Methods

The PubMed database (National Library of Medicine, Bethesda, Maryland, USA) was used to identify the articles published between 1986 and June 2008 containing combinations of the MEdical Subject Headings (MeSH) “human papillomavirus” and “penile carcinoma”, regardless of language of publication. All articles reporting data on HPV prevalence among penile carcinomas where selected and reviewed. The reference lists of identified articles were also reviewed for additional publications. A comprehensive effort to exclude overlapping cases from the final analysis was done. The methods section of each identified article was reviewed, with special interest on those whose first authors belonged to the same research group. When identifying repeated histological samples, the paper which used the most sensitive HPV-DNA detection technique was selected. In case of doubt authors were directly contacted and the article excluded if no answer was received. Publications containing less than five cases were also excluded from the analysis. A final total of 31 articles were included in this review12 ,16 ,17 ,18 ,19 ,20 ,21 ,22 ,23 ,24 ,25 ,26 ,27 ,28 ,29 ,30 ,31 ,32 ,33 ,34 ,35 ,36 ,37 ,38 ,39 ,40 ,41 ,42 ,45 (see table 1).

Identified studies were reviewed by two independent medical readers; information about year of publication, country of origin, sample size, HPV detection methods, sample preparation and histological type was retrieved. Histological groups for analysis were created following the World Health Organization (WHO) histological classification of penile cancers with the expert assessment of the pathologist (AC).

Overall HPV prevalence with the corresponding 95% CI was estimated by dividing the total number of subjects for each histological group by the number of HPV positive cases. HPV type specific contribution for the HPV-6, -11, -16, -18, -31 and -33 types was estimated as a relative proportion of each type among all HPV positive cases. We could not present separate data for HPV-6 and -11 as results are often presented as combined HPV prevalence. Estimates were weighted by the relative histology type contribution to the global percentage of penile malignancies. For this calculation it was assumed that SCC represents 95% of penile malignancies; of SCC, keratinising subtype accounts for 49%,13 ,43 basaloid 4%,16 warty 6%,6 mixed warty–basaloid 17%, verrucous 8%, papillary 7%, other SCC mixed 7%, sarcomatoid carcinomas 1%, and 1% were other non-mixed SCC.43 These results are presented as a closer indicator of the real HPV contribution to overall penile carcinoma. The statistical package STATA V.10.0 (Stata Corporation, Texas, USA) and R free software were used for calculations.

Results

All studies included in the analysis used Southern blot (one study) or PCR assays. The most used PCR primers were MY09/11, GP5+/6+ and type-specific probes for HPV-16/18. Data were analysed for differences on HPV DNA detection between PCR and Southern blot and paraffin embedded or frozen fresh tissue but revealed no significant differences (data not shown).

Data from selected studies were classified according to histological type. Overall HPV prevalence was obtained from a total of 1466 penile carcinomas: 671 were unspecified SCC (45.7%), 448 keratinising SCC (30.5%), 117 typical SCC (7.9%), 57 verrucous SCC (3.8%), 56 warty SCC (3.8%), 46 basaloid SCC (3.1%), 19 SCC with mixed warty and basaloid features (1.3%), 10 papillary SCC (0.6%), and 42 other SCC mixed forms (2.9%) (see table 2).

Few data about pre-invasive lesions were available in the published literature. The identified penile intraepithelial cases (43 PIN1, 18 PIN2 and 4 PIN3) were excluded for subsequent analyses due to the small sample size.

Data were available for Europe, North America, South America and Asia. The overall pooled HPV prevalence was 46.9% (95% CI 44.4 to 49.5), ranging from 40.7% in South America to 57.6% in North America (see table 3).

Relative HPV contribution weighted by histological distribution was 38.92%.The observed specific HPV contribution by histological type was as follows: basaloid SCC 76%; warty SCC 58.9%; non-keratinising/typical SCC 47.8%; keratinising SCC 43.5%; and verrucous SCC 24.5% (see fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Human papillomavirus (HPV) prevalence in penile carcinomas by histological type. SCC, squamous cell carcinoma.

The classification systems of histological subtypes of penile SCC were variable and rarely stated in most studies. There has been much confusion in the literature among verruciform tumours which have been designated using a variety of terms such as verrucous, Buschke–Lowenstein tumour, warty and papillary carcinomas as the same tumour entity. Warty and basaloid carcinomas also caused classification problems. Only articles providing a clear description of the histopathology of the tumour were included in the review.

HPV-16 was the most frequent type among all HPV positive histological types (61.5%). HPV-16 accounted for 75.7% of the HPV positive warty SCC, 75.3% unspecified SCC, 67.8% typical SCC, 65.7% basaloid SCC, 31.7% keratinising SCC and 35.7% verrucous SCC. HPV-18, despite being the second most common type identified, accounted for only 13.2% of the total; its prevalence was lower in all histological groups. The greatest prevalence was observed among keratinising SCC (21.0%) followed by typical SCC (14.2%) and unspecified SCC (12.0%). The lowest rates were observed among basaloid SCC (0%), warty SCC (3.0%) and verrucous SCC (7.1%).

Detection of HPV types other than HPV-16 and -18 was not consistently reported. Most of the earlier studies used HPV-16/18 specific primers; in recent years wide spectrum primers are generally being used to confirm the occasional detection of low risk types.

HPV-6 and -11 were mainly found in non-keratinising (16.0%), verrucous SCC (14.2%), and keratinising (12.3%) histological subtype, while warty (9.1%), unspecified SCC (5.1%), basaloid SCC (2.8%), papillary (0%) and other mixed SCC (0%) had a lower detection rate. Multiple infections accounted for 7.4% of the total, but showed a greater presence in the keratinising subtype (15.9%) (see fig 2 for HPV type distribution in penile carcinoma).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Human papillomavirus (HPV) type prevalence distribution in penile carcinomas.

Discussion

The systematic literature review based on 31 studies included 1466 penile samples and identified a wide variability in HPV DNA detection by histological type. An overall pooled estimate showed that HPV prevalence was found within a range from 24.5% in verrucous SCC to 76% in basaloid SCC, with an overall average percentage of 46.9% and a weighted percentage by histology of 38.9%. HPV-16 was the most common type identified across histological types, accounting for 60.2%, followed by HPV-18 (13.3%) and HPV-6/11 (8.13%).

The high prevalence of HPV-16 is consistent in all studies except for those conducted in Argentina27 and Thailand,34 and is consistent with molecular and serological data.12 ,13 ,26 ,37 ,43 ,44 ,45 Low-risk HPV types have been suggested to be associated with a small subset of penile cancers, but its role is not yet clear.34 ,37 ,46 The relatively high presence of HPV-6/11 (8.13%) is not necessarily causative for cancer. The peneal surface is prone to genital condilomas that might be concomitant with other neoplasic lesions. In some circumstances these types are present in combined lesions, such as giant condiloma with an invasive lesion.52

Our review showed that basaloid SCC and warty SCC are the most common histological types related to HPV penile cancer, which is consistent with previous studies.6 ,13 ,43 However, typical SCC and keratinising SCC, which were thought to be only weakly correlated to HPV, were detected in 50% of the HPV positive cases. This observation suggests that HPV contribution to these histological types could be higher than previously reported.13 Commonly used pathological classifications include the US Armed Forces Institute of Pathology (AFIP)47 and the WHO48 systems. A new classification is to be published in the forthcoming AFIP fascicle fourth series and if widely used it may help in homogenising the histological classification.49

One of the problems in the classification of penile carcinomas is, despite the clear description of its features in a classical paper by Ackerman,53 the not infrequent incorrect classification of verrucous carcinoma, a tumour rarely associated with HPV when correctly diagnosed. A proposal for classification of verruciform tumours was published, categorising them into four distinctive subtypes: verrucous, papillary, warty–condylomatous, and giant condylomas. Mixed verruciform tumours are not unusual.50 ,51 In the articles included in this review, the majority of verrucous and papillary penile carcinomas were reported as negative or presenting a low frequency for HPV in most of the studies.

There have been variations in the reported association of HPV with warty carcinomas; this fact may also reflect the diagnostic difficulties in the classification of some condylomatous tumours. A re-evaluation of the morphological features of warty carcinoma and its HPV contents and a better definition of tissue koilocytosis are desirable. Non-invasive warty carcinomas are also difficult to separate from giant condylomas harbouring atypias, a not unusual finding in large condylomas. In equivocal lesions, HPV typification of the tumour may be required for the distinction, since low-risk HPV types are prevalent in giant condylomas52 and in mostly HPV-16 warty carcinomas, as shown in this review. In addition, giant condylomas may be associated with SCC of the usual type.

Basaloid carcinoma, the penile tumour most consistently associated with HPV, may also cause diagnostic problems when its cell composition is other than the classical uniform small cell. Basaloid growths with larger, more pleomorphic or even spindle cells have been noted; these tumours are difficult to separate from high-grade usual squamous cell carcinomas. The uterine cervical equivalence may be the non-keratinising SCCs. The exact limit or boundary demarcation of the classical basaloid and the large cell basaloidish appearing penile tumour is unknown. The warty–basaloid carcinoma (a mixture of basaloid and condylomatous tumour) is also consistently associated with HPV but has not yet been well reported in the literature.

About a third of penile SCC carcinomas are composed of a mixture of special subtypes and their classification is also problematic. HPV related and HPV non-related tumours may be found mixed with each other. A careful correlation of these mixtures and the presence of HPV are also necessary to weight the role of these neoplasms in the viral pathway of carcinogenesis. Associated studies using HPV serology markers support a strong relationship between HPV and SCC penile cancer, with odds ratios ranging from 1.7 to 6.4.12 ,37 ,44 ,54 ,55

Several molecular techniques with different sensitivity and specificity have been used for HPV detection and genotyping among the different studies included in the review. Prior to 1992, Southern blot was the most common technique used and PCR was mainly performed as type-specific for types HPV-16 and -18 until the mid 1990s. However, the greatest percentage of articles included in this review used PCR consensus primers for HPV DNA detection, such as GP5+/6+ and MY09/11. PCR-SPF10 was used in three studies. Almost all studies used previously stored formalin-fixed and paraffin-embedded samples. This technique leads to DNA degradation and decreases PCR efficiency by diminishing the size of the PCR product.56 Only eight of the studies analysed fresh or frozen tissues, but no significant differences could be detected when compared with paraffin embedded samples. Multiple infections and detection of other than HPV-16 and -18 high-risk HPV genotypes seem to have increased in the most recent reports,13 ,34 ,37 ,40 suggesting that the relative contribution of HPV-16 and -18 should be re-evaluated. An international effort to collect and analyse a large number of penile carcinoma samples from over 17 countries using PCR-SPF10 and INNO-LIPA is underway; its results will provide new evidence to evaluate the contribution of other HPV types and multiple HPV infection.

Provided that identification of HPV implies a causal role of the virus with the carcinogenic process, the attributable fraction of penile cancer related to HPV could be estimated to be 47% (95% CI 44.4 to 49.6). This percentage is similar to cancer of the vulva where the HPV-related fraction is around 40%.8 The aetiology of penile carcinomas is likely to be heterogeneous with a high-risk HPV related and an HPV-independent pathway.13 Based on cervical cancer studies, penile cancer could also arise from an initial HPV infection which persists over time and causes genetic alterations, leading to an interference of the cell division cycle and apoptosis.9 However, more investigation is needed to fully understand the molecular mechanisms behind penile carcinogenesis.

Worldwide variation in penile cancer incidence has been linked to differences in socioeconomic and religious conditions,57 and may be explained by variations in sexual patterns and circumcision rates, but there is need for further research on the contribution of HPV to penile carcinogenesis, particularly in those countries with the greatest incidence rates of penile cancer (ie, the African region)

Efficacy of the HPV prophylactic vaccine in men is still under investigation, but evidence to date suggests safety and immunogenicity.58 Although penile carcinoma is a rare disease, around 7000 cases would be prevented annually by the eradication of HPV-16/18.3 More studies regarding efficacy of the prophylactic vaccine on flat penile lesions (which show a higher HPV prevalence59 ,60) to prevent progression to penile carcinomas are needed.

In conclusion, the present review suggests that around half of the penile carcinomas could be related to HPV infection. However, the existence of a great variability among study designs, HPV DNA detection methods, and the difficulties of stratification of histology are important constraints when evaluating the prevalence of HPV. Systematic international studies are ongoing; they will probably help in reducing uncertainty and provide new evidence on the involvement of HPV in penile carcinoma.