopsyGleason sum �6) [13]. Similarly, the rate of LNI wasas low as 0.7% in a recent low-risk PCa series(defined by PSA�10 ng/ml, biopsy Gleason score�6,and clinical stage T1 or T2a) [59]. Such negligible LNIrates found in the low-risk group significantlycontributed to a continuous decrease in the rate ofPLND performed in this subset of patients [54];however, all of these studies are biased by theinclusion of patients treated with lPLNDs. Interest-ingly, when considering ePLND series, the rate of LNIseems to increase slightly, even in the low-risk PCagroup [40,61,62]. Weckermann et al reported on aretrospective study in which the rate of LNI was 7.4%among patients with PSA <10 ng/ml and biopsyGleason sum �6 who were treated with ePLND [61].
The rate of LNI was even higher (11%) in a recentstudy by Schumacher et al based on a cohort of 231patients with PSA <10 treated with ePLND [62]. Thisrate, however, significantly decreased to 3% whenonly patients with clinical stage T1–T2 and biopsyGleason score�6 were considered [33]. Similarly, therate of LNI was 5.8% in another ePLND seriesincluding patients with PSA <10 ng/ml, T1c PCa,and biopsy Gleason score �6 [40]. Taken together,these data showed that the overall LNI rate in thelow-risk PCa group (PSA <10, clinical stage T1–T2a,and biopsy Gleason sum �6) never exceeded 8%,even among patients treated with more extensivenodal dissections [13,40,57–62]. Based on the resultsof these studies, all of the available PCa guidelinesdo not routinely recommend a staging PLND in thepresence of these preoperatively favourable PCacharacteristics (Table 2) [1,55,56]. Nevertheless, it isstill unknown whether PLND might confer signifi-cant biochemical recurrence (BCR) survival benefitin low-risk PCa due to the lack of prospectiverandomised trials. Indeed, only a few retrospectivestudies to date have assessed the impact of PLND onthe outcome of low-risk PCa patients. Bhatta-Dharet al [57] compared the BCR-free survival of low-riskpatients not randomly assigned to either lPLND orno lPLND. After a mean follow-up of 60 mo, therewas no difference in 6-yr biochemical failure rates inpatients receiving PLND compared with patients nottreated with lPLND (86% and 88%, respectively;p = 0.28). The authors also re-evaluated the samegroups of patients at a longer follow-up [58]. Again,they did not find any difference in the 10-yr BCR-freesurvival rates between the two groups (83.8% vs87.9%, respectively; p = 0.33). Similarly, in anothermulticentre study, the BCR-free survival rates oflow-risk patients were 81% versus 82% in the no-PLND group versus the PLND group, respectively( p = 0.83) [59]. These results, however, must beinterpreted with caution because the studies werelimited by several scientific flaws. First, all patientshad inadequate nodal dissection in that they weretreated with lPLND (mainly an obturator). Second,the vast majority of the patients enrolled wereprobably at very low risk of dying from progressivedisease, even if left untreated. Third, no standar-dised pathologic assessment of lymph nodes wasperformed. Finally, from a statistical perspective,the number of events was too small to allow for anequivalence study.
Therefore, these data do not formally rule outthe possibility that more extensive PLND mightfavourably affect patient survival, even in the low-risk group. Indeed, a significant inverse associationbetween the number of nodes removed and the rate
Stone et al [30] 189 35.9 vs 2 Extended vs limited* 17.8 vs 9.3
Clark et al [31] 123 8.1 vs 2.4 Extended vs limited NA
Briganti et al [32] 963 18.9 vs 7.3 Extended vs limited 17 vs 7
Heidenreich et al [36] 203 8.7 vs 9 Extended vs limited 28 vs 11
Bader et al [38] 365 2.1 Extended 21^
Jeschke et al [43] 71 7 Extended* NA
Schumacher et al [62] 122 4.8 Extended 22^
Herrell et al [63] 68 20 Limited 9.2
Keller et al [64] 90 7.8 Extended 19
Wyler et al [65] 123 4 Extended* 21
Pepper et al [66] 260 3.5 Extended NA
McDowell et al [67] 217 22 Extended NA
Paul et al [68] 150 51 Extended NA
NA = not available.^ Median number of lymph nodes removed.
* Laparoscopic series.
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 5 1257
of BCR has been reported in node-negative patients[41]. Future prospective randomised trials includingpatients treated with ePLND are needed to confirmthese preliminary, potentially biased findings.
3.5. Complications of pelvic lymph node dissection
Surgeons are often deterred from performing anePLND because of the potentially high incidence of
complications. When the cumulative PLND com-plication literature is examined, the rate of compli-cations ranges from 2% to 51% (Table 3) [30–32,36,38,45,63–70], but controversies exist withregard to the rate of PLND-related complicationsaccording to the extent of PLND. Clark et al found anincreased risk of complications attributable to PLNDon the side of extended dissection [31]. Stone et al[30] also reported a strikingly higher complicationrate when they compared laparoscopic ePLND withlaparoscopic lPLND (35.9% vs 2%; p < 0.001). Thelargest contemporary series (n = 963) addressingcomplications after PLND showed that in patientstreated with ePLND, the overall rate of complica-tions was 19.8% versus 8.2% in those treated withlPLND ( p< 0.001) [32]. Alternatively, when individualPLND complications were assessed, only the rate oflymphocele was significantly higher in patientssubjected to ePLND (10.3% vs 4.6%, respectively;p = 0.01). Complications were not invariably high inall ePLND series, as evidenced by Bader et al [38]: Inthis study, an overall complication rate requiringprolonged hospitalisation of 2.1% was recorded.Conversely, a higher complication rate (8.8%) wasreported by Heidenreich et al [36]. Nevertheless, thefrequency and severity of intra- and perioperativecomplications did not differ significantly between
the lPLNDs and the ePLNDs (9% vs 8.7%, respec-tively). Despite the presence of discordant results inthe literature, all of these data seem to suggest thatPLND may not be an entirely innocuous procedure,even in the hands of the most experienced surgeons.To minimise PLND-related morbidity, some keysteps need to be followed. Heidenreich et al [40]suggested that all lymphatics lateral to the externalartery should be saved. Additionally, the distal endsof the lymphatics should be either ligated or clippedwith small clips that exert a higher pressure on thelymphatic vessels than large clips. Two drainsshould also be placed in each side of the pelvisand left in place until <50 ml/d is drained. Finally,low-molecular heparin should be injected into theupper arm. Although it seems logical to think thatsurgical expertise may reduce PLND-associatedmorbidity, this concept still needs to be confirmedin methodologically sound studies.
3.6. Impact of pelvic lymph node dissection on prostate
cancer outcome
The issue of whether PLND might affect PCaoutcome has been an argument of extreme interestin the urologic community. Unfortunately, thequestion remains unanswered because of the lackof prospective randomised trials. Moreover, theimpact of PLND on cancer outcomes remainscontroversial, even in retrospective studies. Mas-terson et al [41] found a significant inverse associa-tion between the number of removed lymph nodesand BCR-free survival in node-negative patients( p = 0.01). These results might be attributable to theremoval of micrometastatic nodal disease, whichmay support the therapeutic role of PLND in this
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 51258
patient category. Patients with nodal micrometas-tases would be those who are more likely to receive apossible curative benefit from PLND. This hypoth-esis is still pending definitive approval, since noimmunohistochemistry evaluation aimed at identi-fying occult nodal disease has been performed in thestudy.
Another retrospective trial found a significantassociation between the extent of PLND and cancer-specific survival. Interestingly, patients undergoingremoval of at least four lymph nodes (node-positiveand node-negative patients) or >10 nodes (onlynode-negative patients) had a lower risk of PCa-specific death at 10 yr compared with patients whodid not undergo PLND [71]. The main limitation ofthis multicentre study is the lack of an homoge-neous and standardised pathologic assessment ofthe removed lymph nodes, which is key fordetermining reliable nodal counts. In contrast, DiMarco et al [72] found no survival benefit associatedwith an increasing number of removed lymph nodesin node-negative patients in a large, single-institu-tion series collected over a 13-yr time span. Patientswho underwent surgery at the beginning of theseauthors’ experience had more nodes removed andshowed an oncologic outcome similar to patientsoperated on 10 yr later. Taking the stage-shift intoaccount, patients operated on earlier should havehad poorer outcomes; as this is apparently not so,one might hypothesise a beneficial role for PLND.This possibility, however, cannot be considered asmore than food for thought. We feel that thequestion of whether a meticulous nodal dissectioncan have an impact on node-negative PCa still needsto be elucidated.
Furthermore, it should be acknowledged that thepositive association between PLND extent andcancer outcome in node-negative patients mightbe based on a misinterpretation of these data causedby the Will Rogers phenomenon [73,74], a well-knownphenomenon in the medical literature. Will Rogers(1879–1935), the great American humorist, drewattention to the apparent mathematical paradoxthat the movements of elements from one set toanother can increase the average value of both sets.In medicine, the Will Rogers phenomenon describesan apparent improvement in outcome for groups ofpatients with no actual improvement for anyindividual patient [73]. In the context of PLND, ifthe number of removed negative lymph nodes isinvestigated as a prognosticator, it is clear thatpatients treated with ePLND have a higher likelihoodof being really node negative without overlookedmetastases. If a patient has a positive node in an areathat is covered by an extended dissection but not by a
limited dissection, this patient is excluded from theanalyses in the group of ePLND patients (as he is nodepositive, and only node-negative patients are left inthe analyses) but is included in the group with alimited dissection. This means that different groupsare compared at a certain disease stage, and thebenefit of the group with an extended dissection canbe explained by the different disease stages. In otherwords, after a limited dissection, the likelihood ofoverlooked metastases is higher, and it is theseoverlooked positive nodes, instead of the removal ofnegative nodes, that influence the prognosis [73,74].Similar results can be achieved when consideringonly patients with positive nodes. Indeed, in patientsin whom many nodes are removed, the incidence offinding positive nodes would be high, and theoutcome of these patients would be relatively goodbecause many patients would have only small-volume metastatic disease. At the same time, whencomparing node-positive patients between a serieswith ePLND or lPLND, the patients with positivenodes would again have a much better outcome inthe series with ePLND because they would containthe patients who had small nodal disease. Theseobservations suggest that the only solution toanswering the question of whether or not removalof the lymph nodes has a role beyond diagnosticpurposes is to conduct a prospective randomised trialin which patients are randomised to either no PLNDor ePLND.
Even in the absence of well-designed trials, dataavailable from large series of patients undergoingPLND have shown that the long-term outcome ofsurgically treated patients with LNI is not invariablypoor (Table 4) [75–85]. Bader et al [76] reported aremarkable 74% 5-yr cancer-specific survival rate in asmaller cohort of patients treated with ePLND and RPand with no adjuvant treatment. Data from the samegroup reported by Schumacher et al indicated a 60%cancer-specific survival rate at 10-yr follow-up [82].Cheng et al [77] reported a 79% 10-yr cause-specificsurvival in a large series of 322 patients treated withRP. Of these patients, 92% received prolongedadjuvant androgen deprivation therapy (ADT). Boor-jan et al [78] recently updated the same institution’sseries, which included 505 patients treated with RPand PLND. Again, roughly 90% of those patientsreceived ADT; the 10-yr cancer-specific survival ratewas as high as 85.8%. In another series of 100 node-positive patients, the 5- and 10-yr disease-specificsurvival rates were 94% and 75%, respectively [83].Interestingly, in the largest node-positive seriesavailable (n = 703) including patients treated with amultimodal, combined approach, the 15-yr cancer-specific survival rate was 78% [79]; however, when
Table 4 – Outcome of patients with lymph node metastases treated with radical prostatectomy (RP) and pelvic lymph node dissection (PLND) with or without adjuvanttreatments in the prostate-specific antigen (PSA) era
^ Disease progression defined by elevation of serum PSA >0.4 ng/ml after surgery, development of local recurrence, or distant metastasis documented by biopsy or radiographic examination.
* 7-yr BCR-free survival rate.
** Freedom from any (systemic, local, or biochemical) progression probabilities.
eu
ro
pe
an
ur
ol
og
y5
5(2
00
9)
12
51
–1
26
51
25
9
Ta
ble
5–
Infl
uen
ceo
fth
eex
ten
to
fn
od
al
inv
asi
on
on
the
ou
tco
me
of
pa
tien
tsw
ith
lym
ph
no
de
meta
sta
ses
trea
ted
wit
hra
dic
al
pro
sta
tect
om
y(R
P)
an
dp
elv
icly
mp
hn
od
ed
isse
ctio
n(P
LN
D)
wit
ho
rw
ith
ou
ta
dju
va
nt
trea
tmen
tsin
the
pro
sta
te-s
peci
fic
an
tig
en
(PS
A)
era
Stu
dy
No
.o
fp
ati
en
tsw
ith
LN
I
Pa
tien
tch
ara
cteri
stic
sM
ed
ian
foll
ow
-u
p,
yr
Ad
juv
an
tth
era
py
Ca
nce
r-sp
eci
fic
surv
iva
lB
CR
-fre
esu
rviv
al
Meta
sta
sis-
free
surv
iva
l
5y
r10
yr
5y
r10
yr
5y
r10
yr
Da
nesh
ma
nd
et
al
[75]
235
<20%
vs�
20%
LN
D11.4
31%
of
pts
––
––
86%y
vs
60%y
72%
vs
47%
Ba
der
et
al
[76]
92
1v
s2
vs�
2
po
siti
ve
no
des
3.7
5N
o93%y
vs
60%y
vs
63%y
84%y *
vs
60%y *
vs
27%y *
32%y
vs
23%y
vs
21%y
14%y *
vs
12%y *
vs
7%y *
52%y
vs
52%y
vs
35%y
35%y *
vs
35%y *
vs
20%y *
Ch
en
get
al
[77]
322
0v
s1
po
siti
ve
no
de
6.3
92%
of
pts
99.3
%v
s99%
97%
vs
94%
77%
^v
s79%y ^
59%
^v
s66%y ^
––
Bo
orj
ian
et
al
[78]
507
0v
s1
vs�
2
po
siti
ve
no
des
10.3
89.7
%o
fp
ts99%
vs
97%
vs
90%
98%
vs
90%
vs
79%
79%
vs
71%
vs
67%
68%
vs
57%
vs
54%
98%
vs
94%
vs
85%
95%
vs
86%
vs
73%
Bri
ga
nti
et
al
[79]
703
�2
vs>
2
po
siti
ve
no
des
9.4
100%
of
pts
93%
vs
81%
85%
vs
73%
75%
vs
60%
61%
vs
45%
––
Sch
um
ach
er
et
al
[82]
122
1v
s2
vs>
2
po
siti
ve
no
des
5.6
No
95%y
vs
100%y
vs
69%y
72.1
%v
s79.1
%
vs
33.4
%
24.7
%v
s11.8
%
vs
4.9
%
11%y *
vs
2%y *
vs
3%y *
––
Pa
lap
att
uet
al
[86]
143
�15%
vs<
15%
LN
D6
No
––
33%y
vs
10%y
25%y *
vs
3%y *
––
BC
R=
bio
chem
ica
lre
curr
en
ce;
LN
D=
lym
ph
no
de
den
sity
(nu
mb
er
of
po
siti
ve
lym
ph
no
des
ov
er
tota
ln
um
ber
of
lym
ph
no
des
rem
ov
ed
);LN
I=
lym
ph
no
de
inv
asi
on
;p
ts=
pa
tien
ts.
yA
pp
rox
ima
tely
.
^D
isea
sep
rogre
ssio
nd
efi
ned
by
ele
va
tio
no
fse
rum
PS
A>
0.4
ng/m
la
fter
surg
ery
,d
ev
elo
pm
en
to
flo
cal
recu
rren
ce,
or
dis
tan
tm
eta
sta
sis
do
cum
en
ted
by
bio
psy
or
rad
iogra
ph
icex
am
ina
tio
n.
*7-y
rfo
llo
w-u
p.
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 51260
considering BCR-free survival rates of node-positivepatients not receiving adjuvant ADT, outcomes seemto be poorer. This finding is expected, since it is wellknown that not all patients with BCR will eventuallydie from prostate cancer. Palapattu et al [86] reportedon a 16% BCR-free survival rate at a median follow-upof 6 yr in patients undergoing RP left untreated untilprogression. Similar results were reported by Mas-terson et al, who showed roughly 19% BCR-freesurvival rates in 175 node-positive patients notreceiving adjuvant therapy [41]. Several studies,however, have shown that not all node-positivepatients are at the same risk of PCa progression anddeath [75–79,82,86,87]. Indeed, patients with a lowvolume of nodal disease have significantly highersurvival rates compared with patients with a highervolume of LNI, regardless of adjuvant treatmentadministration (Table 5) [75–79,82,86]. Daneshmandet al [75] reported on a large retrospective study inwhich patients with one and two positive lymphnodes had an overall survival rate of 94% and 96% at 5yr and 75% and 74% at 10 yr, respectively. Interest-ingly, when stratified by lymph node density (LND),patients with an LND �20% were at higher risk forclinical recurrence compared with those with adensity of <20% (relative risk: 2.31; p < 0.001).Similarly, Palapattu et al found an LND cut-off of15% as a significant predictor of cancer progression(p < 0.0001) [86]. Interestingly, Cheng et al showedthat the 10-yr cancer-specific survival rate was ashigh as 94% in patients with a single node metastasis[77]. This rate was not significantly different from thecancer-specific survival of patients without nodalinvolvement. Similar results seem to be achieved inseries including patients not receiving any adjuvanttreatment. In a study by Bader et al, 39% of patientswith one positive node remained free of clinical orbiochemical progression, compared to 12% ofpatients with two or more positive nodes [76].Schumacher et al recently confirmed these findingsand reported significantly higher 10-yr cancer-spe-cific survival rates in patients with two or fewerpositive nodes compared with patients with three ormore positive nodes not receiving any adjuvanttherapy (78.6% vs 33.4%, respectively) [82]. The cut-off of two positive nodes has also been recentlystudied in a large multi-institutional node-positiveseries ofpatients treated with a multimodalapproach(n = 703) [79]. Patients with two or fewer positivenodes had significantly better cancer-specific survi-val outcome at 15-yr follow-up compared withpatients with more than two positive nodes (84% vs62%; p < 0.001). Patients with more than two positivenodes had a 1.9-fold higher risk of dying of PCacompared with patients with two or fewer positive
e u r o p e a n u r o l o g y 5 5 ( 2 0 0 9 ) 1 2 5 1 – 1 2 6 5 1261
nodes after accounting for all the other predictors(p = 0.002). Moreover, a significant improvement incancer-specific survival prediction was reachedwhen the number of positive nodes was considered[79]; however, the evidence of increased survival ofpatients with low-volume nodal invasion might beexplained by a lead-time bias.
Furthermore, the optimal postoperative manage-ment of patients with nodal metastases is stillcontroversial. Indeed, although a well-designedprospective randomised trial showed a positiveeffect of adjuvant ADT in node-positive patients ofwhom the majority also had positive margins andseminal vesicle invasion [84], it is possible that notall patients with nodal metastases, namely thosewith minimal nodal disease and a slow PSA doublingtime, might benefit from adjuvant ADT [82]. Patientswith a low volume of LNI accurately staged withePLND indeed eventually might be considered forwatch-and-wait protocols, which would reduce therisk of overtreatment of patients at lower risk forcancer progression. Moreover, a recent retrospectivestudy has shown a positive impact of adjuvantradiotherapy in patients with nodal metastases [88].Future prospective studies are needed to clarifythese issues.
Taken together, all of these data show that theimpact of PLND as a curative treatment remainsan open question. Nevertheless, some authorssuggest that the extent of PLND in and of itselfmight have a beneficial effect on symptomaticprogression and PCa-specific survival [41,71].Unfortunately, these assumptions are based onretrospective, uncontrolled trials; nonetheless, itmay be the case that some patients may havebenefited from the removal of micrometastasesthat are eventually only detectable at a molecularlevel. Only future prospective randomised trialscomparing the effect of PLND versus no PLND inhigh-risk patients definitely would assess the roleof PLND on PCa outcomes. In view of thesubstantial amount of indirect evidence thatePLND may benefit, if not cure, particularly thosepatients with low volume of nodal disease, suchstudies are hardly feasible.
4. Conclusions
A number of conclusions can be drawn from thisreview. First, PLND remains the most accurate andreliable approach for detecting the presence oflymph node metastases in PCa. If a PLND isplanned at the time of RP, it should be extended.Increasing the extent of lymph node dissection
results in a more accurate assessment of LNI and ahigher rate of nodal metastases. Limited PLND isassociated with a high rate of false-negativefindings. Second, the downside of more extensivePLND consists of a higher rate of complications, asreported in some studies. Specifically, the rate oflymphoceles might be higher but the higher rate ofcomplications associated with ePLND has notalways been confirmed. Third, previous toolspredicting the rate of LNI are based mostly onlPLND and thus are of limited value. Fourth, therate of LNI is low (<8%) in patients with low-riskPCa (defined as clinical stage T1/T2a, biopsyGleason sum �6, and PSA <10 ng/ml). Despitethe absence of prospective randomised trialsassessing the impact of ePLND in this patientcategory, a staging ePLND might be spared inpatients with low-risk PCa. Fifth, no data fromprospective randomised studies indicate thatthe extent of PLND improves cancer control orsurvival; however, outcome of surgically treatednode-positive patients is not invariably poor. Theextent of lymph node involvement (namely, thenumber of positive lymph nodes) is one of thestrongest predictors of cancer-specific survival.Patients with lymph node metastasis and lownodal burden show excellent long-term outcomes,regardless of the administration of adjuvanttreatments.
Author contributions: Alberto Briganti had full access to all the
data in the study and takes responsibility for the integrity of
the data and the accuracy of the data analysis.
Study concept and design: Briganti, Blute, Eastham, Graefen,
Heidenreich, Karnes, Montorsi, Studer.
Acquisition of data: Briganti, Blute, Eastham, Graefen, Heiden-
reich, Karnes, Montorsi, Studer.
Analysis and interpretation of data: Briganti, Blute, Eastham,
Graefen, Heidenreich, Karnes, Montorsi, Studer.
Drafting of the manuscript: Briganti, Blute, Eastham, Graefen,
Heidenreich, Karnes, Montorsi, Studer.
Critical revision of the manuscript for important intellectual content: