J Clin Oncol 27: by American Society of Clinical Oncology INTRODUCTION

VOLUME 27 NUMBER 12 APRIL 09 JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Effect of HPV-Associated p16 INK4A Expression on Response to Radiotherapy and Survival in Squamous Cell Carcinoma of the Head and Neck Pernille Lassen, Jesper G. Eriksen, Stephen Hamilton-Dutoit, Trine Tramm, Jan Alsner, and Jens Overgaard From the Department of Experimental Clinical Oncology and Institute of Pathology, Aarhus University Hospital, Denmark. Submitted September 25, 08; accepted December 11, 08; published online ahead of print at www.jco.org on March 16, 09. Written on behalf of the Danish Head and Neck Cancer Group. Supported by grants from the Danish Cancer Society, The Danish Council for Strategic Research, and the Lundbeck Foundation Research Centre for Interventional Research in Radiation Oncology. Presented in part at the International Meeting on Innovative Approaches in Head and Neck Oncology, February 22-24, 07, Barcelona, Spain. Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Corresponding author: Pernille Lassen, MD, Department of Experimental Clinical Oncology, Aarhus University Hospital, Noerrebrogade 44, DK-8000 Aarhus C, Denmark; e-mail: pernille@oncology.dk. A B S T R A C T Purpose A subset of head and neck cancers is associated with the human papillomavirus (HPV). Viral infection is closely correlated with expression of p16 INK4A in these tumors. We evaluated p16 INK4A as a prognostic marker of treatment response and survival in a well-defined and prospectively collected cohort of patients treated solely with conventional radiotherapy in the Danish Head and Neck Cancer Group (DAHANCA) 5 trial. Patients and Methods Immunohistochemical expression of p16 INK4A was analyzed in pretreatment paraffin-embedded tumor blocks from 156 patients treated with conventional primary radiotherapy alone. The influence of p16 INK4A status on locoregional tumor control, disease-specific survival, and overall survival after radiotherapy was evaluated. Results p16 INK4A positivity was found in 35 tumors (22%). Tumor-positivity for p16 INK4A was significantly correlated with improved locoregional tumor control (5-year actuarial values 58% v 28%; P.0005), improved disease-specific survival (72% v 34%; P.0006), and improved overall survival (62% v 26%; P.0003). In multivariate analysis, p16 INK4A remained a strong independent prognostic factor for locoregional failure (hazard ratio [HR], 0.35; 95% CI, 0.19 to 0.64), disease-specific death (HR, 0.36; 95% CI, 0. to 0.64), and overall death (HR, 0.44; 95% CI, 0.28 to 0.68). Conclusion Expression of p16 INK4A has a major impact on treatment response and survival in patients with head and neck cancer treated with conventional radiotherapy. J Clin Oncol 27:1992-1998. 09 by American Society of Clinical Oncology 09 by American Society of Clinical Oncology 0732-183X/09/2712-1992/$.00 DOI: 10.10/JCO.08..2853 INTRODUCTION Although tobacco and alcohol consumption are the primary risk factors for development of head and neck squamous cell carcinoma (HNSCC), it has recently become evident that infection with high-risk human papillomavirus (HPV) is etiologically linked to a subgroup of cancers in the head and neck region. 1-3 Some 90% of HPV-positive head and neck tumors are infected with high-risk HPV16. 4,5 Viral DNA is detected in some % of HNSCCs, the strongest association being found in oropharyngeal tumors, where carcinomas of the tonsils are particularly associated with HPV infection. 4-7 Recent studies have indicated that in HNSCC, patients with HPV-infected tumors have a more favorable prognosis compared with patients whose tumors are virus-negative. 8-11 HPV positivity seems to be associated with lower exposure to tobacco and alcohol and with younger age at time of diagnosis, factors that may by themselves positively influence prognosis regardless of tumor biology. However, evidence is accumulating that HPV-positive tumors represent a separate biologic subgroup of HNSCC distinct from those carcinomas that are induced by tobacco and alcohol. The biologic differences between these subtypes of HNSCC may per se have an impact on prognosis. 1,12 The viral oncogene products E6 and E7 play a key role in HPV-associated carcinogenesis, abrogating p53 and retinoblastoma (Rb) tumor suppressor functions, respectively. 13,14 HPV-positive HNSCCs are generally associated with wild-type TP53, as opposed to tobacco-induced tumors, which are characterized by genetic alterations of the TP53 pathway. 12,13 The functional inactivation of Rb by E7 leads in turn to upregulation of the p16 INK4A 1992 09 by American Society of Clinical Oncology

HPV-Associated p16 INK4A Expression in HNSCC protein. p16 INK4A, encoded by the CDKN2A tumor suppressor gene, regulates the activity of Cyclin D-CDK4/6 complexes (CDKs). When not inactivated by p16 INK4A, these CDKs phosphorylate Rb, resulting in release of the transcription factor E2F, which initiates cell cycle progression. 15 In contrast, the bound Rb-E2F protein complex acts as a negative regulator, inhibiting transcription of several genes, including CDKN2A. The functional inactivation of Rb by E7 therefore results in release of the p16 INK4A gene from its transcriptional inhibition. 16 Thus HPV-positive tumors are characterized by high expression of p16 INK4A. 8-10,17 Moreover, because transcription of the E7 oncogene is required for p16 INK4A upregulation, it has been suggested that carcinomas overexpressing p16 INK4A represent those tumors in which HPV has been involved in the carcinogenic process. 18 Thus there is good evidence that p16 INK4A positivity may be regarded as a biomarker for tumors harboring clinically and oncogenetically relevant HPV infections. HNSCC is predominantly a locoregional disease, and achieving local tumor control is essential for survival. 19 Radiotherapy can significantly improve local control and disease-specific survival in patients with HNSCC, either when used as a single treatment modality or in combination with chemotherapy or hypoxic modification.,22 Over the years, radiotherapy has been optimized through modification of tumor hypoxia and by reducing overall treatment time of radiotherapy. -24 Although this has resulted in improved treatment response for patients with HNSCC as a whole, individual patients with HNSCC still show considerable variation in clinical outcome. This indicates that other factors defining tumor response are still to be discovered, 25,26 and HPV infection may be one such factor. With the present study, we aimed to evaluate the impact of HPV-associated tumor cell p16 INK4A expression on response to radiotherapy and survival in a well-characterized and prospectively collected cohort of Danish patients with head and neck cancer treated with conventionally fractionated radiotherapy alone. PATIENTS AND METHODS Patients and Tissues One hundred ninety-five patients with primary squamous cell carcinoma of the pharynx (stages I through IV) or supraglottic larynx (stages II through IV) were enrolled in the placebo arm of the randomized Danish Head and Neck Cancer Group (DAHANCA) 5 protocol from January 1986 to September 1990. 22 Patients in this arm received primary conventionally fractionated radiotherapy (66 to 68 Gy in 33 to 34 fractions, 5 fractions/wk) as the only treatment. Routine paraffin-embedded, formalin-fixed pretreatment tumor tissues were available from the pathology archives in sufficient amounts for study in 156 of the total 195 patients (80%).The DAHANCA 5 study was designed according to the Helsinki Declaration II, and the patients gave written informed consent. Both the main study and the tumor tissue analyses were approved by the local ethics committees according to Danish law and regulations. Evaluation of p16 INK4A Expression Immunohistochemistry for p16 INK4A expression was performed on a BenchMark XT autostainer (Ventana Medical Systems, Illkirch, France) according to the manufacturer s recommendations. Briefly, paraffin sections were cut at 5 m on Superfrost plus charged glass slides (Menzel-Glaser, Braunschweig, Germany), heated at 60 C for 1 hour, and deparaffinized in the instrument with EZ prep solution (Ventana Medical Systems). Heat-induced antigen retrieval was carried out using Cell Conditioning 1 solution (CC1, Ventana Medical Systems). p16 INK4A was detected by incubating sections with antibody clone JC8 (sc-563; Santa Cruz Biotechnology Inc, Santa Cruz, CA) diluted 1:25 for 32 minutes. This is a mouse monoclonal immunoglobulin G2a antibody raised against full-length human p16, particularly suitable for use on formalin-fixed, paraffin-embedded sections. Its specificity has been confirmed by Western blotting. 27 A B C D Fig 1. (A) HPV16 DNA-positive tonsillar carcinoma. Numerous tumor cells show positive nuclear signals. (B) The same tumor showing strong specific tumor cell expression of p16 INK4A. (C) HPV16-negative tonsillar carcinoma. (D) The same tumor showing no expression of p16 INK4A. (A, C) HPV16 DNA in situ hybridization. (B, D) Immunohistochemistry with antip16 INK4A antibody clone JC8. www.jco.org 09 by American Society of Clinical Oncology 1993

Lassen et al HPV16 DNA Table 1. Correlation Between HPV16 DNA In Situ Hybridization and p16 INK4A Immunohistochemistry in Tonsillar Carcinomas Positive p16 INK4A Status Negative Total Positive 19 1 Negative 2 10 12 Total 11 32 Specific reactions were detected using ultraview Universal DAB Detection Kit (Ventana Medical Systems), and the slides were counterstained with hematoxylin. Sections of p16 INK4A -positive cervical carcinoma were used as positive controls. p16 INK4A expression was associated with distinct diffuse nuclear and cytoplasmic staining of epithelial cells. Tumors were classified dichotomously as either p16 INK4A -positive (strong, diffuse nuclear and cytoplasmic staining in 10% of carcinoma cells) or negative (Fig 1). Detection of HPV Infection To demonstrate the correlation between HPV16 infection and tumor cell expression of p16 INK4A, we performed a pilot study on a separate series of 32 tonsillar carcinomas, using in situ hybridization for detection of viral DNA. Briefly, 5- m formalin-fixed, paraffin-embedded tumor sections were deparaffinized and treated with Dako Target Retrieval Solution (Dako, Glostrup, Denmark). Sections were digested with proteinase K (Dako, S04) and rinsed in deionized water. Biotinylated HPV16-type specific DNA probe (Dako, code Y1407) in hybridization mixture was used according to the manufacturer s recommendations. Denaturation and hybridization were carried out using a Hybridizer Instrument (Dako). Sections were denatured for 5 minutes at 92 C and hybridized overnight at 37 C. After stringent washing, hybridized probe was detected using the tyramide-based GenPoint Catalyzed Signal Amplification System (Dako, code K06), with consecutive application of primary peroxidase-conjugated streptavidin, biotinyl tyramide, and secondary peroxidase-conjugated streptavidin. Signals were visualized with the chromogen 3,3-diaminobenzidine. Positive and negative control sections (SiHa, HeLa, bladder cancer J82) were included. Carcinomas were classified as HPV16 positive when a discrete signal was seen localized to the nuclei of tumor cells. Signals were categorized as either punctuate and/or diffuse, representing integrated and episomal virus, respectively. 28- The findings of the pilot study have been previously published in abstract form. We found a strong correlation between HPV16 DNA detection and p16 INK4A expression (Fig 1, Table 1). Statistical Analysis The end points used were locoregional control after radiotherapy, disease-specific survival, and overall survival. Locoregional control was defined as complete and persistent disappearance of the disease in the primary tumor (T site) and regional lymph nodes (N site) after radiotherapy. Failure was recorded in the event of recurrent tumor or if the primary tumor never completely disappeared. In the latter situation, the tumor was then assumed to have failed at the time of randomization. The end point does not, therefore, include the effect of a successful procedure with salvage surgery. Diseasespecific survival was defined as death from or with the actual cancer, and overall survival was defined as death from any cause. Follow-up of the patients was completed in connection with the original publication of the study. 22 All patients were observed for at least 5 years or until death. The survival status has subsequently been tracked until June 04. Features of p16 INK4A -positive and -negative tumors were compared using the 2 test for categoric variables. The actuarial values of the end points were evaluated by the Kaplan-Meier analysis and compared using the log-rank test. A multivariate Cox proportional hazard analysis was used to evaluate prognostic parameters and treatment with respect to the risk of locoregional Table 2. Patient and Tumor Characteristics by p16 INK4A Expression Status All Patients p16 INK4A Positive p16 INK4A Negative Patient/Tumor Data No. % No. % No. % P No. of patients 156 35 1 Age, years Median 60 57 60 NS Range 24-84 34-83 24-84 Sex Female 41 26 12 34 29 24 NS Male 115 74 23 66 92 76 Tumor site Supraglottic larynx 47 9 26 38.01 Hypopharynx 13 0 0 17 Oropharynx 74 47 24 69 50 41 Rhinopharynx 15 10 2 5 13 11 Tumor stage T1-2 71 46 19 54 52 43 NS T3-4 85 54 16 46 69 57 Nodal status N0 63 40 13 37 50 41 NS N1-3 93 60 22 63 71 59 Disease stage I-II 19 6 17 24 NS III-IV 126 81 29 83 97 80 Tumor differentiation Well or moderate 92 59 16 46 75 62 NS Poor 64 41 19 54 46 38 Abbreviation: NS, not significant. International Union of Cancer Research 1982 classification. 1994 09 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

HPV-Associated p16 INK4A Expression in HNSCC failure, disease-specific death, and overall death. Analyses were done with the STATA statistical package, version 9 (STATA Corp, College Station, TX). Comprehensive Meta-Analysis version 2.0 (Biostat, Englewood, NJ) was used to compose the Forest Plot. All results were considered significant at levels less than 5% (two-sided tests), and odds ratios (OR) are presented with 95% CIs. RESULTS Thirty-five (22%) of the 156 tumors expressed p16 INK4A. Table 2 shows patient and tumor characteristics according to p16 INK4A status. The highest frequency of p16 INK4A expression was seen in oropharyngeal tumors, where 24 (32%) of 74 tumors were p16 INK4A -positive. Moreover, 69% of the p16 INK4A -positive tumors originated in the oropharynx compared with only 41% of the p16 INK4A -negative group (P.004). Median age and sex did not differ significantly between the groups, and the vast majority of patients in both groups were diagnosed in disease stage III or IV. Similarly, no statistically significant difference between the groups regarding histopathologic tumor differentiation was found. At the time of evaluation, 97 patients had failed to achieve persistent locoregional control within the irradiated volume. A total of 93 patients had died with or from the cancer in question, and overall, 134 patients had died. The number of locoregional failures was associated with high stage of disease (stage III to IV; P.05) and, to a lesser extent, to nodal spread (P.08) and high tumor classification (T3/4; P.09). Patients with p16 INK4A -positive tumors were significantly less likely to suffer locoregional recurrence than were patients with p16 INK4A -negative tumors (P.001). Figure 2 shows the probability of locoregional tumor control 5 years after radiotherapy according to p16 INK4A status. A statistically significant improvement in locoregional tumor control was found in Locoregional Tumor Control (%) 100 90 80 70 60 50 40 10 No. of patients 0 1 35 Event All 84 1 13 35 OR = 0.26 (0.12 to 0.57) Time Since Treatment (months) 55 28 39 23 34 19 26 18 P =.0005 12 24 36 48 60 Fig 2. Actuarial estimated locoregional tumor control in patients with p16 INK4A - positive and p16 INK4A -negative carcinoma of the pharynx and supraglottic larynx treated with radiotherapy alone. neg, negative; pos, positive. 12 58% 28% patients with p16 INK4A -positive tumors as compared with patients with p16 INK4A -negative tumors (58% v 28%, respectively; P.0005; OR, 0.26; 95% CI, 0.12 to 0.57). Looking in detail at locoregional tumor control after radiotherapy, tumor expression of p16 INK4A is seen to be associated with benefit in all subgroups; however, this benefit is greater in certain subgroups compared with others (Fig 3). The more advanced the disease (N, stage III to IV), the greater the benefit. Patients with histopathologically poorly differentiated tumors or with tumors of oropharyngeal origin profit more from having tumor cell p16 INK4A positivity than do patients with moderately or well-differentiated tumors or with tumors of nonoropharyngeal origin. Moreover, women with p16 INK4A -positive tumors seem to have a rather pronounced advantage over men with p16 INK4A -positive tumors with regard to locoregional tumor control after radiotherapy. The effect of the improved locoregional tumor control in the p16 INK4A -positive group is reflected in the survival outcome parameters. Survival curves demonstrating the disease-specific and overall survival benefits of patients according to p16 INK4A expression status are shown in Figure 4. Patients with p16 INK4A -positive tumors had significantly lower disease-specific mortality (OR, 0.; 95% CI, 0.14 to 0.66) and overall mortality (OR, 0.22; 95% CI, 0.08 to 0.56) than did patients with p16 INK4A -negative tumors. In the final Cox proportional hazards analysis with risk of locoregional failure as the end point, low tumor classification (T1/2, hazard ratio [HR], 0.56; 95% CI, 0.37 to 0.84), negative neck nodes (HR, 0.51; 95% CI, 0.33 to 0.78), and p16 INK4A expression (HR, 0.35; 95% CI, 0.19 to 0.64) were independent factors associated with a good prognosis. Similarly, the same three parameters were independent prognostic factors when analyzed with death from cancer and overall death as end points: T1/2 tumors, HR, 0.61 (95% CI, 0.40 to 0.92) and HR, 0.66 (95% CI, 0.46 to 0.93), respectively; negative nodes, HR, 0.46 (95% CI, 0. to 0.71) and HR, 0.52 (95% CI, 0.37 to 0.75), respectively; and p16 INK4A expression, HR, 0.36 (95% CI, 0. to 0.64) and HR, 0.44 (95% CI, 0.28 to 0.68), respectively. No other parameters were independently associated with treatment outcome. p16 INK4A status Odds ratio (95% CI) Neg Pos All 84/1 13/35 3.8 (1.8 to 8.4) Nodal status N0 29/50 5/13 2.2 (0.6 to 7.7) N1-3 55/71 8/22 6.0 (2.1 to 16.9) T Classification T 1-2 33/52 6/19 3.8 (1.2 to 11.5) T 3-4 51/69 7/16 3.6 (1.2 to 11.2) Stage Stage 1-2 12/24 2/6 2.0 (0.3 to 13.1) Stage 3-4 72/97 11/29 4.7 (2.0 to 11.3) Gender Female /29 2/12 11.1 (2.0 to 61.4) Male 64/92 11/23 2.5 (1.0 to 6.3) Differentiation Low 33/46 5/18 6.6 (1.9 to 22.2) Mod/high 51/75 8/17 2.4 (0.8 to 7.0) Tumor site Oropharynx 36/50 8/24 5.1 (1.8 to 14.7) Nonoropharynx 48/71 5/11 2.5 (0.7 to 9.1) 0.2 0.5 1 2 5 10 p16 INK4A neg better p16 INK4A pos better Fig 3. The relationship between p16 INK4A expression and locoregional tumor control by subgroup. neg, negative; pos, positive. www.jco.org 09 by American Society of Clinical Oncology 1995

Lassen et al A Disease-Specific Survival (%) 100 90 80 70 60 50 40 10 0 No. of patients B Overall Survival (%) 100 90 80 70 60 50 40 10 0 No. of patients 1 35 1 35 Event All 80 1 13 35 OR = 0. (0.14 to 0.66) Time Since Treatment (months) 81 61 DISCUSSION We analyzed the association of tumor p16 INK4A expression with therapeutic response in a prospectively collected cohort of patients with HNSCC treated with radiotherapy alone. This study shows that tumor cell expression of p16 INK4A in this group of patients is associated with markedly improved treatment response and survival and that p16 INK4A expression status is an independent prognostic factor for both outcomes. The potent prognostic value of p16 INK4A expression is 46 23 36 P =.0006 12 24 36 48 60 Event All 110 1 24 35 OR = 0.22 (0.08 to 0.56) Time Since Treatment (months) 81 61 46 23 36 P =.0003 12 24 36 48 60 Fig 4. (A) Actuarial estimated disease-specific and (B) overall survival rates in patients with p16 INK4A -positive and p16 INK4A -negative carcinoma of the pharynx and supraglottic larynx treated with radiotherapy alone. neg, negative; pos, positive. 72% 34% 62% 26% emphasized by the fact that this marker proved to be an even stronger independent prognostic factor than the classical clinical parameters of tumor stage and nodal status, the only other independent variables of prognostic significance in this study. The beneficial locoregional control rate observed in the p16 INK4A -positive group indicates that these tumors respond well to conventionally fractionated radiotherapy. The biologic basis for this observation could reside in intrinsic tumor cell differences that render p16 INK4A -positive cells more susceptible to radiation than p16 INK4A - negative tumor cells. A possible explanation could be that tumor cells expressing p16 INK4A are less hypoxic and potentially respond with less pronounced accelerated repopulation when irradiated, but the precise mechanisms are not known at the moment. Similarly, it is also not clear to what degree patients with p16 INK4A -positive tumors benefit from hypoxic modification and accelerated fractionation treatment schedules, and this is an area that warrants further investigation. In the studies by Kumar et al 11 and by Fakhry et al, 8 patients were treated with induction chemotherapy followed by concomitant chemoradiotherapy. Both groups reported treatment response rates and survival benefits for HPV-positive tumors similar to those found in our study. This indicates that p16 INK4A -positive tumors in general have a better response to treatment than p16 INK4A - negative tumors, regardless of treatment modality, but that the use of radiation either as monotherapy or in combination with chemotherapy seems efficacious. The survival benefits observed for patients with p16 INK4A - positive tumors in this study can partly be ascribed to the improved locoregional control rate in this group. Analysis of known patient and tumor characteristics identified no substantial differences between the two patient groups. In particular, p16 INK4A -positive tumors seemed to be more closely associated with poor histopathologic differentiation (54%) compared with p16 INK4A -negative tumors (38%), but the difference was not significant at the 5% level, possibly because of lack of power. The majority of patients in both groups had nodal spread at time of diagnosis (63% and 59%, respectively), which reflects that patients with HNSCC in general are diagnosed at advanced stages regardless of p16 INK4A status. However, patients with p16 INK4A -positive tumors could potentially have less comorbidity and a different risk profile in terms of exposure to tobacco and alcohol than the patients in the p16 INK4A -negative group. Information on lifetime exposure to tobacco and alcohol and performance status was, however, not available. Therefore, confounding by these variables on the survival analysis cannot be completely ruled out. Recent studies have demonstrated the correlation between HPV and p16 INK4A expression status in HNSCC. 9,10,17,32 Using HPV16 DNA in situ hybridization and p16 INK4A immunohistochemistry on parallel sections in a series of tonsillar carcinomas, we confirmed these findings. We believe that p16 INK4A expression is a reliable biomarker of infection with HPV in HNSCC. However, the importance of p16 INK4A may potentially extend beyond that of merely being a surrogate marker of infection with HPV. From the molecular processes involved in viral carcinogenesis, it is possible to argue the case that p16 INK4A positivity may specifically identify HPV infections in HNSCC that are biologically relevant in carcinogenesis (ie, infections in those tumors in which HPV is transcriptionally active). A similar hypothesis was proposed and tested by Weinberger et al. 10 They demonstrated that double positivity in tumors for both HPV infection and 1996 09 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

HPV-Associated p16 INK4A Expression in HNSCC p16 INK4A expression identified those tumors with a favorable prognosis and that the presence of HPV in the tumors per se did not have a substantial positive impact on prognosis. It should be noted that expression of p16 INK4A is not limited to HPV-positive tumors, and using this marker alone as an indicator of biologically relevant HPV infections inevitably entails the risk of including some (virally) false-positive results. Recently, Smeets et al 17 proposed an algorithm for reliable detection of biologically relevant HPV infections in paraffin-embedded head and neck cancer specimens. Initial evaluation of tumors by p16 INK4A immunohistochemistry, followed by polymerase chain reaction based detection of HPV in the p16 INK4A -positive cases, produced a sensitivity and specificity of 100%. The risk of both assays yielding a false-positive result in the same sample was estimated to be 2%. We believe there is sufficient circumstantial evidence to support the view that the p16 INK4A -positive tumors in our cohort of patients with head and neck cancer represent those carcinomas infected with HPV. The association between HPV and HNSCC is strongest for oropharyngeal carcinomas and, in particular, for carcinomas of the tonsil. Consistently in this cohort, tumors of oropharyngeal origin had the highest frequency of p16 INK4A expression, and a detection rate of 32% in the oropharynx is in good agreement with viral detection rates observed in Swedish tonsillar carcinoma specimens from the late 1980s. 33 According to the Danish cancer registry, there has been a threefold increase in the incidence of tonsillar cancer in Denmark over the past years, and similar findings are reported from other parts of the Western world. 33-36 This increase has occurred at the same time as there has been a pronounced decrease in the classical tobacco-induced carcinomas of the larynx and lung, and it is thought that this increase in tonsillar cancer may be attributable to an increase in HPVassociated tumors. 37 If this hypothesis proves correct, then it suggests that we may soon face a substantial increase in the number of oropharyngeal carcinomas in the future, as the importance of HPV infection becomes apparent. In that setting, it will be essential to acknowledge the importance of the strong and so far not duly appreciated influence of p16 INK4A expression on prognosis for patients with head and neck cancer treated with radiotherapy. In particular, p16 INK4A status needs to be taken into account whenever a clinical trial is conducted and interpreted. Failure to recognize the importance of p16 INK4A could lead to confounding, and p16 INK4A status should be included as a prerandomization stratification factor in future clinical trials conducted on patients with head and neck cancer treated with radiotherapy. The majority of patients with p16 INK4A -positive tumors are diagnosed at advanced clinical stages. Despite having large tumors and nodal involvement, these patients respond extremely well to radiotherapy, which generally seems to make them complete responders and long-term survivors. Thus patients with p16 INK4A -positive tumors are candidates for organ preservative therapeutic procedures, and the approach to elective neck dissection after radiotherapy in these patients should be rather cautious to minimize long-term morbidity of therapy. In conclusion, our study shows that p16 INK4A expression status should be a major determinant of future therapeutic strategies in patients with HNSCC. AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. AUTHOR CONTRIBUTIONS Conception and design: Pernille Lassen, Jesper G. Eriksen, Stephen Hamilton-Dutoit, Jan Alsner, Jens Overgaard Financial support: Jens Overgaard Administrative support: Jens Overgaard Provision of study materials or patients: Pernille Lassen, Jesper G. Eriksen, Jens Overgaard Collection and assembly of data: Pernille Lassen, Jesper G. Eriksen, Jens Overgaard Data analysis and interpretation: Pernille Lassen, Jesper G. Eriksen, Stephen Hamilton-Dutoit, Trine Tramm, Jan Alsner, Jens Overgaard Manuscript writing: Pernille Lassen, Jesper G. Eriksen, Stephen Hamilton-Dutoit, Trine Tramm, Jan Alsner, Jens Overgaard Final approval of manuscript: Pernille Lassen, Jesper G. Eriksen, Stephen Hamilton-Dutoit, Trine Tramm, Jan Alsner, Jens Overgaard REFERENCES 1. Gillison ML, Koch WM, Capone RB, et al: Evidence for a causal association between human papillomavirus and a subset of head and neck cancers. J Natl Cancer Inst 92:709-7, 00 2. Andl T, Kahn T, Pfuhl A, et al: Etiological involvement of oncogenic human papillomavirus in tonsillar squamous cell carcinomas lacking retinoblastoma cell cycle control. Cancer Res 58:5-13, 1998 3. Argiris A, Karamouzis MV, Raben D, et al: Head and neck cancer. Lancet 371:1695-1709, 08 4. Fakhry C, Gillison ML: Clinical implications of human papillomavirus in head and neck cancers. J Clin Oncol 24:2606-2611, 06 5. Kreimer AR, Clifford GM, Boyle P, et al: Human papillomavirus types in head and neck squamous cell carcinomas worldwide: A systematic review. Cancer Epidemiol Biomarkers Prev 14:467-475, 05 6. Begum S, Cao D, Gillison M, et al: Tissue distribution of human papillomavirus 16 DNA integration in patients with tonsillar carcinoma. Clin Cancer Res 11:5694-5699, 05 7. Klussmann JP, Weissenborn SJ, Wieland U, et al: Human papillomavirus-positive tonsillar carcinomas: A different tumor entity? Med Microbiol Immunol (Berl) 192:129-132, 03 8. Fakhry C, Westra WH, Li S, et al: Improved survival of patients with human papillomaviruspositive head and neck squamous cell carcinoma in a prospective clinical trial. J Natl Cancer Inst 100: 261-269, 08 9. Licitra L, Perrone F, Bossi P, et al: High-risk human papillomavirus affects prognosis in patients with surgically treated oropharyngeal squamous cell carcinoma. J Clin Oncol 24:56-5636, 06 10. 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 24:736-747, 06 11. Kumar B, Cordell KG, Lee JS, et al: EGFR, p16, HPV Titer, Bcl-xL and p53, sex, and smoking as indicators of response to therapy and survival in oropharyngeal cancer. J Clin Oncol 26:28-37, 08 12. Braakhuis BJ, Snijders PJ, Keune WJ, et al: Genetic patterns in head and neck cancers that contain or lack transcriptionally active human papillomavirus. J Natl Cancer Inst 96:998-1006, 04 13. zur Hausen H: Papillomaviruses and cancer: From basic studies to clinical application. Nat Rev Cancer 2:342-350, 02 14. Münger K, Howley PM: Human papillomavirus immortalization and transformation functions. Virus Res 89:3-228, 02 15. Nevins JR: The Rb/E2F pathway and cancer. Hum Mol Genet 10:699-703, 01 16. Khleif SN, DeGregori J, Yee CL, et al: Inhibition of cyclin D-CDK4/CDK6 activity is associated with an E2F-mediated induction of cyclin kinase inhibitor activity. Proc Natl Acad Sci U S A 93:4350-4354, 1996 www.jco.org 09 by American Society of Clinical Oncology 1997

Lassen et al 17. Smeets SJ, Hesselink AT, Speel EJ, et al: A novel algorithm for reliable detection of human papillomavirus in paraffin embedded head and neck cancer specimen. Int J Cancer 1:2465-2472, 07 18. von Knebel Doeberitz M: New markers for cervical dysplasia to visualise the genomic chaos created by aberrant oncogenic papillomavirus infections. Eur J Cancer 38:2229-2242, 02 19. Overgaard J, Hansen HS, Jorgensen K, et al: Primary radiotherapy of larynx and pharynx carcinoma: An analysis of some factors influencing local control and survival. Int J Radiat Oncol Biol Phys 12:515-5, 1986. Pignon JP, Bourhis J, Domenge C, et al: Chemotherapy added to locoregional treatment for head and neck squamous-cell carcinoma: Three meta-analyses of updated individual data MACH-NC Collaborative Group. Meta-Analysis of Chemotherapy on Head and Neck Cancer. Lancet 355:949-955, 00. Overgaard J: Hypoxic radiosensitization: Adored and ignored. J Clin Oncol 25:4066-4074, 07 22. Overgaard J, Hansen HS, Overgaard M, et al: A randomized double-blind phase III study of nimorazole as a hypoxic radiosensitizer of primary radiotherapy in supraglottic larynx and pharynx carcinoma: Results of the Danish Head and Neck Cancer Study (DAHANCA) Protocol 5-85. Radiother Oncol 46:135-146, 1998 23. Bourhis J, Overgaard J, Audry H, et al: Hyperfractionated or accelerated radiotherapy in head and neck cancer: A meta-analysis. Lancet 368:843-854, 06 24. Overgaard J, Hansen HS, Specht L, et al: Five compared with six fractions per week of conventional radiotherapy of squamous-cell carcinoma of head and neck: DAHANCA 6 and 7 randomised controlled trial. Lancet 362:933-940, 03 25. Overgaard J, Eriksen JG, Nordsmark M, et al: Plasma osteopontin, hypoxia, and response to the hypoxia sensitiser nimorazole in radiotherapy of head and neck cancer: Results from the DAHANCA 5 randomised double-blind placebo-controlled trial. Lancet Oncol 6:757-764, 05 26. Eriksen JG, Overgaard J: Lack of prognostic and predictive value of CA IX in radiotherapy of squamous cell carcinoma of the head and neck with known modifiable hypoxia: An evaluation of the DAHANCA 5 study. Radiother Oncol 83:383-388, 07 27. Santa Cruz Biotechnology: Datasheet anti-p16 clone JC8. http://www.scbt.com/datasheet-563- p16-jc8-antibody.html 28. Cooper K, Herrington CS, Stickland JE, et al: Episomal and integrated human papillomavirus in cervical neoplasia shown by non-isotopic in situ hybridisation. J Clin Pathol 44:990-996, 1991 29. Kristiansen E, Jenkins A, Holm R: Coexistence of episomal and integrated HPV16 DNA in squamous cell carcinoma of the cervix. J Clin Pathol 47:253-256, 1994. Berumen J, Unger ER, Casas L, et al: Amplification of human papillomavirus types 16 and 18 in invasive cervical cancer. Hum Pathol 26:676-681, 1995. Lassen P, Eriksen JG, Tramm T, et al: Effect of HPV on treatment outcome and survival in oropharyngeal carcinoma: Analysis of the randomized DHANCA 5&7 trials. Presented at the 7th International Conference on Head and Neck Cancer, San Francisco, CA, July 19-23, 08 (abstr S024) 32. Reimers N, Kasper HU, Weissenborn SJ, et al: Combined analysis of HPV-DNA, p16 and EGFR expression to predict prognosis in oropharyngeal cancer. Int J Cancer 1:17-1738, 07 33. Hammarstedt L, Lindquist D, Dahlstrand H, et al: Human papillomavirus as a risk factor for the increase in incidence of tonsillar cancer. Int J Cancer 119:26-2623, 06 34. Hammarstedt L, Dahlstrand H, Lindquist D, et al: The incidence of tonsillar cancer in Sweden is increasing. Acta Otolaryngol 127:988-992, 07 35. Shiboski CH, Schmidt BL, Jordan RC: Tongue and tonsil carcinoma: Increasing trends in the U.S. population ages -44 years. Cancer 103:1843-1849, 05 36. Frisch M, Hjalgrim H, Jaeger AB, et al: Changing patterns of tonsillar squamous cell carcinoma in the United States. Cancer Causes Control 11:489-495, 00 37. Sturgis EM, Cinciripini PM: Trends in head and neck cancer incidence in relation to smoking prevalence: An emerging epidemic of human papillomavirus-associated cancers? Cancer 110: 1429-1435, 07 1998 09 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY