Author's Accepted Manuscript

Human papillomavirus prevalence in invasive penile cancer and association withclinical outcome

Rosa S. Djajadiningrat, Ekaterina S. Jordanova, Bin K. Kroon, Erik van Werkhoven,Jeroen de Jong, Divera T.M. Pronk, Peter J.F. Snijders, Simon Horenblas, DaniëlleA.M. Heideman

PII: S0022-5347(14)04261-XDOI: 10.1016/j.juro.2014.08.087Reference: JURO 11744

To appear in: The Journal of UrologyAccepted Date: 8 August 2014

Please cite this article as: Djajadiningrat RS, Jordanova ES, Kroon BK, van Werkhoven E, de Jong J,Pronk DTM, Snijders PJF, Horenblas S, Heideman DAM, Human papillomavirus prevalence in invasivepenile cancer and association with clinical outcome, The Journal of Urology® (2014), doi: 10.1016/j.juro.2014.08.087.

DISCLAIMER: This is a PDF file of an unedited manuscript that has been accepted for publication. As aservice to our subscribers we are providing this early version of the article. The paper will be copy editedand typeset, and proof will be reviewed before it is published in its final form. Please note that during theproduction process errors may be discovered which could affect the content, and all legal disclaimersthat apply to The Journal pertain.

All press releases and the articles they feature are under strict embargo until uncorrected proof ofthe article becomes available online. We will provide journalists and editors with full-text copies of thearticles in question prior to the embargo date so that stories can be adequately researched and written.The standard embargo time is 12:01 AM ET on that date.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Original Article 1

Human papillomavirus prevalence in invasive penile cancer and association 2

with clinical outcome 3

4

Rosa S. Djajadiningrat (1), Ekaterina S. Jordanova (2), Bin K. Kroon (1), Erik van Werkhoven 5

(3), Jeroen de Jong (4), Divera T.M. Pronk (5), Peter J.F. Snijders (5), Simon Horenblas (1), 6

Daniëlle A.M. Heideman (5) 7

8

From the Departments of Urology (1), Biometrics (3) and Pathology (4), The Netherlands 9

Cancer Institute, Amsterdam, The Netherlands 10

Centre for Gynaecologic Oncology Amsterdam, Department of Obstetrics and Gynaecology 11

(2) and Department of Pathology (5) VU University Medical Center, Amsterdam, The 12

Netherlands 13

14

Address all correspondence to: 15

Simon Horenblas, Department of Urology, The Netherlands Cancer Institute 16

Plesmanlaan 121, NL-1066 CX Amsterdam, the Netherlands, Phone: +31 20 512 9083; Fax: 17

+31 20 512 2554; e-mail: s.horenblas@nki.nl 18

19

20

Conflict of interest or funding: None 21

22

Keywords: penile cancer, squamous cell carcinoma, human papillomavirus, survival 23

Word count text (including abstract): 2186 24

Word count text (excluding abstract): 1974 25

Word count abstract: 212 26

27

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

Abstract 1

Purpose: Penile cancer incidence is rising and suggested to be explained by changes in sexual 2

practice and increased exposure of males to sexually transmitted high-risk human 3

papillomavirus (hrHPV) infection. Previously, we found in penile cancers from a Dutch 4

population treated in the years 1963-2001, a hrHPV prevalence of about 30%. Here, we 5

assessed the prevalence of hrHPV-DNA in a more recent, contemporary penile cancer cohort 6

and association with patient survival. 7

Materials and Methods: hrHPV-DNA presence was assessed by GP5+6+-PCR in 212 formalin-8

fixed, paraffin-embedded invasive penile tumor specimens of patients treated between 9

2001 and 2009. Five-year disease-specific survival (DSS) was calculated using Kaplan-Meier 10

method with log-rank test and Cox regression. 11

Results: hrHPV-DNA was detected in a subset of penile cancer cases (25%; 95%-CI:19-31). 12

HPV16 was the predominant type, representing 79% (42/53) of all hrHPV infections. Five-13

year DSS in the hrHPV-negative group and hrHPV-positive group was 82% and 96%, 14

respectively (log rank test p=0.016). Adjusted for stage, grade, lymphovascular invasion (LVI) 15

and age, HPV status was still prognostic for DSS (p=0.030) with a hazard ratio of 0.2 (95%-16

CI:0.1–0.9). 17

Conclusions: hrHPV DNA was observed in one-fourth of penile cancer cases. No relevant 18

increase in hrHPV prevalence over the past decades was observed. The presence of hrHPV-19

DNA in penile cancer confers a survival advantage. 20

21

22 23

24

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

Introduction 1

The etiology of penile cancer is multifactorial with smoking, phimosis and poor hygiene 2

commonly associated with this tumor1. Other risk factors include the number of sexual 3

partners, and a history of genital warts or other sexually transmitted diseases2. At least part 4

of these risk factors is related to infection with human papillomavirus (HPV). In circumcised 5

men, HPV prevalence is lower than in uncircumcised men3, and penile cancer is rare in 6

populations that routinely practice circumcision4. 7

The reported proportion of penile cancer associated with high-risk HPV (hrHPV) types ranges 8

from 30% to 100%1,5–7

; depending on the population studied, the methods used for HPV 9

detection, and/or histological subtypes analyzed. In the Dutch population (years 1963-2001), 10

we found approximately 30% of penile cancers to be hrHPV-associated8–10

. It was also shown 11

that patients with hrHPV-positive tumors carried a survival advantage over patients with 12

hrHPV-negative tumors. However, other studies concerning the association between HPV 13

status and patient survival report inconsistent results6,11–13

. As such, the exact role of HPV as 14

a prognostic factor in penile cancer remains unclear. Moreover, penile cancer incidence is 15

rising and suggested to be explained by changes in sexual practice and exposure of males to 16

sexually transmitted HPV infection14

. Recent studies on head and neck squamous cell 17

carcinomas (SCC) have shown a rising incidence of hrHPV-associated subtypes over the past 18

decades in several Western countries, including the Netherlands15–17

. It is conceivable that 19

the percentage of hrHPV penile cancers may have increased over time as well. 20

Hence, the current study was set out to assess the prevalence of hrHPV-DNA in invasive 21

penile cancer in a contemporary Dutch cohort and its association with patient survival. 22

Towards this goal, a large series of penile cancer cases of the years 2001-2009 was analyzed 23

for hrHPV-DNA presence by GP5+6+-PCR, and the association between HPV status and 24

patient survival was assessed. 25

26

Materials and Methods 27

Study population 28

Our institutional cohort comprised 487 patients who were diagnosed with penile cancer 29

between 2001 and 2009. To achieve hrHPV-DNA data of patients with primary invasive 30

cancer treated between 2001 and 2009, formalin-fixed paraffin-embedded (FFPE) tissue 31

blocks were retrieved from the archives of the Department of Pathology, the Netherlands 32

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

Cancer Institute, the Netherlands, for additional testing. Patients who were initially treated 1

for their primary tumor elsewhere, neoadjuvantly treated or with carcinoma in situ were 2

excluded. For 212 patients, sufficient material of the primary invasive tumor was left for 3

HPV-DNA testing. The use of clinical material was in compliance with the respective 4

institutional ethical regulations for surplus material18

and approved by the institutional 5

translational research board. 6

All patients were (re)staged according to the TNM 2009 classification for penile cancer, and 7

had not received pre-operative radiotherapy or chemotherapy. The clinical and pathological 8

characteristics were prospectively kept in our institutional penile cancer database. Follow-up 9

until 2013 provided information concerning node-positivity, disease status and disease-10

specific mortality. 11

12

Sample preparation, histopathology and hrHPV-DNA detection and typing 13

Sample preparation and hrHPV-DNA detection and typing have been described before8–10

. 14

Briefly, of each FFPE specimen, a series of consecutive 5-µM sections were cut under safety 15

measures to avoid cross-contamination. The first and last sections were haematoxylin–eosin 16

stained for histopathology, including confirmation of tumor presence, subtyping, and 17

grading according to Broders19

. In-between sections were collected in a reaction vessel for 18

DNA extraction and subsequent PCR analyses. 19

Detection of hrHPV on the DNA extracts from FFPE sections was performed by GP5+/6+-PCR 20

enzyme immunoassay (EIA), using a cocktail of 14 hrHPV types (i.e., HPV 16, 18, 31, 33, 35, 21

39, 45, 51, 52, 56, 58, 59, 66 and 68) as described before8–10

. Beta-globin PCR was performed 22

on each DNA extract as a quality control. hrHPV-positive samples were subsequently 23

genotyped by bead-based array on the Luminex platform20

. 24

25

Statistical analyses 26

Differences in patient characteristics between hrHPV-positive and negative groups were 27

tested for statistical significance. DSS was defined as the time since surgery until death of 28

penile cancer, penile cancer metastasis or complications related to penile cancer treatment. 29

Patients alive at the end of follow-up were censored. DSS rates were calculated using 30

Kaplan-Meier method, with the log-rank test assessing equality of distributions. 31

Multivariable analysis of survival was performed using the Cox proportional hazards model. 32

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

All statistical analyses were performed with SPSS (v20.0, SPSS Inc., Chicago, Illinois) and R, 1

version 3.0.2 (http://www.r-project.org/). Two-sided p-values of <0.05 were considered 2

statistically significant. 3

4

Results 5

HPV prevalence 6

In total, 53/212 (25%, 95%-CI: 19%-31%) of the penile carcinoma cases were positive for 7

hrHPV-DNA; 42 (79%) contained HPV16, 4 (8%) HPV33, 3 (6%) HPV18, 2 (4%) HPV45, 1 (2%) 8

HPV31 and 1 (2%) HPV52. 9

10

Clinicopathological characteristics 11

Clinicopathological variables of patients stratified by tumor HPV status are presented in 12

Table 1. HrHPV-positive patients tended to have smaller tumors than hrHPV-negative 13

patients. In hrHPV-negative patients, 45% of tumors was well differentiated compared to 14

17% of tumors in hrHPV-positive patients. Regarding tumor subtypes, no differences in the 15

distribution of SCC vs. other subtypes were found between HPV-negative and -positive 16

cases (Table 1). Though based on small numbers, warty and subtypes with basaloid 17

features were observed more frequently amongst HPV-positive cancers, reaching 18

significance for the latter group only (p=0.08 and p=0.010, respectively). 19

No differences in age or other clinical and pathological characteristics were observed 20

between the hrHPV-positive and -negative groups. Median follow-up was 5.2 years 21

(interquartile range 4.7-6.1). 22

23

Patient survival 24

Five-year DSS rates according to clinicopathological variables were calculated using Kaplan-25

Meier and log rank test. The status of the lymph nodes (p<0.0001), pT-stage (p<0.0001), 26

lymphovascular invasion (LVI, p=0.0014), grade of differentiation (p=0.0022), and hrHPV 27

status (p=0.0160) were prognostic for DSS. As shown in figure 1, 5-year DSS of hrHPV-28

positive patients (n=53) vs. hrHPV-negative (n=159) patients was 96% vs. 82%, respectively. 29

No differences in survival were found based on age (at cut off 60 years) or tumor diameter. 30

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

In a Cox proportional hazard model, hrHPV (HR 0.2, 95%CI: 0.1-0.9, p=0.034) was still a 1

significant prognostic factor for DSS, after adjustment for lymph nodes status, pT-stage 2

group, grade of differentiation, LVI and age (table 2). 3

4

Discussion 5

In the current study, we showed that the proportion of hrHPV-related invasive penile 6

cancers in the Netherlands Cancer Institute was 25% in the last decade, with HPV16 being 7

the predominant viral type accounting for 79% of all hrHPV infections. Furthermore, we 8

demonstrated that hrHPV-positive tumors appear to provide a significant survival benefit 9

over hrHPV-negative tumors in multivariable analysis. Our findings are consistent with 10

previous observations in Dutch cohorts of penile cancer. Comparing current data to that of a 11

former series of our institute (years 1963-2001) which was evaluated during the course of 12

previous studies8 using the same method for hrHPV-DNA detection, we found no evidence of 13

an increase in the prevalence of HPV in penile cancer over the past decades (i.e., 29% 14

(50/171) in 1963-2001 vs. 25% in 2001-2009). Furthermore, both in the contemporary and 15

former cohorts, HPV16 was the most predominant type and hrHPV-positive tumors showed 16

a better prognosis. 17

The proportion of HPV-related penile cancers has been reported to differ among several 18

areas in the world. In two reviews concerning HPV prevalence and invasive penile cancer, 19

Europe and South-America were found to have lower HPV-related penile cancer numbers 20

than North-America and Asia6,21

. This may explain the observed lower hrHPV prevalence of 21

25-29% in the Dutch population than the reported prevalence in most other studies1,5,6

. A 22

recent, Swedish study reported an HPV prevalence of 82.9% in penile cancer (n=216), with 23

the majority being hrHPV-positive and with multiple HPV infections in almost 30% of HPV-24

positive cases22

. No associations with histopathological parameters were observed and 25

association with patient survival was not studied. The hrHPV prevalence in that study is 26

considerably higher than the prevalence described here. This large difference in hrHPV 27

prevalence might depend on methodological differences for assessing HPV status, or 28

differences among the investigated populations. In the Swedish study almost 30% of cases 29

was carcinoma in situ, while we included only invasive tumors. It is known from vulvar 30

cancer that carcinoma in situ lesions are more often HPV-positive than invasive lesions23

. 31

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

The high prevalence of HPV16 in our study, is consistent with all other penile cancer studies, 1

except for two conducted in Argentina and Thailand21

. Of interest, HPV16 is also the most 2

common type detected in other HPV-associated anogenital cancers, as well as in HPV-3

associated oropharyngeal cancers24–26

. 4

Recent studies on head and neck squamous cell cancers have shown a rising incidence of 5

hrHPV-associated subtypes over the past decades in Europe and North America15–17

. Also in 6

penile cancer, a recent study was performed, where HPV DNA was suggested to be more 7

frequently detected in more recently diagnosed cases27

; being 68% (46/68) in the period 8

2004–2005 vs. 36% (4/11) in 1998–2003 (p = 0.05). Given the limited numbers in early time 9

period, these findings however provide only very limited evidence of a more prominent and, 10

possibly, increasing role of HPV infection in penile carcinogenesis over time. In the current 11

study, we observed no signs of an increase in HPV prevalence in penile carcinoma over time, 12

comprising an extensive period from 1963 till 2009. Due to low numbers in the early years 13

(before 1980, n=18), it is hard to make an accurate estimation in those years, but also 14

between 1980-2009 no clear increase in HPV prevalence was observed. So far, no other 15

studies are available on time trends in HPV prevalence in other HPV-related anogenital 16

tumors, such as anal, vaginal and vulvar cancer. In vulvar cancer however, incidence is rising 17

among young women25

, suggesting a possible causative role of HPV, similar to 18

oropharyngeal cancer. 19

In most other HPV related tumors, a difference in age distribution is observed with HPV-20

positive tumors more often in younger patients. In our penile cancer population – both the 21

contemporary cohort as the population before 2001 – this was not observed. Our study 22

cohort does not include the entire Dutch penile cancer population, but over 60% of all penile 23

cancer patients are annually treated at our institute28

. It is known that basaloid and warty 24

subtypes are more likely to be HPV positive than other subtypes6. Though based on small 25

numbers, we confirm a higher prevalence of basaloid subtypes in HPV-positive patients, 26

but for warty subtypes a non-significant increase was observed. Furthermore, significantly 27

less well differentiated tumors were observed among the HPV-positive tumors. This is in 28

line with some previous studies where virus DNA was also more often associated with 29

high-grade tumors29

. The better prognosis of hrHPV-positive tumors in our cohort is in line 30

with data on head and neck cancer, where HPV is known to be a good prognostic factor for 31

survival and disease progression24

. The survival benefit in the current study was seen over all 32

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

time periods, and confirm our previous analysis8. Treatment differences are probably not 1

the explanation for this survival benefit, as there was no significant difference in primary 2

treatment between HPV-negative and HPV-positive patients. Remarkably, the survival 3

benefit was demonstrated without any association between HPV and lymph node status, 4

and significantly less well differentiated tumors among HPV positive patients. In other 5

penile cancer studies concerning prognosis and HPV status, no differences in survival among 6

HPV positive patients were observed12,13,27

. Also in vulvar cancer studies concerning HPV 7

infection and prognosis, data are contradictive25

. However, in most of these studies limited 8

numbers of patients were included, which may explain the non-discriminative role of HPV. 9

This study is not devoid of limitations. Firstly, by using FFPE material, it may be that the 10

proportion of HPV positivity is underestimated5. Furthermore, we only used tissue from 11

patients not pre-treated with chemotherapy. Therefore it is possible that the proportion of 12

patients with more advanced tumors is underrepresented in this study. 13

14

Conclusion 15

Our results underscore previous observations in penile cancer. hrHPV infection is associated 16

with a subset of invasive penile tumors (25%). The presence of hrHPV confers a survival 17

advantage in multivariable analysis. No evidence of an increase in the prevalence of HPV in 18

penile cancer over the past decades was observed. 19

20

Acknowledgement 21

The authors would like to thank the people from the Molecular Pathology & Biobank Core 22

Facility from the Netherlands Cancer Institute for all their administrative and technical 23

support. 24

25 26 References 27

1. Maden C, Sherman KJ, Beckmann AM, et al: History of circumcision, medical 28 conditions, and sexual activity and risk of penile cancer. J. Natl. Cancer Inst. 1993; 85: 29 19–24. 30

2. Daling JR, Madeleine MM, Johnson LG, et al: Penile cancer: importance of 31 circumcision, human papillomavirus and smoking in in situ and invasive disease. Int. J. 32 Cancer 2005; 116: 606–16. 33

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

3. Castellsagué X, Bosch F, Munoz N, et al: Male circumcision, penile human 1 papillomavirus infection and cervical cancer in female partners. N. Engl. J. Med. 2002; 2 346: 1105–1112. 3

4. Dillner J, von Krogh G, Horenblas S, et al: Etiology of squamous cell carcinoma of the 4 penis. Scand. J. Urol. Nephrol. Suppl. 2000: 189–93. 5

5. Rubin MA, Kleter B, Zhou M, et al: Detection and typing of human papillomavirus DNA 6 in penile carcinoma: evidence for multiple independent pathways of penile 7 carcinogenesis. Am. J. Pathol. 2001; 159: 1211–8. 8

6. Backes DM, Kurman RJ, Pimenta JM, et al: Systematic review of human papillomavirus 9 prevalence in invasive penile cancer. Cancer Causes Control 2009; 20: 449–57. 10

7. McCance DJ, Kalache A, Ashdown K, et al: Human papillomavirus types 16 and 18 in 11 carcinomas of the penis from Brazil. Int. J. Cancer 1986; 37: 55–9. 12

8. Lont AP, Kroon BK, Horenblas S, et al: Presence of high-risk human papillomavirus 13 DNA in penile carcinoma predicts favorable outcome in survival. Int. J. Cancer 2006; 14 119: 1078–81. 15

9. Ferreux E, Lont AP, Horenblas S, et al: Evidence for at least three alternative 16 mechanisms targeting the p16INK4A/cyclin D/Rb pathway in penile carcinoma, one of 17 which is mediated by high-risk human papillomavirus. J. Pathol. 2003; 201: 109–18. 18

10. Heideman DAM, Waterboer T, Pawlita M, et al: Human papillomavirus-16 is the 19 predominant type etiologically involved in penile squamous cell carcinoma. J. Clin. 20 Oncol. 2007; 25: 4550–6. 21

11. Flaherty A, Kim T, Giuliano A, et al: Implications for human papillomavirus in penile 22 cancer. Urol. Oncol. 2013: 1–8. 23

12. Bezerra AL, Lopes A, Santiago GH, et al: Human papillomavirus as a prognostic factor 24 in carcinoma of the penis: analysis of 82 patients treated with amputation and 25 bilateral lymphadenectomy. Cancer 2001; 91: 2315–21. 26

13. Lopes A, Bezerra ALR, Pinto CAL, et al: P53 As a New Prognostic Factor for Lymph 27 Node Metastasis in Penile Carcinoma: Analysis of 82 Patients Treated With 28 Amputation and Bilateral Lymphadenectomy. J. Urol. 2002; 168: 81–6. 29

14. Graafland NM, Verhoeven RHA, Coebergh J-WW, et al: Incidence trends and survival 30 of penile squamous cell carcinoma in the Netherlands. Int. J. cancer J. Int. du cancer 31 2011; 128: 426–432. 32

15. Näsman A, Attner P, Hammarstedt L, et al: Incidence of human papillomavirus (HPV) 33 positive tonsillar carcinoma in Stockholm, Sweden: an epidemic of viral-induced 34 carcinoma? Int. J. Cancer 2009; 125: 362–6. 35

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

16. Nichols A, Palma DA, Dhaliwal SS, et al: The epidemic of human papillomavirus and 1 oropharyngeal cancer in a Canadian population. Curr. Oncol. 2013; 20: 212–9. 2

17. Rietbergen MM, Leemans CR, Bloemena E, et al: Increasing prevalence rates of HPV 3 attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed 4 by a validated test algorithm. Int. J. Cancer 2013; 132: 1565–71. 5

18. Federa: Human Tissue and Medical Research: Code of conduct for responsible use 6 (2011). 2011: 1–36. Available at: 7 http://www.federa.org/sites/default/files/digital_version_first_part_code_of_conduc8 t_in_uk_2011_12092012.pdf. 9

19. Broders A: Squamous cell-epithelioma of the skin. Ann. Surg. 1921; 43: 141–159. 10

20. Schmitt M, Bravo IG, Snijders PJF, et al: Bead-Based Multiplex Genotyping of Human 11 Papillomaviruses Bead-Based Multiplex Genotyping of Human Papillomaviruses. J. 12 Clin. Microbiol. 2006; 44: 504–512. 13

21. Miralles-Guri C, Bruni L, Cubilla AL, et al: Human papillomavirus prevalence and type 14 distribution in penile carcinoma. J. Clin. Pathol. 2009; 62: 870–8. 15

22. Kirrander P, Kolaric A, Helenius G, et al: Human papillomavirus prevalence, 16 distribution and correlation to histopathological parameters in a large Swedish cohort 17 of men with penile carcinoma. BJU Int. 2011; 108: 355–9. 18

23. De Vuyst H, Clifford GM, Nascimento MC, et al: Prevalence and type distribution of 19 human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina 20 and anus: a meta-analysis. Int. J. Cancer 2009; 124: 1626–36. 21

24. O’Rorke MA, Ellison M V, Murray LJ, et al: Human papillomavirus related head and 22 neck cancer survival: a systematic review and meta-analysis. Oral Oncol. 2012; 48: 23 1191–201. 24

25. Smith JS, Backes DM, Hoots BE, et al: Human papillomavirus type-distribution in vulvar 25 and vaginal cancers and their associated precursors. Obstet. Gynecol. 2009; 113: 917–26 24. 27

26. Hoots BE, Palefsky JM, Pimenta JM, et al: Human papillomavirus type distribution in 28 anal cancer and anal intraepithelial lesions. Int. J. Cancer 2009; 124: 2375–83. 29

27. Hernandez BY, Goodman MT, Unger ER, et al: Human papillomavirus genotype 30 prevalence in invasive penile cancers from a registry-based United States population. 31 Front. Oncol. 2014; 4: 9. 32

28. Djajadiningrat RS, Graafland NM, van Werkhoven E, et al: Contemporary Management 33 of Regional Nodes in Penile Cancer-Improvement of Survival? J. Urol. 2014; 191: 68–34 73. 35

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

29. Gregoire L, Cubilla AL, Reuter VE, et al: Preferential association of human 1 papillomavirus with high-grade histologic variants of penile-invasive squamous cell 2 carcinoma. J. Natl. Cancer Inst. 1995; 87: 1705–9. 3

4

Legends 5

Table 1: Patient characteristics 6 a

Wilcoxon rank sum test with continuity correction 7 b

Linear-by-linear Association test 8 c Fisher’s exact test 9

10 Table 2: Multivariable Cox proportional hazards model for disease-specific survival 11 12 Figure 1: Disease-specific survival after primary tumor treatment; hrHPV-positive tumors 13 (n=53) appear to have a significant survival benefit (96%) over hrHPV-negative tumors 14 (n=159, 82%, log rank test p=0.016) 15

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 1: Patient characteristics

hrHPV negative

n=159

hrHPV positive

n=53

Total

N=212

p-value

Age

Median (IQR)

67 (58-74)

63 (54-71)

63 (54-71)

0.23a

Pathological T-stage

pT1a (%)

pT1b (%)

pT2 (%)

pT3 (%)

pT4 (%)

38 (24)

5 (3)

99 (62)

11 (7)

6 (4)

14 (26)

5 (9)

29 (55)

5 (9)

0 (0)

52 (25)

10 (5)

128 (60)

16 (8)

6 (3)

0.29b

Tumor size (cm)

Median (IQR)

3 (2-4.1)

2.5 (1.5-3.5)

3 (2-4)

0.025a

Subtype

SCC (%)

Verruceus (%)

Warty (%)

Basaloid (%)

Basaloid-SCC (%)

Papillar (%)

Cuniculatum (%)

Pseudohyperplastic (%)

Sarcomatoid (%)

139 (87)

5 (3)

2 (1)

1 (1)

1 (1)

8 (5)

1 (1)

1 (1)

1 (1)

44 (83)

0 (0)

3 (6)

4 (8)

1 (2)

1 (2)

0 (0)

0 (0)

0 (0)

183 (86)

5 (2)

5 (2)

5 (2)

2 (1)

9 (4)

1 (0)

2 (1)

1 (0)

0.49c*

Grade of differentiation

Well (%)

Intermediate (%)

Poorly (%)

71 (45)

61 (38)

27 (17)

9 (17)

32 (60)

12 (23)

80 (38)

93 (44)

39 (18)

0.004b

LVI

No (%)

Yes (%)

Missing (%)

135 (86)

22 (14)

2

44 (83)

9 (17)

0

179 (85)

31 (15)

2

0.66c

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Lymph nodes

Tumor negative (%)

Tumor positive (%)

107 (67)

52 (33)

39 (74)

14 (26)

146 (69)

66 (31)

0.49c

Penile treatment

Penis preserving

(partial) amputation

72 (45)

87 (55)

26 (49)

27 (51)

98 (46)

114 (54)

0.64c

IQR = interquartile range

SCC = squamous cell carcinoma

Crude percentages are presented

a Wilcoxon rank sum test with continuity correction

c Fisher’s exact test

b Linear-by-linear Association test *SCC vs. other

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTTable 2: Multivariable Cox proportional hazards model for disease-specific survival

Hazard ratio (HR) (95% CI) p value

hrHPV

Negative

Positive

1

0.2

0.1 – 0.9

0.034*

Lymph node status

Negative

Positive

1

28.1

6.5 – 121.8

<0.001*

pT stage group

pT1a-b

pT2

pT3-4

1

0.6

2.5

0.2 – 1.7

0.8 – 7.8

0.340

0.110

Grade of differentiation

Well / intermediate

Poorly

1

2.8

1.3 – 6.0

0.010*

LVI

No

Yes

1

1.7

0.7 – 3.9

0.250

Age

in years

1.0

1.0-1.1

0.290

*statistically significant

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTD

isea

se s

peci

fic s

urvi

val p

roba

bilit

y

0 1 2 3 4 5

0.0

0.2

0.4

0.6

0.8

1.0

Time (years)

HR−HPV−HR−HPV+

159 142 129 121 99 66 HR−HPV−53 52 50 48 46 31 HR−HPV+

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPTKey of abbreviations

DSS: disease-specific survival

ENE: Extranodal extension

FFPE: formalin-fixed paraffin-embedded

IQR: Interquartile range

HPV: Human Papillomavirus

hrHPV: high-risk Human Papillomavirus

LVI: Lymphovascular invasion

SCC: squamous cell carcinoma