Performance Characteristics of Computed Tomography in Detecting Lymph Node Metastases in Contemporary Patients with Prostate Cancer Treated with Extended Pelvic Lymph Node Dissection

EUROPEAN UROLOGY 61 (2012) 1132–1138

available at www.sciencedirect.com journal homepage: www.europeanurology.com

Platinum Priority – Prostate Cancer Editorial by Steven Joniau et al. on pp. 1139–1141 of this issue

Performance Characteristics of Computed Tomography in Detecting Lymph Node Metastases in Contemporary Patients with Prostate Cancer Treated with Extended Pelvic Lymph Node Dissection Alberto Briganti a,*, Firas Abdollah a, Alessandro Nini a, Nazareno Suardi a, Andrea Gallina a, Umberto Capitanio a, Marco Bianchi a, Manuela Tutolo a, Niccolo` Maria Passoni a, Andrea Salonia a, Renzo Colombo a, Massimo Freschi b, Patrizio Rigatti a, Francesco Montorsi a a

Department of Urology, Vita-Salute University, San Raffaele Scientific Institute, Milan, Italy; b Department of Pathology, Vita-Salute University, San Raffaele

Scientific Institute, Milan, Italy

Article info

Abstract

Article history: Accepted November 3, 2011 Published online ahead of print on November 12, 2011

Background: Computed tomography (CT) is a commonly used noninvasive procedure for prostate cancer (PCa) staging. All previous studies addressing the ability of CT scan to predict lymph node invasion (LNI) were based on historical patients treated with limited pelvic lymph node dissection (PLND). Objective: Assess the value of CT in predicting LNI in contemporary PCa patients treated with extended PLND (ePLND). Design, setting, and participants: We evaluated 1541 patients undergoing radical prostatectomy and ePLND between 2003 and 2010 at a single center. All patients were preoperatively staged using abdominopelvic CT scan. All lymph nodes with a short axis diameter 10 mm were considered suspicious for metastatic involvement. Intervention: All patients underwent preoperative CT scan, radical retropubic prostatectomy, and ePLND, regardless of PCa features at diagnosis. Measurements: The performance characteristics of CT scan were tested in the overall patient population, as well as according to the National Comprehensive Cancer Network (NCCN) classification and according to the risk of LNI derived from a nomogram developed on an ePLND series. Logistic regression models tested the relationship between CT scan findings and LNI. Discrimination accuracy was quantified with the area under the curve. Results and limitations: Overall, a CT scan that suggested LNI was found in 73 patients (4.7%). Of them, only 24 patients (32.8%) had histologically proven LNI at ePLND. Overall, sensitivity, specificity, and accuracy of CT scan were 13%, 96.0%, and 54.6%, respectively. In patients with low-, intermediate-, or high-risk PCa according to NCCN classification, sensitivity was 8.3%, 96.3%, and 52.3%, respectively; specificity was 3.6%, 97.3%, and 50.5%, respectively; and accuracy was 17.9%, 94.3%, and 56.1%, respectively. Similarly, in patients with a nomogram-derived LNI risk 50%, sensitivity, specificity, and accuracy were only 23.9%, 94.7%, and 59.3%, respectively. At multivariable analyses, inclusion of CT scan findings did not improve the accuracy of LNI prediction (81.4% compared with 81.3%; p = 0.8). Lack of a central scan review represents the main limitation of our study. Conclusions: In contemporary patients with PCa, the accuracy of CT scan as a preoperative nodal-staging procedure is poor, even in patients with high LNI risk. Therefore, the need for and the extent of PLND should not be based on the results obtained by CT scan. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Keywords: Prostate cancer Lymph node metastases Computed tomography Preoperative staging

* Corresponding author. Department of Urology, Vita-Salute University, 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 # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.

doi:10.1016/j.eururo.2011.11.008

EUROPEAN UROLOGY 61 (2012) 1132–1138

1.

1133

All patients included in the study underwent an anatomically

Introduction

defined ePLND, regardless of the presence of enlarged lymph nodes at

Prostate cancer (PCa) is the most common cancer in men and the second most common cause of death from tumors in the male population [1]. However, patients with localized PCa have favorable cancer control outcomes, especially when they receive initial active treatment [2]. Conversely, when PCa is locally advanced and/or lymph node invasion (LNI) is present, cancer control outcomes are less favorable [3,4]. Although pelvic lymph node dissection (PLND) is considered the most reliable procedure for assessing the presence of nodal invasion [5,6], computed tomography (CT) is a commonly used noninvasive procedure for nodal staging. However, previous studies reported a limited ability of CT scan to predict LNI [7–9]. This limitation mainly stems from the low sensitivity of CT because of its inability to detect small-volume, micrometastatic nodal disease. A recent meta-analysis showed that the pooled sensitivity of CT scan in predicting LNI was only 42% [7]. However, virtually all previous studies that examined the performance characteristics of CT scan in predicting LNI were based on historic patient cohorts diagnosed before or early on in the prostate-specific antigen (PSA) era [7]. Thus, the results of these studies may not be applicable to contemporary patients, who frequently harbor more favorable tumor characteristics at diagnosis because of the well-known phenomenon of PCa stage migration [10,11]. Moreover, all previous studies used a limited PLND template for PCa nodal staging. Such a template is no longer considered adequate according to virtually all contemporary guidelines [5,6], mainly because of the high rate of false-negative findings associated with limited PLND [12–14]. Indeed, the rate of LNI virtually linearly increases with the extent of PLND [12–14]. As such, limited PLND should be abandoned as a staging procedure [1,6,15]. Finally, previous reports do not allow the assessment of CT scan performance characteristics according to tumor characteristics. This fact is key, since the risk of nodal invasion linearly increases with the aggressiveness of the disease [12–14,16]. To address all these limitations, we aimed at reassessing the ability of CT scan to predict LNI in a contemporary large cohort of PCa patients treated with anatomically defined extended PLND (ePLND).

CT scan. Therefore, the results of CT scan did not influence the extent of nodal dissection, which consisted of excising the fibrofatty tissue along the external iliac vein, the distal limit being the deep circumflex vein and femoral canal. Proximally, PLND was performed up to and including the bifurcation of the common iliac artery. Furthermore, all fibrofatty tissue within the obturator fossa was removed to completely skeletonize the obturator nerve. The lateral limit consisted of the pelvic sidewall, and the medial dissection limit was defined by perivesical fat. Finally, lymph nodes along the internal iliac vessels were also removed. Fat tissue containing lymph nodes was sent in separate containers per region [18] and fixed in 10% buffered formalin. For each anatomic group, the number of nodes, the size of the largest node, and any gross features were described. Macroscopic specimen assessment was based on tactile and visual criteria. Large (>2-cm) nodes were sampled in multiple blocks. If no lymph node was macroscopically detected, all fat tissue was processed. Clearing solution was generally not used for pelvic lymph nodes. All blocks were embedded in paraffin, cut at 3 mm, and stained with hematoxylin-eosin. In selected cases, immunohistochemical stain for cytokeratin and multiple sections were analyzed. The results in terms of clinically suspicious lymph nodes at CT scan were compared with results derived by pathologic nodal assessment.

2.2.

Statistical analyses

Our statistical analysis consisted of two steps. First, we systematically tested the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and discrimination accuracy of CT scan in predicting histologically confirmed LNI. The same analyses were repeated after stratification of patients according to the National Comprehensive Cancer Network (NCCN) practice guidelines risk grouping: low-risk (PSA <10 ng/ml, biopsy Gleason score 6, and cT1 tumor stage), high-risk (PSA >20 ng/ml or biopsy Gleason score 8 or stage cT3), and intermediate-risk (all the remaining patients) groups. Moreover, the risk of LNI was also calculated for each single patient using a previously published ePLND nomogram [16]. Then the performance characteristics of CT scan were also assessed in patients with very high LNI nomogram-calculated risk, namely, 30% and 50%. In the second step of our analysis, univariable and multivariable logistic regression analyses tested the relationship between CT scan findings and the presence of histologically confirmed LNI. Covariates consisted of preoperative PSA, clinical stage, and biopsy Gleason sum. The discrimination accuracy of each covariate included in the model was assessed using the receiver operating characteristics area under the curve (AUC) method. Two hundred bootstrap resamples were used to reduce overfit bias. The AUC analysis was also used to assess the discrimination accuracy of a base LNI prediction model including

2.

Materials and methods

2.1.

Patient population

preoperative PSA, clinical stage, and biopsy Gleason score (base model). Then the discrimination accuracy of a multivariable prediction model that also included CT scan findings as covariate (full model) was also assessed. The difference in the discrimination accuracy between the two

The study included 1541 men with histologically proven PCa treated with radical prostatectomy and ePLND at a single tertiary referral center between 2003 and 2010. All patients signed an informed consent form approved by the institutional review board to be included in our

models was assessed with the Mantel-Haenszel test. All statistical analyses were performed using the R statistical package system (R Foundation for Statistical Computing, Vienna, Austria), with a two-sided significance level set at p < 0.05.

prospectively collected database. Before surgery, all patients were staged with abdominopelvic contrast-enhanced CT scan regardless of PCa characteristics. CT was considered suspicious in the presence of at least one pelvic lymph node with a short axis diameter 10 mm [17]. Data on CT scan results were derived from radiologic medical charts and prospectively included in our database. No patient received hormonal or radiation therapy prior to surgery.

3.

Results

Clinical patient characteristics are summarized in Table 1. Overall, LNI was detected in 184 patients (12.0%). The mean number of nodes removed was 17.3 (median: 16; range:

1134

EUROPEAN UROLOGY 61 (2012) 1132–1138

Table 1 – Descriptive characteristics of 1541 patients treated with radical prostatectomy and extended pelvic lymph node dissection at a single institution between 2003 and 2010y All patients (n = 1541, 100%) Age, yr Mean Median Range Prostate-specific antigen, ng/ml Mean Median Range Clinical tumor stage, no. patients (%) T1c T2 T3 Gleason score at biopsy, no. patients (%) 6 7 8 NCCN risk group, no. patients (%) Low Intermediate High LNI calculated risk, no. patients (%)* <30 30 <50 50 Lymph nodes removed, no. Mean Median Range Positive lymph nodes, no. Mean Median Range CT scan findings, no. patients (%) Negative Positive

Patients without LNI (n = 1357, 88.0%)

Patients with LNI (n = 184, 12.0%)

p value

0.1 64.8 65.6 39.9–85.3

64.9 65.8 39.9–85.3

64.1 64.1 44.4–82.7 <0.001

11.9 7.0 0.9–327.0

9.5 6.8 0.9–327.0

29.3 11.1 0.9–273.6 <0.001

880 (57.1) 449 (29.1) 212 (13.8)

818 (60.3) 393 (29) 146 (10.8)

62 (33.7) 56 (30.4) 66 (35.9)

895 (58.1) 483 (31.3) 163 (10.6)

857 (63.2) 407 (30) 93 (6.9)

38 (20.7) 76 (41.3) 70 (38)

471 (30.6) 689 (44.7) 381 (24.7)

459 (33.8) 634 (46.7) 264 (19.5)

12 (6.5) 55 (29.9) 117 (63.6)

<0.001

<0.001

1348 193 1380 161

(87.5) (12.5) (89.6) (10.4)

1246 (91.8) 111 (8.2) 1263 (93.1) 94 (6.9)

102 82 117 67

(55.4) (44.6) (63.6) (36.4)

<0.001 <0.001 0.2

17.3 16.0 5.0–70.0

17.2 16.0 5.0–70.0

18.1 17.0 5.0–57.0 NA

0.5 0 0–57

NA

3.9 2.0 1–57 <0.001

1468 (95.3) 73 (4.7)

1308 (96.4) 49 (3.6)

160 (87.0) 24 (13.0)

LNI = lymph node invasion; NCCN = National Comprehensive Cancer Network; NA = not applicable; CT = computed tomography. Patients are stratified according to lymph node status. * LNI risk was calculated using a previously published prediction model [16]. y

5–70). Patients with LNI had higher mean PSA level, higher tumor grade at biopsy, and higher clinical stage distribution (all p < 0.001). Conversely, no significant differences were found according to LNI status with regard to age at surgery and number of lymph nodes removed (all p  0.1). Overall, a CT scan suggestive for LNI was found in 73 patients (4.7%). Of them, only 24 patients (32.8%) had histologically proven LNI at ePLND (Table 1), while 49 patients (67.2%) had false-positive findings at CT scan. When stratified according to PCa aggressiveness, a suspicious CT scan was found in 3.8%, 2.8%, and 9.4% of patients with low-, intermediate-, and high-risk PCa, respectively, according to NCCN classification. This rate increased to 11.9% and 13.0% in patients with a nomogram-based LNI calculated risk 30 and 50, respectively (Table 2). Overall, sensitivity, specificity, NPV, PPV, and discrimination accuracy of CT scan in predicting LNI were 13%, 96.0%, 89.1%, 32.9%, and 54.6%, respectively (Table 3). The sensitivity of CT scan in detecting LNI was as low as 8.3%, 3.6%, and 17.9% in patients with low-, intermediate-, and high-risk PCa,

respectively, according to the NCCN classification. The specificity was 96.3%, 97.3%, and 94.3%, respectively, while the NPV was 97.6%, 92.1%, and 72.7%, respectively. Finally, the PPV and discrimination accuracy was only 5.6%, 10.5%, and 58.3% and 52.3%, 50.5, and 56.1%, respectively. Similarly, the sensitivity, specificity, NPV, PPV, and discrimination accuracy were 22.0% and 23.9%, 95.5% and 94.7%, 78.3% and 76.2%, 62.4% and 63.6%, and 58.7% and 59.3% in patients with a nomogram-based LNI calculated risk 30 and 50, respectively (Table 3). At univariable logistic regression analyses, all preoperative variables, including the presence of a suspicious CT scan, were statistically significantly related to the rate of LNI (all p < 0.001; Table 4). Similarly, at multivariable analyses, all the aforementioned covariates were independent predictors of LNI rate (all p  0.02; Table 4). However, inclusion of CT scan findings did not improve the discrimination accuracy of the base model, which included preoperative PSA, clinical tumor stage, and biopsy Gleason score (81.4% compared with 81.3%; 0.1% gain; p = 0.8; Table 4).

1135

EUROPEAN UROLOGY 61 (2012) 1132–1138

Table 2 – Cross-tabulation of computed tomography scan compared with extended pelvic lymph node dissection findings regarding lymph node invasion in 1541 patients treated with radical prostatectomy and extended pelvic lymph node dissection at a single institution between 2003 and 2010y CT scan findings Low-risk patients, n = 471 Negative Positive Intermediate-risk patients, n = 689 Negative Positive High-risk patients, n = 381 Negative Positive LNI calculated risk 30,* n = 93 Negative Positive LNI calculated risk 50,* n = 161 Negative Positive

Patients, overall no. (%)

Patients, no. without LNI (%)

Patients, no. with LNI (%)

453 (96.2) 18 (3.8)

442 (96.3) 17 (3.7)

11 (91.7) 1 (8.3)

670 (97.2) 19 (2.8)

617 (97.3) 17 (2.7)

53 (96.4) 2 (3.6)

345 (90.6) 36 (9.4)

249 (94.3) 15 (5.7)

96 (82.1) 21 (17.9)

170 (88.1) 23 (11.9)

106 (95.5) 5 (4.5)

64 (78.0) 18 (22.0)

140 (87.0) 21 (13.0)

89 (94.7) 5 (5.3)

51 (76.1) 16 (23.9)

p value 0.3

0.6

<0.001

<0.001

0.001

LNI = lymph node invasion; CT = computed tomography. Patients are stratified according to tumor characteristics. * LNI risk was calculated using a previously published prediction model [16]. y

Table 3 – Performance characteristics of computed tomography scan in predicting lymph node invasion in 1541 patients treated with radical prostatectomy and extended pelvic lymph node dissection at a single institution between 2003 and 2010y

Overall population, n = 1541 Low-risk patients, n = 471 Intermediate-risk patients, n = 689 High-risk patients, n = 381 LNI calculated risk 30,* n = 193 LNI calculated risk 50,* n = 161

Sensitivity, %

Specificity, %

NPV, %

PPV, %

Accuracy, %

13.0 8.3 3.6 17.9 22.0 23.9

96.0 96.3 97.3 94.3 95.5 94.7

89.1 97.6 92.1 72.7 78.3 76.2

32.9 5.6 10.5 58.3 62.4 63.6

54.6 52.3 50.5 56.1 58.7 59.3

NPV = negative predictive value; PPV = positive predictive value; LNI = lymph node invasion. Patients are stratified according to tumor characteristics. * LNI risk was calculated using a previously published prediction model [16]. y

Table 4 – Univariable and multivariable logistic regression analyses predicting lymph node invasion in 1541 patients treated with radical prostatectomy and extended pelvic lymph node dissection at a single institution between 2003 and 2010 Predictors

Prostate-specific antigen Clinical tumor stage T1c T2 T3 Biopsy Gleason score 6 7 8 CT scan findings Negative Positive AUC, % Gain in predictive accuracy*

Univariable analyses

Multivariable analyses Base model

Full model

OR (95% CI)

AUC, %

OR (95% CI)

OR (95% CI)

1.04 (1.03–1.05)y

70.5

1.02 (1.01–1.03)y

1.02 (1.01–1.03)y

1.00 (Ref.) 1.88 (1.28–2.75)y 5.96 (4.04–8.8)y

66.7

1.00 (Ref.) 1.27 (0.84–1.92) 2.37 (1.5–3.72)y

1.00 (Ref.) 1.29 (0.85–1.96) 2.33 (1.48–3.68)y

1.00 (Ref.) 4.21 (2.8–6.33)y 16.98 (10.83–26.6)y

75.3

1.00 (Ref.) 3.41 (2.23–5.19)y 9.92 (6.06–16.22)y

1.00 (Ref.) 3.46 (2.27–5.28)y 9.37 (5.7–15.38)y

(Ref.)y 3.1 (1.9–5.1)y – –

54.6 – –

– 81.3 –

OR = odds ratio; CI = confidence interval; AUC = area under curve; Ref. = reference; CT = computed tomography. p < 0.001. z p = 0.02. * p = 0.8. y

1.00 (Ref.) 2.14 (1.1–4.13)z 81.4 0.1*

1136

4.

EUROPEAN UROLOGY 61 (2012) 1132–1138

Discussion

Virtually all contemporary guidelines recommend that CT scan not be used as a preoperative lymph node staging procedure because of the low sensitivity in detecting small-volume nodal invasion [5,6]. However, reports that examined the performance characteristics of CT scan in predicting LNI were mostly based on historical patient series, in which patients were more likely to be affected by advanced, macroscopic, node-positive disease [7–9]. Thus, the results obtained from these reports may not be applicable to contemporary patients, who may harbor less advanced tumor characteristics [10,11]. Moreover, all previous series addressing the performance characteristics of CT scan were based on limited nodal dissections, which are associated with high false-negative findings [12–14]. Therefore, the ability of CT scan to detect nodal metastases in PCa could not be reliably addressed if many patients were artificially understaged to node-negative status because of an inadequate PLND. On the basis of such significant limitations, we aimed at reassessing the ability of CT to predict LNI in a large contemporary patient cohort treated with ePLND. In our cohort, the rate of LNI was 12%. In patients with LNI, the rate of suspicious lymph nodes at CT scan (sensitivity) was only 13%. Although the sensitivity increased according to tumor aggressiveness, sensitivity reached only a maximum of 24% in patients with a very high risk of nodal invasion (LNI calculated risk 50%; Table 3). Conversely, the specificity of CT scan in predicting LNI was very high regardless of tumor characteristics (range: 94–97%). Nevertheless, these sensitivity and specificity figures resulted in low discrimination accuracy values of only 55% (range: 51–59% according to tumor characteristics). Taken together, our results suggest that in the presence of pathologically defined LNI, CT scan will show positive findings only in roughly 1 of 10 patients. This number will increase to a maximum of approximately 2 of 10 patients when the LNI calculated risk is 50%. Conversely, in the absence of pathologically defined LNI, CT scan results will be negative in the vast majority of cases (96%). However, the overall discrimination accuracy of CT scan is generally poor (55%). Therefore, the use of CT scan for preoperative nodal assessment, even in contemporary patients treated with ePLND, is no better than a coin toss. Moreover, inclusion of data derived from CT scan did not increase the predictive accuracy of a model that included only non–imaging-based information, such as preoperative PSA level, clinical tumor stage, and biopsy Gleason score (81.4% compared with 81.3%; 0.1% gain; p = 0.8). This finding implies that CT scan does not provide any further useful information regarding the prediction of nodal invasion when compared with the information provided by these simple and routinely available covariates, even in contemporary patients treated with ePLND. Therefore, the risk of nodal invasion should be based only on clinical preoperative prediction models such as nomograms based on extended nodal dissection series [16]. The use of such easy risk assessment tools may spare the cost and potential morbidity of CT scan without losing any valuable information regarding lymph node staging.

It is noteworthy that 87% of all patients included in our study could have avoided a preoperative CT scan for nodal staging. Indeed, according to the European Association of Urology guidelines, a CT scan may be warranted only in patients with a very high risk of harboring lymph node metastases [6]. In our study, the rationale for also performing CT scan in patients with more favorable PCa characteristics was mainly based on medicolegal issues. Our results confirm previous findings reporting a significant overuse of imaging modalities in patients with low-risk disease [19,20]. Previous studies have assessed the performance characteristics of CT scan in predicting LNI in PCa [7–9]. The methodology and results of these studies have been critically addressed in a recent meta-analysis by Ho¨vels et al, who examined the performance characteristics of CT scan in predicting LNI based on 18 previously published studies [7]. They reported a pooled sensitivity, specificity, and discrimination accuracy of 42%, 82%, and 77%, respectively. These numbers compare unfavorably with the figures of our study, in which overall sensitivity, specificity, and accuracy was 13%, 96%, and 54.6%, respectively. These differences might have several causes. First, the meta-analysis included data from 18 reports limited by their small sample size (n = 7–285), which may undermine the statistical power. Specifically, the total number of patients included in all these reports was 1024, which is smaller than our current cohort sample size (n = 1541). Consequently, the statistical power of our analyses may be considered higher than virtually all previous reports. Second, most of these reports (13 of 18) were based on patients treated between 1981 and 1987. These patients were thus treated in the pre-PSA era. This idea is key, since the rate of patients with clinically grossly enlarged lymph nodes at that time was significantly higher than what is seen today, when the vast majority of men with LNI are diagnosed with micrometastatic, small-volume nodal disease [16,21,3]. Therefore, the performance characteristics of CT scan might have been artificially inflated by the higher prevalence of patients with high nodal tumor load. Thus, data based on historical patient cohorts may not be applicable to contemporary patients. Moreover, even when considering only contemporary patients with very high risk of nodal invasion (ie, 50%), the sensitivity and discrimination accuracy of CT scan are poor (23.9% and 59.3%, respectively). More important, the PPV associated with CT scan in such patients is only 63%. This finding has significant clinical implications. Indeed, patients with suspicious nodes at CT scan and aggressive disease at diagnosis are often spared treatments with radical intent because of high probability of nodal invasion and, thus, of disseminated disease. However, based on our data, 47% of patients with suspicious CT scan and a highly aggressive disease at diagnosis (LNI risk 50%) do not harbor nodal metastases at ePLND. Therefore, nodal information derived by CT scan should not be used in the clinical decision-making process, even in the high-risk setting. Third, virtually all patients included in these reports received a limited PLND, which is associated with high

EUROPEAN UROLOGY 61 (2012) 1132–1138

false-negative results [10–12]. Conversely, all our patients received a standardized anatomically defined ePLND, which is considered the standard of care by virtually all contemporary guidelines [5,6]. All these factors contribute to significantly increase the solidity and generalizability of our findings. Despite several strengths, our study is not without limitations. One limitation is related to differences between the number of lymph nodes removed and the number of lymph nodes that are actually examined by the pathologist. Differences in patient anatomy, as well as in pathologic assessment, account for discrepancies in the number of lymph nodes contained within the fibrofatty tissue specimen. However, this limitation is applicable to virtually all previous reports that addressed a similar subject [12–14]. Moreover, despite potential differences between the number of lymph nodes removed and the number examined, all patients were treated with an anatomically defined ePLND. Furthermore, it may be argued that other imaging techniques, such as carbon 11 (11C)–labeled choline positron emission tomography (PET)/CT and novel magnetic resonance imaging (MRI) approaches, might perform better in predicting LNI [22,23]. However, the only prospective trial addressing the role of 11C-labeled choline PET/CT scan and diffusion-weighted MRI in patients with high-risk PCa has shown limited sensitivity in detecting LNI [24]. Taken together, all these data support the evidence that risk stratification models instead of imaging techniques should be used to assess the risk of LNI in patients with PCa. Furthermore, it is possible that micrometastases in enlarged lymph nodes might have been missed because of the lack of a systematic use of immunohistochemistry at pathologic nodal assessment. In addition, the lack of a central scan review performed by independent radiologists represents another limitation of our study. Finally, it is possible that other CT scan–derived nodal information, such as nodal contrast uptake and lack of hilar fat, might have increased the sensitivity of CT scan. However, the revised Response Evaluation Criteria in Solid Tumors guideline includes only the size of the lymph nodes as an indicator of nodal invasion and response to therapy [17].

1137

Study concept and design: Briganti. Acquisition of data: Bianchi, Tutolo, Passoni, Nini. Analysis and interpretation of data: Abdollah, Briganti, Gallina, Suardi, Capitanio, Freschi. Drafting of the manuscript: Briganti, Abdollah, Nini. Critical revision of the manuscript for important intellectual content: Briganti, Abdollah, Salonia, Gallina, Suardi, Freschi, Montorsi, Rigatti. Statistical analysis: Briganti, Abdollah, Gallina, Suardi, Capitanio. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Montorsi, Rigatti. Other (specify): None. Financial disclosures: I certify 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] Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin 2010;60:277–300. [2] Abdollah F, Sun M, Thuret R, et al. A competing-risks analysis of survival after alternative treatment modalities for prostate cancer patients: 1988–2006. Eur Urol 2011;59:88–95. [3] Briganti A, Karnes JR, Da Pozzo LF, et al. Two positive nodes represent a significant cut-off value for cancer specific survival in patients with node positive prostate cancer: a new proposal based on a two-institution experience on 703 consecutive N+ patients treated with radical prostatectomy, extended pelvic lymph node dissection and adjuvant therapy. Eur Urol 2009;55:261–70. [4] Freedland SJ, Partin AW, Humphreys EB, et al. Radical prostatectomy for clinical stage T3a disease. Cancer 2007;109:1273–8. [5] NCCN clinical practice guidelines in oncology 2011: prostate cancer. http://www.nccn.org/professionals/physician_gls/pdf/prostate.pdf. Accessed June 6, 2011. [6] Heidenreich A, Bellmunt J, Bolla M, et al. EAU guidelines on prostate cancer. Part 1: screening, diagnosis, and treatment of clinically localised disease. Eur Urol 2011;59:61–71. [7] Ho¨vels AM, Heesakkers RA, Adang EM, et al. The diagnostic accuracy of CT and MRI in the staging of pelvic lymph nodes in patients with prostate cancer: a meta-analysis. Clin Radiol 2008;63:387–95.

5.

Conclusions

[8] Engeler CE, Wasserman NF, Zhang G. Preoperative assessment of prostatic carcinoma by computerized tomography: weaknesses

The accuracy of CT scan as a preoperative lymph node staging procedure in contemporary patients treated with ePLND is poor, even in those patients with a very high risk of nodal metastases. Moreover, use of CT scan failed to increase the accuracy for LNI prediction of a model based on routinely available clinical parameters, such as PSA, clinical stage, and biopsy Gleason sum. Therefore, even in contemporary patients, CT scan should not be used to decide on the need for, or extent of, PLND in patients with PCa. Conversely, assessment of the risk of LNI should be based only on preoperative clinical models developed on ePLND series.

and new perspectives. Urology 1992;40:346–50. [9] Flanigan RC, McKay TC, Olson M, et al. Limited efficacy of preoperative computed tomographic scanning for the evaluation of lymph node metastasis in patients before radical prostatectomy. Urology 1996;48:428–32. [10] Moore AL, Dimitropoulou P, Lane A, et al. Population-based prostate-specific antigen testing in the UK leads to a stage migration of prostate cancer. BJU Int 2009;104:1592–8. [11] Gallina A, Chun FK, Suardi N, et al. Comparison of stage migration patterns between Europe and the USA: an analysis of 11 350 men treated with radical prostatectomy for prostate cancer. BJU Int 2008; 101:1513–8. [12] Abdollah F, Sun M, Thuret R, et al. Decreasing rate and extent of

Author contributions: Alberto Briganti had full access to all the data in

lymph node staging in patients undergoing radical prostatectomy

the study and takes responsibility for the integrity of the data and the

may undermine the rate of diagnosis of lymph node metastases in

accuracy of the data analysis.

prostate cancer. Eur Urol 2010;58:882–92.

1138

EUROPEAN UROLOGY 61 (2012) 1132–1138

[13] Heidenreich A, Varga Z, Von Knobloch R. Extended pelvic

[20] Choi WW, Williams SB, Gu X, Lipsitz SR, Nguyen PL, Hu JC. Overuse

lymphadenectomy in patients undergoing radical prostatectomy:

of imaging for staging low risk prostate cancer. J Urol 2011;185:

high incidence of lymph node metastasis. J Urol 2002;167:1681–6.

1645–9.

[14] Briganti A, Chun FK, Salonia A, et al. Critical assessment of ideal

[21] Schumacher MC, Burkhard FC, Thalmann GN, et al. Good outcome

nodal yield at pelvic lymphadenectomy to accurately diagnose

for patients with few lymph node metastases after radical retro-

prostate cancer nodal metastasis in patients undergoing radical retropubic prostatectomy. Urology 2007;69:147–51. [15] Briganti A, Blute ML, Eastham JH, et al. Pelvic lymph node dissection in prostate cancer. Eur Urol 2009;55:1251–65. [16] Briganti A, Chun FK-H, Salonia A, et al. Validation of a nomogram predicting the probability of lymph node invasion among patients undergoing radical prostatectomy and an extended pelvic lymphadenectomy. Eur Urol 2006;49:1019–27. [17] Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228–47. [18] Bochner BH, Cho D, Herr HW, et al. Prospectively packaged lymph

pubic prostatectomy. Eur Urol 2008;54:344–52. [22] Schiavina R, Scattoni V, Castellucci P, et al.

11

C-choline positron

emission tomography/computerized tomography for preoperative lymph-node staging in intermediate-risk and high-risk prostate cancer: comparison with clinical staging nomograms. Eur Urol 2008;54:392–401. [23] Heesakkers RA, Ho¨vels AM, Jager GJ, et al. MRI with a lymph-nodespecific contrast agent as an alternative to CT scan and lymph-node dissection in patients with prostate cancer: a prospective multicohort study. Lancet Oncol 2008;9:850–6. [24] Budiharto T, Joniau S, Lerut E, et al. Prospective evaluation of 11

C-choline positron emission tomography/computed tomography

node dissections with radical cystectomy: evaluation of node count

and diffusion-weighted magnetic resonance imaging for the nodal

variability and node mapping. J Urol 2004;172:1286–90.

staging of prostate cancer with a high risk of lymph node metastases.

[19] Lavery HJ, Brajtbord JS, Levinson AW, et al. Unnecessary imaging for the staging of low-risk prostate cancer is common. Urology 2011; 77:274–8.

Eur Urol 2011;60:125–30.