The Role of Inguinal Lymph Node Dissection in Men with Urethral Squamous Cell Carcinoma

Home , Inguinal lymphadenopathy, Inguinal lymph nodes, Lymph node

Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6 Original Article The role of inguinal lymph node dissection in men with urethral squamous cell carcinoma

RyanD1XX P. Werntz, M.D.D2XX*, ChristopherD3XX B. Riedinger,D4RichardD5XXXX J. Fantus,D6ZacharyD7XXXX L. Smith,D8XX VigneshD9XX T. Packiam,D10MelanieD11XXXX A. Adamsky,D12NormD13XXXX Smith,D14GaryD15XX D.XX SteinbergD16XX Department of Surgery, Section of Urology, The University of Chicago, Chicago, IL

Received 19 April 2018; received in revised form 7 August 2018; accepted 21 September 2018

Abstract Introduction: Urethral squamous cell cancer is a rare disease with limited clinical recommendations regarding management of the inguinal lymph nodes. Despite the similarities to penile cancer in terms of squamous cell carcinoma (SCC) histology and lymphatic drainage, there is not enough evidence to recommend for or against a prophylactic inguinal lymph node dissection (ILND) in patients with clinically negative groins and a primary tumor stage of T1b or higher. The objective of the study was to identify the rate of prophylactic inguinal lymph node dissection, node positive rate, and overall survival in patients with clinical T1 to T4 stage. The patients were separated into clinical N stage and the rates of node positivity were compared. We hypothesize that the node positivity rate would be similar to that observed in penile cancer of similar clinical T and N stage and provide evidence for prophylactic inguinal lymph node dissection in urethral squamous cancer. We also sought to determine the value of ILND in clinically node positive (cN+) and clinically node negative (cN¡) patients. Methods: The National Cancer Database was queried for all cases of primary urethral cancer in men from 2004 to 2014. Patients with other cancer diagnoses, metastasis, nonsquamous histology, female patients, and patients with a history of radiation therapy were excluded. Male patients with urethral squamous cell cancer of the anterior urethra with T1 or higher T stage were included in this study. All-cause mortality was compared using multivariable Cox regression controlling for covariates. Results: The study included 725 men with urethral SCC with T1 or higher clinical T stage. The median age was 63 years (33−83 inter- quartile range). Of the 725 men, 536 men did not receive an ILND and 189 (26%) underwent ILND. Patients who received LND had significantly higher clinical T and clinical N stage. There was no difference in age, sex, or histology between those with ILND versus no ILND. In patients with T1 to T4 and clinical N0, the ILND rate was 21.8% (89/396). The lymph node positive rate in patients with N0 and T1 to T4 primary tumor was 9%. In patients with clinically node positive disease (N1/N2), the overall ILND rate was 76%. The lymph node positive rate for patients with clinical nodal disease was 84%. On multivariable analysis cox regression, lymph node positivity was associated with worse overall survival when controlling for T stage, clinical N stage, and age (HR 1.56, 95% 1.3−1.9, P = 0.000). On multivariable analysis after controlling for T stage, sex, and age, having an ILND was associated with improved OS in patients with clinical N1 or N2 disease (HR 0.46, 95% 0.28−0.78 P = 0.002). Conclusion: The node positivity rate in patients with T1 to T4 and N0 is 9%, much lower than reported in penile cancer with a high-risk primary tumor but clinically negative groins. This argues against routine prophylactic inguinal ILND in patients with urethral SCC who are clinically N0, perhaps suggesting different biological behavior of urethral SCC compared to penile SCC. Performing a lymph node dissection in patients with clinically N1 or N2 disease is associated with improved OS. Published by Elsevier Inc.

Keywords: Squamous cell carcinoma; Urethra; Inguinal lymph node dissection

Disclosures: none. *Corresponding author. Tel.: 574-855-8741. E-mail address: [email protected] (R.P. Werntz). https://doi.org/10.1016/j.urolonc.2018.09.014 1078-1439/Published by Elsevier Inc. R.P. Werntz et al. / Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6 526.e2

1. Introduction 2.2. Study population

Primary urethral carcinoma (PUC) is a rare clinical We analyzed the NCDB participant user files from 2004 entity and accounts for less than 1% of all genitourinary to 2014. During the study period, 725 male patients were malignancies [1]. This has led to a relative paucity of data diagnosed with primary urethral SCC cancer and met the where management is largely informed by small, single- below inclusion criteria (Fig. 1). Patients with other cancer institution studies. Data from several national and interna- diagnoses, female gender, nonurethral primary, non-SCC tional cohorts have been published [1−4], but important histology, history of radiation of chemotherapy (NCDB questions remain with regard to the optimal management of variables RX_SUMM_SURGRAD_SEQ and this condition. RX_SUMM_CHEMO), and metastatic disease at presenta- The optimal management of lymph nodes for PUC tion were excluded from analysis (Fig. 1). For penile cancer remains unclear. In men, lymphatics from the anterior comparative purposes, only male patients with urethral urethra and penis preferentially drain into the superficial SCC of the anterior urethra with clinical stage ≥ T1 were anddeepinguinallymphnodes[5].PUCoftheanterior included for analysis. Patients were included if they urethra presents with inguinal lymphadenopathy 18% to received ILND within 6 months of diagnosis to remove the 30% of the time; [2,3,6,7] 75% to 100% of these repre- possibility of patients being included who had salvage sent true metastatic disease [8−10]. Pathologic lymph ILND. Clinically node positive disease was defined accord- node involvement has been shown to adversely predict ing to the TNM staging system, as defined by the NCDB survival [2,6] and a benefit for inguinal lymph node dis- variable TNM_CLIN_N. Of the 725 patients, 189 under- section (ILND) has been shown when clinical or radio- went a bilateral ILND. Lymphovascular invasion data was graphic evidence of lymphadenopathy is present [9,11]. missing on most patients was not evaluated as part of this However, there is no current data to argue for or against study. This comprised the patients selected for analysis. prophylactic ILND for PUC. In penile cancer, with seemingly the same lymphatic drainage and histology, 2.3. Statistical analyses prophylactic ILND is supported in patients with high- risk primary disease (T1b or greater) [12]. Patients with cT1-4 primary lesions were separated into We hypothesize that the node positivity rate would be an ILND groups versus an observation group after surgery similar to that observed in penilecancerofsimilarclini- on the primary tumor. These patients were then stratified cal T and N stage and provide evidence for prophylactic according to their clinical N stage and descriptive analyses inguinal lymph node dissection in urethral squamous of the groups were performed using Pearson’s chi-squared cancer. We also sought to determine the value of ILND or Fisher’s exact tests for categorical variables and in clinically node positive (cN+) and clinically node Student’s t-test for continuous variables. We performed negative (cN¡) patients. To examine this issue further, propensity core weighted analysis to assess the effect of we performed a comparative effectiveness analysis using ILND on overall survival. The propensity score was deter- a large, contemporary US hospital-based cancer registry mined by logistic regression. We then ran a Cox regression to better understand outcomes associated with this was used to identify if the performance of a ILND procedure. improved OS in patients with who were cN0 and cN+ after adjusting for covariates (age, cT stage, cN status, Charlson comorbidity index). Multivariable Cox regression was used 2. Methods to identify predictors of overall survival after adjusting for 2.1. Data source cN and cT stage, age, cT stage, pN stage, Charlson comorbidity index as covariates. All hypotheses testing was two- The National Cancer Data Base (NCDB) is a joint effort sided and a P-value of < 0.05 was considered statistically between the American Cancer Society and the American significant. All statistical analyses were performed using College of Surgeons’ Commission on Cancer (CoC). By Stata, version 13.1 (StataCorp, College Station, TX). collecting data from more than 1,500 CoC-accredited cancer centers, it is able to capture 70% of all incident malig- 3. Results nancies in the United States. Institutional data is collected using standardized definitions in accordance with the CoC’s Seven hundred twenty five men met inclusion criteria. Facility Oncology Registry Data Standards. All demo- Median age was 63 years (IQR 33−83). Patients with a graphic and clinical patient characteristics are recorded, lower Charlson comorbidity index and higher clinical stage including, but not limited to: cancer staging, pathological were more likely to receive an ILND (Table 1). Of the 725 information, details of first-line treatment administered, men, only 189 (26%) underwent ILND. Patients who under- survival outcomes, hospital length of stay, and readmission went ILND were of significantly higher cT and cN stage. In information. Data quality undergoes routine appraisal to patients with cN0, the ILND rate was 21.8%, with 9% pN+. ensure reliability and quality. In patients with cN1-2, the ILND rate was 76%, with 84% 526.e3 R.P. Werntz et al. / Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6

Fig. 1. Men with urethral SCC exclusion criteria diagram. SCC = squamous cell carcinoma. pN+. On multivariable Cox regression, pN+ was associated positive and pathologically node positive patients, we eval- with worse overall survival when controlling for cT stage, uated the dropout rate in each group to address the issue of cN stage, Charlson comorbidity, age, and sex (HR 1.56, selection bias (Fig. 1 and Table 4). On multivariable analy- 95% CI 1.3−1.9, P < 0.001). In order to determine the sis, ILND was not associated with improved OS in clini- impact of excluding additional therapies in clinically node cally N0 patients (HR 1.4, 95% 0.96−2.00 P = 0.07). On

Table 1 Patient characteristics of men with urethral SCC

Patient characteristics No ILND (n = 536) ILND (n = 189) P value

Mean age (IQR) 66 (24−90) 62 (33−86) P < 0.8 No. RACE (%) White 406 (75.8) 157 (83.1) P < 0.11 Black 106 (19.6) 29 (15.3) Other 14 (2.6) 1 (0.5) cT stage n (%) 1 156 (29.0) 26 (13.8) P < 0.001 2 86 (16.0) 38 (20.1) 3 54 (10.0) 26 (13.7) 4 56 (10.5) 24 (12.7) Unk 184 (38.3) 75 (39.7) cN stage n (%) N0 310 (57.8) 86 (45.5) P < 0.001 N1/2 84 (15.7) 64 (33.8) Unk 142 (26.5) 39 (20.6) Charlson-Deyo score 0 383 (71.5) 148 (78.2) P < 0.001 1 109 (20.3) 35 (18.4) ≥ 2 44 (8.3) 6 (3.3) R.P. Werntz et al. / Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6 526.e4

Table 2 Table 3 Rate of men receiving ILND based on clinical stage Pathologic node positivity stratiﬁed by clinical n stage cT stage No ILND (n, (%)) ILND (n, (%)) pN Status cN0 cN1-2 cN unk

1 156 (85.7) 26 (14.3) pN + n (%) 8 (9) 49 (84) 18 (41.2) 2 86 (69.3) 38 (30.7) pN ¡ n (%) 77 (86.5) 8 (13.8) 19 (43.7) 3 54 (67.5) 26 (32.5) pN £ n (%) 4 (4.5) 1 (1.7) 5 (14.6) 4 56 (70) 24 (30) pN + rate 9 84 41.2 Unk 184 (71) 75 (29) Overall # Pts 89 58 42 Overall 536 189 multivariable analysis, performance of ILND was associ- emission tomography in patients with urethral or penile ated with improved OS in patients with clinical N1 or N2 SCC with clinically negative groins [15]. These data sup- disease (HR 0.46, 95% 0.28−0.78 P = 0.002) (Table 2). port the recommendation by the National Comprehensive Cancer Network (NCCN), American Urological Associa- 4. Discussion tion (AUA), and European Association of Urology (EAU) recommending that patients with a clinical stage of Lymph node involvement is the single greatest prognos- ≥ T1bN0 undergo a bilateral prophylactic ILND or sentinel tic factor for penile cancer survival [13]. Penile SCC has node biopsy [12]. Despite the same apparent lymphatic been well studied with known rates of micrometastatic dis- drainage of the anterior urethra, in urethral SCC with nega- ease to the inguinal nodes based on clinical stage and patho- tive groins, there is no evidence to support prophylactic logic features. In penile SCC, men with clinically negative ILND by the NCCN [16]. Furthermore, there are very few inguinal lymph nodes, the rate of micrometastatic disease studies examining the inguinal nodal management of male to the inguinal nodes ranges from 20% to 50% [14]. Fur- urethral SCC. thermore, imaging cannot reliably predict node positive dis- To our knowledge, this is the largest and ﬁrst population- ease. A recent meta-analysis examined the role of positron based study examining the rate of lymph node positive dis- emission tomography in men with penile SCC and found ease in men with urethral SCC who are clinically N0. We that in patients with clinically negative groins, the sensitiv- found that the rate of pN+ disease in men with any primary ity of detecting metastatic disease on conﬁrmatory ILND cT stage and cN0, was only 9% (Table 3). In this study, per- was only 57% [15]. Therefore, there is no role for positron forming an ILND in cN0 patients was not associated with an

Fig. 2. Overall survival in men with clinical N+ disease: ILND versus no ILND. ILND = inguinal lymph node dissection. 526.e5 R.P. Werntz et al. / Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6

Table 4 Excluding patients who received either neoadjuvant, adju- Impact of the application of exclusion of additional therapies stratified by vant, or salvage chemotherapy or radiation introduces clinical node status another potential element of selection bias. Perhaps when Exclusion of radiation or chemotherapy cN0 cN1-2 cN unk the analysis was completed, a large amount of cN+ patients received additional treatment. This would influence the No (N = 743) 404 155 184 results. However, an analysis was performed prior to apply- Yes (N = 725) 396 148 181 ¡ Pts excluded (%) 8 (1.9) 7 (4.5) 3 (1.6) ing the aforementioned exclusion criteria to both cN+/ and pN+/¡ patients to determine the percentage of patients lost in each group. In Table 4, there was very little differ- improvement in OS on Multivariable Cox regression ence in the percentage of patients receiving additional ther- (P = 0.07). When comparing these data to that of penile can- apy, lessening the potential selection bias. It was difficult to cer who are cN0, the rate of pN+ is much lower [16].In determine timing of ILND. We attempted to mitigate the patients who presented with cN1-N2 disease, the rate of pN+ amount of patients that were “salvage” ILND (patients ini- was 84% (Table 3). These data mirror results reported in a tially cN0 who presented lated with cN+) by including multiinstitutional study by Gakis et al. who sought to deter- patients who received their surgery within 6 months of mine prognostic indicators in urethral cancer. They found diagnosis. However, surely there were likely patients who that the clinical and pathologic node correlation was in did receive salvage ILND that we could not account for. excess of 90% and was associated with poor cancer-specific Another potential limitation is that despite having clinically survival [17]. In cN+ patients, the performance of an ILND node positive disease, not all men elected for treatment. was associated with an improvement in OS after controlling There likely is an element of selection bias involved in the for confounding variables (HR 0.46, P =0.002, Fig. 2). patients with cN positive disease not receiving surgery, per- Together, these data suggest that although the lymphatic haps impacting the overall survival results. Despite appro- drainage is reported to be similar between penile cancer and priate statistic methods to decrease bias and confounding anterior urethral SCC, the biology and disease processes variables, interpretation of OS survival based on a large behave differently. Furthermore, these data do not support observational cohort should be interpreted with caution. prophylactic ILND in urethral SCC with cN0 because of the However, there will not likely be high quality data available low pN+ rate. This is in contrast to the management of penile to answer this question in more detail due to the rarity of cancer of the same stage, which has a much high pN+ rate as this disease. This is the largest study addressing the man- discussed earlier. These data do suggest that ILND should be agement of inguinal lymph nodes in men with urethral SCC considered in men who present with positive groins because and supports a different approach to the management of the of the very high rate of pN+ disease. ILND in this scenario inguinal lymph nodes when compared to penile SCC of the appears to provide a therapeutic advantage. Although the same stage. patients in this study did not receive chemotherapy or radio- therapy, contemporary treatment of patients in this category 5. Conclusion with advanced nodal disease, should consider cisplatin-based chemotherapy followed by consolidative surgery [16].Itis The lymph node positivity rate in patients with clinical important to note, that de novo presentation of widespread T1 to T4 and N0 urethral SCC is 9%, substantially lower metastatic disease is very rare and that aggressive local and than reported in penile cancer, arguing against routine pro- regional control can cure patients [18].Itshouldbecau- phylactic inguinal ILND in this population. ILND in clini- tioned, however, that not all cN0 urethral SCC patients cally N0 patients was not associated with improved OS. should be observed. A young healthy patient may not be However, performance of ILND in clinical N1-2 disease is comfortable with observation in this scenario and a shared associated with improved OS. decision making process should be emphasized. The use of robotic or laparoscopic ILND may make this decision easier for patients. In a single institutional comparative series by References Schwentner et al, the use of minimally invasive approaches to inguinal lymph node dissection appears to be associated [1] Swartz MA, Porter MP, Lin DW, Weiss NS. Incidence of primary urethral carcinoma in the United States. Urology 2006;68:1164–8. with a lower complication rate compared to traditional open https://doi.org/10.1016/j.urology.2006.08.1057. approach (7% vs. 55.3%) [19]. [2] Rabbani F. Prognostic factors in male urethral cancer. Cancer Important limitations of this study include that there was 2011;117:2426–34. https://doi.org/10.1002/cncr.25787. no accurate grade information and lymphovascular invasion [3] Sui W, RoyChoudhury A, Wenske S, Decastro GJ, McKiernan JM, was sparsely recorded in the urethral cancer database. There Anderson CB. Outcomes and prognostic factors of primary urethral cancer. Urology 2017;100:180–6. https://doi.org/10.1016/j.urol- was no central pathologic review and may have been selec- ogy.2016.09.042. tion bias in which cN0 patients went on to receive ILND, [4] Cahn DB, Handorf E, Ristau BT, Geynisman DM, Simhan J, Kutikov which was mitigated by controlling for primary cT stage. A, et al. Contemporary practice patterns and survival outcomes for R.P. Werntz et al. / Urologic Oncology: Seminars and Original Investigations 36 (2018) 526.e1−526.e6 526.e6

locally advanced urethral malignancies: a national cancer database [14] Graaﬂand NM, Lam W, Leijte JAP, Yap T, Gallee MPW, Corbishley analysis. Urol Oncol 2017;35. https://doi.org/10.1016/j.urolonc.2017. C, et al. Prognostic factors for occult inguinal lymph node involve- 07.026:670.e15−670.e21. ment in penile carcinoma and assessment of the high-risk EAU sub- [5] Carroll PR, Dixon CM. Surgical anatomy of the male and female ure- group: a two-institution analysis of 342 clinically node-negative thra. Urol Clin North Am 1992;19:339–46. patients. Eur Urol 2010;58:742–7. https://doi.org/10.1016/j.eur- [6] Dalbagni G, Zhang ZF, Lacombe L, Herr HW. Male urethral carci- uro.2010.08.015. noma: analysis of treatment outcome. Urology 1999;53:1126–32. [15] Ottenhof SR, Leone AR, Horenblas S, Spiess PE, Vegt E. [7] Anderson KA, McAninch JW. Primary squamous cell carcinoma of Advancements in staging and imaging for penile cancer. Curr anterior male urethra. Urology 1984;23:134–40. Opin Urol 2017;27:612–20. https://doi.org/10.1097/MOU.0000000 [8] Mullin EM, Anderson EE, Paulson DF. Carcinoma of the male ure- 000000447. thra. J Urol 1974;112:610–3. [16] Dayyani F, Hoffman K, Eifel P, Guo C, Vikram R, Pagliaro LC, et al. [9] Bracken RB, Henry R, Ordonez N. Primary carcinoma of the male Management of advanced primary urethral carcinomas. BJU Int urethra. South Med J 1980;73:1003–5. 2014;114:25–31. https://doi.org/10.1111/bju.12630. [10] Hopkins SC, Nag SK, Soloway MS. Primary carcinoma of male ure- [17] Gakis G, Morgan TM, Efstathiou JA, Keegan KA, Mischinger J, thra. Urology 1984;23:128–33. Todenhoefer T, et al. Prognostic factors and outcomes in primary ure- [11] Dinney CP, Johnson DE, Swanson DA, Babaian RJ, Eschenbach von thral cancer: results from the international collaboration on primary AC. Therapy and prognosis for male anterior urethral carcinoma: an urethral carcinoma. World J Urol 2016;34:97–103. https://doi.org/ update. Urology 1994;43:506–14. 10.1007/s00345-015-1583-7. [12] Spiess PE. National Comprehensive Cancer Network. New treatment [18] Gakis G, Witjes JA, Comperat E, Cowan NC, De Santis M, Lebret T, guidelines for penile cancer. J Natl Compr Cancer Network et al. EAU guidelines on primary urethral carcinoma. Eur Urol 2013;11:659–62. 2013;64:823–30. https://doi.org/10.1016/j.eururo.2013.03.044. [13] Marconnet L, Rigaud J, Bouchot O. Long-term followup of penile [19] Schwentner C, Todenhofer€ T, Seibold J, Alloussi SH, Mischinger J, carcinoma with high risk for lymph node invasion treated with ingui- Aufderklamm S, et al. Endoscopic inguinofemoral lymphadenectomy nal lymphadenectomy. J Urol 2010;183:2227–32. https://doi.org/ —extended follow-up. J Endourol 2013;27:497–503. https://doi.org/ 10.1016/j.juro.2010.02.025. 10.1089/end.2012.0489.