Original article. Key words: brachytherapy, inguinal metastases, penile carcinoma, prognostic factors, squamous cell, survival.

Transcription

1 Annals of Oncology 8: , 997. O 997 Kluwer Academic Publishers. Printed in the Netherlands. Original article Squamous cell carcinoma of the penis: Multivariate analysis of prognostic factors and natural history in a monocentric study with a conservative policy* J.-C. Soria, K. Fizazi, D. Piron, A. Kramar, 3 A. Gerbaulet, 4 C Haie-Meder, 4 J.-L. Perrin, 5 B. Court, 5 P. Wibault 4 & C. Theodore Department of Medicine, Department of Pathology, 3 Department of Medical Statistics, "Department of Radiotherapy, 3 Department of Urology, Institut Gustave-Roussy, Villejuif France Summary Background: Penile carcinoma is uncommon in Western countries. Here we report on a large series of patients with squamous cell carcinoma (SCC) of the penis, describing prognostic factors, survival and therapeutic results. Patients and methods: From 973 to 993, 0 patients with invasive SCC of the penis were treated at the Institut Gustave- Roussy. Precancerous lesions and conditions associated with penis cancer were analyzed retrospectively. Survival curves were estimated by the Kaplan-Meier method, and groups were compared for outcome by the log rank test for univariate comparisons and by Cox's proportional hazards model for multivariate analysis. Results: The median age at onset was 58 years. Sixty-nine patients presented with Jackson's stage I disease, 7 with stage II and 5 with stage III. The interval between the manifestation of symptoms and the diagnosis was more than a year in 3.7% of cases. Precancerous lesions were found in 7 (6.6%) patients, and a history of phimosis was noted in 5 (4.5%). In situ and invasive carcinoma were observed together in 7 (6.6%) cases and dysplasia was associated with invasive carcinoma in eight (7.8%) further cases. Conservative treatment was administered whenever feasible. Interstitial brachytherapy was performed alone or associated with limited surgery (local excision or circumcision) in 7 (70.6%) patients. Of the 8 patients with a local relapse, nine have died of their neoplasms (3%) compared to of 8 patients with lymph node relapse (75%). The median follow-up was months. Disease-free survival, disease-specific survival and overall survival were, respectively, 56%, 7% and 63% atfiveyears and 4%, 66% and 50% at 0 years. Age (P = 0.0), the N status (P < ) or palpable nodes (P < ), corpus involvement (P ) and a verrucous histology (P = 0.038) had significant prognostic relevance for survival in the univariate analysis whereas the performance status, T status and Broders' grade did not. In the multivariate analysis only two parameters, involvement of the corpus (P < 0.000) and palpable nodes (P = 0.009), were singled out as being independent variables influencing survival. A subgroup of nine patients with verrucous histologies were distinguished by their freedom from node involvement. These patients had an excellent prognosis: all are alive and disease-free. Penile integrity was preserved during follow-up in 54 patients (5.9%), 3 of whom are still alive. Of 7 patients treated by a conservative approach including brachytherapy, long-term penile integrity was maintained in 49 (68%). Conclusion: Corpus involvement and clinically palpable nodes are highly statistically significant independent factors influencing overall survival. Node relapses remain a major cause of death. Thus, better management of lymph nodes is essential for improving survival even when conservative therapy is used to treat the primary. Key words: brachytherapy, inguinal metastases, penile carcinoma, prognostic factors, squamous cell, survival Introduction Penile carcinoma is relatively uncommon in most Western countries []. In France, its incidence is approximately one per 00,000 males per year []. The precise etiology of penis cancer remains obscure; however, an association between the disease and the absence of circumcision or poor hygiene is well established [3]. Neonatal circumcision virtually guarantees immunity from penile carcinoma [4] and the incidence of penis cancer is particularly notable in all underdeveloped regions []. Smegma has been implicated in the carcinogenesis of penile cancer, but its specific carcinogen remains unidentified [3]. The decreased incidence of penile cancer in the United States and Western countries is attributed to increased neonatal circumcision and improved socio-economic conditions with better personal hygiene. However, the last parameter seems sufficient to explain such an evolution according to a recent report by Frish et al. [5]. The role of tobacco has been emphasized recently [6,7] and smoking seems to be an independent risk factor. There is evidence that penis * This work was presented in part at the st ESMO Congress, Vienna, Austria, November -5,996. Downloaded from on 9 November 07

2 090 cancer is associated with the human papilloma virus (HPV) and especially with types 6 and 8 [8-]. Nevertheless, some epidemiological studies have failed to confirm that penile cancer is a sexually transmitted disease [6]. True premalignant lesions, with variable potential for malignancy, are now recognized: leukoplakia, erythroplasia of Queyrat, Bowen's disease, balanitis xerotica obliterans and Buschke-Lowenstein tumor []. Adequate surgical excision of penile cancer provides effective local control and remains the cornerstone of treatment of primary carcinoma of the penis in the United States [3]. However, surgery of the penis means partial or total amputation with subsequent functional and psychosexual morbidity, which is obviously difficult for the patients to accept [4]. Radiation therapy has therefore been used in several institutions to preserve penile function. Since the 960s, with the perfection of afterloading techniques using iridium 9 wires, interstitial implants have become the most frequently indicated treatment in France for primary penile lesions [5, 6]. The precise indications for such a technique are still being discussed in Europe, while this non-surgical therapy is far from popular in the United States. Thus, since 964, the Institut Gustave-Roussy (IGR) has adopted, when possible, a conservative approach (brachytherapy) for the management of penile carcinoma. Here we present an analysis of prognostic factors, survival and therapeutic results of 0 patients treated at the IGR over a 0-year period. Patients and methods The records of all patients with penile neoplasms registered in a computer database at the Institute Gustave-Roussy (IGR) between 973 and 993 were reviewed. Of 5 patients treated for penile cancer during this period at our Institute, 0 had histologically-proven invasive squamous cell carcinoma of the penis. Histologic slides were reviewed by one of the authors (D.P.) when available (9 cases). We searched retrospectively for precancerous lesions or conditions associated with cancer of the penis. In the absence of a complete questionnaire on sexual behaviour in the medical records we classified patients according to their marital status and number of offspring. Most patients were Western Europeans (64% French and 0% Italian). Initially, all patients had at least a careful physical examination by a urologist to determine the extent of local invasion and the status of the inguinal lymph nodes, as well as a chest X-ray and a standard biochemical profile. Disease was staged in all patients according to Jackson's classification [7] and the 987 TNM classification [8]. Response was evaluated according to WHO criteria [9]. Since 964, a conservative approach has been developed at our Institute using interstitial brachytherapy to treat infiltrating or noninfiltrating tumors of less than 4 cm in diameter with minor (^ cm) or no invasion of the corpora cavernosa. Interstitial brachytherapy is performed by transfixing the gjans penis with a series of iridium needles, which allow the delivery of a homogeneous dose of 65 to 70 Gy in five to seven days [6]. This technique requires a generous circumcision before treatment Brachytherapy was used alone and also following a limited incomplete tumor resection (e.g., local excision or therapeutic circumcision). Tumors unsuitable for brachytherapy (diameter greater than 4 cm) were treated by partial or total amputation of the penis. In instances of clinically palpable inguinal nodes, groin dissection is performed. Clinically-negative inguinal nodes are carefully followed and groin dissection is considered only if lymphadenopathy occurs. Cisplatin-based chemotherapy is used in patients with distant metastasis or advanced inguinal nodes, as previously reported [0]. Survival curves were estimated by the Kaplan-Meier method, and groups were compared for outcome by log rank tests for univariate comparisons and by the Cox proportional hazards model for multivariate analysis. All survival times were calculated from the date of the first treatment. Overall survival took all deaths into account (disease-related or not). Failure was defined as the occurrence of a relapse or death regardless of the cause (event-free survival). Disease-free survival was calculated from the date of first treatment to relapse or death from cancer-related causes. Results Riskfactors Previous circumcision, noted in 9.8% of patients, had been performed during adulthood in all but one, in whom it was performed when he was seven years old. A history of venereal disease was observed in only 6.9% of our patients, but no data were recorded in half the records on this item. Eighty-seven percent of our patients were married and had at least one child. No patient was known to be homosexual and none had a positive Human Immunodeficiency Virus (HIV) test result. However, HIV detection was not possible before 984, and after this date was not routine before therapy. Thus, only one patient had an HIV serology during hospitalization. The use of tobacco was noted in only 7 medical records, while tobacco exposure was documented for 6 cases. Four patients had other neoplasms before referral for penis cancer and two of them had received pelvic irradiation delivering about 5 Gy to the penis. A history of phimosis was noted in 5 (4.5%) of our cases. Lesions generally considered as precancerous were found in 7 (6.6%) patients, and are listed in Table. Patient characteristics Age at presentation ranged from 5 to 89 years with a median of 58 years. Most patients had a good performance status according to WHO criteria. Clinical assessment of lymph nodes according to the TNM classification was performed at initial presentation prior to any course of antibiotherapy. Sixty-nine (67.6%) patients presented with Jackson's stage I disease (tumor confined to glans or prepuce), 7 (6.6%) with stage II (tumor limited to Table. Precancerous lesions in 0 patients with invasive penile squamous cell carcinoma (SCQ. Pathological entities Number of patients (%) Bowen's disease Balanitis xerotica obliterans Leukoplakia Erythroplasia of Queyrat Buschke-Lowenstein tumor 3 (.9) (.9) (.9) Downloaded from on 9 November 07

3 09 Table. Characteristics of 0 patients with invasive penile SCC. Performance status (WHO) 0 Not available TNM classification T 3 X N 0 3 X M 0 Jackson's classification I II III Not available Number of patients (%) 3 (.5) 67 (65.6) 7 (6.8) 68 (66.6) 5(4.5) 6(5.8) 3 (.9) 77(75.5) (.7) 8 (7.8) () 4(3.9) 0(00) 0-69 (67.9) 7(6.6) 5(4.7) () shaft or corpora), and 5 (4.7%) with stage III lesions (tumor confined to the penis but with operable nodal metastases). None had stage IV disease (distant metastasis) at the time of diagnosis (Table ). The most common presenting clinical manifestations were a mass (65%) or an ulcer (39%) (Table 3). The lesion affected the glans in 88 (86.%) patients and reached the corpora in 8 (7.6%) cases. However, the lesion was strictly limited to the glans in 3 (3.3%) cases and to the prepuce infive (4.9%) patients. Imaging evaluation Imaging of the lymph nodes was not performed systematically but only when considered necessary. Just 5 patients underwent computed tomography (CT) scan, because this technique was only available at our Institute from the beginning of the 980s. CT scan showed lymph node enlargment in six cases. Neither liver involvement nor lung metastases were present. Lymphangiography was performed in 3 cases and was positive in Table 3. Presenting symptoms of 0 patients with invasive penile SCC. Symptoms Number of patients (%) Mass or lump Ulcer Phimosis Reddish area Pain Lymphadenopathy Bleeding Dysuria Others 66 (64.7) 40 (39.) 34(33.3) 3(3.3) (.7) (.7) 7 (6.8) (0.7) seven. Ultrasound imaging completed CT scan data in five patients. Diagnosis The duration of symptoms before diagnosis was difficult to assess. At least 4 (3.7%) patients had waited more than a year before seeking medical assistance and 3 (.5%) had waited more than six months. The pathological diagnosis was based on tumor biopsy in 43 cases, on the locally excised operative specimen or the removed prepuce in 40 cases, and on the amputated penile sample in the remaining 9 cases. Twenty-two patients presented with palpable nodes. In nine cases cytologic examination was performed on a fine-needle aspirate and lymph node involvement was found in only one patient. In 3 cases, a node biopsy was performed but was positive in only four. Histological data are summarized in Table 4. Verrucous carcinoma, found in nine cases, was characterized as a variant of squamous cell carcinoma of the penis. It is always an extremely well differentiated tumor so that no histological grade, according to Broders' system, was attributed to this subtype. Dysplasia was associated with invasive carcinoma in eight (7.8%) cases, while in situ carcinoma was observed in 7 (6.6%) other cases. The depth of invasion could only be determined in instances of penectomy. Treatment and early toxicity Sixty-three patients underwent tumor resection. The surgical management of the tumor consisted of partial (n = ) or total (n - 8) amputation, particularly when the tumor diameter was more than 4 cm. Limited tumor resection consisted of local excision (n = 4) or therapeutic circumcision (n = 8) and was generally incomplete. Thus, 88% of patients with limited surgery underwent brachytherapy (Table 5). Lymph node management consisted of a 'wait and see' policy in the absence of palpable nodes and when Table 4. Histological data from 0 patients with invasive penile SCC. Pathological data Number of patients (%) Histological subtypes Epidermoid Verrucous Associated lesions In situ carcinoma Dysplasia Grading by the Broders' system 3 4 Not available Level of invasion Corpus cavernosum Corpus spongiosum Urethra Lymphatic embolus 93(9.) 9 (8.9) 7(6.6) 8 (7.8) 7 (69.6) 0(9.8) () 6(5.8) /37 3/36 7/35 5/37 Downloaded from on 9 November 07

4 09 Table 5. Treatment modalities of 0 patients with invasive penile SCC. Table 6. Outcome of patients after a first relapse (n = 47). Site of first relapse Number of patients (%) Died of disease (%) Local Node Local + node Distant metastasis Distant + node metastasis 3 (.5) 6(5.6) 3 (.9) 3 (.9) (.9) 7 (30.4) (75) (33.3) 3(00) (00) Number of patients (%) Initial treatment (n = 0) Surgery alone Interstitial brachytherapy alone Surgery followed by brachytherapy Tumor radiotherapy associated with surgery or brachytherapy Tumor surgical modalities (n = 63) Total amputation Partial amputation Local excision* Therapeutic circumcision* Not available 5 (4.5) 35 (34.3) 37 (36.3) 8(7.8) (0.7) 4(3.7) 8 (7.4) (.9) initial treatment (median five months, range. to 57 months). During the entire follow-up period, ten (9.8%) patients developed distant metastasis involving the bone in six cases, the lung in two and the brain in two cases. Of these patients, five presented initially with involvement of the corpora cavernosa. All but one of the patients with distant metastasis have died of disease (median survival after the development of a metastasis is.7 months). Four patients with a node recurrence are currently disease-free with a median follow-up of 6 months ranging from 34 to 54 months after their lymph node relapse. The final outcome of patients presenting a first relapse is summarized in Table 6. With a median follow-up of months (4-40), the event-free survival at five and 0 years is, respectively, 45% and 9%. Overall survival at five and 0 years is, respectively, 63% and 50% (Figure ). The disease-specific survival is 7% at five years and 66% at 0 years. The five- and 0-year disease-free survivals are, respectively, 56% and 4% (Figure ). The causes of death are presented in Table 7. Penile carcinoma was the cause of death in only 7 patients (54%), and almost all of those were due to uncontrolled loco-regional progression or metastases. Two patients died during initial therapy, one of myocardial infarction during interstitial brachytherapy and the other of septic shock 0 days after brachytherapy. An evaluation of penile anatomic status at the last follow-up examination of the whole population of patients showed that (0.6%) had undergone total penectomy and seven of them were alive, 7 (6.6%) had * Local excision and therapeutic circumcision are frequently an incomplete form of oncologic surgery (usually followed by brachytherapy). a regular follow-up was possible. Inguinal (n - ) or inguino-iliac (n = ) lymph node dissection was performed when nodes initially were, or became, palpable. Groin dissection was bilateral in nine cases and unilateral in the remaining 5. Surgical modalities caused early toxicity in three cases, one inguinal hematoma and two infections of the perineum. Seventy-two patients underwent interstitial brachytherapy as first-line treatment in 35 cases and following limited tumor resection (therapeutic circumcision or local excision) in 37 cases. Interstitial implants allowed the delivery of a homogeneous dose of irradiation ranging from 6 to 70 Gy in four to 0 days. Five patients presented early treatment-related toxicity: one rapid glans necrosis, one urethral stenosis, one orchitis and two treatment-related deaths, to be detailed further. External radiotherapy of the penile tumor was performed in five cases during the 970s and was always associated with surgery or brachytherapy. Lymph node irradiation was performed in nine patients who presented with initial lymph node involvement. Bilateral inguinal radiotherapy was performed in all of these cases at a median dose of 50 Gy and was completed by iliac irradiation in three cases (median dose of 40 Gy). Response to treatment and outcome Response to initial treatment following brachytherap> (n = 7) was satisfactory, with 64 (88.8%) complete responses, three (4.%) partial responses, one (.3%] case of stable disease, two (.7%) of progressive disease and two (.7%) patients were not evaluable. In the whole population of patients a first relapse occurred in 47 (46%) cases, 3 of which were local relapses, 6 secondary lymph node involvement, three distant metastasis, local and concomitant lymph node involvement in three and concomitant lymph node and distant metastasis in the remaining two cases. Eighteen (%) of the 77 patients initially with node-negative disease subsequently developed lymph node metastasis. Seventy-eight percent of these relapses occurred during the first year after Downloaded from on 9 November * ; o ^o - Figure. Overall () and event-free survivals () of 0 patients with squamous cell carcinoma of the penis.

5 093 O.BO O _ Q. Table 8. Median duration of overall and disease-free survival: influence of different parameters. i\ Vanable vv L_ ^ ^ No. of pts Median overall survival (months) P < Median disease free survival (months) ' L f 0 40 " C.0 " Figure. Disease-specific () and disease-free survivals () of 0 patients with squamous cell carcinoma of the penis. Agt (years) < >75 Performance Status (WHO) 0 had partial penectomy and 4 of them alive, and 54 (5.9%) patients had maintained their penile integrity and 3 of them alive. Compared to first-line treatment, these data show that 3 total penectomies were performed at relapse. Sixteen partial penectomies were also necessary after the first treatment, 4 for local progression, one for brachytherapy-related necrosis and one for unconfirmed suspicion of local relapse. Penile function could not be assessed in this retrospective study. Among the 35 patients who had brachytherapy as first-line treatment, local control and penile preservation were achieved in 5 (7.4%) cases. Of the 37 patients who underwent limited tumor surgery followed by brachytherapy, local control was obtained in 9 (78.3%) cases and the penis was conserved in 4 (65%) cases. The actuarial rate of penile preservation following conservative management was 88% at six months, 85% at one year, 8% at two years, 75% at four years, 7% at six and eight years, and 68% at 0 years. Prognosis factors TNM classification T 3 N 0 3 Palpable node No Yes Jackson's classification II III Pain at presentation No Yes Involvement of the corpus No Yes Histological type Epidermoid Verrucous Broders grade i The results of the univariate analysis are summarized in Table 8. Firstly, the simple clinical parameters usually available, such as age at onset and performance status, were tested. Subsequently, variables related to extent of disease were selected (the TNM status, the presence of palpable nodes, stages according to Jackson's classification, involvement of the corpus and pain at presentation as it may reflect locally advanced involvement). Although we are aware that some of these parameters may reflect the same clinical aspects, we chose to test all Table 7. Causes of death (n = 50). Cause Number of patients (%) Carcinoma of the penis Other disease (unrelated) Second tumor Toxicity Suicide Unknown 7(54) 3(6) 4(8) (4) () 3(6) 4 < < of them so that those carrying the highest statistical significance could be identified. Finally, we tested histological differentiation (according to the Broders' system) which is usually reported as well as the recently characterized verrucous form, a variant of squamous cell carcinoma of the penis. Several factors were found to be of prognostic significance for overall survival (OS) and disease-free survival (DFS): the volume or size of the tumor (according to Jackson's classification and the tumor of the TNM classification for OS), the nodal status, involvement of the corpora cavernosa and the histological subtype. Age at presentation reached significance for DFS but not for OS. The performance status as well as pain at presentation had prognostic significance for OS but not for DFS, and histological differentiation was not predictive of OS or DFS. Downloaded from on 9 November 07

6 094 Table 9. Multivariate Cox proportional hazards analysis. Variable Coefficient P value Involvement of the corpus Palpable node Parameters reaching significance in the univariate analysis were taken into account in a multivariate analysis. However, the verrucous form was not included in the model because no event occurred among these patients. Only involvement of the corpus and presence of palpable nodes were highly statistically significant independent prognostic factors for overall survival (Table 9). Jackson's stage III (tumor confined to the penis with operable inguinal nodal metastases) had the same statistical value as the clinical node status. Discussion Since penis cancer is rare in Europe and North America, it is necessary to collect several decades' worth of data to obtain a large series of such patients. Most of the cumulative experience is reported in retrospective series with data collected for periods ranging from nine to 39 years during which therapeutic procedures and attitudes have changed [, 4]. Consequently, analysis of the results of treatment for cancer of the penis is difficult and despite many large published series, several areas of controversy exist about the best way to manage this disease (i.e., lymph node staging and management, nonsurgical therapy for a limited primary lesion). With respect to conditions associated with cancer of the penis, our retrospective analysis could not accurately evaluate tobacco exposure. Nevertheless, smoking is a recognized independent risk factor for penile carcinoma, probably through the secretion of tobacco metabolites by the preputial glands and their concentration in smegma [6, 7]. About 0% of the patients were circumcised during adulthood. A long history of exposure to smegma may account for the development of penile carcinoma in patients circumcised at a late stage, confirming the negligible protective effect of circumcision during adulthood []. A history of phimosis is observed in 5 (4.5%) of our cases. Other authors have reported comparable rates, with the highest being 69% [4,, 3]. Chronic irritation beneath a phimotic foreskin is supposed to explain the development of penile carcinoma. The frequently recognized premalignant lesions were found in 7 (6.6%) patients, which is consistent with results reported by other authors [, 4, 4], Four patients had another neoplasm prior to the penis cancer: two, colorectal cancer, one, bladder cancer and one, cancer of the oral cavity. To our knowledge there are no data in the literature on neoplasms preceding penis cancer and this neoplasm is not integrated among familial neoplasms. However, since most patients are older than 50, it is not surprising that they develop other cancers prior to their penile carcinoma. Two patients had had pelvic irradiation and had received about 5 Gy to the penis. This very low dose might explain the secondary development of penis cancer. However, this point requires further studies for definitive conclusions. The incidence of previous venereal disease evaluated at 6.9% in our study must have been underestimated since half the records in our series gave no information about this item. It is remarkable that 87% of our patients present the same sexual profile, namely, married with at least one child. This is consistent with no evidence of homosexuals in our series. Unfortunately, there are no data on the patients' wives and particulary concerning a history of uterine cervix cancer, since a link has been noted with penis cancer [9, 0]. Some recent studies show that nearly 55% of invasive SCC of the penis present HPV DNA sequences [8]. Patient characteristics concerning age at onset, presenting clinical symptoms, the anatomic location of initial lesion, a delayed diagnosis and the stage of disease are in agreement with most large series. Although penile lesions were noted in men as young as 5 years, most patients were older than 50, with a median age of 58 years, which is consistent with results reported by many authors [3,, 5], but slightly younger than the 68 years reported by Narayana et al. [4]. In our series, the most common presenting symptoms were a mass and an ulcer as noted by other authors [4, 4]. Pain at presentation is infrequent (.7% in our series), and the patient does not feel the need to seek medical attention. Also, node enlargement is only occasionally an initial complaint (.7% in our series, 3.5% for Hanash [4] and.7% for Narayana [4]). Penile cancer originates on the glans, with the next most common sites being the prepuce and the corpus. Our series confirms a long delay before the diagnosis with 3.7% of patients having symptoms which lasted more than a year prior to the initiation of definitive therapy. Such proscrastination stems from ignorance, fear, embarassment, neglect and guilt [4], which can be ameliorated somewhat by educating the public about personal hygiene, including the early treatment of phimosis, and about how to recognize early signs of penis cancer. The stage of the disease in our patients is consistent with results reported by other authors [4,, 3, 5] with a majority of Jackson stage I disease followed by comparable percentages of Jackson stages II and III. Despite the need for a very precise assessment of the extent of the tumor to aid in choosing a therapeutic strategy, imaging of penis cancer has received scant attention [6]. The sensitivity and specificity of most techniques remain unestablished [3]. The IGR's attitude reflects the limited use of imaging in evaluting this disease. In this study, imaging of the primary lesion was performed in only two cases. For determining lymph node involvement, CTscan, ultrasound or lymphangiography were only performed when considered necessary and not systematically. However, it is clear that many of the latest advances in imaging techniques can be success- Downloaded from on 9 November 07

7 095 fully applied to penile carcinoma, and can contribute to optimal staging of these malignancies before treatment decision-making [6]. Currently, patients have a CTscan of the pelvis and abdomen to assess tumor extent prior to therapy. Magnetic resonance imaging (MRI) of the penis may be indicated for better evaluation of the rumor depth to help in determining whether or not conservative treatment is justified. As the therapeutic approach developed in our Institute was conservative, many of the pathological diagnoses were based exclusively on the tumor biopsy specimen. Thus, part of the histological data, and particularly the level of invasion, was only available for patients who underwent penectomy (about 35% of our cases). The specific patterns of verracous carcinoma of the penis have recently been defined [7] and nine (8.9%) of our patients had the verrucous form. All of them presented with signs indicating a favorable outcome, as detailed further. In situ carcinoma and dysplasia were associated with invasive squamous cell carcinoma of the penis in 7 (6.6%) and eight (7.8%) cases, respectively. It is well known that carcinoma of the uterine cervix is almost always associated with a long history of severe dysplasia, chronic cervicitis and carcinoma in situ (CIS). The histogenesis of penile cancer remains unestablished and whether penis carcinoma is preceded by a long period of dysplasia and CIS has not been proven. Time to progression from CIS to invasive carcinoma has not been clearly evaluated in penile cancer. Nevertheless, it is well known that the mean age of patients with cancer of the penis is about 0 years older than that of men with intraepithelial penile neoplasia [8]. The number of patients with poorly or undifferentiated SCC of the penis in our series is 6 (5.9%), lower than that reported by other authors: 8% for Hardner [9], 0% for Fraley [], but comparable to the 4.5% reported by Ornellas []. With respect to the staging of lymph node disease at diagnosis, our results confirmed that fine-needle aspiration of suspect nodes is only reliable when positive (30), and that it has a very low sensitivity. Unilateral biopsy of a clinically positive groin is also an unsatisfactory procedure because only 30.7% of cases were positive. Broad spectrum antibiotics for three to six weeks were not administered in routine practice at IGR during this period, but they appear to be clearly necessary for better staging of lymph nodes. Some authors recommend surgical procedures for staging lymph node disease. In 977 Cabanas proposed bilateral biopsy of the sentinel lymph nodes [3]. However, some studies have indicated that biopsies of sentinel lymph nodes are not reliable indicators of metastasis, mainly because of false negative results [3]. In 988 Catalona proposed a modification of standard lymphadenectomy for staging of patients with clinically negative nodes, but two recent reports regarding the reliability of such a technique for clinical staging of penile carcinoma are contradictory [33, 34]. The treatment modalities reported in our series are linked to the conservative approach adopted by the IGR for the management of penile carcinoma. Whenever possible, conservative treatment was chosen at the IGR, since it limits the extent of the psychological distress associated with amputation [6]. Brachytherapy was practically the only conservative technique used at IGR. Nevertheless, there are many other organ-sparing therapies. External beam radiotherapy is a procedure frequently used in Europe that produces a 75%-80% local success rate [35]. Radiotherapy, like brachytherapy, is not advised for curative treatment of lesions exceeding 4 cm in diameter. To our knowledge, no prospective randomized study comparing brachytherapy and external beam radiotherapy has been reported. Laser treatment can be used to control localized SCC of the penis with good cosmetic and probably functional results. Windhal et al. have reported their experience with over 9 SCC of the penis (4 Tis, 7 Tl, 8 T, all NO) treated with the YAG laser: all patients were alive after a mean follow-up of 3 months, with only two recurrences [36]. Conservative surgery for primary penile carcinoma is represented by cryosurgery and Mohs micrographic chemosurgery. Cryosurgery has been reported to be curative for early-stage disease with minimal infiltration [35]. Mohs et al. have reported their experience with over 3 patients treated between 936 and 986 ( Jackson's stage I, eight stage II, two stage III). The fiveyear cure rate was 74% (86% for patients with Jackson's stage I and 6% for those with stage II) [37]. During the 970s and at the beginning of the 980s the standard bilateral lymphadenopathy techniques were associated with a 3%-0% mortality rate and a 0%-40% morbidity rate [38]. Many authors admitted that radical node dissection in patients with early-stage primaries and nonpalpable nodes was not justifiable since over 80% of patients would be overtreated [35]. Narayana considers that prophylactic inguinal node dissection for carcinoma of the penis is not indicated [4]. The IGR has therefore adopted a 'wait and see' policy for the management of non-palpable lymph nodes. At the beginning of the 970s a few patients in our series had irradiation of the primary since at that time external radiotherapy had not been totally replaced by brachytherapy. Lymph node irradiation was performed in most cases as an adjunct to lymphadenectomy. In our series, post-operative iliac radiotherapy was ineffective in patients with positive iliac nodes, as none of those patients are alive at five years. The five-year overall survival obtained in our series was 63%. It is consistent with the 59% reported by Skinner [38], and compares favorably with the 49% noted by other authors [5, 9, 39]. The disease-specific survival at five years was 7%, comparable to data reported by others [38, 40]. Overall and corrected overall survival at 0 years are not usually reported in penis cancer series. As our median follow-up was longer than nine years ( months) we were able to define a 0-year 05 of 50% and a 0-year disease-specific survival of 66%. The event-free survival at five years is 45%, and directly related to the high incidence of relapses. Local relapses occurred in 6 cases as the first site of recur- Downloaded from on 9 November 07

8 096 rence. Such a result is linked to the wide use of brachytherapy in our series, since the major pitfall of interstitial implants is local failure. Such failures can be salvaged by secondary penectomy. Of the 7 patients who underwent brachytherapy alone or associated with limited tumor surgery, local recurrence occurred in 8 (5%). These results are comparable to the % reported by Mazeron [] and slightly higher than the 6% noted by Rozan [5]. Node relapses occurred in patients as the first site of recurrence and 8 patients developed lymph node metastases during follow-up. Only four patients (4%) with node involvement are alive and free of disease. The median survival after the appearance of a lymph node metastasis is months. It is noteworthy that 8 (%) of the 77 patients classified initially as NO developed lymph node metastases with a median time to relapse of five months (range. to 57 months). Only three (6.6%) of these 8 patients are still alive. The 'wait and see' policy adopted for the management of non-palpable lymph nodes thus appears questionable. Since the median time to relapse is quite short, we suspect that there was in fact lymph node involvement at diagnosis and that prophylactic lymphadenectomy would have been beneficial for some of the patients. It is nonetheless true that prophylactic groin dissection would have comprised overtreatment for 78% of the NO patients. In a recent study, Ornellas et al. support the notion that the earlier groin dissection is performed in patients with clinically undetected but pathologically positive nodes, the better the prognosis []. Indeed, thefive-yeardisease-free survival rate was 6% for patients who underwent systematic lymphadenectomy, whereas for those who underwent delayed lymphadenectomy the percentage fell to 8%. The real problem is determining who is likely to have positive groin nodes when these are not clinically palpable. McDougal advocates proposing prophylactic lymphadenectomy to patients with invasion of the corpora cavernosa or poorly differentiated squamous cell carcinoma since they are highly likely to have positive groin nodes [4]. This opinion is shared by others [4, 43]. Recent publications have reported an improved prognosis when early prophylactic lymphadenectomy is performed [44]. In our series patients had NO disease with involvement of the corpus at diagnosis and four (33.3%) of them developed lymph node metastases. Only six patients in our series presented with poorly- or undifferentiated SCC and one of them (7%) developed lymph node involvement. Surprisingly, one of our patients developed a lymph node metastasis 57 months after initial treatment of its primary. Thus, in our opinion, the disease-free interval after which a patient may be considered cured appears closer to five years than to the three years proposed by Fraley []. Nevertheless, this point clearly requires further studies for definitive conclusions. Distant metastasis is quite rare in the natural history of penile carcinoma: 9.8% in our experience,.3% in the report by Wajsman et al. [45]. The most commonly involved site of metastasis is the lung [3, 45], although bone was the first site of metastasis in our series. In 46% of our patients death was not related to carcinoma of the penis. Of the 7 patients who died of cancer, the majority died of loco-regional recurrence with vessel invasion and sepsis. Death unrelated to carcinoma is frequent among patients with penis cancer since the median age at onset is close to 60 years. Wajsman found that in 6% of cases death was not related to penile carcinoma [45]. Of unrelated deaths, second neoplasms have been noted in 0% [4]. Our results are consistent with such data, with four (8%) deaths related to second neoplasms. Suicide was noted by Hanash [4] in three patients after total penectomy, and one of our patients who underwent partial penectomy also committed suicide. One of the major potential advantages of brachytherapy for penile carcinoma is that it preserves erectile potency and sexual function. It was difficult to obtain accurate data for these items in the present study since the records were reviewed retrospectively. Of the patients treated by brachytherapy alone or brachytherapy associated with limited tumor surgery, anatomic preservation of the penis was achieved in 68%. This percentage probably over-estimates sexual function. However, Opjordsmoen et al. have reported that 0 of patients (70.4%) in whom irradiation was delivered for SCC of the penis have a normal or only slightly reduced sexuality, compared with two of nine patients after partial penectomy [46]. Univariate analysis confirmed the prognostic value of well established parameters such as the size of the tumor (P = 0.00), the nodal status (P < ) and involvement of the corpus (P = ). The excellent prognosis of the verrucous form is consistent with recently published data [7]. In our series, all nine of the patients with a verrucous carcinoma of the penis are currently alive and disease-free, with a median follow-up of 59 months. Seven of these patients were initially classified as NO, while two of them were considered as Nl or N. However, groin dissection in these two patients failed to reveal nodal metastasis. Only one patient had a local relapse, which was treated by partial penectomy. The verrucous subtype appears to be a different form of penile carcinoma in view of its clinical presentation, the absence of nodal or distant metastases and its favorable prognosis. The performance status had a prognostic value for OS. Such a result is probably more closely linked to the better general condition of patients with good performance status than to a specific relation with SCC of the penis. In contrast to others [, 35, 47, 48], histological differentiation failed to confer a prognostic significance on OS or DFS in our series. Other authors also failed to confirm the value of such a parameter [9, 45]. The negativity of our results may be due to the small number of patients with poorly differentiated or undifferentiated tumors. The exclusion of our patients with a verrucous histology from Broders' classification, may be an alternative explanation since these patients were previously classified in the well-differentiated group that carries a good prognosis. In the multivariate analysis, involvement of the corpus (P = 0.009) was clearly an Downloaded from on 9 November 07

9 097 independent variable influencing survival. Some authors have already highlighted the poor prognostic value of the involvement of the corpus [, 6, 4, 4]. However, two recent reports failed to show the independent value of this feature [40, 49]. Only the combination of corpus cavernosum infiltration and poor differentiation portended a poor prognosis and predicted a high probability of lymph node invasion [49]. The prognostic significance for survival of lymph node invasion has been established [4,, 4,, 4, 30, 39]. Despite the well known differences between clinical and pathological lymph node classifications, our analysis shows statistically significant differences in survival for clinical NO and N+ categories. Such a result has already been reported by another multivariate analysis [40]. It is noteworthy that Jackson stage III disease had the same statistical value as clinical node status in the multivariate analysis. In fact, since Jackson stage III disease is a tumor confined to the penis with proven regional node metastases, such a variable is not independent of the clinical nodal status. In conclusion, the conservative approach adopted at the IGR allows good survival in selected patients with invasive SCC of the penis. Multivariate analysis confirms the poor prognostic value of corpus cavernosum involvement and palpable nodes. Improving the management of lymph node disease and searching for factors predictive of nodal metastases are currently of foremost concern as they may now be the key to better survival. Acknowledgements We thank Lorna Saint-Ange for editing the manuscript and Catherine Loge for preparing it. References. Riveros M, Lebron RF. Geographical pathology of cancer of the penis. Cancer 963; 6: Mazeron JJ, Langlois D, Lobo PA et al. Interstitial radiation therapy for carcinoma of the penis using indium 9 wires: The Henri-Mondor experience ( ). Int J Radiat Oncol Biol Phys 984; 0: Burgers JK., Badalament RA, Drago JR. Penile cancer. Clinical presentation, diagnosis and staging. Urol Clin North Am 99; 9: Narayana AS, Olney LE, Loening SA et al. Carcinoma of the penis. Analysis of 9 cases. Cancer 98; 49: Frish M, Friis S, Kruger Kjaer S et al. Falling incidence of penis cancer in an uncircumcised population (Denmark ). BMJ 995; 3: Hellberg D, Valentin J, Eklund Tet al. Penile cancer: Is there an epidemiological role for smoking and sexual behaviour? BMJ 987; 95: Harish K, Ravi R. The role of tobacco in penile carcinoma. Br J Urol 995; 75: Cupp MR, Maiek RS, Goellner JR et al. The detection of human papilloma virus deoxyribonucleic acid in intra-epithelial, in situ, verrucous and invasive carcinoma of the penis. J Urol 995; 54: Martinez I. Relationship of squamous cell carcinoma of the cervix uteri to squamous cell carcinoma of the penis. Among Puerto Rican women married to men with penile carcinoma. Cancer 969; 4: Graham S, Priore R, Graham M et al. Genital cancer in wives of penile cancer patients. Cancer 979; 44: Barrasso R, De Brux J, Croissant O et al. High prevalence of papillomavirus-associated penile intra-epithelial neoplasia in sexual partners of women with cervical intra-epithelial neoplasia. N Engl J Med 987; 37: Persky L, de Kernion J. Carcinoma of the penis. CA Cancer J Clin 986; 36: Das S. Penile amputations for the management of primary carcinoma of the penis. Urol Clin North Am 99; 9: Hanash KA, Furlow WL, Utz DC et al. Carcinoma of the penis: A clinicopathologic study. J Urol 970; 04: Rozan R, AJbuisson E, Giraud B et al. Interstitial brachytherapy for penile carcinoma: A multicentric survey (59 patients). Radiother Oncol 995; 36: Gerbaulet A, Lambin P. Radiation therapy of cancer of the penis. Indications, advantages and pitfalls. Urol Clin North Am 99; 9: Jackson SM. The treatment of carcinoma of the penis. Br J Surg 966; 53: Hermanek P, Sobiri LH. TNM Classification of Malignant Tumors (UICQ, 4th ed., New-York: Springer-Verlag World Health Organization. Handbook for Reporting Results of Cancer Treatment. WHO Offset Publication, no. 48. Geneva: WHO 979; Kattan J, Culine S, Droz JP et al. Penile cancer chemotherapy: Twelve years' experience at Institut Gustave-Roussy. Urology 993; 4: Ornellas AA, Correia Seixas AL, Marota A et al. Surgical treatment of invasive squamous cell carcinoma of the penis: Retrospective analysis of 350 cases. J Urol 994; 5: Fraley EE, Zhang G, Sazama R et al. Cancer of the penis. Prognosis and treatment plans. Cancer 985; 55: Solis WA. Cancer de la verge. Notre experience. Ann Urol 985; 9:338--M. 4. Merrin CE. Cancer of the penis. Cancer 980; 45: Derrick FC, Lynch KM, Kretkowski RC et al. Epidermoid carcinoma of the penis: computer analysis of 87 cases. J Urol 973; 0: Vapnek JM, Hricak H, Carroll PR. Recent advances in imaging studies for staging of penile and urethral carcinoma. Urol Clin North Am 99; 9: Correia Seixas AL, Ornellas AA, Marota A et al. Verrucous carcinoma of the penis: Retrospective analysis of 3 cases. J Urol 994; 5: Aynaud O, Ionesco M, Barrasso R. Penile intra-epithelial neoplasia. Specific clinical features correlate with histologic and virologicfindings.cancer 994; 74: Hardner GJ, Bhanalaph T, Murphy GP et al. Carcinoma of the penis: Analysis of therapy in 00 consecutive cases. J Urol 97; 08: Wajsman Z, Gamarra M, Park JJ. Fine needle aspiration of metastatic lesions and regional lymph nodes in genito-urinary tract. Urology 98; 4: Cabanas RM. An approach for the treatment of penile carcinoma. Cancer 977; 39: Fowler JE. Sentinel lymph node biopsy for staging penile cancer. Urology 984; 3: Parra RO. Accurate staging of carcinoma of the penis in men with non palpable inguinal lymph nodes by modified inguinal lymphadenectomy. J Urol 996; 55: Lopes A, Rossi BM, Fonseca FP et al. Urtrealibility of modified inguinal lymphadenctomy for clinical staging of penile carcinoma. Cancer 996; 77: Jones WG, Fossa SD, Hamers H et al. Penis cancer: A review by the Joint Radiotherapy Committee of the European Organisation for Research and Treatment of Cancer (EORTQ Genito-Urinary and Radiotherapy Groups. J Surg Oncol 989; 40: 7-3. Downloaded from on 9 November 07

10 Windhal T, Hellsten S. Laser treatment of localized squamous cell carcinoma of the penis; J Urol 995; 54: Mohs FE, Stephen NS, Larson PO. Moris micrographic surgery for penile tumors. Urol Clin North Am 99; 9: Skinner DG, Leadbetter WF, Kelley SB. The surgical management of squamous cell carcinoma of the penis. J Urol 97; 07: Baker BH, Spratt JS, Perez-Mesa C et al. Carcinoma of the penis. J Urol 976; 6: Horenblas S, Van Tinteren H. Squamous cell carcinoma of the penis i.v. Prognostic factors of survival: Analysis of tumor, nodes and metastasis classification system. J Urol 994; 5: Me Dougal WS. Editorial: Carcinoma of the penis. J Urol 994; 5: Abi-Aad AS, de Kernion JB. Controversies in ilio-inguinal lymphadenectomy for cancer of the penis. Urol Clin North Am 99; 9: Theodorescu D, Russo P, Zhang ZF et al. Outcomes of initial surveillance of invasive squamous cell carcinoma of the penis and negative nodes. J Urol 996; 55: McDougal WS. Carcinoma of the penis: Improved survival by early regional lymphadenectomy based on histological grade and depth of invasion of the primary lesion. J Urol 995; 54: Wajsman Z, Moore RH, Merrin CE et al. Surgical treatment of penile cancer. A follow-up report. Cancer 977; 40: Opjordsmoen S, Waehre H, Aass N et al. Sexuality in patients treated for penile cancer: Patients' experience and doctors' judgement. Br J Urol 994; 73: Fraley EE, Zhang G, Manivel C et al. The role of ilio-inguinal lymphadenectomy and significance of histological differentiation in treatment of carcinoma of the penis. J Urol 989; 4: Maiche AG, Pyrhonen S, Karkinen M. Histological grading of squamous cell carcinoma of the penis: A new scoring system. Br J Urol 99; 67: Solsona E, Iborra I, Ricos JVet al. Corpus carvenosum invasion and tumor grade in the prediction of lymph node condition in penile carcinoma. Eur Urol 99; : 5-8. Received 30 May 997; accepted August 997. Correspondence to: Dr. K. Fizazi Department of Medicine Institut Gustave-Roussy 39, Rue Camille-Desmoulins VHlejuifCedex France Downloaded from on 9 November 07