Clinical significance of preoperative C-reactive protein and squamous cell carcinoma antigen levels in patients with penile squamous cell carcinoma

Transcription

1 Clinical significance of preoperative C-reactive protein and squamous cell carcinoma antigen levels in patients with penile squamous cell carcinoma Zai-Shang Li*,,, Kai Yao*,,, Yong-Hong Li*,,, Jie-Ping Chen*,,, Chuang-Zhong Deng*,,, Qi Zhao, Peng Chen, Bin Wang**, Qi-Wu Mi, Zhuo-Wei Liu*,,, Zi-Ke Qin*,,, Hui Han*,, and Fang-Jian Zhou*,, *Department of Urology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in Southern China, Collaborative Innovation Center of Cancer Medicine, School of Life Science, Sun Yat-sen University, Guangzhou, Department of Urology, Affiliated Tumor Hospital of Xinjiang Medical University, Urumchi, **Department of Urology, Cancer Center of Guangzhou Medical University, Guangzhou, and Department of Urology, Dong Guan People s Hospital, Guang Dong, China Z.-S. L. and K. Y. contributed equally to this work. Objective To evaluate the relevance of C-reactive protein (CRP) and squamous cell carcinoma antigen (SCC-Ag) levels in relation to clinicopathological factors and prognosis in penile cancer. Patients and Methods A total of 124 Chinese patients with penile squamous cell carcinoma (SCC), treated between November 2007 and October 2014, were analysed retrospectively. Receiveroperating characteristic curves were used to identify the combination of markers with the best sensitivity and specificity for prognosis prediction. Statistical data analysis was performed using a non-parametric method, and survival analysis was performed using the log-rank test and Cox proportional hazard model. Results Levels of CRP 4.5 mg/l and SCC-Ag 1.4 ng/ml were both significantly associated with lymph node metastasis (LNM) laterality (chi-squared trend test, P = 0.041), extranodal extension (chi-squared trend test, P < 0.001), pelvic LNM (chi-squared trend test, P = 0.024), pathological tumour status (chi-squared trend test, P = 0.002), pathological nodal status (chi-squared trend test, P < 0.001), and diseasespecific survival (DSS; log-rank test, P < 0.001). Moreover, the influence of CRP and SCC-Ag levels on DSS (P = 0.033, hazard ratio 3.390, 95% confidence interval ) remained after adjusting for smoking history, phimosis, tumour status, tumour cell differentiation and nodal status. Conclusions The present study shows that the combined measurement of preoperative CRP and SCC-Ag levels may serve as an independent biomarker for LNM, advanced tumour stage and DSS in patients with penile SCC. Keywords penile neoplasm, staging system, lymph node metastasis, lymph node excision, prognosis Introduction More than 95% of penile malignancies are classified as squamous cell carcinoma (SCC) [1,2]. As this genitourinary cancer is typically characterized by stepwise regional lymph node spread prior to the formation of distant metastases [3 5], the lymph node metastasis (LNM) stage is the most important predictor of survival in penile cancer [1 3,6 8]. Inguinal lymphadenectomy is currently the most useful and commonly performed surgery for penile cancer staging, and it can be curative in cases with minimal metastases [1 3,5]; however, a significant number of patients with enlarged lymph nodes do not present with metastatic spread; consequently, up to 80% of patients are overtreated [9,10], leading to high morbidity rates [11,12]. Preoperative diagnosis and decreased intervention without the loss of positive results is therefore of paramount importance. Because classical molecular markers are not clinically useful for SCC of the penis, a tumour marker related to disease burden and activity would be helpful for the optimum management of this disease. BJU Int 2016; 118: wileyonlinelibrary.com BJU International 2015 BJU International doi: /bju Published by John Wiley & Sons Ltd.

2 Preoperative CRP and SCC antigen levels and penile cancer C-reactive protein (CRP) is an indicator of acute and chronic inflammation [13]. An elevated CRP concentration is independently correlated with tumour load and disease progression, which is also correlated with survival in various cancers [14,15], and two recent studies have shown that CRP correlates with tumour burden in penile cancer [16,17]. Another potential prognostic marker in penile cancer is SCC antigen (SCC-Ag), a serum tumour marker used for various human SCCs [18,19], and an elevated level of SCC-Ag has been reported to be associated with clinical outcome in patients with penile cancer [19 21]; however, to the best of our knowledge, the relationship between CRP and SCC-Ag levels and their potential value in combination as a prognostic marker of survival in penile cancer have not been previously explored. In the present study, we retrospectively analysed 124 patients with penile SCC to evaluate the combined prognostic value of elevated CRP and SCC-Ag levels. Patients and Methods We retrospectively reviewed the charts of 124 consecutive patients diagnosed with primary penile SCCs between November 2007 and October 2014 at our institution. Pretreatment CRP and SCC-Ag serum levels were measured using fresh blood samples obtained at the time of diagnosis before any medical intervention (before 3 5 days). The eligibility criteria were histologically confirmed penile SCC and treatment (lesions and lymph nodes were proven by pathological analysis). The exclusion criteria were as follows: neoadjuvant chemotherapy; previous surgery or radiotherapy of the inguinal region; clinical evidence of distant metastasis; and loss to follow-up. Based on guideline recommendations, treatment protocols were discussed with each patient [1 3,22,23]. All histopathology reports were based on the 2015 American Joint Committee on Cancer TNM system [1]. The deadline for follow-up data to be included in this analysis was March The follow-up period for each patient began at the time of cancer diagnosis and ended at death or the deadline. All patients underwent follow-up examinations every 3 months for the first 2 years after surgery, every 6 months in the 3rd and 4th years after surgery, and once yearly thereafter [2]. Statistical Analyses The chi-squared test was used for multiple comparisons. The optimum CRP and SCC-Ag cut-off values for prognosis prediction were calculated using receiver-operating characteristic curve analysis with reference to cancer-specific death [24]. Linear correlation between variables was calculated using linear regression analysis. Patients with both an elevated CRP level and elevated SCC-Ag were given a score of 3. Patients with either elevated CRP ( 4.5 mg/l or elevated SCC-Ag 1.4 g/ml) levels only received a score of 2, and patients with no elevation received a score of 1. The logrank test was used to generate Kaplan Meier plots to estimate disease-specific survival (DSS); however, we did not evaluate the role of adjuvant therapy by multivariate analysis because such therapies were not routinely administered to all of the enrolled patients. Several risk factors for penile SCC, which were identified according to European Association of Urology guidelines, were included in the regression analysis. Because LNM laterality, extranodal extension (ENE) and pelvic LNMs are included in the 7th nodal staging system, only additional factors were included in the regression analysis. Multivariable Cox regression models were fitted to test DSS predictors. Because of the small sample size and number of relationships examined and because CRP and SCC-Ag levels are not accepted as prognostic predictors in penile cancer, all analyses were considered exploratory, and results were considered as hypothesis-generating rather than hypothesis-testing; therefore, in our model, a positive signal for each marker was combined into a single category. The likelihood ratio, Akaike s information criterion and C-index value were investigated to evaluate the accuracy of the models. All statistical analyses were performed with R ( and a two-sided P value <0.05 was taken to indicate statistical significance. Results A total of 124 patients were retrospectively reviewed. Of these, 16 (12.9%) whose treatment plan was penile preservation underwent local excision with circumcision, 81 (65.3%) underwent partial penile amputation, and 37 (29.8%) underwent total penile amputation. A total of 108 patients (87.1%) met the eligibility criteria for the study and underwent bilateral inguinal lymph node dissection, and 60 of those patients (48.4%) had LNM. The median (interquartile range) numbers of positive and removed lymph nodes were 2 (1 13) and 27 (5 55), respectively, and 45 patients received adjuvant chemotherapy therapy. According to the receiveroperating characteristic curve analysis, the potential preoperative levels of CRP and SCC-Ag were 4.5 mg/l and 1.4 g/ml, respectively. During the follow-up period, 17 patients died from penile cancer at a mean (median; range) of 17 (12; 2 117) months. The clinicopathological characteristics of the patients are presented in Table 1. A close association was observed between an increased CRP level and ENE (P < 0.001), pelvic LNM (P = 0.007), pathological tumour status (P = 0.002), and pathological nodal status (P < 0.001; Table 2). DSS was significantly lower in the positive CRP group compared with the negative CRP group (P < 0.001; Fig. 1A). A multivariate model that included smoking history, phimosis, tumour status, tumour BJU International 2015 BJU International 273

3 Li et al. Table 1 Clinicopathological characteristics and univariate log-rank test of prognostic covariates in 124 patients. Variable 3-year DSS (95% CI) Age at surgery (years), median (range) 50 (25 86) n (%) 63 (50.8) 78.9 ( ) <50 n (%) 61 (49.2) 86.6 ( ) BMI (kg/m²), median (range) 22.7 ( ) n (%) 62 (50.0) 84.6 ( ) <22.7 n (%) 62 (50.0) 80.6 ( ) Smoking history, n (%) Yes 74 (59.7) 80.2 ( ) No 50 (40.3) 86.4 ( ) Phimosis, n (%) Yes 92 (74.2) 83.4 ( ) No 32 (25.8) 80.1 ( ) LNM laterality, n (%) Unilateral 35 (28.2) 83.2 ( ) <0.001 Bilateral 25 (20.2) 35.6 ( ) ENE, n (%) Yes 30 (24.2) 40.8 ( ) <0.001 No 94 (75.8) 94.1 ( ) Pelvic LNM, n (%) Yes 16 (12.9) 73.1 ( ) No 108 (87.1) 84.2 ( ) Tumour stage, n (%) T1 42 (33.9) 97.6 ( ) T2 68 (54.8) 78.4 ( ) T3 14 (11.3) 51.1 ( ) Tumour grade, n (%) G1 79 (62.9) 89.6 ( ) G2 35 (28.2) 74.5 ( ) G3 10 (8.1) 58.3 ( ) Node status, n (%) N0 64 (51.6) <0.001 N1 8 (6.5) 75.0 ( ) N2 15 (12.1) 93.3 ( ) N3 37 (29.8) 49.7 ( ) CRP, n (%) Negative: <4.5 mg/l 70 (56.5) 96.9 ( ) <0.001 Positive: 4.5 mg/l 54 (43.5) 67.8 ( ) SCC-Ag, n (%) Negative: <1.4 g/ml 57 (46.0) 97.8 ( ) <0.001 Positive: 1.4 g/ml 67 (54.0) 69.8 ( ) CRP and SCC-Ag*, n (%) Score 1 45 (36.3) <0.001 Score 2 37 (29.8) 90.6 ( ) Score 3 42 (33.9) 60.2 ( ) BMI, body mass index; CRP, C-reactive protein; DSS, disease-specific survival; ENE, extranodal extension; LNM, lymph node metastases; SCC-Ag, squamous cell carcinoma antigen. * CRP( ), SCC-Ag( ) = score 1; CRP(+), SCC-Ag( )/CRP( ), SCC-Ag(+) = score 2; CRP(+), SCC-Ag(+) =score 3. cell differentiation, nodal status and CRP found no statistical significance for any of these factors as prognostic indicators for penile cancer (Table 3 ; model 1). An elevated level of SCC-Ag was also positively correlated with ENE (P < 0.001), pathological tumour status (P = 0.001) and pathological nodal status (P < 0.001); however, despite the fact that the positive SCC-Ag group had a higher incidence of pelvic LNM compared with the negative SCC-Ag group (9.6 vs 3.2%; P = [Table 2]), presurgical SCC-Ag P levels did not correlate with pelvic LNM. Our results showed that SCC-Ag positivity was a valuable prognostic factor for penile cancer (P < 0.001, Fig. 1B). Similarly to the CRP level, in a Cox multivariate model that included the SCC-Ag level and other risk factors, no factor retained statistical significance as a prognostic indicator for penile cancer (Table 3, model 2). The CRP level was found to be negatively correlated with the SCC-Ag level (r 2 = 0.193, P < 0.01). To examine our hypotheses (see Statistical Analyses), CRP and SCC-Ag levels were combined as a single variable, and the patients were divided into three groups for prognosis analysis using this variable (no marker positive, one marker positive, and two markers positive). A close association was observed between the coexistence of a positive status for CRP and SCC-Ag and LNM laterality (P = 0.041), ENE (P < 0.001), pelvic LNM (P = 0.024), pathological tumour status (P = 0.002), and pathological nodal status (P < 0.001; Table 4). When survival rates were compared, DSS in the positive CRP and SCC-Ag group was significantly reduced compared with the other two groups (P < 0.001; Fig. 1C). In a multivariate analysis, the combination of positivity for CRP and SCC-Ag had a significant influence on DSS, with a high hazard ratio (hazard ratio 3.39, 95% CI ; P = 0.033) after adjusting for possible confounding variables (Table 3; model 3). Discussion Elevated CRP levels before surgery can serve as a diagnostic marker and a marker of poor prognosis in penile cancer [16,17]. Similarly, several studies have shown that SCC-Ag level, which might also serve as a sensitive marker of disease progression, is correlated with tumour burden in penile cancer [19 21]. In the present study, we first show the relationship between preoperative serum CRP and SCC-Ag levels in penile SCC. C-reactive protein is an acute-phase reactant that responds to the tumour microenvironment, causing tumour cell death and local tissue damage [25], and high levels of this protein have been linked to cancer risk and/or progression in various malignancies [13 15]. In the present study, elevated CRP levels were associated with progressive tumour characteristics, including ENE (P < 0.001), pelvic LNM (P = 0.007), pathological tumour status (P = 0.002), and pathological nodal status (P < 0.001; Table 2), consistent with several previous studies [16,17]. Because these pathological characteristics are of utmost prognostic importance in penile carcinoma [1 3,22,23], an elevated CRP level may be predictive of more aggressive and progressive disease [16,17]. Furthermore, univariate analysis showed a significant correlation between the CRP level and DSS (P < 0.001, Fig. 1A). Interestingly, when the level of CRP was included in a multivariate model with previously identified risk factors, 274 BJU International 2015 BJU International

4 Preoperative CRP and SCC antigen levels and penile cancer Table 2 Associations between preoperative CRP and SCC-Ag levels and clinicopathological characteristics (n = 124). Variable CRP SCC-Ag Negative, n (%) Positive, n (%) P Negative, n (%) Positive, n (%) P Age at surgery (years) (25.8) 31 (25.0) (20.2) 38 (30.6) <50 38 (30.6) 23 (18.5) 32 (25.8) 29 (23.4) BMI (kg/m²) (31.5) 23 (18.5) (25.8) 30 (24.2) < (25.0) 31 (25.0) 25 (20.2) 37 (29.8) Smoking history Yes 40 (32.3) 34 (27.4) (28.2) 39 (31.5) No 30 (24.2) 20 (16.1) 22 (17.7) 28 (22.6) Phimosis Yes 52 (41.9) 40 (32.3) (33.1) 51 (41.1) No 18 (14.5) 14 (11.3) 16 (12.9) 16 (12.9) LNM laterality Unilateral 17 (13.7) 18 (14.5) (8.1) 25 (20.2) Bilateral 6 (4.8) 19 (15.3) 4 (3.2) 21 (16.9) ENE Yes 7 (5.6) 23 (18.5) < (4.0) 25 (20.2) <0.001 No 63 (50.8) 31 (25.0) 52 (41.9) 42 (33.9) Pelvic LNM Yes 4 (3.2) 12 (9.6) (3.2) 12 (9.6) No 66 (53.2) 42 (33.9) 53 (42.7) 55 (44.4) Tumour stage T1 32 (25.8) 10 (8.1) (23.4) 13 (10.4) T2 34 (27.4) 34 (27.4) 24 (19.4) 44 (35.5) T3 4 (3.2) 10 (8.1) 4 (3.2) 10 (8.1) Tumour grade G1 48 (38.7) 31 (25.0) (32.3) 39 (31.5) G2 18 (14.5) 17 (13.7) 12 (9.7) 23 (18.5) G3 4 (3.2) 6 (4.8) 5 (4.0) 5 (4.0) Node status N0 47 (37.9) 17 (13.7) < (35.5) 20 (16.1) <0.001 N1 4 (3.2) 4 (3.2) 1 (0.8) 7 (5.6) N2 9 (7.3) 6 (4.8) 5 (4.0) 10 (8.1) N3 10 (8.1) 27 (21.8) 7 (5.6) 30 (24.2) CRP Negative 45 (36.3) 25 (20.2) Positive 12 (9.7) 42 (33.9) BMI, body mass index; CRP, C-reactive protein; ENE, extranodal extension; LNM, lymph node metastases; SCC-Ag, squamous cell carcinoma antigen. none of those risk factors retained significance as an independent prognostic indicator for penile SCC (Table 3; model 1). This finding may be attributable to the following factors: (i) the small sample size; (ii) the strong association between CRP level and metastasis; and (iii) the fact that nodal disease was the only significant predictor of cancerrelated death in the multivariate analyses [17]. We also acknowledge that the small sample size restricted the number of variables in the multivariate analysis and the power of model comparison. The optimum cut-off for these predictors needs further evaluation in a large multicentre study. Accordingly, all of the analyses reported in the present paper may be considered exploratory rather than hypothesis-testing. Squamous cell carcinoma antigen is a tumour-associated protein that was first identified in the cervix [26]. As an increasing number of studies have since shown the clinical significance of SCC-Ag in human cancers [18,19,27 29], serum SCC-Ag level is useful for evaluating response to therapy and tumour prognosis. Univariate analysis in the present study also confirmed positive relationships between elevated SCC-Ag and ENE (P < 0.001), pathological tumour status (P = 0.001), pathological nodal status (P < 0.001; Table 2) and DSS (P < 0.001, Fig. 1B). SCC-Ag may promote tumorigenesis by inhibiting apoptosis, promoting tumour cell survival, and increasing cell migration [30], and SCC-Ag may accordingly play a role in tumour invasion and metastasis. Unfortunately, SCC-Ag level did not retain significance as an independent prognostic indicator of penile cancer in our multivariate analysis (Table 3; model 2). This result may be attributed to the fact that the level of SCC-Ag does not have sufficient sensitivity for the detection of tumours with a small tumour burden and has minimal prognostic significance with regard to survival after surgery [19]. To the best of our knowledge, the present study is the first to show the relationship between CRP and SCC-Ag levels and penile SCC. Larger internal and external validation is BJU International 2015 BJU International 275

5 Li et al. Fig. 1 Kaplan Meier estimates for disease-specific survival stratified by (A) C-reactive protein (CRP) level, (B) squamous cell carcinoma antigen (SCC-Ag) level and (C) CRP and SCC-Ag level. A Disease Specific Survival (%) Disease Specific Survival (%) Disease Specific Survival (%) B 100 C CRP(-) CRP(+) P< Time(months) SCC-Ag(-) SCC-Ag(+) P< Time(months) Score1 Score2 Score Time(months) necessary to confirm the prognostic value of elevations in these levels. Furthermore, combined CRP and SCC-Ag level measurement was more significantly correlated with clinical status than either level alone. Indeed, positivity for increases in both CRP and SCC-Ag levels had a profoundly significant association with LNM laterality (P = 0.041), ENE (P < 0.001), pelvic LNM (P = 0.024), pathological tumour status (P = 0.002), pathological nodal status (P < 0.001) and poorer DSS (P < 0.001) compared with positivity for one or neither of those factors (Tables 1 and 4 and Fig. 1C). Interestingly, the correlation of these two markers in combination with survival was confirmed by multivariate analysis after adjusting for the same factors (Table 3; model 3), indicating that the combination of these two markers can serve as an independent prognostic marker in patients with penile SCC. Various studies have shown that para-carcinoma, carcinoma and LNM tissues are generally accompanied by inflammation, which promotes inflammatory cell infiltration and cytokine production and results in increased CRP and SCC-Ag serum levels; thus, accelerated tumour growth is associated with elevated CRP and SCC-Ag levels [28,29]. Clinically, Huang et al. [29] demonstrated that elevated SCC-Ag and CRP levels are associated with a high metabolic rate as well as proliferative activity in oral SCC. Radical surgery is curative in patients with penile SCC with minimal metastases but represents overtreatment in >80% of cases if applied unselectively to all patients [9]. Nonetheless, the value of properly performed inguinal lymphadenectomy in the treatment of penile cancer is undisputed. Because classic molecular markers are not clinically useful in SCC of the penis, in the present study we explored the use of the combination of CRP and SCC-Ag levels as an independent prognostic marker and a biomarker in penile SCC to stratify patients for timely individualization of treatment without overtreatment. Further analysis of the molecular mechanisms of the observed interaction between CRP and SCC-Ag in penile SCC may help clarify the results of this study. The present study has some limitations. First, the data collection was retrospectively performed, and the follow-up period was relatively short, therefore, further research and longer follow-up periods are required. Second, the diversity of treatment in the present study, i.e. adjuvant therapy and pelvic lymphadenectomy, may potentially affect other variables. Pelvic lymphadenectomy was not routinely performed before 2009 because it was not recommended by the guidelines as standard treatment for penile cancer. In addition, adjuvant therapies as well as the course of treatment varied. Unfortunately, some patients who should have been treated with adjuvant therapies declined to receive them. A large proportion of Chinese patients present with advanced cancer when they see a doctor, and these individuals are typically poor and undereducated and live in outlying regions of major cities. The results of the present study may also be 276 BJU International 2015 BJU International

6 Preoperative CRP and SCC antigen levels and penile cancer Table 3 Multivariate Cox regression model of prognostic covariates in 124 patients with penile squamous cell carcinoma. Variable Multivariate analysis model 1 Multivariate analysis model 2 Multivariate analysis model 3 HR 95% CI P HR CI (95%) P HR CI (95%) P Smoking history No vs Yes Phimosis No vs Yes Tumour stage T1 vs T2 vs T Tumour grade G1 vs G2 vs G Node status N0 vs N1 vs N2 vs N CRP Negative vs positive SCC-Ag Negative vs positive CRP and SCC-Ag Score 1 vs Score 2 vs Score CRP, C-reactive protein; HR, hazard ratio; SCC-Ag, squamous cell carcinoma antigen. Table 4 Associations between preoperative C-reactive protein and squamous cell carcinoma antigen levels and clinicopathological variables (n = 124). Variable Score 1*, n (%) Score 2*, n (%) Score 3*, n (%) Age at surgery (years) (15.3) 19 (15.3) 25 (20.2) <50 26 (21.0) 18 (14.5) 17 (13.7) BMI (kg/m²) (21.8) 17 (13.7) 18 (14.5) < (14.5) 20 (16.1) 24 (19.4) Smoking history Yes 29 (23.4) 17 (13.7) 28 (22.6) No 16 (12.9) 20 (16.1) 14 (11.3) Phimosis Yes 33 (26.6) 27 (21.8) 32 (25.8) No 12 (9.7) 10 (8.1) 10 (8.1) LNM laterality Unilateral 6 (4.8) 15 (12.1) 14 (11.3) Bilateral 3 (2.4) 4 (3.2) 18 (14.5) ENE Yes 2 (1.6) 8 (6.5) 20 (16.1) <0.001 No 43 (34.7) 29 (23.4) 22 (17.7) Pelvic LNM Yes 2 (1.6) 4 (3.2) 10 (8.1) No 43 (34.7) 33 (26.6) 32 (25.8) Tumour stage T1 26 (21.0) 9 (7.3) 7 (5.6) <0.001 T2 18 (14.5) 22 (17.7) 28 (22.6) T3 1 (0.8) 6 (4.8) 7 (5.6) Tumour grade G1 33 (26.6) 22 (17.7) 24 (19.4) G2 8 (6.5) 14 (11.3) 13 (10.5) G3 4 (3.2) 1 (0.8) 5 (4.0) Node status N0 37 (29.8) 17 (13.7) 10 (8.1) <0.001 N1 0 5 (4.0) 3 (2.4) N2 4 (3.2) 6 (4.8) 5 (4.0) N3 4 (3.2) 9 (7.3) 24 (19.4) BMI, body mass index; CRP, C-reactive protein; ENE, extranodal extension; LNM, lymph node metastases; SCC-Ag, squamous cell carcinoma antigen. * CRP( ), SCC-Ag ( ) = score 1; CRP(+), SCC-Ag( )/CRP( ), SCC-Ag(+) = score 2; CRP(+), SCC-Ag (+) = score 3. P affected by selection bias, which is inherent to retrospective studies. Dynamic sentinel node biopsy has been suggested as an alternative for low-risk patients in a few centres; however, such multidisciplinary teamwork requires ultrasonography, nuclear imaging, surgical exploration and meticulous pathological examination, which has limited its widespread use in developing countries such as China. Dynamic sentinel node biopsy was therefore not considered in the present study, potentially adding bias. Third, the correlation between CRP/SCC-Ag and recurrence rates should be assessed in future research. Given that penile cancer is a relatively rare malignancy, the sample size of 124 is a fairly large cohort and well reflects our intent. The limited number of patients and other limitations described above, however, also inhibited our ability to perform multivariate analyses and to select patients for inguinal lymph node dissection vs surveillance. For example, the inclusion of tioncancers as one group leads to an inherent selection bias, and authors should consider subgrouping these patients. All analyses can be considered exploratory rather than hypothesis-testing. Overall, the relevance of the markers examined in the present study and their ultimate usefulness in patients should be validated externally using larger, independent datasets. In conclusion, the present study showed that the combined measurement of both preoperative CRP and SCC-Ag levels may serve as an independent biomarker for LNM, advanced tumour stage and DSS in patients with penile SCC. The optimum cut-off for these predictors requires further evaluation in a large multicentre study, and the long-term follow-up also needs external validation. Conflict of Interest None declared. BJU International 2015 BJU International 277

7 Li et al. References 1 Clark PE. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) Penile Cancer (2015) Available at: professionals/physician_gls/f_guidelines.asp#penile (Accessed March 2015) 2 Hakenberg OW, Comperat EM, Minhas S, Necchi A, Protzel C, Watkin N. EAU Guidelines on Penile Cancer: 2014 Update. Eur Urol 2015 [E-pub ahead of print], DOI: /j.eururo Edge SB, Byrd D, Compton C et al. AJCC Cancer Staging Manual, 7th edn. New York: Springer, 2009: Leijte JA, Valdes OR, Nieweg OE, Horenblas S. Anatomical mapping of lymphatic drainage in penile carcinoma with SPECT-CT: implications for the extent of inguinal lymph node dissection. Eur Urol 2008; 54: Pettaway C. Tumors of the penis. In Wein AJ ed. Campbell-Walsh Urology, 10th edn, Chapt 34. St. Louis: WB Saunders, 2012: Leijte JA, Gallee M, Antonini N, Horenblas S. Evaluation of current TNM classification of penile carcinoma. J Urol 2008; 180: Li ZS, Yao K, Chen P et al. Disease-specific survival after radical lymphadenectomy for penile cancer: prediction by lymph node count and density. Urol Oncol 2014; 32: Zhu Y, Ye DW, Yao XD et al. New N staging system of penile cancer provides a better reflection of prognosis. J Urol 2011; 186: Hegarty PK, Kayes O, Freeman A et al. A prospective study of 100 cases of penile cancer managed according to European Association of Urology guidelines. BJU Int 2006; 98: 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: Spiess PE, Hernandez MS, Pettaway CA. Contemporary inguinal lymph node dissection: minimizing complications. World J Urol 2009; 27: Yao K, Tu H, Li YH et al. Modified technique of radical inguinal lymphadenectomy for penile carcinoma: morbidity and outcome. J Urol 2010; 184: Ledue TB, Weiner DL, Sipe JD et al. Analytical evaluation of particleenhanced immunonephelometric assays for C-reactive protein, serum amyloid A and mannose-binding protein in human serum. Ann Clin Biochem 1998; 35 (Pt 6): Chen HH, Chen IH, Liao CT et al. Preoperative circulating C-reactive protein levels predict pathological aggressiveness in oral squamous cell carcinoma: a retrospective clinical study. Clin Otolaryngol 2011; 36: Motoyama S, Nakatsu T, Miura M et al. A CRP genetic polymorphism associated with the tumoral expression of CRP in esophageal cancer. Surg Today 2013; 43: Al GA, Steffens S, Steinestel J et al. Elevated C-reactive protein values predict nodal metastasis in patients with penile cancer. BMC Urol 2013; 13: Steffens S, Al GA, Steinestel J et al. High CRP values predict poor survival in patients with penile cancer. BMC Cancer 2013; 13: Strauss HG, Laban C, Lautenschlager C et al. SCC antigen in the serum as an independent prognostic factor in operable squamous cell carcinoma of the cervix. Eur J Cancer 2002; 38: Zhu Y, Ye DW, Yao XD et al. The value of squamous cell carcinoma antigen in the prognostic evaluation, treatment monitoring and followup of patients with penile cancer. J Urol 2008; 180: Hungerhuber E, Schlenker B, Schneede P, Stief CG, Karl A. Squamous cell carcinoma antigen correlates with tumor burden but lacks prognostic potential for occult lymph node metastases in penile cancer. Urology 2007; 70: Touloupidis S, Zisimopoulos A, Giannakopoulos S et al. Clinical usage of the squamous cell carcinoma antigen in patients with penile cancer. Int J Urol 2007; 14: Hegarty PK. EAU guidelines for management of penile cancer. Indian J Urol 2007; 23: Pizzocaro G, Algaba F, Horenblas S et al. EAU penile cancer guidelines Eur Urol 2010; 57: Foucher Y, Danger R. Time dependent ROC curves for the estimation of true prognostic capacity of microarray data. Stat Appl Genet Mol Biol 2012; 11: 1 25 McMillan DC, Canna K, McArdle CS. Systemic inflammatory response predicts survival following curative resection of colorectal cancer. Br J Surg 2003; 90: Kato H, Torigoe T. Radioimmunoassay for tumor antigen of human cervical squamous cell carcinoma. Cancer 1977; 40: Kawaguchi H, Ohno S, Miyazaki M et al. CYFRA 21-1 determination in patients with esophageal squamous cell carcinoma: clinical utility for detection of recurrences. Cancer 2000; 89: Chen HH, Wang HM, Fan KH et al. Pre-treatment levels of C-reactive protein and squamous cell carcinoma antigen for predicting the aggressiveness of pharyngolaryngeal carcinoma. PLoS One 2013; 8: e Huang SF, Wei FC, Liao CT et al. Risk stratification in oral cavity squamous cell carcinoma by preoperative CRP and SCC antigen levels. Ann Surg Oncol 2012; 19: Takeda A, Kajiya A, Iwasawa A, Nakamura Y, Hibino T. Aberrant expression of serpin squamous cell carcinoma antigen 2 in human tumor tissues and cell lines: evidence of protection from tumor necrosis factormediated apoptosis. Biol Chem 2002; 383: Correspondence: Fang-Jian Zhou, Department of Urology, Sun Yat-sen University Cancer Center, Guangzhou , China. zhoufj@sysucc.org.cn Abbreviations: CRP, C-reactive protein; SCC, squamous cell carcinoma; SCC-Ag, squamous cell carcinoma antigen; LNM, lymph node metastasis; DSS, disease-specific survival; ENE, extranodal extension. 278 BJU International 2015 BJU International