- Email: [email protected]
Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer
EUF-133; No. of Pages 8 EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
available at www.sciencedirect.com journal homepage: www.europeanurology.com/eufocus
Platinum Priority – Prostate Cancer Editorial by XXX on pp. x–y of this issue
Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer Philipp Mandel a,b, Maximilian C. Kriegmair c, Katharina Bogdan a, Katharina Boehm b, Lars Buda¨us a, Markus Graefen a, Hartwig Huland a, Derya Tilki a,b,* a
Martini-Klinik Prostate Cancer Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany;
b
Department of Urology, University Hospital
c
Hamburg-Eppendorf, Hamburg, Germany; Department of Urology, University Medical Centre Mannheim, Mannheim, Germany
Article info
Abstract
Article history: Accepted February 28, 2016
Background: While the diagnostic value of pelvic lymph node dissection (PLND) during radical prostatectomy (RP) is undisputed, its therapeutic benefit remains a matter of debate. Objective: To investigate the association between total and positive lymph node (LN) counts and oncological outcomes in pN1 patients at RP. Design, setting, and participants: The records of 706 LN-positive patients undergoing RP and PLND between 1998 and 2012 in a single center were analyzed. The median follow-up was 47.8 mo. Intervention: RP and PLND. Outcome measurements and statistical analysis: The number of harvested and positive LNs and their influence on biochemical recurrence-free survival, metastasis-free survival, and cancer-specific survival was evaluated using univariate and multivariate Cox regressions. The number of harvested LNs was used as a continuous and dichotomous variable with a cut-off of 13 LNs. Results and limitations: The mean (median) number of removed LNs was 15.0 (13) and the mean (median) number of positive LNs was 2.4 (1). In multivariate analysis, the number of removed LNs did not significantly influence oncologic outcomes (biochemical recurrence, metastasis-free survival, or cancer-specific survival). The occurrence of metastasis and cancer-specific mortality significantly increased with higher number of positive LNs. The main limitation was the retrospective nature of the study. Conclusions: While a higher number of positive LNs were significantly associated with worse oncological outcomes, the number of removed LNs was not a significant predictor. Patient summary: We found that only the number of positive lymph nodes (LNs) but not the total number of removed LNs was a factor adversely influencing oncological outcomes in LN-positive patients undergoing radical prostatectomy. # 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Associate Editor: James Catto Keywords: Prostate cancer Radical prostatectomy Lymph node dissection Positive lymph nodes
* Corresponding author. Martini-Klinik Prostate Cancer Center, Martinistrasse 52, 20246 Hamburg, Germany. Tel. +49 4074100; Fax: +49 4074100. E-mail address: [email protected] (D. Tilki).
1.
Introduction
Prostate cancer is the most frequent cancer among men in Europe and other developed countries and among the
leading causes of death [1,2]. For its treatment, radical prostatectomy (RP) is one of the recommended treatment options for all levels of risk. Pelvic lymph node dissection (PLND) is the most precise staging procedure for possible
http://dx.doi.org/10.1016/j.euf.2016.02.018 2405-4569/# 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 2
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
lymph node (LN) metastasis; however, its therapeutic benefit concerning oncologic outcome remains controversial [3–6]. According to the guidelines of the European Urological Association (EAU), whenever a PLND is indicated, an anatomically extended PLND should be performed [7]. To present, Level IB evidence from randomized controlled trials on the oncological role of PLND is lacking. The Prospective Study to Compare a Limited Versus Extended Pelvic Lymphadenectomy During Prostatectomy (NCT01555086) is currently recruiting participants. Several retrospective studies could not show that removing more LNs during RP results in improved oncological outcome in patients with pN0 [8], pN0 or pN1 [9], and pN1 [10] only. In contrast, Abdollah et al. [11] recently reported that a higher number of removed LNs is an independent predictor of a more favorable cancer-specific survival (CSS) in LN-positive patients [11]. All patients included in the study received immediate adjuvant hormonal treatment (ADT). This study was followed by a controversial discussion on potential biases and confounders involved in this type of analysis, including the Will Rogers phenomenon possibly explaining the results [12,13]. It remains questionable whether a higher number of removed LNs, regardless if these additional removed LNs harbor metastases or not, impacts oncologic outcome. The aim of the present study was to analyze the influence of the extent of PLND according to the number of removed LNs in pN1 patients at RP on biochemical recurrence (BCR), metastasis, and cancer-specific mortality (CSM) in a large single center and to add further information to the controversial ongoing discussion using a dataset of more than 700 LN-positive patients. 2.
Materials and methods
treatment was defined as immediate treatment administered within 6 mo after RP with an undetectable PSA (< 0.2 ng/ml), while salvage treatment was defined as treatment administered at a postoperative PSA 0.2 ng/ml.
2.2.
Surgical procedure and histologic evaluation
RP was performed using an open retropubic approach or robotic-assisted laparoscopic approach by high volume surgeons, as previously described [14]. Pathological outcome was assessed using the American Joint Cancer Committee 2002 staging system and tumor grading was classified using the Gleason Grading system [15]. From 2004, the template of PLND was in accordance with the EAU guidelines for anatomically extended PLND and included removal of the nodes overlying the external iliac artery and vein, the nodes within the obturator fossa located cranially and caudally to the obturator nerve, and the nodes medial and lateral to the internal iliac artery [7]. Before 2004, bilateral PLND was not uniformly performed with extended template and mainly included the region of the obturator fossa and along the external iliac artery. As the pathologic examination of LNs has changed during the study period due to advances in LN metastasis detection via the introduction of complete embedding of fat tissue, an increase in the detection of harvested LNs may have occurred. We introduced dummy variables to control for both the change in PLND template and change in pathologic examination in all multivariate regressions. To control for further unobserved time-dependent variation we included the year of surgery in all regressions.
2.3.
Statistical analyses
To check for differences in our respective subsamples Pearson’s Chi-square tests and Kruskall-Wallis tests were performed. To evaluate the association between removed lymph nodes and oncologic outcome (MFS and CSS) Kaplan-Meier curves and uniand multivariate Cox proportional hazard models adjusting for pre- and postoperative prognosticators were assessed. These prognosticators included age at surgery, preoperative PSA, pathologic Gleason score, pT-stage, surgical margin status, number of positive LNs, year of surgery,
2.1.
Patient population
neoadjuvant treatment, adjuvant ADT, salvage ADT, adjuvant RT, and salvage RT. Statistical software STATA was used (version 14 for
From our institutional Martini-Klinik database, we identified 969 out of
Windows, StataCorp LP, College Station, TX, USA). LN count was used
15 947 patients who underwent RP at our institution between January
as both, continuous and dichotomous variable. For dichotomization all
1998 and December 2012 with positive LNs in the final pathology.
cut-offs between 5 and 25 removed LNs were analyzed. As all cut-off
Information on the number of harvested LNs at PLND was missing in
values of removed LNs showed only insignificant differences on
75 men. Moreover patients who were lost to follow up within 24 mo
oncological outcomes, we presented the results at a cut-off value of
after RP were excluded (188), which left 706 patients for our analysis.
13 (median of removed LNs).
Furthermore, for analysis of BCR, patients with ADT were excluded from regressions (242 patients). All data were collected prospectively into our review board-approved database. Patients were stratified according to
3.
Results
3.1.
Patient and tumor characteristics
the number of removed LNs. Indication to perform PLND was in accordance to respective EAU guidelines applicable at the relevant time and partly based on an individual approach [7]. BCR was defined as a prostate-specific antigen (PSA) level 0.2 ng/ml and rising after RP in patients without ADT. Metastasis was confirmed with imaging. The cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone. After RP, patient follow up consisted of periodical PSA testing and digital rectal examination. Postoperative imaging studies to detect local recurrence or systemic spread were performed according to PSA level, patients’ preferences, and further symptoms. Adjuvant hormonal and/ or radiation therapy (RT) was recommended based on further patient and tumor characteristics and according to respective EAU guidelines applicable at the relevant time [7]. Adjuvant
Patient and tumor characteristics of the entire cohort and stratified by the number of removed LNs (high LN count [> 13] and low LN count [ 13]) are summarized in Table 1. In the entire population of LN positive patients, the mean (median) number of removed LN was 15.0 (13) and the mean (median) number of positive LN was 2.4 (1). Next to number of removed LNs and number of positive LNs, LN density (ratio of the number of metastatic LNs to the total number of LNs removed; p < 0.001) and Gleason score (p = 0.002) were significantly different between patients with a high and low LN count.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 3
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 1 – Demographics and tumor characteristics stratified by number of removed lymph nodes Entire cohort (n = 706) Age (yr), mean PSA (ng/ml), mean Removed LNs Mean Median IQR Positive LNs Mean Median IQR LN density Mean Median IQR Gleason score 7 8 Pathologic stage pT2 pT3a pT3b pT4 Surgical margins Negative Positive Hormonal therapy Neoadjuvant Adjuvant Salvage Radiation therapy Adjuvant Salvage
64.5 21.0
LNs removed 13 (n = 376)
LNs removed >13 (n = 330)
64.4 19.0
15.0 13 9–19
64.6 23.3
8.6 9 6–11
p value
0.448 0.068 <0.001
22.3 20 16–26 <0.001
2.4 1 1–3
1.8 1 1–2
3.1 2 1–3
0.189 0.133 0.083–0.250
0.223 0.167 0.111–0.272
0.151 0.087 0.053–0.174
<0.001
0.002 62.6% 37.4%
68.8% 31.2%
55.6% 44.4%
6.9% 22.7% 67.3% 3.1%
6.9% 25.8% 63.6% 3.7%
7.0% 19.1% 71.5% 2.4%
49.5% 50.5%
52.8% 47.2%
45.7% 54.3%
14.0% 20.5% 33.0%
13.3% 20.7% 34.3%
14.9% 20.3% 31.5%
0.554 0.885 0.431
23.6% 32.9%
23.7% 30.3%
23.6% 35.8%
0.992 0.125
0.159
0.068
IQR = interquartile range; LN = lymph node; PSA = prostate-specific antigen.
The overall percentage of patients undergoing hormonal treatment after RP was 53.5% (20.5% adjuvant and 33.0% salvage hormonal treatment). The overall rate of RT was 56.5%, whereas patients underwent salvage RT more often (32.9%) than adjuvant RT (23.6%). 3.2.
Oncological outcome
The median follow-up was 47.8 mo (interquartile range: 27.3–68.8). Within the follow-up period 75.0%, 22.4%, and 6.2% of the patients showed BCR, metastasis, or cancerspecific death, respectively. In univariate analysis, the number of removed LNs was significantly associated with BCR (hazard ratio [HR] 1.015, p = 0.019) and occurrence of metastasis (HR 1.023, p = 0.026; Table 2). In contrast, in multivariate analysis, the number of removed LNs did not significantly influence oncologic outcomes (BCR, MFS, and CSM). When stratifying patients by LN count (> 13 vs 13), patients with 13 or less removed LNs had a significantly better BCR-free (p = 0.003) and MFS (p = 0.006), while there was no difference in CSS between both patient groups (Fig. 1A–C). In multivariate analysis this association was not significant. The number of positive LNs was significantly associated with BCR in univariate analysis (HR: 1.091, p < 0.001), but not in multivariate analysis (HR: 1.029, p = 0.194; Table 2).
Occurrence of metastasis and cancer-specific mortality significantly increased with a higher number of positive LNs in both uni- (HR: 1.157, p < 0.001 for metastasis and HR: 1.187, p < 0.001 for CSM) and multivariate (HR: 1.110, p < 0.001 for metastasis and HR: 1.137, p = 0.001 for CSM; Tables 3 and 4) analyses. Oncologic outcome stratified by the number of positive LNs is depicted in Figure 2A–C. A higher LN density was a significant factor adversely affecting BCR-free survival (HR: 2.84, p = 0.001), MFS (HR: 7.60, p < 0.001), and CSS (HR: 11.67, p < 0.001) in univariate and MFS (HR: 4.01, p = 0.001) and CSS (HR: 4.71, p = 0.014) multivariate regressions. 4.
Discussion
While the diagnostic value of a LN dissection during RP is undisputed, its therapeutic benefit still remains a matter of debate. The evidence with regards to the impact of a higher number of removed LNs during PLND on oncological outcomes in prostate cancer patients undergoing RP is limited [12,13]. Several retrospective studies have been published on the topic, which have shown conflicting results. Masterson et al. [9] reviewed 4611 patients undergoing RP and showed that an increased number of removed LNs was associated with higher BCR-free probabilities in pN0 patients, but not in the
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 4
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 2 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Biochemical recurrence Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
1.001 1.011 1.392
0.98–1.02 1.01–1.02 1.23-1.57
0.936 <0.001 <0.001
1.000 1.008 1.244
0.98–1.02 1.00–1.01 1.09–1.42
0.945 0.001 0.001
Ref. 1.973 1.004
1.54–2.53 0.32–3.17
<0.001 0.995
1.511 0.509
1.14–2.00 0.16–1.66
0.004 0.263
Ref. 1.721 1.015 1.091 1.048
1.36–2.18 1.00–1.03 1.05–1.13 1.01–1.09
<0.001 0.019 <0.001 0.013
1.351 1.011 1.029 0.992
1.04–1.76 1.00–1.03 0.99–1.07 0.91–1.09
0.026 0.153 0.194 0.872
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
1.00
B
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
A
without PLND between 1988 and 1991 with a minimum follow-up of 10 yr. In line with Masterson et al. [9], the authors reported that pN0 patients with removal of at least 10 LNs had a lower CSM compared with patients without a
1.00
entire group. The median number of removed LNs was nine in this study [9]. Joslyn and Konety [16] studied the data of almost 13 000 patients from the Surveillance, Epidemiology, and End Results (SEER) database undergoing RP with or
0
50
Analysis time
150
50
0
More than 13 LNs removed
150
13 or less LNs removed
More than 13 LNs removed
0.00
0.25
0.50
0.75
C
100
Mo after surgery
1.00
13 or less LNs removed
100
0
50
100
150
Mo after surgery 13 or less LNs removed
More than 13 LNs removed
Fig. 1 – Survival stratified by number of lymph nodes (LNs) removed in (A) biochemical recurrence-free survival (= 13 vs > 13; p = 0.003), (B) metastasisfree survival (= 13 vs > 13; p = 0.006), and (C) cancer-specific survival (= 13 vs > 13; p = 0.760).
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 5
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 3 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Occurrence of metastasis Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery Adjuvant treatment No Yes Salvage treatment No Yes
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
0.967 1.005 1.625
0.94–0.99 1.00–1.01 1.36–1.95
0.009 0.091 <0.001
0.972 1.000 1.475
0.95–1.00 0.99–1.01 1.20–1.82
0.040 0.932 <0.001
Ref. 2.795 6.542
1.74–4.48 2.89–14.8
<0.001 <0.001
1.959 4.459
1.17–3.27 1.82–10.9
0.010 0.001
Ref. 1.464 1.023 1.157 1.070
1.03–2.08 1.00–1.04 1.11–1.20 1.00–1.15
0.034 0.026 <0.001 0.054
1.006 0.997 1.110 1.047
0.67–1.52 0.97–1.02 1.06–1.17 0.89–1.23
0.982 0.809 <0.001 0.566
Ref. 0.801
0.55–1.16
0.238
0.587
0.37–0.92
0.021
Ref. 1.504
1.06–2.13
0.021
1.313
0.89–1.94
0.173
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
PLND and therefore unknown pN status. Bivalacqua and colleagues [6] analyzed the data of 94 LN-positive patients undergoing anatomically extended or limited PLND at the time of RP and showed that extended PLND was associated with a trend toward improved BCR-free survival (p = 0.064), which was particularly seen in the group of patients with <15% positive LNs (p = 0.016). Moreover, the authors found a significant improvement in MFS in the group of patients with extended PLND (p = 0.045), while no difference was seen in CSS. The study was limited by its small sample size
and the fact that RP was performed by only two surgeons, whereby one routinely performed a limited PLND, while the second performed an extended PLND. Using a large single center dataset with more than 700 pN1 patients at RP, we could not show a significant positive effect of a higher number of LNs removed on oncological outcomes (BCR, MFS, or CSM) in uni- and multivariate analyses, neither when using the number of removed LNs as a continuous nor as a dichotomous variable. Moreover, a variation of the cut-off for high versus low
Table 4 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Cancer-specific mortality Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery Adjuvant treatment No Yes Salvage Treatment No Yes
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
1.005 0.992 2.412
0.96–1.05 0.97–1.01 1.74–3.34
0.845 0.435 <0.001
1.004 0.987 2.133
0.95–1.06 0.97–1.01 1.42–3.20
0.872 0.211 <0.001
Ref. 4.780 19.82
1.69–13.5 5.74–68.4
0.003 <0.001
2.393 5.977
0.78–7.30 1.45–24.7
0.125 0.013
Ref. 2.239 1.019 1.187 0.998
1.20–4.19 0.98–1.06 1.12–1.26 0.89–1.11
0.012 0.325 <0.001 0.975
1.580 0.999 1.137 0.970
0.76–3.28 0.95–1.06 1.05–1.23 0.73–1.29
0.220 0.982 0.001 0.838
Ref. 0.912
0.48–1.72
0.776
0.465
0.18–1.19
0.110
Ref. 1.032
0.57–1.88
0.918
1.073
0.44–2.64
0.879
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 6
1.00
B
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
A
1.00
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
0
50
Analysis time
150
0
50
2 positive LNs 4 or more positive LNs
150
1 positive LN 3 positive LNs
2 positive LNs 4 or more positive LNs
0.00
0.25
0.50
0.75
C
100
Mo after surgery
1.00
1 positive LN 3 positive LNs
100
0
50
100
Mo after surgery
1 positive LN 3 positive LNs
150
2 positive LNs 4 or more positive LNs
Fig. 2 – Survival stratified by number of positive lymph nodes (LNs) in (A) biochemical recurrence-free survival (1 vs 2 vs 3 vs 4+; p < 0.001), (B) metastasis-free survival (1 vs 2 vs 3 vs 4+; p < 0.001), and (C) cancer-specific survival (1 vs 2 vs 3 vs 4+; p = 0.001).
counts of removed LNs did not qualitatively alter our results. In contrast, both higher number of positive LNs and higher LN density had a negative impact on oncological outcomes. Patients with more removed LNs had adverse tumor characteristics (higher Gleason score) compared with patients with a lower LN count, which explains the significant negative impact of a higher number of removed LNs on oncological outcome seen in the univariate regressions. When controlling for tumor characteristics in multivariate analysis, this effect was not seen. Similarly, utilizing the Johns Hopkins Radical Prostatectomy Database (1982–2011), Pierorazio et al. [10] showed the number and percent of positive LNs to be an independent predictor of oncologic outcome in 505 LN positive patients who underwent RP and either extended or standard PLND. Furthermore, again conforming to our results, the number of removed LNs did not predict oncological outcomes in their patient cohort [10]. However, less than 12% had more than two positive LNs in this study compared with 26% in our patient population and results of this study are mainly derived from pre- (1982–1990) and early (1991–2000) PSA-era patients.
These results are in contrast to a recently published study by Abdollah et al. [11] who reported their data of 315 LN positive patients who underwent RP and PLND in the contemporary PSA-era from 2000 to 2012. The authors observed that the removal of a higher number of LNs during RP was associated with improvement in CSS rate. All patients included in the study received immediate ADT [13]. Therefore, the results of the present study are not directly comparable to those of Abdollah and colleagues [11]. In the present study, only 53.5% (20.5% adjuvant and 33.0% salvage hormonal treatment) received ADT, enabling us to control for this potential confounder. Moschini et al. [17] reported similar results using data from patients with pT3 and pT4 cancers with or without LN invasion. The removal of a higher number of LNs during RP in these patients was associated with higher CSS rates [17]. In another study, Abdollah et al. [18] developed a nomogram to predict CSM in LN-positive patients using a large dataset of more than 1100 patients from two tertiary centers. The results of this study were in line with our reported results. While the number of positive LNs was a significant prognosticator influencing CSM, the number of
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
LNs removed had no significant influence on CSM and therefore was omitted for the nomogram. When applying our cohort to the nomogram, the predicted mean 10-yr CSS is very close to the observed one, especially in patients receiving adjuvant RT (predicted rate of approximately 93%, observed rate of 95%). In patients without adjuvant RT, however, the rate of predicted and observed 10-yr CSS differed (85% vs 73%). Besides the number of positive LNs, higher LN density, a surrogate for the extent of metastatic spread, was demonstrated to be a prognostic factor adversely influencing oncologic outcomes in the present study. As in our population number of negative nodes removed did not influence oncologic outcome, the association of LN density with oncological outcome is mainly driven by number of positive LNs. Counts of removed LNs, which may be influenced by other factors than extent such as patient characteristics, surgical, and pathologic expertise [19,20], varied from three to 67 documented removed LNs. The mean and median number as well as the range of the number of removed LNs is comparable to other studies including patients with anatomically extended PLND [6,11,18,21–23]. Patients may benefit from a removal of all positive LNs and the likelihood for a complete removal may increase with increasing LN counts. However, a more extensive PLND with a yet to be proven therapeutic advantage is associated with a higher complication rate and prolongation of operative time [24]. Currently, the effect of an extended versus limited PLND is investigated in the prospective randomized Prospective Study to Compare a Limited Versus Extended Pelvic Lymphadenectomy During Prostatectomy study (NCT01555086), which is recruiting participants at present. There are several further important limitations to our study. First and foremost are the limitations inherent to retrospective analyses. Secondly, the population in this study underwent surgery by multiple surgeons. This might have had an effect on the number of removed LNs and oncological outcome. Moreover the follow-up period of 48 mo was relatively short. Furthermore, patients with adjuvant and salvage therapy were not excluded in the present study, which may have affected our results even after controlling for additional treatment in multivariate analysis. The type and duration of hormonal treatment differed among patients and was not standardized. Additionally, the number of removed LNs may have been influenced by tumor characteristics. Finally, the study period spans 15 yr with several diagnostic and surgical changes and a variation in adjuvant treatment, which may have led to certain heterogeneity of the study group. However, we accounted for this by including the year of surgery in the multivariate analyses. Despite these limitations, this study with more than 700 pN1 patients adds important information to the ongoing debate by showing no influence of the number of removed LNs in LN-positive patients on BCR, metastasis, and CSM. Nevertheless, for a final evaluation of the therapeutic effect of a more extensive PLND, the results
7
of the above mentioned randomized trial need to be awaited. 5.
Conclusions
The number of removed LNs did not significantly influence BCR, occurrence of metastases, and CSM. A higher number of positive LNs and higher LN density were prognostic factors adversely affecting oncological outcomes in patients with LN metastases. To finally assess a possible therapeutic effect of a more extensive PLND, the results of a randomized study are awaited. Author contributions: Derya Tilki 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: Tilki, Mandel. Acquisition of data: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Analysis and interpretation of data: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Drafting of the manuscript: Mandel, Tilki. Critical revision of the manuscript for important intellectual content: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Statistical analysis: Mandel. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Tilki. Other: None. Financial disclosures: Derya Tilki certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Arnold M, Karim-Kos H, Coebergh J, et al. Recent trends in incidence of five common cancers in 26 European countries since 1988: Analysis of the European Cancer Observatory. Eur J Cancer 2015;51:1164–87. [2] WHO, I. GLOBACON 2012: estimated cancer incidence, mortality, and prevalence worldwide in 2012. http://globocan.iarc.fr. [3] Briganti A, Giannarini G, Karnes R, Gandaglia G, Ficarra V, Montorsi F. What evidence do we need to support the use of extended pelvic lymph node dissection in prostate cancer? Eur Urol 2015;67: 597–8. [4] Briganti A, Abdollah F, Nini A, et al. Performance characteristics of computed tomography in detecting lymph node metastases in contemporary patients with prostate cancer treated with extended pelvic lymph node dissection. Eur Urol 2012;61:1132–8. [5] Gakis G, Boorjian S, Briganti A, et al. The role of radical prostatectomy and lymph node dissection in lymph node-positive prostate cancer: a systematic review of the literature. Eur Urol 2014;66: 191–9. [6] Bivalacqua T, Pierorazio P, Gorin M, Allaf ME, Carter HB, Walsh PC. Anatomic extent of pelvic lymph node dissection: impact on longterm cancer-specific outcomes in men with positive lymph nodes at time of radical prostatectomy. Urology 2013;82:653–9.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 8
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
[7] Mottet N, Bellmunt J, Briers E, et al. Guidelines on Prostate Cancer
[17] Moschini M, Fossati F, Abdollah F, et al. Determinants of long-term
2015. European Association of Urology, 2015. http://uroweb.org/
survival of patients with locally advanced prostate cancer: the role
wp-content/uploads/09-Prostate-Cancer_LR.pdf.
of extensive pelvic lymph node dissection. Prostate Cancer Pros-
[8] Murphy A, Berkman D, Desai M, Benson MC, McKiernan JM, Badani
tatic Dis 2016;19:63–7.
KK. The number of negative pelvic lymph nodes removed does not
[18] Abdollah F, Karnes R, Suardi N, et al. Predicting survival of patients
affect the risk of biochemical failure after radical prostatectomy.
with node-positive prostate cancer following multimodal treat-
BJU Int 2010;105:176–9. [9] Masterson T, Bianco Jr F, Vickers A, et al. The association between
ment. Eur Urol 2014;65:554–62. [19] Briganti A, Capitanio U, Chun FK, et al. Impact of surgical volume
total and positive lymph node counts, and disease progression in
on the rate of lymph node metastases in patients undergoing
clinically localized prostate cancer. J Urol 2006;175:1320–4.
radical prostatectomy and extended pelvic lymph node dissection
[10] Pierorazio P, Gorin M, Ross A, et al. Pathological and oncologic outcomes for men with positive lymph nodes at radical prostatectomy: The Johns Hopkins Hospital 30-year experience. Prostate 2013;73:1673–80. [11] Abdollah F, Gandaglia G, Nazareno N, et al. More extensive pelvic lymph node dissection improves survival in patients with nodepositive prostate cancer. Eur Urol 2015;67:212–9. [12] Stief C. Mere extension of the field of resection cannot be the
for clinically localized prostate cancer. Eur Urol 2008;54: 794–804. [20] Mazzola C, Savage C, Ahallal Y, et al. Nodal counts during pelvic lymph node dissection for prostate cancer: an objective indicator of quality under the influence of very subjective factors. BJU Int 2012;109:1323–8. [21] Eden G, Zacharakis E, Bott S. The learning curve for laparoscopic extended pelvic lymphadenectomy for intermediate- and high-risk
answer to surgery for metastatic spread: We need individualized
prostate cancer: implications for compliance with existing guide-
approaches based on modern imaging techniques. Eur Urol 2015;
lines. BJU Int 2013;112:346–54.
67:220–2. [13] Bogdanovic´ J, Sekulic´ V. Re: Firas Abdollah, Giorgio Gandaglia,
[22] Sagalovich D, Calaway A, Srivastava A, Sooriakumaran P, Tewari AK.
Nazareno Suardi, et al. More extensive pelvic lymph node dissec-
extended pelvic lymphadenectomy in robotic-assisted radical pros-
Assessment of required nodal yield in a high risk cohort undergoing
tion improves survival in patients with node-positive prostate
tatectomy and its impact on functional outcomes. BJU Int 2012;
cancer. Eur Urol 2015;67:212–9, Eur Urol 2015;68:e35–5.
111:85–94.
[14] Budaus L, Isbarn H, Schlomm T, et al. Current technique of open
[23] Touijer K, Mazzola C, Sjoberg D, Scardino PT, Eastham JA. Long-term
intrafascial nerve-sparing retropubic prostatectomy. Eur Urol
outcomes of patients with lymph node metastasis treated with
2009;56:317–24. [15] Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 1974;111:58–64. [16] Joslyn S, Konety BR. Impact of extent of lymphadenectomy on survival after radical prostatectomy for prostate cancer. Urology
radical prostatectomy without adjuvant androgen-deprivation therapy. Eur Urol 2014;65:20–5. [24] Briganti A, Chun F, Salonia A, et al. Complications and other surgical outcomes associated with extended pelvic lymphadenectomy in men with localized prostate cancer. Eur Urol 2006;50: 1006–13.
2006;68:121–5.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
available at www.sciencedirect.com journal homepage: www.europeanurology.com/eufocus
Platinum Priority – Prostate Cancer Editorial by XXX on pp. x–y of this issue
Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer Philipp Mandel a,b, Maximilian C. Kriegmair c, Katharina Bogdan a, Katharina Boehm b, Lars Buda¨us a, Markus Graefen a, Hartwig Huland a, Derya Tilki a,b,* a
Martini-Klinik Prostate Cancer Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany;
b
Department of Urology, University Hospital
c
Hamburg-Eppendorf, Hamburg, Germany; Department of Urology, University Medical Centre Mannheim, Mannheim, Germany
Article info
Abstract
Article history: Accepted February 28, 2016
Background: While the diagnostic value of pelvic lymph node dissection (PLND) during radical prostatectomy (RP) is undisputed, its therapeutic benefit remains a matter of debate. Objective: To investigate the association between total and positive lymph node (LN) counts and oncological outcomes in pN1 patients at RP. Design, setting, and participants: The records of 706 LN-positive patients undergoing RP and PLND between 1998 and 2012 in a single center were analyzed. The median follow-up was 47.8 mo. Intervention: RP and PLND. Outcome measurements and statistical analysis: The number of harvested and positive LNs and their influence on biochemical recurrence-free survival, metastasis-free survival, and cancer-specific survival was evaluated using univariate and multivariate Cox regressions. The number of harvested LNs was used as a continuous and dichotomous variable with a cut-off of 13 LNs. Results and limitations: The mean (median) number of removed LNs was 15.0 (13) and the mean (median) number of positive LNs was 2.4 (1). In multivariate analysis, the number of removed LNs did not significantly influence oncologic outcomes (biochemical recurrence, metastasis-free survival, or cancer-specific survival). The occurrence of metastasis and cancer-specific mortality significantly increased with higher number of positive LNs. The main limitation was the retrospective nature of the study. Conclusions: While a higher number of positive LNs were significantly associated with worse oncological outcomes, the number of removed LNs was not a significant predictor. Patient summary: We found that only the number of positive lymph nodes (LNs) but not the total number of removed LNs was a factor adversely influencing oncological outcomes in LN-positive patients undergoing radical prostatectomy. # 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Associate Editor: James Catto Keywords: Prostate cancer Radical prostatectomy Lymph node dissection Positive lymph nodes
* Corresponding author. Martini-Klinik Prostate Cancer Center, Martinistrasse 52, 20246 Hamburg, Germany. Tel. +49 4074100; Fax: +49 4074100. E-mail address: [email protected] (D. Tilki).
1.
Introduction
Prostate cancer is the most frequent cancer among men in Europe and other developed countries and among the
leading causes of death [1,2]. For its treatment, radical prostatectomy (RP) is one of the recommended treatment options for all levels of risk. Pelvic lymph node dissection (PLND) is the most precise staging procedure for possible
http://dx.doi.org/10.1016/j.euf.2016.02.018 2405-4569/# 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 2
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
lymph node (LN) metastasis; however, its therapeutic benefit concerning oncologic outcome remains controversial [3–6]. According to the guidelines of the European Urological Association (EAU), whenever a PLND is indicated, an anatomically extended PLND should be performed [7]. To present, Level IB evidence from randomized controlled trials on the oncological role of PLND is lacking. The Prospective Study to Compare a Limited Versus Extended Pelvic Lymphadenectomy During Prostatectomy (NCT01555086) is currently recruiting participants. Several retrospective studies could not show that removing more LNs during RP results in improved oncological outcome in patients with pN0 [8], pN0 or pN1 [9], and pN1 [10] only. In contrast, Abdollah et al. [11] recently reported that a higher number of removed LNs is an independent predictor of a more favorable cancer-specific survival (CSS) in LN-positive patients [11]. All patients included in the study received immediate adjuvant hormonal treatment (ADT). This study was followed by a controversial discussion on potential biases and confounders involved in this type of analysis, including the Will Rogers phenomenon possibly explaining the results [12,13]. It remains questionable whether a higher number of removed LNs, regardless if these additional removed LNs harbor metastases or not, impacts oncologic outcome. The aim of the present study was to analyze the influence of the extent of PLND according to the number of removed LNs in pN1 patients at RP on biochemical recurrence (BCR), metastasis, and cancer-specific mortality (CSM) in a large single center and to add further information to the controversial ongoing discussion using a dataset of more than 700 LN-positive patients. 2.
Materials and methods
treatment was defined as immediate treatment administered within 6 mo after RP with an undetectable PSA (< 0.2 ng/ml), while salvage treatment was defined as treatment administered at a postoperative PSA 0.2 ng/ml.
2.2.
Surgical procedure and histologic evaluation
RP was performed using an open retropubic approach or robotic-assisted laparoscopic approach by high volume surgeons, as previously described [14]. Pathological outcome was assessed using the American Joint Cancer Committee 2002 staging system and tumor grading was classified using the Gleason Grading system [15]. From 2004, the template of PLND was in accordance with the EAU guidelines for anatomically extended PLND and included removal of the nodes overlying the external iliac artery and vein, the nodes within the obturator fossa located cranially and caudally to the obturator nerve, and the nodes medial and lateral to the internal iliac artery [7]. Before 2004, bilateral PLND was not uniformly performed with extended template and mainly included the region of the obturator fossa and along the external iliac artery. As the pathologic examination of LNs has changed during the study period due to advances in LN metastasis detection via the introduction of complete embedding of fat tissue, an increase in the detection of harvested LNs may have occurred. We introduced dummy variables to control for both the change in PLND template and change in pathologic examination in all multivariate regressions. To control for further unobserved time-dependent variation we included the year of surgery in all regressions.
2.3.
Statistical analyses
To check for differences in our respective subsamples Pearson’s Chi-square tests and Kruskall-Wallis tests were performed. To evaluate the association between removed lymph nodes and oncologic outcome (MFS and CSS) Kaplan-Meier curves and uniand multivariate Cox proportional hazard models adjusting for pre- and postoperative prognosticators were assessed. These prognosticators included age at surgery, preoperative PSA, pathologic Gleason score, pT-stage, surgical margin status, number of positive LNs, year of surgery,
2.1.
Patient population
neoadjuvant treatment, adjuvant ADT, salvage ADT, adjuvant RT, and salvage RT. Statistical software STATA was used (version 14 for
From our institutional Martini-Klinik database, we identified 969 out of
Windows, StataCorp LP, College Station, TX, USA). LN count was used
15 947 patients who underwent RP at our institution between January
as both, continuous and dichotomous variable. For dichotomization all
1998 and December 2012 with positive LNs in the final pathology.
cut-offs between 5 and 25 removed LNs were analyzed. As all cut-off
Information on the number of harvested LNs at PLND was missing in
values of removed LNs showed only insignificant differences on
75 men. Moreover patients who were lost to follow up within 24 mo
oncological outcomes, we presented the results at a cut-off value of
after RP were excluded (188), which left 706 patients for our analysis.
13 (median of removed LNs).
Furthermore, for analysis of BCR, patients with ADT were excluded from regressions (242 patients). All data were collected prospectively into our review board-approved database. Patients were stratified according to
3.
Results
3.1.
Patient and tumor characteristics
the number of removed LNs. Indication to perform PLND was in accordance to respective EAU guidelines applicable at the relevant time and partly based on an individual approach [7]. BCR was defined as a prostate-specific antigen (PSA) level 0.2 ng/ml and rising after RP in patients without ADT. Metastasis was confirmed with imaging. The cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone. After RP, patient follow up consisted of periodical PSA testing and digital rectal examination. Postoperative imaging studies to detect local recurrence or systemic spread were performed according to PSA level, patients’ preferences, and further symptoms. Adjuvant hormonal and/ or radiation therapy (RT) was recommended based on further patient and tumor characteristics and according to respective EAU guidelines applicable at the relevant time [7]. Adjuvant
Patient and tumor characteristics of the entire cohort and stratified by the number of removed LNs (high LN count [> 13] and low LN count [ 13]) are summarized in Table 1. In the entire population of LN positive patients, the mean (median) number of removed LN was 15.0 (13) and the mean (median) number of positive LN was 2.4 (1). Next to number of removed LNs and number of positive LNs, LN density (ratio of the number of metastatic LNs to the total number of LNs removed; p < 0.001) and Gleason score (p = 0.002) were significantly different between patients with a high and low LN count.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 3
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 1 – Demographics and tumor characteristics stratified by number of removed lymph nodes Entire cohort (n = 706) Age (yr), mean PSA (ng/ml), mean Removed LNs Mean Median IQR Positive LNs Mean Median IQR LN density Mean Median IQR Gleason score 7 8 Pathologic stage pT2 pT3a pT3b pT4 Surgical margins Negative Positive Hormonal therapy Neoadjuvant Adjuvant Salvage Radiation therapy Adjuvant Salvage
64.5 21.0
LNs removed 13 (n = 376)
LNs removed >13 (n = 330)
64.4 19.0
15.0 13 9–19
64.6 23.3
8.6 9 6–11
p value
0.448 0.068 <0.001
22.3 20 16–26 <0.001
2.4 1 1–3
1.8 1 1–2
3.1 2 1–3
0.189 0.133 0.083–0.250
0.223 0.167 0.111–0.272
0.151 0.087 0.053–0.174
<0.001
0.002 62.6% 37.4%
68.8% 31.2%
55.6% 44.4%
6.9% 22.7% 67.3% 3.1%
6.9% 25.8% 63.6% 3.7%
7.0% 19.1% 71.5% 2.4%
49.5% 50.5%
52.8% 47.2%
45.7% 54.3%
14.0% 20.5% 33.0%
13.3% 20.7% 34.3%
14.9% 20.3% 31.5%
0.554 0.885 0.431
23.6% 32.9%
23.7% 30.3%
23.6% 35.8%
0.992 0.125
0.159
0.068
IQR = interquartile range; LN = lymph node; PSA = prostate-specific antigen.
The overall percentage of patients undergoing hormonal treatment after RP was 53.5% (20.5% adjuvant and 33.0% salvage hormonal treatment). The overall rate of RT was 56.5%, whereas patients underwent salvage RT more often (32.9%) than adjuvant RT (23.6%). 3.2.
Oncological outcome
The median follow-up was 47.8 mo (interquartile range: 27.3–68.8). Within the follow-up period 75.0%, 22.4%, and 6.2% of the patients showed BCR, metastasis, or cancerspecific death, respectively. In univariate analysis, the number of removed LNs was significantly associated with BCR (hazard ratio [HR] 1.015, p = 0.019) and occurrence of metastasis (HR 1.023, p = 0.026; Table 2). In contrast, in multivariate analysis, the number of removed LNs did not significantly influence oncologic outcomes (BCR, MFS, and CSM). When stratifying patients by LN count (> 13 vs 13), patients with 13 or less removed LNs had a significantly better BCR-free (p = 0.003) and MFS (p = 0.006), while there was no difference in CSS between both patient groups (Fig. 1A–C). In multivariate analysis this association was not significant. The number of positive LNs was significantly associated with BCR in univariate analysis (HR: 1.091, p < 0.001), but not in multivariate analysis (HR: 1.029, p = 0.194; Table 2).
Occurrence of metastasis and cancer-specific mortality significantly increased with a higher number of positive LNs in both uni- (HR: 1.157, p < 0.001 for metastasis and HR: 1.187, p < 0.001 for CSM) and multivariate (HR: 1.110, p < 0.001 for metastasis and HR: 1.137, p = 0.001 for CSM; Tables 3 and 4) analyses. Oncologic outcome stratified by the number of positive LNs is depicted in Figure 2A–C. A higher LN density was a significant factor adversely affecting BCR-free survival (HR: 2.84, p = 0.001), MFS (HR: 7.60, p < 0.001), and CSS (HR: 11.67, p < 0.001) in univariate and MFS (HR: 4.01, p = 0.001) and CSS (HR: 4.71, p = 0.014) multivariate regressions. 4.
Discussion
While the diagnostic value of a LN dissection during RP is undisputed, its therapeutic benefit still remains a matter of debate. The evidence with regards to the impact of a higher number of removed LNs during PLND on oncological outcomes in prostate cancer patients undergoing RP is limited [12,13]. Several retrospective studies have been published on the topic, which have shown conflicting results. Masterson et al. [9] reviewed 4611 patients undergoing RP and showed that an increased number of removed LNs was associated with higher BCR-free probabilities in pN0 patients, but not in the
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 4
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 2 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Biochemical recurrence Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
1.001 1.011 1.392
0.98–1.02 1.01–1.02 1.23-1.57
0.936 <0.001 <0.001
1.000 1.008 1.244
0.98–1.02 1.00–1.01 1.09–1.42
0.945 0.001 0.001
Ref. 1.973 1.004
1.54–2.53 0.32–3.17
<0.001 0.995
1.511 0.509
1.14–2.00 0.16–1.66
0.004 0.263
Ref. 1.721 1.015 1.091 1.048
1.36–2.18 1.00–1.03 1.05–1.13 1.01–1.09
<0.001 0.019 <0.001 0.013
1.351 1.011 1.029 0.992
1.04–1.76 1.00–1.03 0.99–1.07 0.91–1.09
0.026 0.153 0.194 0.872
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
1.00
B
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
A
without PLND between 1988 and 1991 with a minimum follow-up of 10 yr. In line with Masterson et al. [9], the authors reported that pN0 patients with removal of at least 10 LNs had a lower CSM compared with patients without a
1.00
entire group. The median number of removed LNs was nine in this study [9]. Joslyn and Konety [16] studied the data of almost 13 000 patients from the Surveillance, Epidemiology, and End Results (SEER) database undergoing RP with or
0
50
Analysis time
150
50
0
More than 13 LNs removed
150
13 or less LNs removed
More than 13 LNs removed
0.00
0.25
0.50
0.75
C
100
Mo after surgery
1.00
13 or less LNs removed
100
0
50
100
150
Mo after surgery 13 or less LNs removed
More than 13 LNs removed
Fig. 1 – Survival stratified by number of lymph nodes (LNs) removed in (A) biochemical recurrence-free survival (= 13 vs > 13; p = 0.003), (B) metastasisfree survival (= 13 vs > 13; p = 0.006), and (C) cancer-specific survival (= 13 vs > 13; p = 0.760).
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 5
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
Table 3 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Occurrence of metastasis Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery Adjuvant treatment No Yes Salvage treatment No Yes
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
0.967 1.005 1.625
0.94–0.99 1.00–1.01 1.36–1.95
0.009 0.091 <0.001
0.972 1.000 1.475
0.95–1.00 0.99–1.01 1.20–1.82
0.040 0.932 <0.001
Ref. 2.795 6.542
1.74–4.48 2.89–14.8
<0.001 <0.001
1.959 4.459
1.17–3.27 1.82–10.9
0.010 0.001
Ref. 1.464 1.023 1.157 1.070
1.03–2.08 1.00–1.04 1.11–1.20 1.00–1.15
0.034 0.026 <0.001 0.054
1.006 0.997 1.110 1.047
0.67–1.52 0.97–1.02 1.06–1.17 0.89–1.23
0.982 0.809 <0.001 0.566
Ref. 0.801
0.55–1.16
0.238
0.587
0.37–0.92
0.021
Ref. 1.504
1.06–2.13
0.021
1.313
0.89–1.94
0.173
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
PLND and therefore unknown pN status. Bivalacqua and colleagues [6] analyzed the data of 94 LN-positive patients undergoing anatomically extended or limited PLND at the time of RP and showed that extended PLND was associated with a trend toward improved BCR-free survival (p = 0.064), which was particularly seen in the group of patients with <15% positive LNs (p = 0.016). Moreover, the authors found a significant improvement in MFS in the group of patients with extended PLND (p = 0.045), while no difference was seen in CSS. The study was limited by its small sample size
and the fact that RP was performed by only two surgeons, whereby one routinely performed a limited PLND, while the second performed an extended PLND. Using a large single center dataset with more than 700 pN1 patients at RP, we could not show a significant positive effect of a higher number of LNs removed on oncological outcomes (BCR, MFS, or CSM) in uni- and multivariate analyses, neither when using the number of removed LNs as a continuous nor as a dichotomous variable. Moreover, a variation of the cut-off for high versus low
Table 4 – Uni- and multivariate analysis predicting oncologic outcome in patients with pN1 Cancer-specific mortality Univariate analysis
Age (yr) PSA (ng/ml) Gleason score pT-stage pT2/pT3a pT3b pT4 Surgical margin Negative Positive Removed LNs Positive LNs Yr of surgery Adjuvant treatment No Yes Salvage Treatment No Yes
Multivariate analysis
HR
95% CI
p value
HR
95% CI
p value
1.005 0.992 2.412
0.96–1.05 0.97–1.01 1.74–3.34
0.845 0.435 <0.001
1.004 0.987 2.133
0.95–1.06 0.97–1.01 1.42–3.20
0.872 0.211 <0.001
Ref. 4.780 19.82
1.69–13.5 5.74–68.4
0.003 <0.001
2.393 5.977
0.78–7.30 1.45–24.7
0.125 0.013
Ref. 2.239 1.019 1.187 0.998
1.20–4.19 0.98–1.06 1.12–1.26 0.89–1.11
0.012 0.325 <0.001 0.975
1.580 0.999 1.137 0.970
0.76–3.28 0.95–1.06 1.05–1.23 0.73–1.29
0.220 0.982 0.001 0.838
Ref. 0.912
0.48–1.72
0.776
0.465
0.18–1.19
0.110
Ref. 1.032
0.57–1.88
0.918
1.073
0.44–2.64
0.879
CI = confidence interval; HR = hazard ratio; LN = lymph node; PSA = prostate-specific antigen; Ref. reference.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 6
1.00
B
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
A
1.00
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
0
50
Analysis time
150
0
50
2 positive LNs 4 or more positive LNs
150
1 positive LN 3 positive LNs
2 positive LNs 4 or more positive LNs
0.00
0.25
0.50
0.75
C
100
Mo after surgery
1.00
1 positive LN 3 positive LNs
100
0
50
100
Mo after surgery
1 positive LN 3 positive LNs
150
2 positive LNs 4 or more positive LNs
Fig. 2 – Survival stratified by number of positive lymph nodes (LNs) in (A) biochemical recurrence-free survival (1 vs 2 vs 3 vs 4+; p < 0.001), (B) metastasis-free survival (1 vs 2 vs 3 vs 4+; p < 0.001), and (C) cancer-specific survival (1 vs 2 vs 3 vs 4+; p = 0.001).
counts of removed LNs did not qualitatively alter our results. In contrast, both higher number of positive LNs and higher LN density had a negative impact on oncological outcomes. Patients with more removed LNs had adverse tumor characteristics (higher Gleason score) compared with patients with a lower LN count, which explains the significant negative impact of a higher number of removed LNs on oncological outcome seen in the univariate regressions. When controlling for tumor characteristics in multivariate analysis, this effect was not seen. Similarly, utilizing the Johns Hopkins Radical Prostatectomy Database (1982–2011), Pierorazio et al. [10] showed the number and percent of positive LNs to be an independent predictor of oncologic outcome in 505 LN positive patients who underwent RP and either extended or standard PLND. Furthermore, again conforming to our results, the number of removed LNs did not predict oncological outcomes in their patient cohort [10]. However, less than 12% had more than two positive LNs in this study compared with 26% in our patient population and results of this study are mainly derived from pre- (1982–1990) and early (1991–2000) PSA-era patients.
These results are in contrast to a recently published study by Abdollah et al. [11] who reported their data of 315 LN positive patients who underwent RP and PLND in the contemporary PSA-era from 2000 to 2012. The authors observed that the removal of a higher number of LNs during RP was associated with improvement in CSS rate. All patients included in the study received immediate ADT [13]. Therefore, the results of the present study are not directly comparable to those of Abdollah and colleagues [11]. In the present study, only 53.5% (20.5% adjuvant and 33.0% salvage hormonal treatment) received ADT, enabling us to control for this potential confounder. Moschini et al. [17] reported similar results using data from patients with pT3 and pT4 cancers with or without LN invasion. The removal of a higher number of LNs during RP in these patients was associated with higher CSS rates [17]. In another study, Abdollah et al. [18] developed a nomogram to predict CSM in LN-positive patients using a large dataset of more than 1100 patients from two tertiary centers. The results of this study were in line with our reported results. While the number of positive LNs was a significant prognosticator influencing CSM, the number of
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
LNs removed had no significant influence on CSM and therefore was omitted for the nomogram. When applying our cohort to the nomogram, the predicted mean 10-yr CSS is very close to the observed one, especially in patients receiving adjuvant RT (predicted rate of approximately 93%, observed rate of 95%). In patients without adjuvant RT, however, the rate of predicted and observed 10-yr CSS differed (85% vs 73%). Besides the number of positive LNs, higher LN density, a surrogate for the extent of metastatic spread, was demonstrated to be a prognostic factor adversely influencing oncologic outcomes in the present study. As in our population number of negative nodes removed did not influence oncologic outcome, the association of LN density with oncological outcome is mainly driven by number of positive LNs. Counts of removed LNs, which may be influenced by other factors than extent such as patient characteristics, surgical, and pathologic expertise [19,20], varied from three to 67 documented removed LNs. The mean and median number as well as the range of the number of removed LNs is comparable to other studies including patients with anatomically extended PLND [6,11,18,21–23]. Patients may benefit from a removal of all positive LNs and the likelihood for a complete removal may increase with increasing LN counts. However, a more extensive PLND with a yet to be proven therapeutic advantage is associated with a higher complication rate and prolongation of operative time [24]. Currently, the effect of an extended versus limited PLND is investigated in the prospective randomized Prospective Study to Compare a Limited Versus Extended Pelvic Lymphadenectomy During Prostatectomy study (NCT01555086), which is recruiting participants at present. There are several further important limitations to our study. First and foremost are the limitations inherent to retrospective analyses. Secondly, the population in this study underwent surgery by multiple surgeons. This might have had an effect on the number of removed LNs and oncological outcome. Moreover the follow-up period of 48 mo was relatively short. Furthermore, patients with adjuvant and salvage therapy were not excluded in the present study, which may have affected our results even after controlling for additional treatment in multivariate analysis. The type and duration of hormonal treatment differed among patients and was not standardized. Additionally, the number of removed LNs may have been influenced by tumor characteristics. Finally, the study period spans 15 yr with several diagnostic and surgical changes and a variation in adjuvant treatment, which may have led to certain heterogeneity of the study group. However, we accounted for this by including the year of surgery in the multivariate analyses. Despite these limitations, this study with more than 700 pN1 patients adds important information to the ongoing debate by showing no influence of the number of removed LNs in LN-positive patients on BCR, metastasis, and CSM. Nevertheless, for a final evaluation of the therapeutic effect of a more extensive PLND, the results
7
of the above mentioned randomized trial need to be awaited. 5.
Conclusions
The number of removed LNs did not significantly influence BCR, occurrence of metastases, and CSM. A higher number of positive LNs and higher LN density were prognostic factors adversely affecting oncological outcomes in patients with LN metastases. To finally assess a possible therapeutic effect of a more extensive PLND, the results of a randomized study are awaited. Author contributions: Derya Tilki 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: Tilki, Mandel. Acquisition of data: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Analysis and interpretation of data: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Drafting of the manuscript: Mandel, Tilki. Critical revision of the manuscript for important intellectual content: Mandel, Kriegmair, Bogdan, Boehm, Buda¨us, Graefen, Huland, Tilki. Statistical analysis: Mandel. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Tilki. Other: None. Financial disclosures: Derya Tilki certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Arnold M, Karim-Kos H, Coebergh J, et al. Recent trends in incidence of five common cancers in 26 European countries since 1988: Analysis of the European Cancer Observatory. Eur J Cancer 2015;51:1164–87. [2] WHO, I. GLOBACON 2012: estimated cancer incidence, mortality, and prevalence worldwide in 2012. http://globocan.iarc.fr. [3] Briganti A, Giannarini G, Karnes R, Gandaglia G, Ficarra V, Montorsi F. What evidence do we need to support the use of extended pelvic lymph node dissection in prostate cancer? Eur Urol 2015;67: 597–8. [4] Briganti A, Abdollah F, Nini A, et al. Performance characteristics of computed tomography in detecting lymph node metastases in contemporary patients with prostate cancer treated with extended pelvic lymph node dissection. Eur Urol 2012;61:1132–8. [5] Gakis G, Boorjian S, Briganti A, et al. The role of radical prostatectomy and lymph node dissection in lymph node-positive prostate cancer: a systematic review of the literature. Eur Urol 2014;66: 191–9. [6] Bivalacqua T, Pierorazio P, Gorin M, Allaf ME, Carter HB, Walsh PC. Anatomic extent of pelvic lymph node dissection: impact on longterm cancer-specific outcomes in men with positive lymph nodes at time of radical prostatectomy. Urology 2013;82:653–9.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018
EUF-133; No. of Pages 8 8
EUROPEAN UROLOGY FOCUS XXX (2016) XXX–XXX
[7] Mottet N, Bellmunt J, Briers E, et al. Guidelines on Prostate Cancer
[17] Moschini M, Fossati F, Abdollah F, et al. Determinants of long-term
2015. European Association of Urology, 2015. http://uroweb.org/
survival of patients with locally advanced prostate cancer: the role
wp-content/uploads/09-Prostate-Cancer_LR.pdf.
of extensive pelvic lymph node dissection. Prostate Cancer Pros-
[8] Murphy A, Berkman D, Desai M, Benson MC, McKiernan JM, Badani
tatic Dis 2016;19:63–7.
KK. The number of negative pelvic lymph nodes removed does not
[18] Abdollah F, Karnes R, Suardi N, et al. Predicting survival of patients
affect the risk of biochemical failure after radical prostatectomy.
with node-positive prostate cancer following multimodal treat-
BJU Int 2010;105:176–9. [9] Masterson T, Bianco Jr F, Vickers A, et al. The association between
ment. Eur Urol 2014;65:554–62. [19] Briganti A, Capitanio U, Chun FK, et al. Impact of surgical volume
total and positive lymph node counts, and disease progression in
on the rate of lymph node metastases in patients undergoing
clinically localized prostate cancer. J Urol 2006;175:1320–4.
radical prostatectomy and extended pelvic lymph node dissection
[10] Pierorazio P, Gorin M, Ross A, et al. Pathological and oncologic outcomes for men with positive lymph nodes at radical prostatectomy: The Johns Hopkins Hospital 30-year experience. Prostate 2013;73:1673–80. [11] Abdollah F, Gandaglia G, Nazareno N, et al. More extensive pelvic lymph node dissection improves survival in patients with nodepositive prostate cancer. Eur Urol 2015;67:212–9. [12] Stief C. Mere extension of the field of resection cannot be the
for clinically localized prostate cancer. Eur Urol 2008;54: 794–804. [20] Mazzola C, Savage C, Ahallal Y, et al. Nodal counts during pelvic lymph node dissection for prostate cancer: an objective indicator of quality under the influence of very subjective factors. BJU Int 2012;109:1323–8. [21] Eden G, Zacharakis E, Bott S. The learning curve for laparoscopic extended pelvic lymphadenectomy for intermediate- and high-risk
answer to surgery for metastatic spread: We need individualized
prostate cancer: implications for compliance with existing guide-
approaches based on modern imaging techniques. Eur Urol 2015;
lines. BJU Int 2013;112:346–54.
67:220–2. [13] Bogdanovic´ J, Sekulic´ V. Re: Firas Abdollah, Giorgio Gandaglia,
[22] Sagalovich D, Calaway A, Srivastava A, Sooriakumaran P, Tewari AK.
Nazareno Suardi, et al. More extensive pelvic lymph node dissec-
extended pelvic lymphadenectomy in robotic-assisted radical pros-
Assessment of required nodal yield in a high risk cohort undergoing
tion improves survival in patients with node-positive prostate
tatectomy and its impact on functional outcomes. BJU Int 2012;
cancer. Eur Urol 2015;67:212–9, Eur Urol 2015;68:e35–5.
111:85–94.
[14] Budaus L, Isbarn H, Schlomm T, et al. Current technique of open
[23] Touijer K, Mazzola C, Sjoberg D, Scardino PT, Eastham JA. Long-term
intrafascial nerve-sparing retropubic prostatectomy. Eur Urol
outcomes of patients with lymph node metastasis treated with
2009;56:317–24. [15] Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 1974;111:58–64. [16] Joslyn S, Konety BR. Impact of extent of lymphadenectomy on survival after radical prostatectomy for prostate cancer. Urology
radical prostatectomy without adjuvant androgen-deprivation therapy. Eur Urol 2014;65:20–5. [24] Briganti A, Chun F, Salonia A, et al. Complications and other surgical outcomes associated with extended pelvic lymphadenectomy in men with localized prostate cancer. Eur Urol 2006;50: 1006–13.
2006;68:121–5.
Please cite this article in press as: Mandel P, et al. Association between Lymph Node Counts and Oncological Outcomes in Lymph Node Positive Prostate Cancer. Eur Urol Focus (2016), http://dx.doi.org/10.1016/j.euf.2016.02.018