Oral Oncology 45 (2009) Contents lists available at ScienceDirect. Oral Oncology. journal homepage:

Transcription

1 Oral Oncology 45 (2009) Contents lists available at ScienceDirect Oral Oncology journal homepage: Oropharyngeal carcinoma in non-smokers and non-drinkers: A role for HPV Elizabeth Andrews a, William T. Seaman c, Jennifer Webster-Cyriaque b,c,d, * a College of Dental Medicine, Western University of Health Sciences, Pamona, CA , USA b School of Dentistry, University of North Carolina at Chapel Hill, NC , USA c Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC , USA d School of Medicine, Department of Microbiology and Immunology, CB#7455, University of North Carolina at Chapel Hill, NC , USA article info summary Article history: Received 30 April 2008 Received in revised form 8 July 2008 Accepted 9 July 2008 Available online 21 November 2008 Keywords: Oropharynx Cancer HPV Virus Transmission Sequence analysis Incidence of oropharyngeal squamous cell carcinoma (OSCC) increased 3% annually from 1973 to OSCC s can be attributed to tobacco and alcohol, but 25% are unlinked to typical risks. Case control studies on HPV detection in non-smoking/non-drinking (NS/ND) OSCC patients have not previously been performed. The primary objective of this study was to determine whether high-risk HPV infection was significantly associated with development of oral squamous malignancy in non-smokers/non-drinkers. A chart review of 802 OSCC patients from the UNC Pathology Archives ( ) yielded 40 nonsmoker/non-drinker subjects. Utilizing a case control design, 18 cancer cases and 22 benign biopsy controls were consecutively identified. Biopsy tissue was subjected to (i) HPV-L1 consensus PCR and sequencing (ii) real-time PCR. Chi-square and logistic regression analysis was employed. Logistic regression analysis determined that cases were 6.1 (OR 95% CI, ) times more likely to have HPV infection in their tumors than controls. High-risk HPV-DNA was readily detected in the tonsils and base of tongue (oropharynx) of 14/18 cases and 6/22 controls by both consensus and real-time PCR. Of high-risk HPV containing lesions, 85% (17/20) originated in the oropharynx (chi-square, p = 0.03). High risk HPV was also detected in benign biopsies of the oropharynx in 30% (3/10) of individuals who had a previous oral cancer (chi-square, p = 0.006). The infectious nature of OSCC in NS/ND was revealed by consistent detection of HPV, suggesting HPV s potential role in transforming oral epithelium, providing further evidence of the need to screen the oropharynx for HPV in NS/ND. Ó 2008 Published by Elsevier Ltd. Introduction Head and neck squamous cell carcinoma (HNSCC) is a significant cause of morbidity and mortality with 30,000 new cases in the US each year. While the incidence of HNSCC has remained stable, the incidence of oropharyngeal squamous cell carcinoma (OSCC), a subset of HNSCC, increased 2 3% annually ( ). 1 A recent study of SEER data analyzed cancer incidence in sites of consistent Human Papillomavirus (HPV) detection including base of tongue, tonsils, lingual tonsil, and Waldeyer s ring. From 2000 to 2004, the incidence of malignancy at these sites increased annually by 5.22% (p < 0.05). 2 Traditionally, OSCC is characterized by a poor prognosis, with an overall 5 year survival rate of 59.4%, accounting for 8000 deaths/year in the US. Many OSCCs can be attributed to exposures such as tobacco and alcohol, but there is a subset of patients with neither of these risk factors * Corresponding author. Address: School of Medicine, Department of Microbiology and Immunology, CB#7455, University of North Carolina at Chapel Hill, NC , USA. Tel.: +1 (919) ; fax: +1 (919) address: jennifer_cyriaque@dentistry.unc.edu (J. Webster-Cyriaque). who still develop cancer. 3 Also, only a fraction of smokers and drinkers develop cancer, suggesting cofactors including oncogenic viruses or other infectious agents as risk factors. 2,3 HPV s are epitheliotropic, oncogenic DNA viruses with greater than 120 different types identified to date. Low-risk, episomal HPV s (HPV6/11) induce benign proliferation of epithelium. In contrast, high-risk oncogenic types, HPV16/18, are commonly, but not always, found integrated into host DNA. The role of HPV16/18 has been firmly established, by strong epidemiologic association, as the etiologic agent in cervical cancer. 4 High-risk type HPV-E6 and E7onco-proteins target the p53 and prb tumor suppressor pathways rendering these pathways dysfunctional in the majority of HPV related human cancers. High-risk HPV DNA has been consistently detected in 20% of HNSCCs overall and in 20 72% of the OSCC subset. 2,4,5 It has been proposed that high-risk type HPV infection has contributed to the increase in OSCC. To date, no studies have addressed HPV-associated OSCC in non-smokers and non-drinkers (NS/ND). The purpose of this case control study was to determine whether high-risk HPV infection was significantly associated with oral malignancy in the absence of tobacco/alcohol risk. A secondary /$ - see front matter Ó 2008 Published by Elsevier Ltd. doi: /j.oraloncology

2 E. Andrews et al. / Oral Oncology 45 (2009) objective of the following studies was to determine the sensitivity of a novel, HPV-based real-time assay as compared to consensus PCR within tissues from this case control population. Methods and materials Subjects This study was approved by the School of Medicine, Institutional Review Board University of North Carolina, Chapel Hill (IRB# 05-DENT-1263-ORC). Study subjects were identified through the UNC Healthcare Cancer Registry of >22,000 cases (1995-present). This Registry maintains a database of all patients diagnosed and/or treated for malignant neoplasm s at UNC Healthcare. Of all registered cases, 802 had a cancer in any oral cavity and perioral site. There were 458 subjects with oropharynx or oral cavity cancer. All subjects were selected based on having no prior or current use of tobacco and/or alcohol, confirmed by chart review. Sixty-one subjects were excluded due to undocumented habit history in the database. Sixty-eight were confirmed to be NS/ND by chart review, and 47 of these had a lesion in the oropharynx or various oral cavity sites. Tissue samples were consecutively retrieved based on site of tumor and on date ( ), starting with subjects who were more recently biopsied. Benign control tissues were defined as those not demonstrating any properties associated with malignancy (i.e. mitotic figures, hyperchromasia, pleomorphism, and increased nuclear/cytoplasmic ratio). Additional control benign tonsil tissue was assayed from previous sleep apnea subjects without a cancer history. Other benign tissues came from UNC Oral and Maxillofacial Pathology Archives ( ). Cases, comprising 2.5% (18/734) of this registry subset, were subjects with a histologically diagnosed cancer (confirmed by two independent pathologists). Controls were chosen based on biopsies histologically confirmed as benign. A subset of these benign controls had a previous history of OSCC. Sample size calculation Sample size was calculated utilizing a two-sample comparison of proportions estimation. Based on previous studies we estimated the proportion of exposed controls to be 5% and exposed cases to be P60%. From this estimation, it was determined that there should be 18 cancer cases and 18 histologically benign controls. There were a total of 22 controls, 10 benign controls had a previous cancer history and 12 benign controls had no cancer history. Preparation of specimens Paraffin blocks containing tumor tissue from 40 patients were retrieved, 5u sections were cut from the block. Blades were disinfected with alcohol for each sample and the microtome was adjusted to insure a new region of the cutting surface was being used for each block. Five to ten of resultant sequential scrolls from each individual block were placed in an eppendorff tube, deparaffinized and DNA extracted using DNAeasy protocol (Qiagen, Valencia, Ca). Amplifiable quality of DNA was confirmed by PCR assay targeting the human beta-actin, or 18S genes. 6 HeLa cell, an immortalized HPV-18(+) cell line, was used as a positive control. All control samples were prepared and analyzed in parallel with clinical specimens to ensure proper reaction conditions were maintained. Special care was taken to control contamination. To detect possible DNA contamination, all tests included a negative control amplification containing double distilled, autoclaved, filtered water and PCR reagents. PCR assays were carried out using dedicated PCR area for reaction setup, dedicated positive displacement pipettors and single use aliquots of all reagents including primers and enzyme. At least two replicates/sample were analyzed, and DNA from HPV( ) fibromas interspersed as another negative control. HPV- L1 PCR assay All samples were screened for the presence of HPV using a standard nested PCR assay, described elsewhere, 7 10 consisting of the MY09/11 primer set (primary PCR) and GP5+/6+ primer set (secondary PCR). These primers are consensus degenerate primers complementary to the conserved L1 region of HPV. PCR products were separated by gel electrophoresis. In initial amplification, MY PCR detected a product size of 450 bp. Second step MY/GP PCR detected a product size of 140 bp. To insure accuracy, all samples were tested in replicate (2 3x 0 s). HPV-16/18 L1 multiplex real-time PCR assay DNA was employed for real-time PCR amplification of the L1 region of the HPV genome. Briefly, DNA, Lightcycler TaqMan Master Mix (Roche) and the Roche Lightcycler 480 detection system were used for amplification. Detection of HPV16L1 DNA was performed using HPV16L1F (5 0 TTGTTGGGGTAACCAACTATTTGTTACTGTT3 0 ) and HPV16L1R (5 0 CCTCCCCATGTCGTAGGTACTCCTTAAAG3 0 ) as primers and HPV16L1PROBE (5 0 6-FAM-GTCATTATGTGCTGCCA- TATCTACTTC-TAM3 0 ) as a TAqMan probe. Detection of HPV18L1 DNA was performed using HPV18L1F (5 0 ATCAATTATTTGTTAC- TGTGGTAGATACCACTCC3 0 ) and HPV18L1R (5 0 GCTATACTGCTTA- AATTTGGTAGCATCATATTGC3 0 ) as primers and HPV18L1PROBE (5 0 HEX-GTCATTATGTGCTGCCATATCTACTTC-TAM3 0 as a TAqMan probe. Cellular Human APOB gene was amplified to ensure the quality of the DNAs using HAPBF (5 0 TGAAGGTGGAGGACATTCCTCTA3 0 ) and HAPBR (5 0 CTGGAATTGCGATTTCTGGTAA3 0 ) as primers and HAP- BPROBE (5 0 CY5-CGAGAATCACCCTGCCAGACTTCCGT-BHQ23 0 ) as a TAqMan probe. Analysis of the data was performed with the associated LightCycler 480 software (Roche). DNA sequencing Specimens with amplification products of the expected size were considered HPV(+), were purified with ExoSap-IT reagent (USB, Cleveland, Ohio, USA) followed by direct sequencing of the targeted HPV region using ABI Big Dye Ò terminator reagents (Applied Biosystems, Foster City, CA) and an ABI PRISM 3730 DNA Sequencer (UNC Genome Analysis Facility). Amplimers were sequenced in both forward and reverse orientations and compared using NCBI Blast2. Sequencing analysis was performed with Vector NTI (Invitrogen) utilizing DNA extracted sequences from (+) subjects, HeLa, HPV-18 (Genbank-U89349), HPV-16 (Genbank- K02718) and Siha (Genbank-U89348). A Guide Tree (VectorNTI), which resembles a phylogenetic tree, was calculated after sequence alignment using the Neighbor-Joining method which works on a matrix of distances between all pairs of sequences to be analyzed and relates degree of divergence between sequences. Statistical analysis Comparisons of general characteristics were made by Pearson s Chi-square test for categorical variables. In all evaluations p values were considered statistically significant. Factors utilized for analysis include the categorical variables sex, age at diagnosis, race, high-risk HPV presence, and tumor/biopsy location. Traditionally, OSCC s arise in males >60 years of age. Because of this, age was analyzed as a dichotomous variable with the cut point of 60 years. A logistic regression (LR) model was used to evaluate the association between several factors and HPV status in NS/ND

3 488 E. Andrews et al. / Oral Oncology 45 (2009) cancer (cases) and NS/NDs with benign biopsies (controls). To measure the association with HPV, the odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Because of the small sample size, the LR model could only assess two variables in the analysis. The variables utilized in this analysis were the main outcome of interest, HPV infection and location of tumor/biopsy, as these are hypothesized to be associated with malignancy. Computations were performed using STATA statistical software (Version 9.2, Statistics/Data Analysis, Stata Corp, College Station, Texas, USA). Results Table 1 Study subject characteristics Characteristic Cases Controls Cancer (18/40)% (n) Benign W/prev CA (10/40) % (n) Benign W/out CA (12/40) % (n) Mean age 56 (13.9 SD) 59 (15.8 SD) 56 (17 SD) Male 44 (8/18) 40 (4/10) 58 (7/12) White 89 (16/18) 80 (8/10) 75 (9/12) Location of BX Tonsils 56 (10/18) 40 (4/10) 33 (4/12) Base of tongue 33 (6/18) 20 (2/10) 0 Oral tongue 11 (2/18) 40 (4/10) 67 (8/12) HPV+ 78 (14/18) 30 (3/10) 25 (3/12) SD, standard deviation; BX, biopsy. In this study of 40 NS/ND, there were 18 subjects with cancer (cases) and 22 subjects with a benign biopsy (controls), 10 subjects with a previous cancer history and subsequent benign biopsy and 12 with a benign biopsy and no previous cancer history. Comparative analysis of subject characteristics determined that the mean age of the cases was 56, while the mean age of the controls was 57.5 (mean age with a previous cancer history was 59, and 56 years in the subjects with no previous cancer history) (Table 1). With regard to distribution of sites biopsied, 89% of cancer cases had their cancer originate in the oropharynx and 45% of controls had a benign biopsy from this region. In order to detect multiple HPV types simultaneously within a single subject sample, we developed a novel real-time PCR assay using multiplexed HPV16 and HPV18 specific probes/primers targeting the L1 region as well as primer/probe for an internal standard human APO-B. This has significant advantages over consensus PCR in that the assay is single tube reaction that is type specific and quantifiable. DNA from paraffin embedded tissue was analyzed by both the novel multiplex real-time PCR assay and consensus PCR. With the exception of 7 benign controls that were not tested with real-time PCR due to lack of tissue, all other tissues were analyzed in this way (Table 2). By real-time, of 18 cancer/ cases, 11 were found to harbor HPV16 and 3 were HPV18-positive. Of the 10 controls with previous cancer history, 3/10 were HPV-positive, 2 with HPV16 and 1 with HPV18. Consensus degenerate PCR of the HPV-L1 region in this population detected HPV in half of the subjects (20/40). Amplimers from all of those that were positive were sequenced and results of the alignment analysis are found in Fig. 1 and Table 2. Highly conserved HPV signature sequences were consistently detected and are highlighted. High-risk HPV16 was detected in 13 of the subjects. High-risk HPV18 was detected in 7 of the subjects. The relationship between clinical isolates was assessed by phylogenetic analysis (NJ method) (Fig. 2). Consensus PCR found 78% (14/18) of the cancer/cases were HPV-positive, all of which were high-risk type (Table 3). In controls with previous cancer history, 30% had HPV in base of tongue biopsy tissue, all of high-risk type. In controls without previous cancer history, 25% (3/12) were HPV18-positive and these biopsies were from the oral cavity and tonsils (Table 3). A secondary objective was to validate this novel, HPV-based multiplex real-time assay. Comparison of consensus PCR, the gold standard, to real-time targeting the same region detected complete concordance between real-time and consensus PCR for HPV16 and HPV18 detection in both cases and controls. Of the 14 cases that were HPV-positive with consensus PCR, all 14 showed concordant results for the high-risk HPV type detected by consensus PCR. Three of four cancer/cases were consensus PCR-negative and were negative with real-time PCR. In the benign/controls, real-time PCR detected the same type of HPV in 75% of controls which were HPVpositive with consensus PCR. Of the 11 benign/controls, HPV-negative by consensus PCR, 4 were HPV-positive with real-time PCR, perhaps reflecting differences in assay sensitivity level. From the bivariate chi-square analysis, characteristics that were significantly associated with high-risk HPV infection included (1) having a cancer diagnosis and (2) a biopsy located in the oropharynx (p value < 0.05) (Table 4). A logistic regression model that analyzed the variables HPV status and biopsy/tumor site, was utilized to determine the odds of high-risk HPV infection in cancers of NS/ ND. The analysis determined that NS/ND cancer/cases were 6.1 (95% C.I ) times more likely to have HPV infection than patients with a benign diagnosis, adjusted for oropharynx. The confidence interval was wide due to the sample size, but was found to be significant p = 0.02 (Table 4). Discussion Today, it is generally agreed that viruses are implicated in 10 20% of all cancers, with DNA oncogenic viruses like HPV, often targeting the p53 and prb tumor suppressor pathways resulting in cell cycle alteration Greater than 90% of cervical cancers are related to HPV infection and in a majority of these cervical cancers, high-risk HPV types 16 and 18 are implicated. 14 Oral mucosa, especially oropharyngeal mucosa, is similar histologically to cervical mucosa. The presence of HPV in oral mucosa suggests that, as in cervical cancer, HPV infection may also play a similar role in the transformation of oral epithelium. While it has been postulated that the path to cancer formation is a distinct multi-step process, it also appears that there may be two major pathways leading to oral carcinogenesis: p53 mutation and DNA tumor viral pathogenesis. 9 The transformation to cancer in the setting of HPV infection is distinct when risk factors, such as smoking and or drinking, are not present. The significance of infectious agents in oral cancers has not been fully demonstrated in persons without any known risk factors. This study supports the hypothesis that high-risk HPV infection is a predominant risk factor in the development of oropharyngeal squamous cell carcinoma, in patients who do not smoke or drink. Of 741 individuals from the UNC Registry with a documented (ever/never) history of alcohol and tobacco use, with oral cavity or oropharynx cancer, 70% were smoker and drinkers, 17% were smokers, 4% were drinkers and only 9% were NS/ND, without typical risk. The mean age of this study population was well below that of average, the age of most oropharyngeal cancer patients in the US is over 65 years. 15 Similar to other case control studies we consistently detected HPV in OSCC. 9,16,17 Oral cancer patients with no exposure to smoking and/or drinking were 6.1 times more likely to have HPV DNA in there tumors than the NS/ND benign controls, when adjusting for location. This odds ratio is lower than those reported by case studies that examined HPV in association with traditional risk factors. Previous studies have shown a strong association of high-risk sexual practices with a resultant OSCC. 9 These practices were not examined by this study, but may be an important risk factor in the possible transmission of these

4 E. Andrews et al. / Oral Oncology 45 (2009) Table 2 HPV L1 presence detected by consensus and real-time PCR Cancer Sample Location L1 PCR Multiplex L1 real-time HPV MY Nested MY/GP Sequence/HPV type Cases 1A Tonsil B Tonsil A Tonsil B Tonsil BOT BOT BOT NONE 18 9 BOT BOT Tonsil NONE NONE 21 Tongue /18 23 Tongue Tonsil Tonsil None None 38 Tonsil None None 39 Tonsil Tonsil BOT Controls Benign prev CA 52 BOT BOT A Tonsil None NT 2A Tonsil None NT 2B Tonsil None NT 15 Tongue None None 20 Tongue None None 27 Tongue None None 32 Tongue None None 37 Tonsil Tongue None None Benign no CA HX 24 Tongue None None 28 Tongue None None 42 Tongue None Tongue None Tongue None Tongue None Tongue Tonsil None NT 54 Tonsil None NT 55 Tonsil + 18 NT 56 Tonsil + 18 NT NT (not tested); None (no HPV detected). Figure 1 Alignments of HPV L1 DNA sequences from oral cancer and benign biopsies of NS/ND. Seven regions of HPV L1 signature sequence, highlighted in yellow, were detected in all positive samples. A hallmark of HPV16 was a 3 base pair (BP) and 4 BP deletion in only the HPV 16 positive subjects. (For interpretation of color in Fig.1, the reader is referred to the web version of this article.)

5 490 E. Andrews et al. / Oral Oncology 45 (2009) Figure 2 Phylogenetic analysis of HPVL1 sequences from NS/ND. Phylogenetic analysis shows distinct ancestry by HPV type. HPV16 types originate from one ancestral node and the HPV18 sequences are similar to each other and least related to the 16 types. Table 3 Relationship of high risk HPV presence by location and diagnosis Characteristic Cases Controls HPV + cancer N =14 HPV + benign W/CA N =3 % High/low risk 100/0 100/0 100/0 % Oropharynx HPV + benign W/out N =3 infectious agents and should be addressed in future studies. Other limitations of this study include small sample size. We have shown for the first time that 78% (N = 18) of NS/ND OSCC cases had highrisk HPV in their tumors (p < 0.006), predominantly HPV16, as determined by sequence and subsequent phylogenetic analysis and also using a novel HPV-based real-time assay. Consistent with the secondary aims of this study, we were able to validate the use of multiplex real-time PCR as a potential diagnostic tool. The ability to multiplex provides a significant advantage when dealing with limited amounts of tissue or nucleic acid. Further, the assay was equivalent to the gold standard in its detection of HPV in tumor tissues. High-risk HPV types were consistently detected in the oropharynx of cancer cases. As previously shown, the base of tongue and tonsil, which comprise the oropharynx, were often sites of malignancy. 18 Hence, location of infection is significant to development of malignancy in this population. Of the cases in this study, HPV infection was detected in 5/6 (83%) cases with a cancer originating in the base of tongue and in 12/14 (86%) of all oropharyngeal cancers. D Souza et al. detected high-risk HPV in 72% (N = 100) of their oropharyngeal cancer cases and determined that high-risk HPV was strongly associated with OSCC, regardless of exposure to established risk factors. 9 Many oral cancers originate in the base of the tongue and often go undetected, due to the location being difficult to view. Consistent detection of HPV infection in the base of tongue highlights the importance of targeting other oropharyngeal regions for preventive screening/intervention even in those persons without traditional risk factors. High-risk HPV was detected in 30% of benign controls with a previous cancer history, in follow up benign biopsies from the oropharynx. Interestingly, all had a previous cancer in the same site as the subsequent negative biopsy. High-risk HPV infection in the same location as the original cancer may heighten the potential for future recurrence in these individuals, warranting close follow up. High-risk HPV18 was detected in the tonsils of 3 benign controls without a previous cancer history. Others have reported that HPV16 presence is related to oral malignancy in 95% of HPV associated HNSCC. 18 The detection of HPV18 in this subset was interesting, as HPV18 does not share the same oral oncogenic potential as HPV16. Approximately 20% of adults become infected with HPV16 in their lifetime. 19 Prophylactic HPV vaccines based on virus-like particles (VLPs) have been successfully tested in clinical trials. These vaccines are safe and effective in the prevention of persistent infection and subsequent pre-malignant cervical disease, and are now being introduced onto the market. 19 The presence of HPV in orophayngeal cancers may warrant the use of vaccination in boys as well as girls 9 15 years of age, for oral cancer prophylaxis. As HPV is associated with 1/5 of head and neck cancers, prophylaxis may result in a significant decrease in the incidence of disease. Table 4 Characteristics associated with HPV infection and logistic regression model for OSCC in NS/ND Chi-square analysis LR analysis Characteristic HPV+ % (n) P-value Logistic regression model unadjusted or adjusted or (95% CI) Age < 60 (30) 57 (17/30) 0.1 a Age > 60 (10) 30 (3/10) Male (19) 63 (12/19) 0.1 a Female (21) 38 (8/21) White (33) 48 (16/33) 0.7 a Other (7) 57 (14/7) Oropharynx (26) 65 (13/26) a 9.6 ( ) Oral cavity (14) 21 (3/14) 3.6 ( ) Cases Cancer (18) 78 (14/18) a Controls Benign W/CA (10) 30 (3/10) Benign W/out CA (12) 25 (33/12) HPV+ 9.3 ( ) 6.1 (1.3 28) OR, odds ratio; CI, confidence interval. a Chi-square analysis.

6 E. Andrews et al. / Oral Oncology 45 (2009) In this small case control study we have identified close links between HPV and OSCC in NS/ND. Early identification of this virus and prevention is desirable as this oncovirus may act as a crucial marker of malignancy susceptibility. Identifying patients with persistent infection and treating them with close follow up may also be important to reduce future cancers. The findings in this report have significant public health implications and will be important to the management and counseling of non-smoking/non-drinking oral cancer patients. Conflict of interest statement None declared. Acknowledgements We are indebted to study participants, to Webster-Cyriaque Lab Members for helpful discussion, to Carol Shores, to Marion Couch and to the NIAID ACTG/NIDCR OHARA 1 U01 AI /BRS- ACURE-Q T References 1. Shiboski CH, Schmidt BL, Jordan RC. Tongue and tonsil carcinoma: increasing trends in the US population ages years. Cancer 2005;103(9): Chaturvedi AK, Engels EA, Anderson WF, Gillison ML. Incidence trends for human papillomavirus related and unrelated oral squamous cell carcinomas in the United States. J Clin Oncol 2008;26: Blot WJ, McLaughlin JK, Winn DM, et al. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 1988;48(11): Fakhry C, Gillison ML. Clinical implications of human papillomavirus in head and neck cancers. J Clin Oncol 2006;24(17): Begum S, Cao D, Gillison M, Zahurak M, Westra WH. Tissue distribution of human papillomavirus 16 DNA integration in patients with tonsillar carcinoma. Clin Cancer Res 2005;11(16): Nazarenko I, Lowe B, Darfler M, Ikonomi P, Schuster D, Rashtchian A. Multiplex quantitative PCR using self-quenched primers labeled with a single fluorophore. Nucleic Acids Res 2002;30(9):e Fuessel Haws AL, He Q, Rady PL, et al. Nested PCR with the PGMY09/11 and GP5(+)/6(+) primer sets improves detection of HPV DNA in cervical samples. J Virol Meth 2004;122(1): Harwood CA, Spink PJ, Surentheran T, et al. Degenerate and nested PCR: A highly sensitive and specific method for detection of human papillomavirus infection in cutaneous warts. J Clin Microbiol 1999;37(11): D Souza G, Kreimer AR, Viscidi R, et al. Case control study of human papillomavirus and oropharyngeal cancer. N Engl J Med 2007;356(19): Pizzighella S, Pisoni G, Bevilacqua F, Vaona A, Palu G. Simultaneous polymerase chain reaction detection and restriction typing for the diagnosis of human genital papillomavirus infection. J Virol Meth 1995;55(2): Talbot SJ, Crawford DH. Viruses and tumours an update. Eur J Cancer 2004;40(13): Baseman JG, Koutsky LA. The epidemiology of human papillomavirus infections. J Clin Virol 2005;32(Suppl 1):S Sellers WR, Kaelin Jr WG. Role of the retinoblastoma protein in the pathogenesis of human cancer. J Clin Oncol 1997;15(11): Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189(1): Weinberger PM, Yu Z, Haffty BG, et al. Molecular classification identifies a subset of human papillomavirus-associated oropharyngeal cancers with favorable prognosis. J Clin Oncol 2006;24(5): Pintos J, Black MJ, Sadeghi N, et al. Human papillomavirus infection and oral cancer: a case control study in Montreal, Canada. Oral Oncol. 17. Dahlstrom KR, Adler-Storthz K, Etzel CJ, et al. Human papillomavirus type 16 infection and squamous cell carcinoma of the head and neck in never-smokers: a matched pair analysis. Clin Cancer Res 2003;9(7): Gillison ML. Human papillomavirus-associated head and neck cancer is a distinct epidemiologic, clinical, and molecular entity. Semin Oncol 2004;31(6): Koutsky LA, Ault KA, Wheeler CM, et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 2002;347(21):