- Email: [email protected]
Accuracy of Sentinel Lymph Node Biopsy for Inguinal Lymph Node Staging of Penile Squamous Cell Carcinoma: Systematic Review and Meta-Analysis of the Literature
Accuracy of Sentinel Lymph Node Biopsy for Inguinal Lymph Node Staging of Penile Squamous Cell Carcinoma: Systematic Review and Meta-Analysis of the Literature Ramin Sadeghi,* Hassan Gholami, Seyed Rasoul Zakavi, Vahid Reza Dabbagh Kakhki, Kamyar Tavakkoli Tabasi and Simon Horenblas From the Nuclear Medicine Research Center (RS, SRZ, VRDK) and Department of Urology (KTT), Mashhad University of Medical Sciences, and Faculty of Medicine, Imam Reza Hospital, and Evidence Based Medicine Committee (RS, HG), Mashhad University of Medical Sciences, Mashhad, Iran, and Department of Urology, The Netherlands Cancer Institute – Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands (SH)
Purpose: Sentinel lymph node biopsy is emerging as a promising method for inguinal lymph node staging of penile squamous cell carcinoma. In the current systematic review we evaluated the accuracy of sentinel lymph node biopsy for inguinal lymph node staging of penile squamous cell carcinoma and studied possible influential factors. Materials and Methods: MEDLINE®, Scopus®, ISI®, Ovid SP®, Springer, ScienceDirect® and Google™ Scholar were searched by the key words “(penile OR penis) AND sentinel”. No date or language limitation was imposed on the search and meeting abstracts were not excluded from analysis. A random effects model was used for statistical pooling. Results: A total of 17 studies suitable for meta-analysis were detected. Three articles had 2 different subgroups of patients and each subgroup was considered as a separate study. Overall 18 studies (including the subgroups) were used for detection rate meta-analysis and 19 for sensitivity meta-analysis. The pooled detection rate was 88.3% (95% CI 81.9 –92.6). Pooled detection rate of 90.1% (95% CI 83.6 –94.1) was calculated for the studies using blue dye and radiotracer. The pooled sensitivity was 88% (95% CI 83–92). The highest pooled sensitivity (92% [95% CI 86 –96]) was in the studies using radiotracer and blue dye, and recruiting only cN0 cases. Conclusions: Sentinel lymph node mapping in penile squamous cell carcinoma is a method with a high detection rate and sensitivity. Using radiotracer and blue dye for sentinel lymph node mapping and including only cN0 disease ensures the highest detection rate and sensitivity.
Abbreviations and Acronyms FNA ⫽ fine needle aspiration SCC ⫽ squamous cell carcinoma SLN ⫽ sentinel lymph node Submitted for publication March 9, 2011. Supported by the Vice Chancellery of Research of Mashhad University of Medical Sciences. Nothing to disclose. Supplementary material for this article can be obtained at http://www.mums.ac.ir/shares/nmrc/ nmrc/Tables.pdf. * Correspondence: Nuclear Medicine Research Center, Mashhad University of Medical Sciences, Faculty of Medicine, Imam Reza Hospital, Mashhad, Iran (telephone: ⫹98(511)8022729; FAX: ⫹98(511) 8599359; e-mail: [email protected]).
Editor’s Note: This article is the second of 5 published in this issue for which category 1 CME credits can be earned. Instructions for obtaining credits are given with the questions on pages 364 and 365.
Key Words: meta-analysis; penile neoplasms; sentinel lymph node biopsy; carcinoma, squamous cell; review SENTINEL lymph node biopsy is a fairly new technique in medical practice and is becoming the standard of care for regional lymph node staging of many solid tumors. This technique is based on the hypothesis of orderly distribution of malignant cells in the lymphatic system. It proposes the first lymph
node(s) in the lymphatic drainage of a tumor (called SLN) are representative of the regional lymph node basin(s).1 SLN biopsy is actually the preferred method of lymph node staging in breast cancer2 and melanoma.3 This field is also rapidly extending into other urological malignancies.4
0022-5347/12/1871-0025/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
Vol. 187, 25-31, January 2012 Printed in U.S.A. DOI:10.1016/j.juro.2011.09.058
AND
RESEARCH, INC.
www.jurology.com
25
26
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
It is worth mentioning that the SLN concept was applied to penile cancer for the first time by Cabanas.5 Since then several groups have reported the accuracy of SLN biopsy in penile cancer as an alternative to inguinal lymph node dissection and this procedure was included in the 2009 European Association of Urology guidelines on penile cancer.6 Inguinal lymph node dissection is used in patients with penile cancer for 2 reasons. 1) Staging has strong prognostic implications in penile cancer. Node-free patients have excellent survival while those with lymph node involvement have a significantly worse prognosis.7 Although most patients presenting with penile carcinoma are clinically node negative, approximately 20% to 25% have occult regional lymph node metastasis, which indicates the same grave prognosis.8 2) Compared to surgical removal when metastases become clinically palpable, regional lymph node dissection improves survival considerably.8 –10 However, regional lymph node dissection in penile cancer results in significant morbidity and, as mentioned, is not necessary in approximately 75% to 80% of patients.10 The main advantage of SLN biopsy in penile cancer is to decrease this morbidity without compromising the survival benefit. In 2009 results of the largest study on sentinel node biopsy of penile cancer were published.11 Several other research groups also reported their experience in this field, and systematic review and meta-analysis is useful, especially given the scarcity of this malignancy as well as the small sample size of many published studies. Therefore, we performed a comprehensive search of the literature to evaluate the sensitivity and detection rate of SLN biopsy in inguinal lymph node staging of penile squamous cell carcinoma using the meta-analysis of the included studies.
MATERIALS AND METHODS Search Strategy, Selection Criteria and Data Abstraction Several databases were comprehensively searched for relevant articles including MEDLINE, Scopus, ISI, OVID SP, Springer, ScienceDirect and Google Scholar (last search December 2010). The search terms (free text not MeSH [Medical Subject Headings] terms) were “(penile OR penis) AND sentinel”. No date or language limitation was imposed on the search and meeting abstracts were also not excluded. Only studies performed on SCC of the penis were included in the study. The reference lists of the retrieved studies were hand searched for possible missing relevant articles. For more complete data we contacted the corresponding authors when necessary. Studies were included for sensitivity evaluation provided they 1) used inguinal lymph node dissection for all patients or followup of those SLN negative as the gold
standard, 2) recruited at least 5 patients, 3) reported the number of groins with involved inguinal lymph nodes (identified by inguinal lymph node dissection or inguinal recurrence during followup) as well as the number of false-negative groins, and 4) used modern techniques of SLN mapping (blue dyes and/or radiotracers). Studies were included for detection rate evaluation provided they 1) recruited at least 5 patients, 2) had data on the total number of groins and number of groins with detected sentinel nodes during surgery, and 3) used modern techniques of SLN mapping (blue dyes and/or radiotracers). Two of the authors reviewed the retrieved articles independently. The third author opinion was used in case of any controversy. Possible duplicate studies were discussed and the most recent reports were included in the metaanalysis. Not all studies that were included had the highest quality. The quality of the included studies was evaluated using the Oxford Centre for Evidence Based Medicine checklist for diagnostic studies.12 This checklist has several sections outlining the best possible quality for each. The best method for patient recruitment would be prospective consecutive inclusion (or random inclusion). The best reference standard is radical inguinal lymph node dissection for all patients regardless of SLN biopsy results. Alternatively, followup of the patients with negative SLN biopsy results would also be acceptable. Followup and/or pathological evaluation of lymph nodes should be done blindly regarding the SLN biopsy results. The best spectrum of the patients would be all cN0 (including those with palpable inguinal lymph nodes with negative FNA results). Data abstraction was done by 2 authors independently, and data on authors, publication year, method, patient characteristics, detection and false-negative rates were recorded.
Statistical Analysis Considering the different methods used in the included studies, we applied a random effects model (Der-Simonian and Laird method13) for the meta-analysis of the proportions without any transformation. The random effects model is a statistical method in which the between study variability is accounted for. This method is used especially in pooling data across studies which are different in terms of design, included patients etc. The Cochrane Q test was used for heterogeneity evaluation and the significance level was set at p ⬍0.05. The I2 index was also calculated to quantify the extent of heterogeneity.14 The I2 index shows the amount of variability that is caused by real heterogeneity and is not spurious. Funnel plots, Egger’s regression intercept,15 and Duval-Tweedie’s trim and fill16 method were used for publication bias evaluation. The Funnel plot is the plot of the standard errors of the included studies on the y-axis and the effect size on the x-axis. Asymmetry of this plot may be due to publication bias. Egger’s regression intercept is the mathematical counterpart of this visual assessment. Statistically significant results of this test indicate a large asymmetry in the funnel plot. Duval-Tweedie’s trim and fill method uses an iterative method to omit the smallest studies from the funnel plot until a symmetrical plot
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
27
ies including clinically node positive disease and the effect of inguinal sonography in all patients before SLN biopsy.
RESULTS
Figure 1. Process of study inclusion in meta-analysis
results. The new measured effect size by this method shows the impact of possible publication bias. For statistical analyses Comprehensive Meta-Analysis (version 2) and Meta-DiSc (version 1.4) were used.17 Major indices of interest were detection rate and sensitivity. Specificity was not used because it is always 100% in studies of sentinel node biopsy accuracy. Subgroup analysis or sensitivity analysis was used to evaluate the lymphatic mapping method (blue dye and/or tracer), number of recruited patients (more than 30 or less than 30), stud-
Studyname
Year of publication
Figure 1 shows the diagram of the study search strategy. A total of 5,590 studies seemed to be potentially relevant in the first search. However, 5,376 studies were excluded (irrelevant subjects) by screening the titles and/or abstracts. The number of evaluated citations was high due to the inclusion of Google Scholar in the search strategy, which has a low specificity in any search. Full text of the remaining 214 studies was evaluated in more depth. In addition, 197 studies were excluded as narrative review articles, duplicates, letters to editors, recruiting less than 5 patients, or using methods other than blue dye and/or radiotracers. The remaining 17 articles were included in the meta-analysis.11,18 –33 Of the included studies 3 were meeting abstracts.19,31,33 Three articles had 2 different subgroups of patients and each subgroup was considered as a separate study.11,25,27 Overall 18 studies (including the subgroups mentioned) met the criteria of detection rate meta-analysis and 19 met the criteria of sensitivity meta-analysis. Detection Rate Figure 2 shows the forest plot of the detection rate meta-analysis from the 18 included studies. The
Statistics for each study Detection Event
rate
Lower limit
Upper limit
Z-Value
Detection rate and 95% CI p-Value
Akduman et al
2001
0.950
0.525
0.997
2.029
0.042
Perdona et al
2005
0.795
0.651
0.890
3.634
0.000
Brennhovd et al
2006
0.820
0.689
0.904
4.119
0.000
Gonzaga-Silva et al
2007
0.537
0.405
0.664
0.544
0.587
Leijte et al
2007
0.897
0.843
0.934
8.704
0.000
Spiess et al
2007
0.976
0.713
0.999
2.594
0.009
Spiess et al
2007
0.762
0.611
0.867
3.211
0.001
Hernandez-Toris Hernandez Toris et al
2007
0 944 0.944
0 693 0.693
0 992 0.992
2 753 2.753
0 006 0.006
Rubi et al
2008
0.900
0.533
0.986
2.084
0.037 0.004
Rubi et al
2008
0.950
0.718
0.993
2.870
Ferreira et al
2008
0.813
0.641
0.913
3.238
0.001
Heyns et al
2008
0.913
0.790
0.967
4.494
0.000
Jensen et al
2008
0.969
0.908
0.990
5.873
0.000
Leijte et al
2009
0.958
0.930
0.975
11.448
0.000
Leijte et al
2009
0.962
0.931
0.979
10.021
0.000
Vinayak et al
2009
0.886
0.755
0.952
4.324
0.000
Lemstra et al
2010
0.760
0.623
0.858
3.481
0.000
von Forstner et al
2010
0.813
0.677
0.899
3.965
0.000
0.883
0.819
0.926
7.786
0.000
Summary effect
-1.00 0.00
Figure 2. Forrest plot of studies included for detection rate pooling
0.50 -0.50
0.00 1.00
28
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
detection rate meta-analysis. Egger’s regression intercept was 1.21 (p ⫽ 0.20124), which indicates a fairly symmetrical funnel plot. The adjusted pooled detection rate after trimming 6 studies (using Duval-Tweedie’s trim and fill method) was 82.6% (95% CI 73.5– 89). This shows that the best estimate of unbiased effect size is 5.7% lower than the calculated pooled detection rate. The subgroup analysis of studies using blue dye and radiotracers for SLN mapping vs those using only 1 method demonstrated pooled detection rates of 90.1% (95% CI 83.6 –94.1) vs 85.1% (95% CI 71.3–92.9), respectively.
detection Funnel Plot of Standard Error by Logit event raterate 0.0
Standard Errorr S
0.5
1.0
1.5
2.0 20 -4
-3
-2
-1
0
1
2
3
4
detection Logit event raterate
Figure 3. Funnel plot of studies included for detection rate pooling. Note asymmetry of plot which indicates possible important publication bias.
pooled detection rate was 88.3% (95% CI 81.9 –92.6). The Cochrane Q value was 105.942 (p ⬍0.0001) and I2 ⫽ 83.95%. Figure 3 shows the funnel plot of the
Sensitivity Figure 4 shows the forest plot of the sensitivity meta-analysis from 19 included studies. The pooled sensitivity was 88% (95% CI 83–92). The Cochrane Q value was 31.04 (p ⫽ 0.0285) and I2 ⫽ 42%. Figure 5 shows the funnel plot of the included studies for the sensitivity meta-analysis. The Egger’s regression intercept was -0.65 (p ⫽ 0.1743), which indicates a symmetrical funnel plot. Duval-Tweedie’s trim and fill method was unable to find any study for trimming and the adjusted pooled sensitivity was the
Figure 4. Forrest plot of studies included for sensitivity pooling
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
sensitivity Funnel Plot of Standard Error by Logit event rate 0.0
Standard Errorr S
0.5
1.0
1.5
20 2.0 -4
-3
-2
-1
0
1
2
3
4
sensitivity Logit event rate
Figure 5. Funnel plot of studies included for sensitivity pooling. Note asymmetry of plot is less than that of funnel plot of detection rate.
same as before trimming. Therefore, the best estimate of unbiased effect size was not different from the calculated pooled sensitivity. Subgroup analysis of the SLN mapping method (blue dye and/or radiotracer) showed a pooled sensitivity of 60% (95% CI 15–95) for blue dye alone, 84% (95% CI 72–92) for radiotracer alone and 90% (95% CI 85–95) for studies using both methods. Excluding 2 studies in which patients with palpable inguinal node regardless of FNA results were recruited21,29 resulted in a pooled sensitivity of 90% (95% CI 85–94). Considering the SLN mapping method, the pooled sensitivity of studies using radiotracer and blue dye for SLN mapping and recruiting only patients with cN0 disease was 92% (95% CI 86 –96). Only 2 studies (2 subgroups of the study by Leijte et al11) used inguinal ultrasonography to detect suspicious inguinal lymph nodes and performed FNA if necessary. Pooling the sensitivity of these 2 studies revealed a 93% sensitivity (95% CI 85–97).
DISCUSSION The condition of regional lymph nodes is an important factor affecting the survival of patients with several solid tumors, and penile cancer is no exception in this regard.34,35 An important aspect of penile SCC in contrast to many other solid tumors is the therapeutic benefit of regional lymph node dissection in this tumor.7 Immediate resection of occult lymph node metastases improves the survival of patients with penile cancer considerably compared to the delayed resection of palpable inguinal lymph node involvement.8 –10 The relative ineffectiveness of chemotherapy and radiotherapy compared to the surgical approach is also important in this regard.36
29
However, prophylactic inguinal lymph node dissection for penile cancer is not required in those without regional lymph node involvement, which constitutes almost 75% to 80% of cN0 disease.7,10 In addition to the significant morbidity of inguinal lymph node dissection,37 this is the main impetus behind the search of urologists for an alternative to inguinal lymph node dissection. Several alternatives to inguinal lymph node dissection are currently under study and several nomograms for predicting inguinal lymph node involvement are currently in use.6,38 However, their performance is not perfect enough and needs to be externally validated.39,40 Several imaging techniques such as computerized tomography or conventional magnetic resonance imaging can also be used for inguinal lymph node staging. However, their accuracy is also suboptimal, especially for cN0 disease.41– 43 Although novel techniques such as positron emission tomography/ computerized tomography44 and nanoparticle enhanced magnetic resonance imaging45 seem to be promising in this regard, their sensitivity for detecting micrometastases is actually low.43,45 SLN biopsy is another technique which is promising for inguinal lymph node staging and our current meta-analysis determines the accuracy of this approach. The overall detection rate (per groin, not per patient) of 88.3% (95% CI 81.9 –92.6) was fairly high. However, the I2 value was 83.95%, meaning that most of the observed variance between the detection rates of the included studies was real and could be attributed to the between study differences (and not only to sampling errors of the included studies). To address this high level of heterogeneity we performed a subgroup analysis of the SLN mapping method. The pooled detection rate of the studies using radiotracer and blue dye was 5% more than the studies using only 1 of the methods (90.1% [95% CI 83.6 –94.1] vs 85.1% [95% CI 71.3–92.9]). This shows the importance of methodology in SLN mapping and this issue should be considered carefully while performing SLN biopsy for penile cancer. The sensitivity of SLN mapping for inguinal lymph node staging is the more important issue in our meta-analysis. The pooled sensitivity of the included studies was 88% (95% CI 83–92), which was also fairly high. However, the heterogeneity of the included studies (I2 ⫽ 42%) warranted evaluation in more depth. The spectrum of the patients recruited in the studies seemed to be important. Excluding 2 studies which recruited patients with palpable inguinal lymph nodes (without performing FNA) increased the pooled sensitivity to 90% (95% CI 85–94).21,29 This was also supported by Leijte et al, who reported rerouting of lymphatic drainage in 76% of the groins
30
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
with palpable lymph nodes with resultant false-negative SLN biopsy.46 Another aspect of the included studies was the method of SLN mapping. The highest sensitivity was in the studies using blue dye and radiotracer (90% [95% CI 85–95]) and the lowest sensitivity was in those using only blue dye for SLN mapping (60% [95% CI 15–95]). Excluding the studies that recruited patients with palpable inguinal lymph nodes as well as those using only blue dye or radiotracers increased the sensitivity to 92% (95% CI 86 –96), which is high enough to be considered safe and is comparable to recommendations for patients with breast cancer.47 A method for decreasing the false-negative rate proposed by Leijte et al is performing inguinal ultrasonography to detect suspicious inguinal nodes requiring FNA for better evaluation.24 They reported a decreased false-negative rate by this method. We also performed sensitivity analysis and the pooled sensitivity of the 2 studies with inguinal ultrasonography in all patients was 93%, which is in agreement with the study by Leijte et al. The sensitivity of SLN mapping, with proven safety and efficacy in many cancers such as breast,2 prostate,4 melanoma48 etc, is not different from what we found for penile cancer in our meta-analysis. In fact, SLN mapping is included in the 2009 European Association of Urology guidelines on penile cancer (level of evidence 2b).6 Our study did have some limitations. Penile SCC is one of the rare carcinomas in humans, which results in a low sample size in many studies evaluating this malignancy. This is especially true when it comes to studies on SLN biopsy for penile SCC. For statistical reasons we excluded studies with a sample size of less than 5 patients (10 groins). In fact, pooling all relevant studies improves overall precision, and the fairly narrow confidence intervals of our meta-analysis results are also in agreement with this idea. Another related limitation is the possibility of a learning curve for the performance of SLN biopsy in penile cancer. This is a major concern since this malignancy is rare and the urologist would not have the chance to pass the learning curve similar to surgeons treating breast cancer. Leijte et al addressed this issue (by evaluating the first 30 cases for false-negative rate and complications) and they could not demonstrate the presence of this phenomenon.11 However, as they mentioned this could be due to pure chance and should be evaluated more
in future studies. Actually only 3 of all included studies in the current meta-analysis recruited more than 30 patients.11,24,30 Subgroup analysis using only these 3 studies demonstrated a pooled detection rate and sensitivity of 94.9% (95% CI 90.9 –97.2) and 90.9% (95% CI 84.7–95.2), respectively, which are higher than overall pooled indices, and show the effect of a high volume of patients on the success of the sentinel node procedure. In introducing SLN biopsy a main obstacle could be the necessity of multidisciplinary teams, especially considering that the geographic areas with the highest rates of penile cancer have limited resources and would have more limitations in this regard.28 A particular aspect of many studies evaluating SLN mapping accuracy in penile SCC is the reliance on followup to determine the false-negative rate. Most inguinal recurrence in penile SCC occurs within approximately 2 years after surgery.11,24 However, not all studies included in the current meta-analysis followed all patients that long, which is a major limitation of our meta-analysis. An important issue in all meta-analyses is publication bias. We included meeting abstracts as well as nonEnglish studies in the meta-analysis, which can decrease publication bias. Although Egger’s tests were not statistically significant in sensitivity or detection rate pooling, this test has low power and visual assessment of the funnel plots, and shows some asymmetry, especially for the detection rate pooling. This was also supported by Duval-Tweedie’s trim and fill method, which showed an almost 6% decrease in the pooled detection rate after trimming while no study could be trimmed for sensitivity pooling. Thus, another limitation of our study is that publication bias, if present, affects the results of the current meta-analysis (especially for detection rate, not sensitivity).
CONCLUSIONS The current meta-analysis showed that SLN mapping in penile SCC is a method with a high detection rate and sensitivity. However, the technical aspects of this procedure are important. Using radiotracer and blue dye for SLN mapping and including only cN0 disease ensures the highest detection rate and sensitivity. However, the possible presence of a learning curve as well as limited resources in many areas with a high incidence of this malignancy could reduce the usefulness of this procedure.
REFERENCES 1. Purdie CA: Sentinel lymph node biopsy: review of the literature and guidelines for pathological handling and reporting. Curr Diagn Pathol 2007; 13: 106.
2. Miltenburg DM, Miller C, Karamlou TB et al: Meta-analysis of sentinel lymph node biopsy in breast cancer. J Surg Res 1999; 84: 138.
3. Warycha MA, Zakrzewski J, Ni Q et al: Metaanalysis of sentinel lymph node positivity in thin melanoma (⬍/⫽1 mm). Cancer 2009; 115: 869.
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
4. Sadeghi R, Tabasi KT, Bazaz SM et al: Sentinel node mapping in the prostate cancer. Meta-analysis. Nuklearmedizin 2011; 50: 107. 5. Cabanas RM: An approach for the treatment of penile carcinoma. Cancer 1977; 39: 456. 6. Pizzocaro G, Algaba F, Horenblas S et al: EAU penile cancer guidelines 2009. Eur Urol 2010; 57: 1002.
21. Hungerhuber E, Schlenker B, Frimberger D et al: Lymphoscintigraphy in penile cancer: limited value of sentinel node biopsy in patients with clinically suspicious lymph nodes. World J Urol 2006; 24: 319. 22. Brennhovd B, Johnsrud K, Berner A et al: Sentinel node procedure in low-stage/low-grade penile carcinomas. Scand J Urol Nephrol 2006; 40: 204.
31
35. Dechev J, Banchev A and Georgieva-Mateva N: Overall survival in carcinoma of the penis–a single institution study. J BUON 2007; 12: 377. 36. Lubke WL and Thompson IM: The case for inguinal lymph node dissection in the treatment of T2-T4, N0 penile cancer. Semin Urol 1993; 11: 80.
7. Wespes E: The management of regional lymph nodes in patients with penile carcinoma and reliability of sentinel node biopsy. Eur Urol 2007; 52: 15.
23. Gonzaga-Silva LF, Tavares JM, Freitas FC et al: The isolated gamma probe technique for sentinel node penile carcinoma detection is unreliable. Int Braz J Urol 2007; 33: 58.
37. Tobias-Machado M, Tavares A, Ornellas AA et al: Video endoscopic inguinal lymphadenectomy: a new minimally invasive procedure for radical management of inguinal nodes in patients with penile squamous cell carcinoma. J Urol 2007; 177: 953.
8. Kroon BK, Horenblas S, Lont AP et al: Patients with penile carcinoma benefit from immediate resection of clinically occult lymph node metastases. J Urol 2005; 173: 816.
24. Leijte JA, Kroon BK, Valdes Olmos RA et al: Reliability and safety of current dynamic sentinel node biopsy for penile carcinoma. Eur Urol 2007; 52: 170.
38. Ficarra V, Zattoni F, Artibani W et al: Nomogram predictive of pathological inguinal lymph node involvement in patients with squamous cell carcinoma of the penis. J Urol 2006; 175: 1700.
9. McDougal WS, Kirchner FK Jr, Edwards RH et al: Treatment of carcinoma of the penis: the case for primary lymphadenectomy. J Urol 1986; 136: 38.
25. Spiess PE, Izawa JI, Bassett R et al: Preoperative lymphoscintigraphy and dynamic sentinel node biopsy for staging penile cancer: results with pathological correlation. J Urol 2007; 177: 2157.
39. Ficarra V, Novara G, Boscolo-Berto R et al: How accurate are present risk group assignment tools in penile cancer? World J Urol 2009; 27: 155.
26. Hernandez-Toris N, Quintero-Becerra J, GallegosHernandez JF et al: Lymphatic mapping and sentinel node biopsy in penis cancer. Feasibility study and preliminary report. Cir Cir 2007; 75: 87.
40. Novara G, Artibani W, Cunico SC et al: How accurately do Solsona and European Association of Urology risk groups predict for risk of lymph node metastases in patients with squamous cell carcinoma of the penis? Urology 2008; 71: 328.
27. Rubi S, Vidal-Sicar S, Ortega M et al: Localization of sentinel node in squamous cell carcinoma of the penis. Initial experience. Rev Esp Med Nucl 2008; 27: 3.
41. Heyns CF, Fleshner N, Sangar V et al: Management of the lymph nodes in penile cancer. Urology 2011; 76: S43.
10. Protzel C, Alcaraz A, Horenblas S et al: Lymphadenectomy in the surgical management of penile cancer. Eur Urol 2009; 55: 1075. 11. Leijte JA, Hughes B, Graafland NM et al: Twocenter evaluation of dynamic sentinel node biopsy for squamous cell carcinoma of the penis. J Clin Oncol 2009; 27: 3325. 12. Oxford Centre for Evidence-Based Medicine: Levels of Evidence. Available at http://www.cebm. net/index.aspx?o⫽1025. Accessed October 5, 2010. 13. DerSimonian R and Laird N: Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177. 14. Higgins JP and Thompson SG: Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21: 1539. 15. Egger M, Davey Smith G, Schneider M et al: Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629. 16. Duval S and Tweedie R: A nonparametric ‘trim and fill’ method of accounting for publication bias in meta-analysis. J Am Stat Assoc 2000; 95: 89. 17. Zamora J, Abraira V, Muriel A et al: Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6: 31. 18. Akduman B, Fleshner NE, Ehrlich L et al: Early experience in intermediate-risk penile cancer with sentinel node identification using the gamma probe. Urology 2001; 58: 65.
28. Ferreira U, Ribeiro MA, Reis LO et al: Sentinel lymph node biopsy in penile cancer: a comparative study using modified inguinal dissection. Int Braz J Urol 2008; 34: 725. 29. Heyns CF and Theron PD: Evaluation of dynamic sentinel lymph node biopsy in patients with squamous cell carcinoma of the penis and palpable inguinal nodes. BJU Int 2008; 102: 305. 30. Jensen JB, Jensen KM, Ulhoi BP et al: Sentinel lymph-node biopsy in patients with squamous cell carcinoma of the penis. BJU Int 2009; 103: 1199. 31. Pytel A, Damasdi M, Frick A et al: Surgical management of low and medium risk penile cancers, with isotope guided sentinel lymph node biopsy technique. Urology 2009; 74: S63. 32. Vinayak RS and Rajen TA: Sentinel node biopsy in carcinoma penis using methylene blue dye technique. Urol Ann 2009; 1: 18.
19. Lemstra C, de Jong I, Leliveld A et al: Evaluation of lymphoscintigraphy and sentinel node biopsy in patients with penile carcinoma. J Nucl Med 2010; 51: 2004.
33. von Forstner C, Naumann M, Henze E: Improvement of the surgical procedure for penile carcinoma by sentinel lymph node scintigraphy with SPECT and SPECT/CT. Eur J Nucl Med Mol Imaging 2010; 37: S270.
20. Perdona S, Autorino R, De Sio M et al: Dynamic sentinel node biopsy in clinically node-negative penile cancer versus radical inguinal lymphadenectomy: a comparative study. Urology 2005; 66: 1282.
34. Zini L, Cloutier V, Isbarn H et al: A simple and accurate model for prediction of cancer-specific mortality in patients treated with surgery for primary penile squamous cell carcinoma. Clin Cancer Res 2009; 15: 1013.
42. Kochhar R, Taylor B and Sangar V: Imaging in primary penile cancer: current status and future directions. Eur Radiol 2011; 20: 36. 43. Hughes B, Leijte J, Shabbir M et al: Non-invasive and minimally invasive staging of regional lymph nodes in penile cancer. World J Urol 2009; 27: 197. 44. Graafland NM, Leijte JA, Valdes Olmos RA et al: Scanning with 18F-FDG-PET/CT for detection of pelvic nodal involvement in inguinal node-positive penile carcinoma. Eur Urol 2009; 56: 339. 45. Mueller-Lisse UG, Scher B, Scherr MK et al: Functional imaging in penile cancer: PET/computed tomography, MRI, and sentinel lymph node biopsy. Curr Opin Urol 2008; 18: 105. 46. Leijte JA, van der Ploeg IM, Valdes Olmos RA et al: Visualization of tumor blockage and rerouting of lymphatic drainage in penile cancer patients by use of SPECT/CT. J Nucl Med 2009; 50: 364. 47. MacNeill F, Mansel R, Horgan K et al: NEW START: the United Kingdom sentinel lymph biopsy training programme. A model of how multiprofessional training can achieve competent national performance whilst maintaining patient safety. Breast Cancer Res Treat 2007; 106: S233. 48. Mattsson J, Bergkvist L, Abdiu A et al: Sentinel node biopsy in malignant melanoma: Swedish experiences 1997-2005. Acta Oncol 2008; 47: 1519.
Purpose: Sentinel lymph node biopsy is emerging as a promising method for inguinal lymph node staging of penile squamous cell carcinoma. In the current systematic review we evaluated the accuracy of sentinel lymph node biopsy for inguinal lymph node staging of penile squamous cell carcinoma and studied possible influential factors. Materials and Methods: MEDLINE®, Scopus®, ISI®, Ovid SP®, Springer, ScienceDirect® and Google™ Scholar were searched by the key words “(penile OR penis) AND sentinel”. No date or language limitation was imposed on the search and meeting abstracts were not excluded from analysis. A random effects model was used for statistical pooling. Results: A total of 17 studies suitable for meta-analysis were detected. Three articles had 2 different subgroups of patients and each subgroup was considered as a separate study. Overall 18 studies (including the subgroups) were used for detection rate meta-analysis and 19 for sensitivity meta-analysis. The pooled detection rate was 88.3% (95% CI 81.9 –92.6). Pooled detection rate of 90.1% (95% CI 83.6 –94.1) was calculated for the studies using blue dye and radiotracer. The pooled sensitivity was 88% (95% CI 83–92). The highest pooled sensitivity (92% [95% CI 86 –96]) was in the studies using radiotracer and blue dye, and recruiting only cN0 cases. Conclusions: Sentinel lymph node mapping in penile squamous cell carcinoma is a method with a high detection rate and sensitivity. Using radiotracer and blue dye for sentinel lymph node mapping and including only cN0 disease ensures the highest detection rate and sensitivity.
Abbreviations and Acronyms FNA ⫽ fine needle aspiration SCC ⫽ squamous cell carcinoma SLN ⫽ sentinel lymph node Submitted for publication March 9, 2011. Supported by the Vice Chancellery of Research of Mashhad University of Medical Sciences. Nothing to disclose. Supplementary material for this article can be obtained at http://www.mums.ac.ir/shares/nmrc/ nmrc/Tables.pdf. * Correspondence: Nuclear Medicine Research Center, Mashhad University of Medical Sciences, Faculty of Medicine, Imam Reza Hospital, Mashhad, Iran (telephone: ⫹98(511)8022729; FAX: ⫹98(511) 8599359; e-mail: [email protected]).
Editor’s Note: This article is the second of 5 published in this issue for which category 1 CME credits can be earned. Instructions for obtaining credits are given with the questions on pages 364 and 365.
Key Words: meta-analysis; penile neoplasms; sentinel lymph node biopsy; carcinoma, squamous cell; review SENTINEL lymph node biopsy is a fairly new technique in medical practice and is becoming the standard of care for regional lymph node staging of many solid tumors. This technique is based on the hypothesis of orderly distribution of malignant cells in the lymphatic system. It proposes the first lymph
node(s) in the lymphatic drainage of a tumor (called SLN) are representative of the regional lymph node basin(s).1 SLN biopsy is actually the preferred method of lymph node staging in breast cancer2 and melanoma.3 This field is also rapidly extending into other urological malignancies.4
0022-5347/12/1871-0025/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
Vol. 187, 25-31, January 2012 Printed in U.S.A. DOI:10.1016/j.juro.2011.09.058
AND
RESEARCH, INC.
www.jurology.com
25
26
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
It is worth mentioning that the SLN concept was applied to penile cancer for the first time by Cabanas.5 Since then several groups have reported the accuracy of SLN biopsy in penile cancer as an alternative to inguinal lymph node dissection and this procedure was included in the 2009 European Association of Urology guidelines on penile cancer.6 Inguinal lymph node dissection is used in patients with penile cancer for 2 reasons. 1) Staging has strong prognostic implications in penile cancer. Node-free patients have excellent survival while those with lymph node involvement have a significantly worse prognosis.7 Although most patients presenting with penile carcinoma are clinically node negative, approximately 20% to 25% have occult regional lymph node metastasis, which indicates the same grave prognosis.8 2) Compared to surgical removal when metastases become clinically palpable, regional lymph node dissection improves survival considerably.8 –10 However, regional lymph node dissection in penile cancer results in significant morbidity and, as mentioned, is not necessary in approximately 75% to 80% of patients.10 The main advantage of SLN biopsy in penile cancer is to decrease this morbidity without compromising the survival benefit. In 2009 results of the largest study on sentinel node biopsy of penile cancer were published.11 Several other research groups also reported their experience in this field, and systematic review and meta-analysis is useful, especially given the scarcity of this malignancy as well as the small sample size of many published studies. Therefore, we performed a comprehensive search of the literature to evaluate the sensitivity and detection rate of SLN biopsy in inguinal lymph node staging of penile squamous cell carcinoma using the meta-analysis of the included studies.
MATERIALS AND METHODS Search Strategy, Selection Criteria and Data Abstraction Several databases were comprehensively searched for relevant articles including MEDLINE, Scopus, ISI, OVID SP, Springer, ScienceDirect and Google Scholar (last search December 2010). The search terms (free text not MeSH [Medical Subject Headings] terms) were “(penile OR penis) AND sentinel”. No date or language limitation was imposed on the search and meeting abstracts were also not excluded. Only studies performed on SCC of the penis were included in the study. The reference lists of the retrieved studies were hand searched for possible missing relevant articles. For more complete data we contacted the corresponding authors when necessary. Studies were included for sensitivity evaluation provided they 1) used inguinal lymph node dissection for all patients or followup of those SLN negative as the gold
standard, 2) recruited at least 5 patients, 3) reported the number of groins with involved inguinal lymph nodes (identified by inguinal lymph node dissection or inguinal recurrence during followup) as well as the number of false-negative groins, and 4) used modern techniques of SLN mapping (blue dyes and/or radiotracers). Studies were included for detection rate evaluation provided they 1) recruited at least 5 patients, 2) had data on the total number of groins and number of groins with detected sentinel nodes during surgery, and 3) used modern techniques of SLN mapping (blue dyes and/or radiotracers). Two of the authors reviewed the retrieved articles independently. The third author opinion was used in case of any controversy. Possible duplicate studies were discussed and the most recent reports were included in the metaanalysis. Not all studies that were included had the highest quality. The quality of the included studies was evaluated using the Oxford Centre for Evidence Based Medicine checklist for diagnostic studies.12 This checklist has several sections outlining the best possible quality for each. The best method for patient recruitment would be prospective consecutive inclusion (or random inclusion). The best reference standard is radical inguinal lymph node dissection for all patients regardless of SLN biopsy results. Alternatively, followup of the patients with negative SLN biopsy results would also be acceptable. Followup and/or pathological evaluation of lymph nodes should be done blindly regarding the SLN biopsy results. The best spectrum of the patients would be all cN0 (including those with palpable inguinal lymph nodes with negative FNA results). Data abstraction was done by 2 authors independently, and data on authors, publication year, method, patient characteristics, detection and false-negative rates were recorded.
Statistical Analysis Considering the different methods used in the included studies, we applied a random effects model (Der-Simonian and Laird method13) for the meta-analysis of the proportions without any transformation. The random effects model is a statistical method in which the between study variability is accounted for. This method is used especially in pooling data across studies which are different in terms of design, included patients etc. The Cochrane Q test was used for heterogeneity evaluation and the significance level was set at p ⬍0.05. The I2 index was also calculated to quantify the extent of heterogeneity.14 The I2 index shows the amount of variability that is caused by real heterogeneity and is not spurious. Funnel plots, Egger’s regression intercept,15 and Duval-Tweedie’s trim and fill16 method were used for publication bias evaluation. The Funnel plot is the plot of the standard errors of the included studies on the y-axis and the effect size on the x-axis. Asymmetry of this plot may be due to publication bias. Egger’s regression intercept is the mathematical counterpart of this visual assessment. Statistically significant results of this test indicate a large asymmetry in the funnel plot. Duval-Tweedie’s trim and fill method uses an iterative method to omit the smallest studies from the funnel plot until a symmetrical plot
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
27
ies including clinically node positive disease and the effect of inguinal sonography in all patients before SLN biopsy.
RESULTS
Figure 1. Process of study inclusion in meta-analysis
results. The new measured effect size by this method shows the impact of possible publication bias. For statistical analyses Comprehensive Meta-Analysis (version 2) and Meta-DiSc (version 1.4) were used.17 Major indices of interest were detection rate and sensitivity. Specificity was not used because it is always 100% in studies of sentinel node biopsy accuracy. Subgroup analysis or sensitivity analysis was used to evaluate the lymphatic mapping method (blue dye and/or tracer), number of recruited patients (more than 30 or less than 30), stud-
Studyname
Year of publication
Figure 1 shows the diagram of the study search strategy. A total of 5,590 studies seemed to be potentially relevant in the first search. However, 5,376 studies were excluded (irrelevant subjects) by screening the titles and/or abstracts. The number of evaluated citations was high due to the inclusion of Google Scholar in the search strategy, which has a low specificity in any search. Full text of the remaining 214 studies was evaluated in more depth. In addition, 197 studies were excluded as narrative review articles, duplicates, letters to editors, recruiting less than 5 patients, or using methods other than blue dye and/or radiotracers. The remaining 17 articles were included in the meta-analysis.11,18 –33 Of the included studies 3 were meeting abstracts.19,31,33 Three articles had 2 different subgroups of patients and each subgroup was considered as a separate study.11,25,27 Overall 18 studies (including the subgroups mentioned) met the criteria of detection rate meta-analysis and 19 met the criteria of sensitivity meta-analysis. Detection Rate Figure 2 shows the forest plot of the detection rate meta-analysis from the 18 included studies. The
Statistics for each study Detection Event
rate
Lower limit
Upper limit
Z-Value
Detection rate and 95% CI p-Value
Akduman et al
2001
0.950
0.525
0.997
2.029
0.042
Perdona et al
2005
0.795
0.651
0.890
3.634
0.000
Brennhovd et al
2006
0.820
0.689
0.904
4.119
0.000
Gonzaga-Silva et al
2007
0.537
0.405
0.664
0.544
0.587
Leijte et al
2007
0.897
0.843
0.934
8.704
0.000
Spiess et al
2007
0.976
0.713
0.999
2.594
0.009
Spiess et al
2007
0.762
0.611
0.867
3.211
0.001
Hernandez-Toris Hernandez Toris et al
2007
0 944 0.944
0 693 0.693
0 992 0.992
2 753 2.753
0 006 0.006
Rubi et al
2008
0.900
0.533
0.986
2.084
0.037 0.004
Rubi et al
2008
0.950
0.718
0.993
2.870
Ferreira et al
2008
0.813
0.641
0.913
3.238
0.001
Heyns et al
2008
0.913
0.790
0.967
4.494
0.000
Jensen et al
2008
0.969
0.908
0.990
5.873
0.000
Leijte et al
2009
0.958
0.930
0.975
11.448
0.000
Leijte et al
2009
0.962
0.931
0.979
10.021
0.000
Vinayak et al
2009
0.886
0.755
0.952
4.324
0.000
Lemstra et al
2010
0.760
0.623
0.858
3.481
0.000
von Forstner et al
2010
0.813
0.677
0.899
3.965
0.000
0.883
0.819
0.926
7.786
0.000
Summary effect
-1.00 0.00
Figure 2. Forrest plot of studies included for detection rate pooling
0.50 -0.50
0.00 1.00
28
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
detection rate meta-analysis. Egger’s regression intercept was 1.21 (p ⫽ 0.20124), which indicates a fairly symmetrical funnel plot. The adjusted pooled detection rate after trimming 6 studies (using Duval-Tweedie’s trim and fill method) was 82.6% (95% CI 73.5– 89). This shows that the best estimate of unbiased effect size is 5.7% lower than the calculated pooled detection rate. The subgroup analysis of studies using blue dye and radiotracers for SLN mapping vs those using only 1 method demonstrated pooled detection rates of 90.1% (95% CI 83.6 –94.1) vs 85.1% (95% CI 71.3–92.9), respectively.
detection Funnel Plot of Standard Error by Logit event raterate 0.0
Standard Errorr S
0.5
1.0
1.5
2.0 20 -4
-3
-2
-1
0
1
2
3
4
detection Logit event raterate
Figure 3. Funnel plot of studies included for detection rate pooling. Note asymmetry of plot which indicates possible important publication bias.
pooled detection rate was 88.3% (95% CI 81.9 –92.6). The Cochrane Q value was 105.942 (p ⬍0.0001) and I2 ⫽ 83.95%. Figure 3 shows the funnel plot of the
Sensitivity Figure 4 shows the forest plot of the sensitivity meta-analysis from 19 included studies. The pooled sensitivity was 88% (95% CI 83–92). The Cochrane Q value was 31.04 (p ⫽ 0.0285) and I2 ⫽ 42%. Figure 5 shows the funnel plot of the included studies for the sensitivity meta-analysis. The Egger’s regression intercept was -0.65 (p ⫽ 0.1743), which indicates a symmetrical funnel plot. Duval-Tweedie’s trim and fill method was unable to find any study for trimming and the adjusted pooled sensitivity was the
Figure 4. Forrest plot of studies included for sensitivity pooling
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
sensitivity Funnel Plot of Standard Error by Logit event rate 0.0
Standard Errorr S
0.5
1.0
1.5
20 2.0 -4
-3
-2
-1
0
1
2
3
4
sensitivity Logit event rate
Figure 5. Funnel plot of studies included for sensitivity pooling. Note asymmetry of plot is less than that of funnel plot of detection rate.
same as before trimming. Therefore, the best estimate of unbiased effect size was not different from the calculated pooled sensitivity. Subgroup analysis of the SLN mapping method (blue dye and/or radiotracer) showed a pooled sensitivity of 60% (95% CI 15–95) for blue dye alone, 84% (95% CI 72–92) for radiotracer alone and 90% (95% CI 85–95) for studies using both methods. Excluding 2 studies in which patients with palpable inguinal node regardless of FNA results were recruited21,29 resulted in a pooled sensitivity of 90% (95% CI 85–94). Considering the SLN mapping method, the pooled sensitivity of studies using radiotracer and blue dye for SLN mapping and recruiting only patients with cN0 disease was 92% (95% CI 86 –96). Only 2 studies (2 subgroups of the study by Leijte et al11) used inguinal ultrasonography to detect suspicious inguinal lymph nodes and performed FNA if necessary. Pooling the sensitivity of these 2 studies revealed a 93% sensitivity (95% CI 85–97).
DISCUSSION The condition of regional lymph nodes is an important factor affecting the survival of patients with several solid tumors, and penile cancer is no exception in this regard.34,35 An important aspect of penile SCC in contrast to many other solid tumors is the therapeutic benefit of regional lymph node dissection in this tumor.7 Immediate resection of occult lymph node metastases improves the survival of patients with penile cancer considerably compared to the delayed resection of palpable inguinal lymph node involvement.8 –10 The relative ineffectiveness of chemotherapy and radiotherapy compared to the surgical approach is also important in this regard.36
29
However, prophylactic inguinal lymph node dissection for penile cancer is not required in those without regional lymph node involvement, which constitutes almost 75% to 80% of cN0 disease.7,10 In addition to the significant morbidity of inguinal lymph node dissection,37 this is the main impetus behind the search of urologists for an alternative to inguinal lymph node dissection. Several alternatives to inguinal lymph node dissection are currently under study and several nomograms for predicting inguinal lymph node involvement are currently in use.6,38 However, their performance is not perfect enough and needs to be externally validated.39,40 Several imaging techniques such as computerized tomography or conventional magnetic resonance imaging can also be used for inguinal lymph node staging. However, their accuracy is also suboptimal, especially for cN0 disease.41– 43 Although novel techniques such as positron emission tomography/ computerized tomography44 and nanoparticle enhanced magnetic resonance imaging45 seem to be promising in this regard, their sensitivity for detecting micrometastases is actually low.43,45 SLN biopsy is another technique which is promising for inguinal lymph node staging and our current meta-analysis determines the accuracy of this approach. The overall detection rate (per groin, not per patient) of 88.3% (95% CI 81.9 –92.6) was fairly high. However, the I2 value was 83.95%, meaning that most of the observed variance between the detection rates of the included studies was real and could be attributed to the between study differences (and not only to sampling errors of the included studies). To address this high level of heterogeneity we performed a subgroup analysis of the SLN mapping method. The pooled detection rate of the studies using radiotracer and blue dye was 5% more than the studies using only 1 of the methods (90.1% [95% CI 83.6 –94.1] vs 85.1% [95% CI 71.3–92.9]). This shows the importance of methodology in SLN mapping and this issue should be considered carefully while performing SLN biopsy for penile cancer. The sensitivity of SLN mapping for inguinal lymph node staging is the more important issue in our meta-analysis. The pooled sensitivity of the included studies was 88% (95% CI 83–92), which was also fairly high. However, the heterogeneity of the included studies (I2 ⫽ 42%) warranted evaluation in more depth. The spectrum of the patients recruited in the studies seemed to be important. Excluding 2 studies which recruited patients with palpable inguinal lymph nodes (without performing FNA) increased the pooled sensitivity to 90% (95% CI 85–94).21,29 This was also supported by Leijte et al, who reported rerouting of lymphatic drainage in 76% of the groins
30
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
with palpable lymph nodes with resultant false-negative SLN biopsy.46 Another aspect of the included studies was the method of SLN mapping. The highest sensitivity was in the studies using blue dye and radiotracer (90% [95% CI 85–95]) and the lowest sensitivity was in those using only blue dye for SLN mapping (60% [95% CI 15–95]). Excluding the studies that recruited patients with palpable inguinal lymph nodes as well as those using only blue dye or radiotracers increased the sensitivity to 92% (95% CI 86 –96), which is high enough to be considered safe and is comparable to recommendations for patients with breast cancer.47 A method for decreasing the false-negative rate proposed by Leijte et al is performing inguinal ultrasonography to detect suspicious inguinal nodes requiring FNA for better evaluation.24 They reported a decreased false-negative rate by this method. We also performed sensitivity analysis and the pooled sensitivity of the 2 studies with inguinal ultrasonography in all patients was 93%, which is in agreement with the study by Leijte et al. The sensitivity of SLN mapping, with proven safety and efficacy in many cancers such as breast,2 prostate,4 melanoma48 etc, is not different from what we found for penile cancer in our meta-analysis. In fact, SLN mapping is included in the 2009 European Association of Urology guidelines on penile cancer (level of evidence 2b).6 Our study did have some limitations. Penile SCC is one of the rare carcinomas in humans, which results in a low sample size in many studies evaluating this malignancy. This is especially true when it comes to studies on SLN biopsy for penile SCC. For statistical reasons we excluded studies with a sample size of less than 5 patients (10 groins). In fact, pooling all relevant studies improves overall precision, and the fairly narrow confidence intervals of our meta-analysis results are also in agreement with this idea. Another related limitation is the possibility of a learning curve for the performance of SLN biopsy in penile cancer. This is a major concern since this malignancy is rare and the urologist would not have the chance to pass the learning curve similar to surgeons treating breast cancer. Leijte et al addressed this issue (by evaluating the first 30 cases for false-negative rate and complications) and they could not demonstrate the presence of this phenomenon.11 However, as they mentioned this could be due to pure chance and should be evaluated more
in future studies. Actually only 3 of all included studies in the current meta-analysis recruited more than 30 patients.11,24,30 Subgroup analysis using only these 3 studies demonstrated a pooled detection rate and sensitivity of 94.9% (95% CI 90.9 –97.2) and 90.9% (95% CI 84.7–95.2), respectively, which are higher than overall pooled indices, and show the effect of a high volume of patients on the success of the sentinel node procedure. In introducing SLN biopsy a main obstacle could be the necessity of multidisciplinary teams, especially considering that the geographic areas with the highest rates of penile cancer have limited resources and would have more limitations in this regard.28 A particular aspect of many studies evaluating SLN mapping accuracy in penile SCC is the reliance on followup to determine the false-negative rate. Most inguinal recurrence in penile SCC occurs within approximately 2 years after surgery.11,24 However, not all studies included in the current meta-analysis followed all patients that long, which is a major limitation of our meta-analysis. An important issue in all meta-analyses is publication bias. We included meeting abstracts as well as nonEnglish studies in the meta-analysis, which can decrease publication bias. Although Egger’s tests were not statistically significant in sensitivity or detection rate pooling, this test has low power and visual assessment of the funnel plots, and shows some asymmetry, especially for the detection rate pooling. This was also supported by Duval-Tweedie’s trim and fill method, which showed an almost 6% decrease in the pooled detection rate after trimming while no study could be trimmed for sensitivity pooling. Thus, another limitation of our study is that publication bias, if present, affects the results of the current meta-analysis (especially for detection rate, not sensitivity).
CONCLUSIONS The current meta-analysis showed that SLN mapping in penile SCC is a method with a high detection rate and sensitivity. However, the technical aspects of this procedure are important. Using radiotracer and blue dye for SLN mapping and including only cN0 disease ensures the highest detection rate and sensitivity. However, the possible presence of a learning curve as well as limited resources in many areas with a high incidence of this malignancy could reduce the usefulness of this procedure.
REFERENCES 1. Purdie CA: Sentinel lymph node biopsy: review of the literature and guidelines for pathological handling and reporting. Curr Diagn Pathol 2007; 13: 106.
2. Miltenburg DM, Miller C, Karamlou TB et al: Meta-analysis of sentinel lymph node biopsy in breast cancer. J Surg Res 1999; 84: 138.
3. Warycha MA, Zakrzewski J, Ni Q et al: Metaanalysis of sentinel lymph node positivity in thin melanoma (⬍/⫽1 mm). Cancer 2009; 115: 869.
META-ANALYSIS OF SENTINEL NODE BIOPSY FOR PENILE CANCER
4. Sadeghi R, Tabasi KT, Bazaz SM et al: Sentinel node mapping in the prostate cancer. Meta-analysis. Nuklearmedizin 2011; 50: 107. 5. Cabanas RM: An approach for the treatment of penile carcinoma. Cancer 1977; 39: 456. 6. Pizzocaro G, Algaba F, Horenblas S et al: EAU penile cancer guidelines 2009. Eur Urol 2010; 57: 1002.
21. Hungerhuber E, Schlenker B, Frimberger D et al: Lymphoscintigraphy in penile cancer: limited value of sentinel node biopsy in patients with clinically suspicious lymph nodes. World J Urol 2006; 24: 319. 22. Brennhovd B, Johnsrud K, Berner A et al: Sentinel node procedure in low-stage/low-grade penile carcinomas. Scand J Urol Nephrol 2006; 40: 204.
31
35. Dechev J, Banchev A and Georgieva-Mateva N: Overall survival in carcinoma of the penis–a single institution study. J BUON 2007; 12: 377. 36. Lubke WL and Thompson IM: The case for inguinal lymph node dissection in the treatment of T2-T4, N0 penile cancer. Semin Urol 1993; 11: 80.
7. Wespes E: The management of regional lymph nodes in patients with penile carcinoma and reliability of sentinel node biopsy. Eur Urol 2007; 52: 15.
23. Gonzaga-Silva LF, Tavares JM, Freitas FC et al: The isolated gamma probe technique for sentinel node penile carcinoma detection is unreliable. Int Braz J Urol 2007; 33: 58.
37. Tobias-Machado M, Tavares A, Ornellas AA et al: Video endoscopic inguinal lymphadenectomy: a new minimally invasive procedure for radical management of inguinal nodes in patients with penile squamous cell carcinoma. J Urol 2007; 177: 953.
8. Kroon BK, Horenblas S, Lont AP et al: Patients with penile carcinoma benefit from immediate resection of clinically occult lymph node metastases. J Urol 2005; 173: 816.
24. Leijte JA, Kroon BK, Valdes Olmos RA et al: Reliability and safety of current dynamic sentinel node biopsy for penile carcinoma. Eur Urol 2007; 52: 170.
38. Ficarra V, Zattoni F, Artibani W et al: Nomogram predictive of pathological inguinal lymph node involvement in patients with squamous cell carcinoma of the penis. J Urol 2006; 175: 1700.
9. McDougal WS, Kirchner FK Jr, Edwards RH et al: Treatment of carcinoma of the penis: the case for primary lymphadenectomy. J Urol 1986; 136: 38.
25. Spiess PE, Izawa JI, Bassett R et al: Preoperative lymphoscintigraphy and dynamic sentinel node biopsy for staging penile cancer: results with pathological correlation. J Urol 2007; 177: 2157.
39. Ficarra V, Novara G, Boscolo-Berto R et al: How accurate are present risk group assignment tools in penile cancer? World J Urol 2009; 27: 155.
26. Hernandez-Toris N, Quintero-Becerra J, GallegosHernandez JF et al: Lymphatic mapping and sentinel node biopsy in penis cancer. Feasibility study and preliminary report. Cir Cir 2007; 75: 87.
40. Novara G, Artibani W, Cunico SC et al: How accurately do Solsona and European Association of Urology risk groups predict for risk of lymph node metastases in patients with squamous cell carcinoma of the penis? Urology 2008; 71: 328.
27. Rubi S, Vidal-Sicar S, Ortega M et al: Localization of sentinel node in squamous cell carcinoma of the penis. Initial experience. Rev Esp Med Nucl 2008; 27: 3.
41. Heyns CF, Fleshner N, Sangar V et al: Management of the lymph nodes in penile cancer. Urology 2011; 76: S43.
10. Protzel C, Alcaraz A, Horenblas S et al: Lymphadenectomy in the surgical management of penile cancer. Eur Urol 2009; 55: 1075. 11. Leijte JA, Hughes B, Graafland NM et al: Twocenter evaluation of dynamic sentinel node biopsy for squamous cell carcinoma of the penis. J Clin Oncol 2009; 27: 3325. 12. Oxford Centre for Evidence-Based Medicine: Levels of Evidence. Available at http://www.cebm. net/index.aspx?o⫽1025. Accessed October 5, 2010. 13. DerSimonian R and Laird N: Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177. 14. Higgins JP and Thompson SG: Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21: 1539. 15. Egger M, Davey Smith G, Schneider M et al: Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629. 16. Duval S and Tweedie R: A nonparametric ‘trim and fill’ method of accounting for publication bias in meta-analysis. J Am Stat Assoc 2000; 95: 89. 17. Zamora J, Abraira V, Muriel A et al: Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6: 31. 18. Akduman B, Fleshner NE, Ehrlich L et al: Early experience in intermediate-risk penile cancer with sentinel node identification using the gamma probe. Urology 2001; 58: 65.
28. Ferreira U, Ribeiro MA, Reis LO et al: Sentinel lymph node biopsy in penile cancer: a comparative study using modified inguinal dissection. Int Braz J Urol 2008; 34: 725. 29. Heyns CF and Theron PD: Evaluation of dynamic sentinel lymph node biopsy in patients with squamous cell carcinoma of the penis and palpable inguinal nodes. BJU Int 2008; 102: 305. 30. Jensen JB, Jensen KM, Ulhoi BP et al: Sentinel lymph-node biopsy in patients with squamous cell carcinoma of the penis. BJU Int 2009; 103: 1199. 31. Pytel A, Damasdi M, Frick A et al: Surgical management of low and medium risk penile cancers, with isotope guided sentinel lymph node biopsy technique. Urology 2009; 74: S63. 32. Vinayak RS and Rajen TA: Sentinel node biopsy in carcinoma penis using methylene blue dye technique. Urol Ann 2009; 1: 18.
19. Lemstra C, de Jong I, Leliveld A et al: Evaluation of lymphoscintigraphy and sentinel node biopsy in patients with penile carcinoma. J Nucl Med 2010; 51: 2004.
33. von Forstner C, Naumann M, Henze E: Improvement of the surgical procedure for penile carcinoma by sentinel lymph node scintigraphy with SPECT and SPECT/CT. Eur J Nucl Med Mol Imaging 2010; 37: S270.
20. Perdona S, Autorino R, De Sio M et al: Dynamic sentinel node biopsy in clinically node-negative penile cancer versus radical inguinal lymphadenectomy: a comparative study. Urology 2005; 66: 1282.
34. Zini L, Cloutier V, Isbarn H et al: A simple and accurate model for prediction of cancer-specific mortality in patients treated with surgery for primary penile squamous cell carcinoma. Clin Cancer Res 2009; 15: 1013.
42. Kochhar R, Taylor B and Sangar V: Imaging in primary penile cancer: current status and future directions. Eur Radiol 2011; 20: 36. 43. Hughes B, Leijte J, Shabbir M et al: Non-invasive and minimally invasive staging of regional lymph nodes in penile cancer. World J Urol 2009; 27: 197. 44. Graafland NM, Leijte JA, Valdes Olmos RA et al: Scanning with 18F-FDG-PET/CT for detection of pelvic nodal involvement in inguinal node-positive penile carcinoma. Eur Urol 2009; 56: 339. 45. Mueller-Lisse UG, Scher B, Scherr MK et al: Functional imaging in penile cancer: PET/computed tomography, MRI, and sentinel lymph node biopsy. Curr Opin Urol 2008; 18: 105. 46. Leijte JA, van der Ploeg IM, Valdes Olmos RA et al: Visualization of tumor blockage and rerouting of lymphatic drainage in penile cancer patients by use of SPECT/CT. J Nucl Med 2009; 50: 364. 47. MacNeill F, Mansel R, Horgan K et al: NEW START: the United Kingdom sentinel lymph biopsy training programme. A model of how multiprofessional training can achieve competent national performance whilst maintaining patient safety. Breast Cancer Res Treat 2007; 106: S233. 48. Mattsson J, Bergkvist L, Abdiu A et al: Sentinel node biopsy in malignant melanoma: Swedish experiences 1997-2005. Acta Oncol 2008; 47: 1519.