Overexpression of p63 is associated with radiation resistance and prognosis in oral squamous cell carcinoma

Oral Oncology 46 (2010) 667–671

Contents lists available at ScienceDirect

Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology

Overexpression of p63 is associated with radiation resistance and prognosis in oral squamous cell carcinoma Maximilian Moergel a,*, Eva Abt a, Marcus Stockinger b, Martin Kunkel c a

Department of Oral and Maxillofacial Surgery, University of Mainz, Medical Center, Augustusplatz 2, 55131 Mainz, Germany Department of Radiation Oncology, University of Mainz, Medical Center, Langenbeckstr. 1, 55101 Mainz, Germany c Department of Oral and Maxillofacial Surgery, University of Bochum Medical Center, Knappschaftskrankenhaus Bochum Langendreer, In der Schonau 23-25, 44892 Bochum, Germany b

a r t i c l e

i n f o

Article history: Received 29 April 2010 Received in revised form 24 June 2010 Accepted 24 June 2010 Available online 24 July 2010 Keywords: p63 Head and neck cancer Radiation Radiation resistance Oral cancer

s u m m a r y Background: The tumor suppressor homologue p63 is expressed in basal and parabasal layers of intraoral mucosa. Full length transcripts with transactivational domain (TA forms) present with homology to p53 and implicate functions governing cell proliferation, differentiation and apoptosis control. To date studies show an increase of p63 expression in oral dysplasia and additionally high expression levels correlated with poor prognosis for patients with OSCC, whereas a possible link to radiation resistance of tumors has not been investigated yet. In the present study we tested the hypothesis for p63 being a marker of radioresistance and overall survival in OSCC. Methods: p63 Expression was labeled by immunohistochemistry in pre-treatment biopsies collected from 33 patients with OSCC. Quantification of the labeling index (Li) based on the relation of p63 positive cells to overall tumor cell count. Histological examination of the resection specimen allowed categorization of the radiation response. Statistical analyses of the association between Li and radiation response were performed. Survival analysis utilized Kaplan–Meier estimates and additionally a Cox regression model was built for p63 (Li), T stage, N-stage and chemotherapy and presented as hazard ratios. Results: All tumors had enhanced p63 expression. The median Li was 63.1% (range 36–87%). Tumors with a p63 positive cell count >63.1% showed increased resistance to radiation (p = 0.027). Overall survival was higher (p = 0.001) for patients with low Li (median value) and multivariate Cox regression analysis confirmed the significance of p63 as a prognostic marker of survival. Conclusions: The results of this analysis advocate p63 expression in pre-treatment tumor tissue to be a marker of radiation resistance in OSCC, with high expression levels being associated with poor radiation response and shorter survival. The promising results of this biomarker should now be confirmed by a study with larger patient counts. Ó 2010 Elsevier Ltd. All rights reserved.

Introduction The p63 gene located on chromosome 3q27–29 encodes a group of six proteins with structural homology to the tumor suppressor p53.1 Physiologically p63 is expressed in progenitor cells within the basal lamina of epithelial tissues and p63-/- Knock-out mice revealed crucial involvement of p63 in epithelial proliferation and differentiation.2 Further analysis of p53- and p63-sequences Abbreviations: p63, protein 63; CAI-X, carboanhydrase X; Li, labeling index; DÖSAK, deutsch österreichisch schweizerischer Arbeitskreis für Tumoren im Kieferund Gesichtsbereich; OSSC, oral squamous cell carcinoma; H&NSSC, squamous cell carcinoma of the head and neck; PBS, phosphate buffered saline; CR, complete response; GPR, good partial response; MPR, minor partial response. * Corresponding author. Address: Department of Oral and Maxillofacial Surgery, Johannes Gutenberg-University, Medical Center, Augustusplatz 2, 55131 Mainz, Rheinland-Pfalz, Germany. Tel.: +49 6131 175459; fax: +49 6131 176602. E-mail address: [email protected] (M. Moergel). 1368-8375/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.oraloncology.2010.06.012

revealed similar evolutionary conserved DNA binding domains, referring to important biological functions as the target genes govern cell cycle control, apoptosis and cell proliferation.3 In human solid tumors mutation of p53 and p73 resulting in functional deficiency are very common.4 By contrast, the p63 gene most often remains unaffected.5 Its expression is regulated by two different promotor regions entailing two opposing protein groups. TAp63 isoforms with highest analogy to p53 consist of a transactivational domain, a DNA binding domain, and an oligomerization domain. Transcription from a second promotor region in intron three results in three proteins lacking the NH2-terminal transactivational domain (DNp63). Investigations on p73 and p63 interaction with the p53 promoter regions suggest a functional role in control of tumor progression in malignancies of squamous cell origin. A model of murine embryonic fibroblasts advocates that p63 and p73 support p53 in pro-apoptotic function and may drive apoptosis in T-cell populations.6,7 To date no studies about distinct upstream

668

M. Moergel et al. / Oral Oncology 46 (2010) 667–671

regulatory elements for p63 in human oral squamous cell carcinoma exist. Solely a hamster model detected a shift from full length transcripts to the DNp63 isoforms with increasing grade of dysplasia and carcinoma induction.8 Similarities of p63 and p53 gene code and the restricted expression in basal and parabasal epithelial cell layers of differentiated epithelia point at a role for control of proliferation and differentiation.9 Recent studies revealed an association of clinical aggressiveness and an unfavourable outcome for overexpression of p63 in ovarial carcinoma and squamous cell carcinoma of the head and neck region.10,11 Although TAp63 forms may induce mitochondrial cytochrome c release in vitro12 and therefore may have a potential influence on apoptosis, to the best of our knowledge the linkage of p63 expression and clinical radioresistance has not been explored so far. Thus, we compared p63 expression in oral squamous cell carcinoma to effectiveness of radiation therapy in a clinical setting. Materials and methods Patients 33 patients (6 females and 27 males), age 34–72 years (mean 50.8) were enrolled in this retrospective study. Criteria for inclusion were: primary OSCC; pre-operative radiotherapy (36 Gy) followed by surgical resection with clinical safety margin of 1 cm; available follow-up data; antigen preservation in the biopsy confirmed by vimentin control.13 Retrospective investigation of paraffin-embedded biopsy material for the study was performed according to the regulations of the local ethical committee. Treatment protocol All patients were treated from 1996–2002 at the Department of Oral and Maxillofacial Surgery, University of Mainz Medical Center. In contrast to modern altered fractionation radiotherapy the patients within the present study underwent multimodality therapy according to the former ‘‘Essen” protocol14: pre-operative radiation therapy (36 Gy) was followed by radical surgical resection with a clinical safety margin of 1 cm. Lymph node dissection followed the recommendations of the DÖSAK cooperative group (Deutsch Österreichisch Schweizerischer Arbeitskreis für Tumoren im Kiefer- und Gesichtsbereich)15: Specifically, the clinically negative neck (N0) was treated by selective anterolateral neck dissection (Level I/II/III) when the primary tumor was situated anterior or lateral within the oral cavity. Neck dissection was extended to Level IV and V when metastases were found in frozen sections. Tumors at the retromolar trigone were treated by a complete lymph node dissection (Level I–V). The primary tumors were located as follows: one tumor of the maxilla; 17 tumors of the floor of the mouth; eight tumors of the tongue; three gingival tumors, of the mandible and the retromolar trigone and four in other locations. The TNM staging categories were determined according to the criteria established by the AJCC/UICC.16 Stage grouping of the patients was as follows: Stage I: one patient; Stage II: four patients; Stage III: three patients; Stage IV: 25 patients). Radiation protocol Radiotherapy treatment planning started with confirmation of a malignancy in the biopsy and was based on CT sections using computerized planning systems (Helax-TMSÒ, Nucleotron B.V., Ax Veenendaal, Netherlands or OSSÒ, Royal Philips Electronics, Eindhoven, Netherlands). External radiotherapy was delivered using a cobalt-60 machine (Gammatron 3, Siemens AG, Munich,

Germany) or/and linear accelerators (Type KD-2 or type Mevatron-77, both manufactured by Siemens AG, Munich, Germany). Ink markings ensured proper patient positioning in every radiation session. A pre-operative total dose of 36 Gy was administered by bilateral opposing beams (1.3 MV photons), covering tumor site and bilateral draining lymphatics of the neck, including the supraclavicular region. The dose was given in fractions of 2 Gy per day through 5 days a week. In cases of positive resection margins, lymphangiosis carcinomatosa or extracapsular spread of involved nodes, additional postoperative radiotherapy was given up to a complete dose of 60 Gy in fractions of 2 Gy every day through 5 days a week. Classification of radiation response was categorized based on histologic examination of the resection specimen. Patients with minimal or partial response (gross residual tumor) were classified as group MPR. In case of complete clinical response, specimens were further subdivided according to the additional histological findings into group GPR (good partial response) with remaining viable tumor cells on histological evaluation and group CR (complete response with no residual vital tumor on histological evaluation). At the time of survival analysis, 17 patients had died, three of causes unrelated to their tumors. The follow-up period for the surviving patients ranged from three to 106 months (median 44 months).

Immunohistology Specimen and staining Paraffin blocks of formalin-fixed biopsies were retrieved from the pathology archives of Johannes Gutenberg-University Mainz for immunohistological evaluation. The quality of antigen preservation in each specimen was confirmed by immunostaining for vimentin as described by Battifora.13 Immunostaining of p63 in the nuclei followed a standard protocol using an avidin–biotin technique and incubation with the monoclonal mouse anti-human p63-antibody (clone 4A4; Code M7247 DAKO Diagnostics, Hamburg, Germany) at a final dilution of 1:50 (60 min, room temperature). The LSAB IIÒ Kit (DAKO Diagnostics, Hamburg Germany) and 3,3-diamino benzidine (DAB) was utilized for visualization. A light counterstaining was performed with hematoxylin. Negative controls used the same protocol excluding the primary antibody. Basal epithelia, which were present in every section, served as internal control for positive p63 reaction and to assure constant immunostaining intensity.

Evaluation of stained sections All sections were assessed by the same investigator blinded to the patients´ clinical data. Quantitative evaluation of p63 expression based on the percentage of tumor cells that expressed the protein establishing a labeling index (Li) for each tumor. Therefore, a grid of 100 squares (0.025  0.025 mm each) was projected into the field of vision. Ten randomly selected high power fields (magnification 400) were photographed with a LeicaÒ microscope equipped with a digital camera (Leica WildÒ MPS52). For photoanalysis the image software (IM50Ò image manager V 1.2) was used. The grid allowed determination of the total tumor cell count and the number of p63 positive stained cells for each 10  10 matrix of the high power fields. The tumor cells recorded in 1000 squares were counted for each specimen. Tumors were categorized according to the Li: ‘‘Low” Li was defined by the number of p63 positive tumor cells being less the median value for all tumors; ‘‘high” Li was defined as the number of p63 positive tumor cells being higher the median value for all tumors.

669

M. Moergel et al. / Oral Oncology 46 (2010) 667–671

Statistical analysis The Mann and Whitney U-test17 was used to compare distributions of Li (p63 expression) between clinical responder (groups CR + GPR) and incomplete responder (group MPR). To assess potential differences between all groups (CR, GPR and MPR), the nonparametric Kruskal–Wallis test18 was employed. All reported p-values are two-sided. Survival probabilities were estimated by the Kaplan–Meier method.19 The endpoint of interest was time from tumor resection to tumor-related death. The patients’ data were first dichotomized by the median labeling index (Li) of the p63 expression to evaluate association with the survival outcome. The log-rank test was used to obtain p-values for two-group comparisons without adjustment for other potential prognostic factors.

In addition p63 Li was considered as a continuous variable in the proportional hazards model including the following covariates as independent variables: p63 Li, T-stage (T1 + T2 vs. T3 + T4), Nstage (N0 vs. N > 0), chemotherapy (administered vs. not administered), and grading (G1 + G2 vs. G3). Two sided p-values were obtained from the Wald-tests for testing any statistical significance of the regression coefficients.20 Results p63 expression in OSCC All the squamous cell carcinomas analyzed as part of this study were positive for p63 on immunohistological evaluation. Neither correlation to sex and gender nor to TNM stage or grading was found. The p63 labeling index was highly variable, ranging from 36% to 87%. Representative examples of tumors showing ‘‘high” and ‘‘low” p63-labeling indices are given in Figure 1. Association between p63 expression and radiation resistance

p63- LI [%]

After a pre-operative radiation dose of 36 Gy, 22 tumors showed only minimal or partial shrinkage, as the specimen were classified as MPR based on the macroscopically detected presence of residual tumors. Complete clinical response was noticed in 11 tumors. Upon histological evaluation of these cases, seven specimens were classified as GPR (good partial response: positive for residual tumor cells on histological assessment) and four specimens were classified as CR (complete response: no residual tumor cells on histological assessment). The average Li was 48% (±10) in group CR, 61% (±6) in group GPR and 68% (±10) in group MPR (p = 0.005, Kruskal–Wallis-Test). Clinical responders (groups CR + GPR) showed a significantly lower expression of p63 when compared to the non- (or minimal) responders classified as MPR (p = 0.003). The boxplot illustrates the distribution of the p63 expression when grouped according to radiation resistance (Figure 2). These results suggest that p63 expression might be suitable as a marker of radioresistance in OSCC, with high p63 expression levels being associated with poor radiation response and vice versa.

group CR

Figure 1 Two examples of p63 expression in OSCC as determined by immunohistology. The paraffin sections are representative for the low Li subgroup (
group GPR

group MPR

Figure 2 p63 Expression in the group of 33 OSCC patients with mainly T3 and T4 tumors. Resistance to radiation was associated with an increased p63 labeling index and vice versa (p = 0.005). The group definition followed histological evaluation of resection specimen as: Group CR (n = 4): complete response with no residual vital tumor; Group GPR (n = 7): good partial response with remaining viable tumor cells; Group MPR (n = 22): minor partial response with bulky residual tumor. The box plots represent median % of Li (black bar); the boxes are Quartile ranges (25–75%), and the whiskers extend to the maximum and minimum values.

670

M. Moergel et al. / Oral Oncology 46 (2010) 667–671

Discussion

Overall Survival

LI < 63% (n=16)

LI > 63% (n=17)

time [months] Figure 3 Kaplan–Meier survival analysis as univariate analysis and dichotomized according to the median Li for high Li (>63.1) vs. low Li (<63.1) expression pattern. The data suggest that the p63 Li has a significant predictive value for tumor related survival (p = 0.0077).

Association between p63 expression in the tumor and patient survival During the follow-up period, three of 16 patients (19%) in the ‘‘low Li” group (Li < median) and 11 of 17 patients (65%) in the ‘‘high Li” group (Li P median) died. The difference in survival between the two groups was statistically significant (p = 0.0077). The Kaplan Meier plot in Figure 3 suggests higher p63 levels to be associated with a poor clinical outcome. When considered as a continuous variable, p63 expression was confirmed as a significant independent prognostic marker in a proportional hazards regression analysis, when the data were controlled for T- and N-stage, grading, and administration of chemotherapy. Age and gender showed no significant hazard ratio, neither as a single variable nor as part of the model. The overall significance of the model was p = 0.007 (-2-log-likelihood function: 49.6). High p63 expression was assigned an increased hazard of death due to tumor (p = 0.017, Wald test, incremental hazard ratio 1.082 for every% of p63 Li). Thus, the Li of 65% can be considered to be a worse prognostic characteristic than the Li of 35%, for example. Hazard ratios, 95% confidence intervals and p-values of the Wald test for all variables are given in Table 1. Both the impact of p63 expression on radiation response and the association between p63 expression and survival suggest a biologically and clinically relevant influence of p63 on radiation responses in OSCC.

Squamous cell carcinoma of the head and neck (H&NSSC) is a common disease with high mortality rates for patients with stage III and IV cancers.21–24 Apart from radical surgical resection, irradiation is the most important treatment modality in curative and palliative management of these tumors. However, in spite of highly sophisticated methods to define target regions and to ensure proper dose delivery, the effectiveness of radiation therapy is still highly variable. Pre-treatment assessment of radiation resistance may, thus, improve the management of H&NSSC in terms of primary allocation of the patient to an ‘‘efficient” treatment modality. Effectiveness of radiation therapy is addressed by the ‘‘target cell theory” which states that radiation response reflects cell death in a stem cell population in the first place and that surviving cells are capable of proliferation and possible tissue regeneration, secondly.25 Thus, tumor cell survival by apoptosis control and up-regulation of proliferation are imminent influencing factors on tumor maintenance and may lead to limited tumor control under irradiation. Since p63 ruled out to be a major factor of epithelial tissue proliferation and differentiation and studies revealed regular overexpression of DNp63-levels in carcinoma of the head and neck, we hypothesized a pivotal role of p63 in radiation response. Consistent with previous studies we found marked overexpression of p63 as quantified by a labelling index with obvious variability between different tumors (range: 36–87%). p63 expression pattern was independent from age and gender and p63 positive cells were found disseminated throughout the tumor whereas in normal mucosa p63 expression was restricted to the basal lamina. These histological findings are in line with studies revealing frequent amplification of the p63 gene locus (3q27–28) in oral squamous cell carcinoma.26 Furthermore a study reported p63 overexpression as one of the most common molecular abnormalities in HNSCC with increased p63 expression in more than 80% of the cases.27 Hence, upregulation of p63 is presumably initiated early in carcinogenesis preserving the carcinoma in a stem cell manner and providing a high proliferation potential.28 The clinical impact of this finding may be additionally reflected by investigations of different epithelial derived cancer types (e.g. OSCC, Merkel zell Carcinoma, lung, ovary) reporting p63 overexpression regularly characterize a group of patients with poor overall outcome.29–32 With regard to irradiation, inhibition of apoptotic regulatory elements may allow tumor cells to survive and a high proliferative potential may promote faster tumor recovery after cessation of irradiation. Unfortunately, few reliable parameters are available for measurement of radiation effectiveness (e.g. reduction of tumor volume, time to recurrence or survival rates). Superior to split course irradiation modern radiotherapy employs altered fractionation to improve effectiveness of irradiation. The main objective of accelerated fractionation or hyperfractionation radiotherapy is to increase a high total dose in a shorter overall time with benefit in survival and local control as shown in different trials (e.g. DAHANCA, MARCH or GORTEC).33–35 The patients of the present inves-

Table 1 Survival analysis based on p63 expression in the tumor and controlled for other variables in OSCCa.

a

Variable

Definition of groups (group 1 vs. group 2)

95% Confidence interval for the hazard ratio

Hazard ratio (group 2 relative to group 1)

p-Value (Wald)

T-stage N-stage Grade of differentiation Chemotherapy (cisplatin) p63 (Li)

T1 + T2 vs. T3 + T4 N0 vs. N > 0 G1 + 2 vs. G3 Administered vs. not administered Continuous variable

0.48–16.9 0.35–26.3 0.5–9.5 1.2–32.2 1.01–1.16

2.9 3.0 2.3 6.3 1.1 (per %)

0.25 0.32 0.26 0.027 0.017

Cox regression analysis.

M. Moergel et al. / Oral Oncology 46 (2010) 667–671

tigation were treated (in the late 1990th) according to the meanwhile outdated ‘‘Essen protocol” wherein all patients received equal neoadjuvant irradiation dosages and subsequently underwent complete tumor resection. This cohort allows both the direct correlation of p63 levels obtained in pre-treatment biopsies, to radiation response as verified by the histological workup of the tumor specimen and a correlation of p63 levels to survival data. Tumors susceptible to radiation (groups CR + GPR) presented a significantly (p = 0.005) lower p63 labeling-index (Li) compared to non-responders (group MPR). These results support the assumption of p63 expression being a potential marker of radiation resistance in OSCC, with high p63 expression levels being correlated to poor radiation response. In general, radiation leads to induction of radicals and reactive oxygen/nitrogen species followed by stimulation of signalling cascades of intrinsic and extrinsic apoptotic pathways. p63 isoforms consist of a highly p53-homologous DNAbinding domain and may govern critical downstream targets of apoptotic signalling pathways similar to p53 as gatekepper of apoptosis control. A recent study on ovarian cancers linked a dominant DNp63/TAp63 ratio to poor survival in uni- and multivariate analysis as also to reduced sensibility for platinum based chemotherapy.10 Likewise, Liefer and coworkers demonstrated a 40% decrease of apoptosis in epidermal cells under UVB radiation in a ML.DNp63a transgenic mouse model stressing the influence of DNp63 isoforms on apoptosis control.36,8,37 Hence, predominance of DNp63 isoforms may reflect a potential biological advantage for the tumor by inhibition of pro-apoptotic down-stream elements contributing to advanced radioresistance in OSCC. Conclusion Our comparative analysis of p63 expression and therapeutic effectiveness of irradiation, as measured by histology and survival, suggests, that p63 may be used for rapid and effective identification of patients with radioresistant OSCC, which could not be expected to be cured by radio(chemo)therapy alone. The promising results of this biomarker should now be confirmed in a clinical setting with larger patient counts. Conflicts of interest statement None declared. Acknowledgement This work was supported in part by a grant of the Stiftung Tumorforschung Kopf-Hals to Maximilian Moergel and Martin Kunkel. References 1. Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dötsch V, et al. p63, a p53 Homolog at 3q27–29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Molecular Cell 1998;2: 305–16. 2. Mills A, Zheng B, Wang XJ, Vogel H, Roop DR, Bradley A. p63 is a p53 homologue required for limb and epidermal morphogenesis. Letters to Nature 1999;22(4):708–13. 3. Yang A, Zhu Z, Kapranov P, McKeon F, Church GM, Gingeras TR, et al. Relationships between p63 binding, DNA sequence, transcription activity, and biological function in human cells. Molecular Cell 2006;24(17):593–602. 4. Nomoto S, Haruki N, Kondo M, Konishi H, Takahashi T. Search for mutations and examination of allelic expression imbalance of the p73 gene at one p36.33 in human lung cancers. Cancer Res 1998;7(58):1380–3. 5. Osada M, Ohba M, Kawahara C, Ishioka C, Kanamaru R, Katoh I, et al. Cloning and functional analysis of human p51, which structurally and functionally resembles p53. Nat Med 1998;4(7):839–43. 6. Flores ER, Tsai KY, Crowley D, Sengupta S, Yang A, McKeon F, et al. p63 and p73 are required for p53-dependent apoptosis in response to DNA damage. Nature 2002;4(416):560–4.

671

7. Senoo M, Manis JP, Alt FW, McKeon F. p63 and p73 are not required for the development and p53-dependent apoptosis of T cells. Cancer Cell 2004;6(1):85–9. 8. Chen YK, Hsue SS, Lin LM. Immunohistochemical demonstration of p63 in DMBA-induced hamster buccal pouch squamous cell carcinogenesis. Oral Dis 2003;9(5):235–40. 9. Pellegrini G, Dell’ambra E, Molisano O, Martinelli E, Fantozzi I, Bondanza S, et al. p63 Identifies keratinocyte stem cells. Proc Natl Acad Sci USA 2001;13(98): 3156–61. 10. Marchini S, Marabese M, Marrazzo E, Mariani P, Cattaneo D, Fossati R, et al. DeltaNp63 expression is associated with poor survival in ovarian cancer. Ann Oncol 2008;19(3):501–7. 11. Lo Muzio L, Santarelli A, Caltabiano R, Rubini C, Pieramici T, Trevisiol L, et al. P63 overexpression associates with poor prognosis in head and neck squamous cell carcinoma. Human Pathol 2005;36(2):187–94. 12. Gressner O, Schilling T, Lorenz K, Schulze Schleithoff E, Koch A, SchulzeBergkamen H, et al. TAp63alpha induces apoptosis by activating signaling via death receptors and mitochondria. EMBO J 2005;24(13):2458–71. 13. Battifora H. Assessment of antigen damage in immunohistochemistry. The vimentin internal control. Am J Clin Pathol. 1991;96:669–71. 14. Mohr C, Bohndorf W, Gremmel H. Preoperative radiochemotherapy and radical surgery of advanced head and neck cancers–results of a prospective, multicenter DOSAK study. Recent Results Cancer REs 1994;134:155–63. 15. Bier J. Definitionen zum radikalchirurgischenVorgehen bei Plattenepithelkarzinomen der Mundhöhle. Dtsch Z Mund Kiefer GesichtsChir 1982;6:369–72. 16. Wittekind C, Wagner G. TNM-Klassifikation maligner Tumoren. fifth. Berlin: Springer; 1997.. p. 213. 17. Mann HB, Whitney DR. On a test whether one or two random variables is stochastically larger than the other. Annals Math Stat 1947;18:50. 18. Kruskal WH, Wallis WA. Use of ranks in one criterion variance analysis. J Amer Stat Assoc 1952;57:583–621. 19. Kaplan EL, Meier P. Nonparametric estimator from incomplete observations. J Amer Stat Assoc 1958;53:457–81. 20. Cox DR. Regression models and life-tables (with discussion). J Royal Stat SocietySeries B 1972;34:187–202. 21. Mignogna MD, Fedele S, Lo Russo L. The world cancer report and the burdon of oral cancer. Eur J Cancer Prev 2004;13(2):139–42. 22. Suntharalingam M, Haas ML, Van Echo DA, Haddad R, Jacobs MC, Levy S, et al. Predictors of response and survival after concurrent chemotherapy and radiation for locally advanced squamous cell carcinoma of the head and neck. Cancer 2001;91(3):548–54. 23. Lavaf A, Genden EM, Cesaretti JA, Packer S, Kao J. Adjuvant radiotherapy improves overall survival for patients with lymph node-positive head and neck squamous cell carcinoma. Cancer 2008;112(3):535–43. 24. Gourin CG, Conger BT, Porubsky ES, Sheils WC, Bilodeau PA, Coleman TA. The effect of occult nodal metastasis on survival and regional control in patients with head and neck squamous cell carcinoma. Laryngoscope 2008;118(7): 1191–1194. 25. Stewart FA, Dörr W. Milestones in normal tissue radiation biology over the past 50 years: from clonogenic cell survival to cytokine networks and back to stem cell recovery. Int J Radiat Biol 2009;85(7):574–86. 26. Lo Muzio L, Campisi G, Farina A, Rubini C, Pastore L, Giannone N, et al. Effect of p63 expression on survival in oral squamous cell carcinoma. Cancer Invest 2007;25(6):464–9. 27. Rocco JW, Leong CO, Kuperwasser N, DeYoung MP, Ellisen LW. p63 Mediates survival in squamous cell carcinoma by suppression of p73-dependent apoptosis. Cancer Cell 2006;9(1):45–56. 28. Moll UM, Slade N. p63 and p73: roles in development and tumor formation. Mol Cancer Res 2004;2(7):371–86. 29. Muzio L, Campisi G, Farina A, Rubini C, Pastore L, Giannone N, et al. Effect of p63 expression on survival in oral squamous cell carcinoma. Cancer Invest 2007;25(6):464–9. 30. Asioli S, Righi A, Volante M, Eusebi V, Bussolati G. P63 expression as a new prognostic marker in Merkel cell carcinoma. Cancer 2007;110(3):640–7. 31. Narahashi T, Niki T, Wang T, Goto A, Matsubara D, Funata N, et al. Cytoplasmic localization of p63 is associated with poor patient survival in lung adenocarcinoma. Histopathology 2006;49(4):349–57. 32. Marchini S, Marabese M, Marrazzo E, Mariani P, Cattaneo D, Fossati R, et al. DeltaNp63 expression is associated with poor survival in ovarian cancer. Ann Oncol 2008;19(3):501–7. 33. Bourhis J, Lapeyre M, Tortochaux J, Rives M, Aghili M, Bourdin S, et al. Ph III randomized trial of very accelerated radiation therapy compared with conventional radiation therapy in squamous cell head and neck cancer: a GORTEC trial. J Clin Oncol 2006;24(18):2873–8. 34. Bourhis J, Overgaard J, Audry H, Ang KK, Saunders M, Bernier J, et al. Hyperfractionated or accelerated radiotherapy in head and neck cancer: a meta-analysis. Lancet 2006;368(9538):843–54. 35. Overgaard J, Hansen HS, Specht L, Overgaard M, Grau C, Andersen E, et al. Five compared with six fractions per week of conventional radiotherapy of squamous-cell carcinoma of head and neck: DAHANCA six and seven randomised controlled trial. Lancet 2003;362(9388):933–40. 36. Liefer KM, Koster MI, Wang XJ, Yang A, McKeon F, Roop DR. CRA- downregulation of p63 is required for epidermal UV-B-induced apoptosis. Cancer Res 2000;60(15):4016–20. 37. Westfall MD, Pietenpol JA. p63: Molecular complexity in development and cancer. Carcinogenesis 2004;25(6):857–64.