Although SCCP is a rare disease in the Brazilian male

Transcription

1 Prognostic Factors in Invasive Squamous Cell Carcinoma of the Penis: Analysis of 196 Patients Treated at the Brazilian National Cancer Institute Antonio Augusto Ornellas,* Bernardo Lindenberg Braga Nóbrega, Eduardo Wei Kin Chin, Aristóteles Wisnescky, Paulo Cesar Barbosa da Silva and Aline Barros de Santos Schwindt From the Departments of Urology and Pathology, Brazilian National Cancer Institute, Rio de Janeiro, Brazil Purpose: We evaluated the role of primary tumor histopathological features for predicting regional metastasis and the prognosis in patients with penile squamous cell carcinoma. Materials and Methods: From April 1996 to January 2007, 202 consecutive patients with penile carcinoma underwent surgical treatment at our institution. Of these patients 196 were studied to identify prognostic factors. All histological specimens were examined by the same pathologist. We considered certain histological parameters, including histological grade, invasion depth, lymphovascular embolization, perineural infiltration, infiltration of the corpus cavernosum or spongiosum, urethral infiltration and koilocytosis. Results: Variables significantly associated with regional metastasis on univariate analysis were stage stratification (p ), histological grade (p ), invasion depth (0.0114), lymphovascular embolization (p ), perineural infiltration (p ), corpora cavernosa infiltration (p ) and koilocytosis (p ). In the multivariable model lymphovascular embolization and absent koilocytosis were independent risk factors for lymphatic metastasis (p and 0.009, respectively). We also found a better survival rate in patients with koilocytosis and without lymphovascular embolization (p and 0.005, respectively). Conclusions: Lymphovascular embolization and absent koilocytosis were independent prognostic factors for the risk of lymphatic metastasis. Patients with koilocytosis and without lymphovascular embolization had better 5-year survival. Key Words: penis; penile neoplasms; carcinoma, squamous cell; neoplasm metastasis; lymph nodes Although SCCP is a rare disease in the Brazilian male population, it is a significant health problem in some regions of the country. Penile carcinoma metastasis usually spreads through penile lymphatic vessels to regional nodes, especially the superficial and deep inguinal nodes, and subsequently to the iliac nodes in the pelvis. Tumor involvement of regional nodes is the best indicator of longterm survival in patients with invasive SCCP. 1 Physical examination is not a reliable predictor of lymph node status and, therefore, another clinicopathological indicator of inguinal lymph node involvement is needed. 2 Therefore, the only reliable staging information may be acquired through a surgical procedure with subsequent histological examination of the inguinal lymph nodes. However, patients with clinically nonsuspicious nodes and a significant number with enlarged nodes do not have metastatic spread and, thus, they would undergo an unnecessary procedure. Pathological factors with a known prognostic value, other than lymph node metastasis, are tumor thickness, grade, histological type, lymphovascular embolization and stage. 3,4 We evaluated the clinical and pathological parameters of the primary tumor regarding the incidence of metastasis and overall survival rates in 196 patients treated at our institution. Submitted for publication February 10, * Correspondence: Instituto Nacional de Câncer, Praça da Cruz Vermelha 23, Rio de Janeiro, Brazil ( ornellasa@hotmail.com). PATIENTS AND METHODS Clinical and pathological data on 202 patients with invasive SCCP treated at our institution between 1996 and 2007 were prospectively evaluated. Mean and median patient age was 57 years at diagnosis (range 25 to 98). Disease was staged in all patients according to the 1978 TNM system. Six patients with unknown clinical T and/or N stage were excluded from statistical analysis. The distribution of T category in all 196 cases was clinical stage T1 in 29 (14.8%), clinical stage T2 in 94 (47.9%), clinical stage T3 in 58 (29.6%) and clinical stage T4 in 15 (7.7%). The distribution of N category in the 196 cases was stage N0 in 103 (52.6%), stage N1 in 22 (11.2%), stage N2 in 46 (23.5%) and stage N3 in 25 (12.7%). We stratified stage into group 1 T1N0, group 2 T1N1 and T2N0-1, group 3 T1N2, T2N2 and T3N0-2, and group 4 T1-3N3 and T4N0-3. Median followup was 74 months (range 1 to 93). Followup was evaluated in 193 of 196 patients. We excluded from the analysis of survival curves 3 patients with less than 1 month of clinical followup. Pathological material was reviewed and all tumors were classified according to the Broders grading system. A single pathologist examined all specimens. The pathological variables studied were histological differentiation grade, invasion depth, infiltration of the corpus cavernosum or spongiosum, perineural infiltration, urethral infiltration, lymphovascular embolization and koilocytosis. Lymphovascular embolization was de /08/ / Vol. 180, , October 2008 THE JOURNAL OF UROLOGY Printed in U.S.A. Copyright 2008 by AMERICAN UROLOGICAL ASSOCIATION DOI: /j.juro

2 PROGNOSTIC FACTORS IN INVASIVE SQUAMOUS CELL CARCINOMA OF PENIS 1355 FIG. 1. SCCP. A, area of lymphatic and venous embolization with tumor nests invading lymphatic and venous vascular spaces in lamina propria. B, area of dyskeratotic cells with atypical koilocytes showing enlarged hiperchromatic nuclei, irregular contours and prominent perinuclear halo. H & E, reduced from 400. fined as tumor emboli in endothelium lined spaces bounding a wall in the presence or absence of parietal smooth muscle fibers and red blood cells. Koilocytosis was identified by a large halo around cell nuclei. According to histopathological evaluation in 196 patients risk stratification was separated into low T1G1 and T1G2, intermediate T2G1, T2G2, T3G1 and T3G2, and high risk tumors T1-3G3 and T4G1-3. Primary treatment for penile lesions varied from local excision to total removal of the penis, testicles and scrotum. Local excision was performed in 14 patients, circumcision was done in 8, partial amputation of the penis was performed in 144, total amputation was done in 27 and total amputation with removal of the testicles and scrotum was performed in 3. Of 196 patients 85 underwent 1 surgical procedure (different techniques) as primary treatment for penile cancer. Immediate lymphadenectomy was not initially indicated due to advanced age or clinical conditions in these patients. An additional 24 patients required subsequent bilateral lymphadenectomy due to clinically suspicious lymph nodes and another 64 underwent amputation with immediate inguinal lymphadenectomy as the only treatment. Cultures were performed in collected secretions from penile lesions to test against different antibiotics to see which drug would best destroy the bacteria growing in tumors. Antibiotics were given postoperatively according to the results of these cultures. Immediate lymph node dissection was performed simultaneously with penectomy or a few days later. Palliative treatment, including surgical debulking combined with reconstruction techniques, was performed in 23 patients because of extensive locoregional disease. To reconstruct large defects tissue was obtained from the tensor fascia lata myocutaneous flap and/or from abdominal flaps. TABLE 1. Univariate chi-square and multivariate linear regression analysis of prognostic factors predictive of lymph nodes metastasis p Value Prognostic Factors No. Pts No. Metastasis (%) Chi-Square Univariate Multivariate Stage stratification: Groups I II (27.6) Groups III IV (42.2) Risk stratification: Low, intermediate (33.3) High (54.5) Histological grade: G (21.8) G (41.1) Invasion depth (mm): Less than (17.1) 5 or Greater (39.7) Lymphovascular embolization: Absent (28.7) Present (64.1) Perineural infiltration: Absent (30.9) Present (52.3) Corpora cavernosa infiltration: Absent (25) Present (48.7) Corpus spongiosum infiltration: Absent (33.8) Present (39.1) Urethra infiltration: Absent (35) Present (38) Koilocytosis: Absent (44.8) Present (22.5)

3 1356 PROGNOSTIC FACTORS IN INVASIVE SQUAMOUS CELL CARCINOMA OF PENIS TABLE 2. Lymph node involvement according to presence or absence of lymphovascular invasion and koilocytosis Lymphovascular Invasion Total No. Pts/No. Metastasis (%) Present: Koilocytosis 15/8 (53.3) No koilocytosis 24/17 (70.8) Absent: Koilocytosis 65/10 (15.4) No koilocytosis 92/35 (38) Patient followup data were gathered from medical charts at our institution and when necessary through contact with the patient family. The data obtained were recorded on a standard research form and filed in a database. Analysis was performed using SPSS. The associations of qualitative factors with lymphatic metastasis variables were done with the chi-square test. Independent factors related to lymphatic metastasis were identified by linear regression analysis with p 0.05 considered statistically significant. The Kaplan- Meier technique was used to evaluate overall survival and the log rank test was used to compare survival curves with p 0.05 as the significance cutoff. RESULTS Histopathological findings confirmed invasive SCCP in all patients. Of the 196 cases 21 (10.7%) were included in stage stratification group 1, 66 (33.7%) were in group 2, 75 (38.3%) were in group 3 and 34 (17.3%) were in group 4. Amputation and prophylactic inguinal lymphadenectomy were performed in 64 patients. Five of 30 patients (16.7%) without palpable adenopathy or stage N0 tumors had microscopic metastases, while 10 of 15 (66.7%) with unilateral palpable adenopathy (stage N1) had regional metastases and 10 of 19 (52.6%) with bilateral palpable adenopathy (stage N2) had inguinal metastatic disease. Penile surgery and delayed lymphadenectomy were performed in 24 patients. Nine of 18 patients (50%) who were initially staged as having N0 disease had delayed nodal metastases, as did 1 of 1 (100%) with initial stage N1 tumors and 3 of 4 (75%) with initial stage N2 cancer. One patient with initial stage N3 tumors who underwent delayed lymph node dissection had regional metastases. Palliative (hygienic) lymph node dissection was performed in 23 patients because of extensive locoregional disease. Overall tumors from 55 patients were well differentiated, 134 were moderately differentiated and 7 were poorly differentiated. Risk stratification revealed 28 low (14.3%), 146 intermediate (74.5%) and 22 high (11.2%) risk cases. We noted a tendency of patients in the high risk group to have lymph node involvement (p 0.05). Depth invasion was 5 mm or less in 35 patients (17.9%) and 5 mm or greater in 161 (82.1%). Lymphovascular embolization was present in the penectomy specimens of 39 patients (19.9%) (fig. 1, A). Local extension of the primary tumor involved the corpora cavernosa in 88 patients (44.9%), the corpus spongiosum in 69 (35.2%) and the urethra in 42 (21.4%). Koilocytosis was found in 80 patients (40.8%) (fig. 1, B). In these 80 cases a FIG. 2. Of patients 154 without lymphovascular embolization had higher 5-year survival than 39 with lymphovascular embolization (Kaplan-Meier log rank test p 0.005).

4 PROGNOSTIC FACTORS IN INVASIVE SQUAMOUS CELL CARCINOMA OF PENIS 1357 total of 35 tumors were well differentiated, 44 were moderately differentiated and 1 was poorly differentiated. We noted a correlation between koilocytosis and histological grade with our analysis showing higher koilocytosis rates in patients with well differentiated tumors than in patients with moderately and poorly differentiated tumors (p ). Lymph node metastasis results significantly correlated with stage stratification groups 3 and 4 (p ), histological grade (p ), lymphovascular embolization (p ), invasion depth (p ), perineural infiltration (p ), corpora cavernosa infiltration (p ) and absent koilocytosis (p ). Only lymphovascular embolization and absent koilocytosis proved to be independent predictive variables of lymph node involvement on multivariate analysis (p and 0.009, respectively, table 1). Table 2 lists the percent of lymph node involvement in a comparison of lymphovascular invasion (yes/no) vs koilocytosis (yes/no). Patients without lymphovascular embolization had better survival rates at 5 years than those with lymphovascular embolization (p 0.005, fig. 2). Log rank analysis also showed differences in 5-year disease survival curves according to the presence of koilocytosis (p 0.001, fig. 3). Palliative surgical treatment led to short-term improved quality of life with no surgical complications or mortality. Only 1 of 23 patients who underwent palliative surgery survived for 18 months. Distant metastases developed in 1 of 196 patients (0.5%). This patient had inguinal metastases and distant metastatic lesions were found in the lung. DISCUSSION The presence and the extent of metastasis to the inguinal region were the most powerful prognostic factors for survival in patients with SCCP. The rate of metastatic lymph node involvement in patients with clinical stages N0, N1 and N2 were 24%, 70% and 42%, respectively. Patients with negative lymphadenectomy findings had a better 10-year survival rate than those with positive lymphadenectomy results. Stage stratification showed a higher survival rate in patients in groups 1 and 2 than in those in groups 3 and 4 (p and vs p and , respectively). 1 In the current study when we evaluated lymph node metastasis stage stratification was statistically significant on univariate analysis (p ), although not on multivariate analysis. Most tumors of the penis are of lower grade. 5,6 In our series most penile tumors were diagnosed as well and moderately differentiated squamous cell carcinoma. Patients with well differentiated carcinoma have a better survival rate at 10 years than those with moderately and poorly differentiated carcinoma (p and 0.006, respectively). 1 Several studies have also emphasized the association between high grade disease and regional nodal metastasis. 2,7 9 Solsona et al reported that the risk of lymph node metastasis in patients with grade 1 differentiation was 15% vs 67% in those with grade 2 and 75% in those with grade In this study because of the small number of patients with tumors classified as grade 3 (only 7), they were grouped together with patients classified as having grade 2 disease. Histolog- FIG. 3. Five-year disease-free survival according to koilocytosis presence and absence in 115 and 78 patients, respectively (Kaplan-Meier log rank test p 0.001).

5 1358 PROGNOSTIC FACTORS IN INVASIVE SQUAMOUS CELL CARCINOMA OF PENIS ical grade was a significant variable for the risk of metastasis on univariate analysis (p ), although not on multivariate analysis. In a previous study stratification showed a better survival rate in patients in the low risk groups (p and , respectively). 1 Differences between patients with intermediate and high risk tumors were also significant (p ). Rempelakos et al found an 80% 10-year survival rate for stage T1-3N0, 42% for stage T1-3N1-2 and 0% for stage T1-3N3. 11 Hungerhuber et al found a correlation between risk stratification and nodal involvement, although no patient with a stage T4 tumor underwent lymph node dissection. 12 Theodorescu et al found that 2 of 8 patients (25%) with a pt1g1 primary tumor subsequently had inguinal nodal metastasis. 2 When we compared patients in the low and intermediate groups with patients in the high risk group, we only found a trend toward inguinal metastasis in the latter group (p 0.05). The results of measuring tumor depth of invasion to determine the prognosis of penile cancer are contradictory. From the pathological point of view invasion depth and tumor stage are not the same because tumors of the same stage can show different depths of invasion. Slaton et al reported no difference in the lymph node metastatic rate between patients with penile cancer with a tumor invasion depth of less than 3 vs 3 mm or greater. 13 Hall et al reported that tumors with a depth of more than 5 mm showed a greater incidence of lymph node metastasis. 14 Lopes et al observed that tumors that were 5 mm thick carried a greater risk of lymph node metastasis, although only on univariate analysis (p 0.02). 15 In our series a greater than 5 mm depth of invasion was only statistically significant for the risk of metastasis on univariate analysis (p ). Vascular invasion by cancerous cells was a significant prognostic factor in the studies by Slaton 13 and Lopes 15 et al. In the large multi-institutional study by Ficarra et al venous and lymphatic embolizations were independent predictors of lymph node metastasis on multivariate analysis. 16 In our study lymphovascular embolization correlated with lymph node metastasis and proved to be an independent predictive variable of lymph node involvement on multivariate analysis (p and 0.001, respectively). Patients without lymphovascular embolization had a better survival rate at 5 years than those with lymphovascular embolization. We agree with Ficarra et al concerning the relevant roles of venous and/or lymphatic embolization for predicting pathological lymph node involvement in patients with penile carcinoma. Venous and/or lymphatic embolization should be considered an important parameter for determining which patients with clinically negative lymph nodes should undergo immediate lymphadenectomy. Vascular invasion is a more important predictor of metastasis than perineural invasion. 17 In our study perineural invasion was only statistically significant for the risk of metastasis on univariate analysis (p ). Deeply invasive tumors in the corpus cavernosum, or in the glans or skin of the foreskin have been associated with a high risk of metastasis, while superficially invasive tumors in the lamina propria, corpus spongiosum or dartos tend not to be associated with lymph node involvement. 17 Ficarra et al also noted that corpora cavernosa infiltration, corpus spongiosum infiltration and urethra infiltration were associated with regional node metastasis on univariate analyses. 16 However, in our patients only infiltration of the corpus cavernosum was associated with lymph node metastasis on univariate analysis (p ). The association of koilocytosis and HPV varies from 30% to 60% in the literature. 18,19 Virus DNA has been associated with high grade penile tumors showing aggressive growth and theoretically a worse prognosis. 20 However, in a series of 72 patients with invasive SCCP treated at our institution we found no statistical correlation between HPV status and histopathological subtypes (p 0.51), while patients who were HPV positive had better 5-year disease-free survival than those with negative HPV results, although differences between the 2 groups were not significant (p 0.779). 21 The prognostic factor of koilocytosis in penile carcinoma was first studied by de Paula et al. 22 Koilocytosis was found in 91 of the 144 patients studied (63.1%). Overall 35 of 49 patients (71.4%) with histological grade I, 55 of 85 (64.7%) with grade II and 1 of 10 (10%) with grade III had koilocytosis. The prognostic usefulness of koilocytosis could not be confirmed at any level. In our study koilocytosis was found in 80 of 196 patients (40.8%). Overall 35 of 55 grade I (63.6%), 44 of 134 grade II (32.8%) and 1 of 7 (14.3%) grade III lesions were associated with koilocytosis. However, we noted a correlation between koilocytosis and histological grade (GI vs GII-III, p ). It was possible to correlate absent koilocytosis with regional metastasis (p ). In addition, multivariate analysis (linear regression) confirmed the prognostic importance of absent koilocytosis and patients with koilocytosis had better survival rates at 5 years than those without koilocytosis (p and 0.001, respectively). CONCLUSIONS TNM stage (p ), histological grade (p ), invasion depth (p ), lymphovascular embolization (p ), perineural infiltration (p ), corpora cavernosa infiltration (p ) and absent koilocytosis (p ) were significantly associated with regional metastasis on univariate analysis. However, only lymphovascular embolization and absent koilocytosis were independent prognostic factors for the risk of lymphatic metastasis (p and 0.009, respectively). Patients without lymphovascular embolization and with koilocytosis also had better 5-year survival. Abbreviations and Acronyms SCCP squamous cell carcinoma of penis REFERENCES 1. Ornellas AA, Chin EWK, Nóbrega BLB, Wisnescky A, Koifman N and Quirino R: Surgical treatment of invasive squamous cell carcinoma of the penis: Brazilian National Cancer Institute long-term experience. J Surg Oncol 2008; 97: Theodorescu D, Russo P, Zhang ZF, Morash C and Fair WR: Outcomes of initial surveillance of invasive squamous cell carcinoma of the penis and negative nodes. J Urol 1996; 155: Soria JC, Fizazi K, Piron D, Kramar A, Gerbaulet A, Haie- Meder C et al: Squamous cell carcinoma of the penis: multivariate analysis of prognostic factors and natural history

6 PROGNOSTIC FACTORS IN INVASIVE SQUAMOUS CELL CARCINOMA OF PENIS 1359 in a monocentric study with a conservative policy. Ann Oncol 1997; 8: Lopes A, Bezerra AL, Pinto CAL, Serrano SV, DeMello CA and Villa LL: p53 as a new prognostic factor for lymph node metastasis in penile carcinoma: analysis of 82 patients treated with amputation and bilateral lymphadenectomy. J Urol 2002; 168: Maiche AG, Pyrhönen S and Karkinen M: Histological grading of squamous cell carcinoma of the penis: a new scoring system. Br J Urol 1991; 67: Staubitz WJ, Melbourne HL and Oberkircher OJ: Carcinoma of the penis. Cancer 1955; 8: Ravi R: Correlation between the extent of nodal involvement and survival following groin dissection for carcinoma of the penis. Br J Urol 1993; 72: McDougal WS: Carcinoma of the penis: Improved survival by early regional lymphadenectomy based on the histological grade and depth of invasion of the primary lesion. J Urol 1995; 154: Heyns CF, Van Vollenhoven P, Steenkamp JW, Allen FJ and van Velden DJ: Carcinoma of the penis appraisal of a modified tumor-staging system. Br J Urol 1997; 80: Solsona E, Iborra I, Rubio J, Casanova JL, Ricos JV and Calabuig C: Prospective validation of the association of local tumor stage and grades as a predictive factor for occult lymph node micrometastasis in patients with penile carcinoma and clinically negative inguinal nodes. J Urol 2001; 165: Rempelakos A, Bastas E, Lymperakis CH and Thanos A: Carcinoma of the penis: experience from 360 cases. J BUON 2004; 9: Hungerhuber E, Schlenker B, Karl A, Frimberger D, Rothenberger KH, Stief CG et al: Risk stratification in penile carcinoma: 25-year experience with surgical inguinal lymph node staging. Urology 2006; 68: Slaton JW, Morgenstern N, Levy DA, Santos MW Jr, Tamboli P, Ro JY et al: Tumor stage, vascular invasion and the percentage of poorly differentiated cancer independent prognosticators for inguinal lymph node metastasis in penile squamous cancer. J Urol 2001; 165: Hall MC, Sanders JS, Vuitch F, Ramirez E and Pettaway CA: Deoxyribonucleic acid flow cytometry and traditional pathologic variables in invasive penile carcinoma assessment of prognostic significance. Urology 1998; 52: Lopes A, Hidalgo GS, Kowalski LP, Torloni H, Rossi BM and Fonseca FP: Prognostic factors in carcinoma of penis multivariate analysis of 145 patients treated with amputation and lymphadenectomy. J Urol 1996; 156: Ficarra V, Zattoni F, Cunico SC, Galetti TP, Luciani L, Fandella A et al: Lymphatic and vascular embolizations are independent predictive variables of inguinal lymph node involvement in patients with squamous cell carcinoma of the penis. Gruppo Uro-Oncologico del Nord Est (Northeast Uro-Oncological Group) Penile Cancer Project. Cancer 2005; 103: Cubilla Antonio L, Velazquez Elsa F, Ayala Gustavo E, Chaux Alcides, Torres J and Reuter V: Identification of prognostic pathologic parameters in squamous cell carcinoma of the penis: significance and difficulties. Pathol Case Rev 2005; 10: Hørding U, Rygaard C, Ruge S, Felding C, Lundvall F and Junge J: Cervical koilocytosis and high risk HPV types: the benefit of laser vaporization. Eur J Obstet Gynecol Reprod Biol 1993; 51: Hippeläinen MI, Syrjänen S, Hippeläinen MJ, Saarikoski S and Syrjänen K: Diagnosis of genital human papillomavirus (HPV) lesions in the male: correlation of peniscopy, histology and in situ hybridization. Genitourin Med 1993; 69: Gregoire L, Cubilla AL, Reuter VE, Haas GP and Lancaster WD: Preferential association of human papillomavirus with high-grade histologic variants of penile-invasive squamous cell carcinoma. J Natl Cancer Inst 1995; 87: Scheiner MAM, Campos MM, Ornellas AA, Chin EWK, Ornellas MH and Andrada-Serpa MJ: Human papillomavirus and penile cancers in Rio de Janeiro, Brazil: HPV typing and clinical features. Unpublished data. 22. de Paula AA, Netto JC, Freitas R Jr, de Paula LP, Mota ED and Alencar RC: Penile carcinoma: the role of koilocytosis in groin metastasis and the association with disease specific survival. J Urol 2007; 177: 1339.