SEARCH

SEARCH BY CITATION

Keywords:

  • penile neoplasms;
  • lymphatic metastasis;
  • biological markers

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

OBJECTIVE

To evaluate the prognostic significance of p53, Ki-67, epithelial cadherin (E-cadherin) and matrix metalloproteinase-9 in primary penile cancer, as the presence of lymph node metastasis and long-term survival are hard to define in penile squamous cell carcinoma.

PATIENTS AND METHODS

Paraffin-embedded primary tumour samples were obtained from 73 Chinese patients who had penile amputation and regional lymphadenectomy. The expression of molecular markers was determined by immunohistochemistry. Logistic regression was used to identify factors associated with lymph node metastasis, and a Cox proportional-hazards model was used to measure cancer-specific survival (CSS).

RESULTS

Thirty (41%) patients presented with nodal disease and the 3-year CSS rate for all patients was 72%. Lymph node metastasis was significantly correlated with tumour stage, histological grade, lymphatic and vascular embolization, and the expression of p53, Ki-67 and E-cadherin. By multivariate analysis, tumour embolization and the expression of p53 were independent predictors of metastasis. Survival analysis showed that the expression of p53 was an independent prognostic factor for CSS. In stage T1 tumours, high expression of p53 was significantly associated with metastasis and poor survival.

CONCLUSION

Lymphatic and vascular embolization, and p53 immunoreactivity, are helpful in establishing the probability of lymph node metastasis. The expression of p53 is an independent predictor of CSS in Chinese patients with penile cancer. In stage T1 tumours, p53 staining is an important variable determining the prognosis and treatment outcome.


Abbreviations
SCC

squamous cell carcinoma

CSS

cancer-specific survival

LN(M)

lymph node (metastasis)

LVE

lymphatic and vascular embolization

E-cadherin

epithelial cadherin

MMP-9

matrix metalloproteinase-9.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Penile cancer is a rare disease in urban Shanghai, accounting for <1% of all male malignancies [1]. But in certain areas where hygiene and health conditions are poor, it is still a substantial health problem, constituting up to 10% of cancers in men [2]. The presence and extent of lymph node metastases (LNMs) are the main prognostic factors for survival in penile squamous cell carcinoma (SCC) [3–5]. Regional lymphadenectomy has been regarded as the most effective treatment for LNMs [6], but noninvasive staging strategies to predict pathologically involved LNs are unreliable [7], and prophylactic lymphadenectomy is associated with a high morbidity rate [8,9]. Therefore, several studies investigated the prognostic significance of histopathological features to identify those patients who are at high risk of LNM and should have a therapeutic lymphadenectomy [3,5,10,11]. For example, tumour stage, grade and the presence of lymphatic or vascular invasion are well established predictors [3,5,10,11], but a prospective study showed that 82% of patients had an unnecessary prophylactic lymphadenectomy according to the risk group categorized by tumour stage and grade [12]. Thus, more accurate prognostic factors are needed for better patient selection.

The current model of carcinogenesis suggests that the formation of metastasis is a multi-step process governed by successive changes of certain genes and their encoded products. Early detection of molecules that enable cancer cells to metastasize might help to develop more targeted therapeutic strategies. Furthermore, tumour marker profiles including several markers with different biological roles are more helpful in assessing a patient’s prognosis. Recent evidence suggested that p53 and epithelial cadherin (E-cadherin) might be promising predictors of metastasis in penile cancer [13–15]. In the present study we examined various molecular markers (p53, Ki-67, E-cadherin, and matrix metalloproteinase-9, MMP-9) in patients with primary penile cancer to investigate their prognostic value for LNM and cancer-specific survival (CSS).

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The study included 73 Chinese patients who had surgery for penile SCC from 1990 to 2005; we excluded a further 18 who had penile-sparing procedures, including local excision and radiation therapy. Sixty-two patients had partial amputation of the penis and the other 11 had total amputation. All patients had a bilateral inguinal lymphadenectomy; the borders of the inguinal dissection included 2 cm above the inguinal ligament superiorly, the apex of the femoral triangle inferiorly, the adductor longus medially and the sartorius muscles laterally [16]. All superficial and deep inguinal LNs within the field were removed. A vertical incision from the middle of the inguinal ligament down to the apex of the femoral triangle is the approach preferred at our institution. Iliac procedures, with removal of the external iliac and obturator LNs, were used in 13 patients with radiologically suspected positive pelvic nodes or involvement of Cloquet’s node on frozen-section analysis. None of the patients had had any adjuvant treatment before or after surgery. The patients had a routine follow-up assessment every 3 months in the first 2 years, every 6 months in the third, and annually thereafter, including a physical examination, abdominal ultrasonography and chest radiography.

For the histopathological evaluation two observers, unaware of the clinical data, independently reviewed the slides, and all discrepancies were resolved by joint review of the slides in question. The primary lesion was staged according to the 2002 TNM system [17]. Histological grade was classified according to the Broders system as G1 (well), G2 (moderately) and G3 (poorly) differentiated [18]. Lymphatic and vascular embolization (LVE) was defined by the presence of tumour emboli within endothelium-lined spaces bounded by a thin wall [10].

For the immunohistochemical staining and evaluation, 10 4 µm-thick sections were cut from each designated formalin-fixed, paraffin-embedded tissue block. To facilitate epitope retrieval, the sections were incubated in heated 0.01 m citrate buffer (pH 6.0) for 10 min. Endogenous peroxidase was blocked by 3% H2O2 for 10 min and nonspecific binding of immunoglobulin was blocked by 10% normal goat serum for 10 min. The sections were then incubated with appropriately diluted primary antibodies at 4 °C overnight. Monoclonal antibodies used were clone DO-7 (anti-p53, at 1 : 50 dilution, Dakocytomation, Glostrup, Denmark), clone MIB-1 (anti-Ki-67, at 1 : 75 dilution, Dakocytomation), clone 4A2C7 (anti-E-cadherin, at 1 : 50 dilution, Zymed Laboratories Inc., CA, USA), and clone GE-213 (anti-MMP-9, at 1 : 50 dilution, Oncogene Research Products, CA, USA). The slides were incubated with goat anti-mouse immunoglobulin conjugated to peroxidase (EnVision + System, Dakocytomation) for 30 min, followed by application of 3, 3′-diaminobenzidine and haematoxylin.

Normal skin epithelium was used as a positive control for E-cadherin, while sections of breast cancer known to express p53, Ki-67 and MMP-9 were used as the positive controls for the remaining markers. Sections treated with no primary antibodies were used as negative controls. The percentage of p53- or Ki-67-positive tumour cells was determined by calculating the percentage of stained tumour nuclei in the most highly stained area on each slide. p53 expression was categorized as high when >20% of the tumour nuclei were stained, according to a previous report [13]. For Ki-67 staining the overall mean expression score (20%) was used as the threshold. E-cadherin and MMP-9 immunoreactivity were evaluated semiquantitatively and classified into four categories: 0 (absent), no staining; 1 (weak), staining in <10% of cells; 2 (moderate), moderate homogenous staining or intense localized staining in 10–75% of cells; 3 (intense), intense homogeneous staining in >75% of cells. For statistical analyses, these categories were pooled (0 and 1, low expression; 2 and 3, high expression).

Patients were classified into four categories at the end of the follow-up, i.e. alive, dead from penile cancer, dead from causes other than penile cancer, and lost to follow-up. The chi-square test was used to assess differences among the variables. A logistic regression model was used in a multivariate analysis of the potential factors, and the Kaplan–Meier method to plot the survival function, using the log-rank test to compare survival curves. A Cox proportional hazard model was used to identify the best predictors. Two-sided P values are reported for all analyses and considered statistically significant when <0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The median (range) age of the patients was 55 (27–75) years; 63 (86%) presented with a redundant prepuce or phimosis, and seven had been circumcised as adults. The median (range) number of LNs removed in the groin dissection and iliac procedure were 21 (11–33) and 8 (6–14), respectively.

Of the 73 tumour sections, 24 (33%) had high expression of p53, 26 (36%) had high expression of Ki-67, 33 (45%) had low expression of E-cadherin and 34 (47%) had high expression of MMP-9 (Fig. 1). Comparing the expression of molecular markers with pathological features showed no significant association.

image

Figure 1. Immunohistochemical detection of p53, Ki-67, E-cadherin and MMP-9 in penile SCC. × 50.

Download figure to PowerPoint

The pathological evaluation showed that 30 (41%) patients had LNM; the relationships between the evaluated variables and nodal status are shown in Table 1. The presence of LVE (odds ratio 5.35; 95% CI 1.009–28.313) and high expression of p53 (6.01; 1.402–25.764) were independent risk factors for LNM. Combining risk factors was informative in this subset of patients; when there was tumour embolization and high expression of p53 all 10 patients had metastasis. As 56% of patients had stage T1 tumours and the incidence of LNM was 20%, we evaluated whether any of the variables could identify the T1 tumours with a high risk of metastasis (Table 1). High expression of p53 and low expression of E-cadherin were significant for stratifying the risk of metastasis for patients with T1 tumours (both P = 0.013). LVE was rare in T1 tumours (three of 41), and statistical analysis showed no significant association with LNM within this subset (P = 0.092). In the 32 patients with ≥ T2 tumours p53 expression did not correlate with metastasis (P = 0.446). The incidence of LNM was 81% (17 of 21) for patients with tumour embolization and five of 11 for patients without LVE in those with ≥ T2 stage, but Fisher’s exact test showed no significant association (P = 0.056).

Table 1.  Prognostic factors for LNMs in penile SCC, for all patients and T1 only, and for the 3-year CSS
CategoryNn (%) LNMsP for analysisT1 only3-year CSS, %P for analysis
univariate*multivariateNn (%) LMNsPunivariatemultivariate§
  • *

    Chi-square test;

  • †logistic regression test;

  • Fisher’s exact test;

  • ¶Log-rank test;

  • §

    §Cox regression test; N, number of patients.

Age, years:
 ≤55740.5000.704
 >55       70  
Tumour stage:
 T141 8 (20)<0.0010.079870.0010.090
 ≥T23222 (69)     49  
Histological grade:
 G138 9 (24)0.0020.118263 (12)0.11794<0.0010.153
 ≥G23521 (60)  155 49  
LVE:
 Absent49 11 (22)<0.0010.049386 (16)0.09289<0.0010.203
 Present2419 (79)   32 36  
p53:
 Low4914 (29)0.0020.016313 (10)0.01387<0.0010.011
 High2416 (67)  105 41  
Ki-67:
 Low4719 (40)0.8760.861265 (19)1.0750.2610.323
 High26 11 (42)  153 64  
E-cadherin:
 Low3319 (58)0.0090.089187 (39)0.013530.0020.176
 High40 11 (28)  231 (4) 92  
MMP-9:
 Low3913 (33)0.1490.852212 (10)0.130840.0060.079
 High3417 (50)  206 (30) 58  

At the time of analysis, 17 (23%) patients had died from penile cancer, four (6%) from other causes, and two (3%) were lost to follow-up. At a median (range) of 23 (8–174) months, the 3-year CSS was 72%. As shown in Table 1, tumour stage (P = 0.001), histological grade (P < 0.001), LVE (P < 0.001), the expression of p53 (P = 0.002), E-cadherin (P = 0.002), and MMP-9 (P = 0.006) were significantly associated with poor CSS in the univariate analysis. Multivariate analysis showed that the expression of p53 (relative risk 5.31; 95% CI 1.477–19.086) was the only independent predictor of CSS. In the subgroup of patients with T1 disease, high expression of p53 was significantly associated with a poor CSS (P = 0.005). By contrast, p53 had no significant predictive value for survival in those with ≥ T2 stage (P = 0.051) (Fig. 2A,B). In those patients with positive LNs, high expression of p53 was significantly correlated with poor survival (P = 0.002; Fig. 2C).

image

Figure 2. Kaplan–Meier survival curves for CSS according to the expression of p53: A , for stage T1 (P = 0.005, log-rank test); B , for stage ≥ T2 (P = 0.051, log-rank test); C , for patients with positive LNs (P = 0.002, log-rank test).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

We evaluated the prognostic significance of several molecular markers involved in carcinogenesis to better define strategies for treating penile cancer. The expression of p53, Ki-67, E-cadherin and MMP-9 were previously shown to have prognostic importance in head and neck SCC [19,20]. Two reports from a Brazilian group suggested that p53 and E-cadherin were associated with LNM in penile cancer [13,15]. The present study comprised many Chinese patients with a definitive pathological stage and strict follow-up. Because 56% of the patients had stage T1 disease, our results might be helpful in evaluating the prognostic role of molecular markers in a relatively low-risk group.

To improve the selection of candidates for lymphadenectomy several clinicopathological variables have been examined in penile cancer. However, according to a prospective study, the risk groups categorized by tumour stage and grade were unsatisfactory [12]. In the present study, multivariate analysis showed that the presence of LVE, and high expression of p53, were independent risk factors for LNMs. Of 24 patients with tumour embolization, 19 (79%) had LNM, in contrast to 22% (11 of 49) in those without. Previous studies examined the role of LVE as a prognostic factor of metastasis [3,10,11,21]. In 1996, Lopes et al.[3] reported on the prognostic value of LVE in 145 patients with penile cancer. Multivariate analysis showed that only LVE remained significant predictors of positive LNs (P < 0.001 and 0.041, respectively). In 2001, Slaton et al.[11] reported that vascular invasion could be used to stratify patients with stage T2 to receive early lymphadenectomy. In the subgroup with T2 disease, the risk of metastasis was 25% in patients with no and 71% in those with vascular invasion. Recently, a multicentre study including 175 patients further showed that LVE was an important predictor of nodal involvement in patients with clinically negative LNs [10]. In the present series, although LVE remained an independent prognostic factor for metastasis, the statistical significance was marginal (P = 0.049). This difference could be attributable to a high percentage of stage T1 disease. Slaton et al.[11] reported that LVE was absent in T1 tumours (15 cases). In this patient subset, tumour embolization was present in three T1 tumours, and was not associated with nodal disease (P = 0.092). In tumours of ≥ T2, the present data suggested that LVE might be an important predictor, but the relatively few high-stage patients limited the statistical significance (P = 0.056).

Relatively few studies have evaluated the prognostic significance of molecular markers in penile cancer [13–15]. Lopes et al.[13] reported 82 patients with penile cancer staged according to the 1978 TNM system; all had ≥ T2 disease and the overall incidence of metastasis was 51%. Univariate analysis showed that clinical LN stage, tumour embolization, p53 positivity and p53 grade were significantly associated with LNM. Multivariate analysis showed that only lymphatic embolization and p53 positivity were independent prognostic factors. Martins et al.[14] evaluated p53 expression in 50 patients with penile cancer, staged according to 1999 TNM system; 22 of 50 (44%) had stage T1 disease, and 36 (72%) with negative groin nodes at penectomy were followed. Their results confirmed that p53 staining correlated with lymphatic spread, and the prognostic significance remained in those patients with negative LNs at penectomy. In the present series, immunostaining of p53 was detected in 24 of 73 (33%) patients, and multivariate analysis confirmed p53 expression as an independent predictor for LNM (P = 0.016). Patients with high expression of p53 had 2.3 times the risk of metastasis than those with low p53 staining. In the subgroup of T1 tumours, half the patients with high p53 staining had metastasis, while only 10% of those with low expression had nodal disease (P = 0.013). Thus, p53 staining instead of tumour embolization should be emphasized in stage T1 tumours. In the present series, the incidence of LNM in T1 and ≥ T2 tumours was 20% and 69%, respectively. p53 staining might better stratify the relatively low-risk group, sparing most of the unnecessary dissections.

We also investigated the biological markers involved in tumour invasion and metastasis. E-cadherin is one of the epithelial cell adhesion molecules and is expressed in normal epithelia. The defect in E-cadherin might reduce intercellular adhesion, which allows cells to detach from the primary site. MMP-9 is a secreted enzyme that degrades the interstitial collagens in the modulation of cell–matrix interactions. There was a significant correlation between reduced E-cadherin expression and nodal disease in the present univariate analysis (P = 0.009). In T1 tumours, E-cadherin expression also had a significant association with metastasis (P = 0.013). Recently, Campos et al.[15] reported that low expression of E-cadherin was significantly correlated with LNM. In their group, 59.5% patients with low expression of E-cadherin harboured metastases, vs 38.3% in those with high expression (P = 0.032). However, multivariate analysis showed that E-cadherin expression was not an independent predictor of LN involvement. In the present study there was a trend to a relationship between E-cadherin expression and metastasis (P = 0.089), and thus further evaluation of metastasis-related markers is needed.

The present results showed that the expression of p53 was an important predictor for CSS in Chinese patients with penile cancer. Furthermore, the predictive role of p53 was more informative in patients with stage T1 disease. Three of seven patients with T1 and high expression of p53 died from penile cancer during the follow-up, while only one died from cancer in those with low expression of p53. Martins et al.[14] suggested that p53 reactivity was associated with disease progression and cause-specific survival in those patients who had surveillance of regional LNs. A greater risk of metastasis in the group with high p53 expression might contribute to the poor survival. We also found that p53 was a prognostic factor in patients with positive LNs. In this subgroup, the 3-year CSS was 71% in patients with low expression of p53 and only 11% in those with high expression. Therefore, p53 expression might be suitable for identifying patients at high risk of a poor prognosis and who might benefit from adjuvant therapy.

Although p53 had promising prognostic value in penile cancer, significant controversy remains. Because of the difference in sample characteristics and immunostaining methods used, a standard threshold of p53 positivity is hard to define. Thus, cooperative studies with many patients and the same protocol are needed to validate molecular markers prospectively.

In conclusion, LVE and p53 immunoreactivity are helpful in establishing the probability of LNM. The expression of p53 is an independent predictor of CSS in Chinese patients with penile cancer. In stage T1 tumours, p53 staining is an important variable for determining the prognosis and outcome of treatment.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

We thank Li Xiao Mei and Jing Su Ping for excellent technical assistance, Fei Yang for advice on statistical analysis.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES