The Prognostic Significance of the Biomarker p16 in Oropharyngeal Squamous Cell Carcinoma

Clinical Oncology 25 (2013) 630e638 Contents lists available at SciVerse ScienceDirect

Clinical Oncology journal homepage: www.clinicaloncologyonline.net

Original Article

The Prognostic Significance of the Biomarker p16 in Oropharyngeal Squamous Cell Carcinoma K.K. Oguejiofor *, J.S. Hall *, N. Mani *, C. Douglas *, N.J. Slevin y, J. Homer z, G. Hall z, C.M.L. West * * Translational

Radiobiology Group, Institute of Cancer Sciences, Manchester Academic Health Science Centre, The University of Manchester, Manchester, UK y Christie NHS Foundation Trust, Manchester, UK z Central Manchester University Hospitals, Manchester, UK Received 22 February 2013; received in revised form 9 May 2013; accepted 14 May 2013

Abstract Aims: There is an increasing incidence of human papillomavirus (HPV)-positive oropharyngeal squamous cell cancers (OPSCC) mostly associated with favourable outcomes. p16 immunohistochemistry is a surrogate marker for HPV positivity in OPSCC. The prognostic strength of p16 over traditional prognostic factors is not fully characterised. In this study, we evaluated the clinical and demographic differences between p16-positive and -negative OPSCC and characterised its prognostic strength versus traditional prognostic factors. Materials and methods: Formalin-fixed, paraffin-embedded blocks and clinical information from 217 OPSCC patients, treated with radiotherapy (alone or in combination with other therapies) between 2000 and 2010 were collected retrospectively. Immunohistochemistry for p16 protein was carried out; cancerspecific survival (CSS), recurrence-free survival (RFS) and locoregional control (LRC) were calculated for both univariate and multivariate analyses. Results: Ninety-two per cent of the OPSCC originated from tonsil and tongue base sites, 61% were p16 positive. Patients with p16-positive OPSCC were younger (P < 0.0001), with lower alcohol (P ¼ 0.0002) and tobacco (P ¼ 0.0001) exposure. The tumours were less differentiated (P ¼ 0.0069), had a lower T stage (P ¼ 0.0027), higher nodal status (P ¼ 0.014) and higher American Joint Committee on Cancer (AJCC) prognostic group (P ¼ 0.0036). AJCC prognostic group was significant for RFS (P ¼ 0.0096) and CSS (P ¼ 0.018) in patients with p16-negative OPSCC, but not those with p16-positive tumours (P ¼ 0.30 and 0.54). Other significant factors for CSS and RFS in univariate analysis were: pretreatment haemoglobin (P < 0.0001 and <0.0001), chemoradiotherapy (P ¼ 0.005 and 0.03) and P16 status (P < 0.0001 and 0.0001). In multivariate analysis, p16 positivity was the strongest independent prognostic variable for both CSS, RFS and LRC (P < 0.0001, hazard ratio 4.15; 95% confidence interval 2.43e7.08), (P < 0.0001, hazard ratio 6.15; 95% confidence interval 3.57e10.61) and (P ¼ 0.001, hazard ratio 3.74; confidence interval 1.76e7.95). Conclusion: This study shows that p16 is the single most important prognostic variable in OPSCC, surpassing traditional prognostic factors for both CSS and RFS. Furthermore, disease stage has no prognostic significance in p16-positive patients, highlighting the need for routine p16 assessment in OPSCC. Ó 2013 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Chemoradiotherapy; head and neck cancer; human papillomavirus (HPV); oropharyngeal cancer; p16; prognosis

Introduction Human papillomaviruses (HPV) are non-enveloped double-stranded DNA viruses commonly associated with cervical cancers [1]. High-risk types are also linked with cancers of the oropharynx (oropharyngeal squamous cell Author for correspondence: C.M.L. West, The University of Manchester, Christie Hospital NHS Trust, Manchester Academic Health Science Centre, Wilmslow Road, Manchester M20 4BX, UK. Tel: þ44-161-446-8275; Fax: þ44-161-446-8111. E-mail address: [email protected] (C.M.L. West).

cancers; OPSCC) affecting primarily tonsil and tongue base sites [2]. Emerging trends suggest an increase in the prevalence of HPV-associated OPSCC mostly in young people with lower exposure to tobacco and alcohol [3e7]. Similarly, reports suggest that these cancers are distinct from other head and neck cancers with differing biology, morphology and prognosis [8]. HPV-positive OPSCC tend to be poorly differentiated with basaloid morphology, presenting with a lower T stage and higher nodal status [2], but paradoxically better prognosis [9e11]. Presently, there is considerable interest in detecting these cancers for risk stratification, epidemiological studies, de-intensification of current

0936-6555/$36.00 Ó 2013 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clon.2013.07.003

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

treatment regimens and for possible initiation of viraltargeted therapy (vaccination and/or antiviral agents). Currently, available methods detect DNA, RNA or surrogate markers of HPV infection. HPV DNA is detected using polymerase chain reaction and/or in situ hybridisation, but DNA-based analysis is imperfect [12]. p16 immunohistochemistry (IHC) is used as an alternative and/or complementary test, the p16 protein is encoded by CDKN2A, a tumour suppressor gene located on chromosome 9p21 that is frequently lost in head and neck squamous cell cancer [13]. p16 is involved in maintaining the integrity of the G1/S cell cycle check point through binding to cyclin-dependent kinases, which are involved in the release of the E2F transcription factor by the phosphorylation of retinoblastoma protein. Inactivation of retinoblastoma by the HPV E7 leads to the subsequent up-regulation of the p16. As p16 is usually lost early in HPV-unrelated OPSCC [14,15], tests detecting p16 are used as a surrogate marker of HPV-induced oncogenesis. In most studies, p16 IHC alone is a strong prognostic factor consistently linked to differences in overall, recurrence-free (RFS) and disease-specific survival that is independent of treatment type [10]. Presently it is not known if all p16-positive OPSCCs share similar outcomes [16e18], with considerable effort being made to identify tumours with discordant HPV/p16 results. Similarly, treatment options in OPSCC have generated considerable interest. In general, OPSCC is treated similarly despite available evidence suggesting biological, morphological and survival differences. This has led to de-escalation trials for HPVpositive cases that aim to improve survival while limiting morbidity and increasingly to surgical approaches for fitter patients with HPV-negative disease. In this study we evaluated the differences in the clinicodemographic, survival and treatment options between p16-positive and -negative OPSCC in a large cohort of patients. The prognostic strength of p16 was compared with other traditional prognostic factors in OPSCC and the importance of prognostic stage was compared in p16positive and -negative patients.

Materials and Methods Patients The 217 patients in this study were treated at The Christie NHS Foundation Trust Hospital. The inclusion criteria were: histologically confirmed OPSCC; treatment between January 2000 and December 2010; and the use of radiotherapy as one or the only modality of treatment. Patients treated with palliative intent were excluded. Patient information was collected in two cohorts 1999e2004 and 2005e2010. Treatment involved: radiotherapy alone (intensity-modulated radiotherapy or conventional radiotherapy [60e66 Gy over 6e6.6 weeks or 55 Gy in 4 weeks]), primary neck surgery and radiotherapy, or chemoradiotherapy (induction and/or concurrent chemoradiotherapy). Induction and concurrent chemotherapy

631

comprised: docetaxel, cisplatin, 5-fluorouracil and cisplatin or carboplatin respectively. Ethical approval for the study was obtained from the local research ethics committee (reference number 03/TG/076; Christie reference number 03_RADIO_22). Immunohistochemistry Formalin-fixed, paraffin-embedded blocks were obtained from referring hospitals. Sections were cut and stained with haematoxylin and eosin. Blocks with confirmed tumour presence were selected for IHC, which was carried out using the CINtec histology kit (MTM Laboratories, Heidelberg, Germany) as per manufacturer’s instructions. Briefly, 4 mm sections were deparaffinised and rehydrated. This was followed by epitope retrieval in a water bath using epitope retrieval solution (CINtec). The solution and slides were then heated to 97  C for 10 min and cooled at room temperature for 20 min. The staining was carried out on a Biogenix 16000 autostainer as per the manufacturer’s protocol. The slides were then counterstained with Gill’s haematoxylin and cover-slipped. Controls were included with each batch, and were HPV-positive and -negative samples. No batch-to-batch variation was encountered between controls. Scoring The stained slides were scanned using the Leica SCN 400. p16 expression was scored as positive if there was a strong and diffuse brown staining of the nucleus and cytoplasm in 70% of the tumour specimen [19]. The slides were scored by a single scorer with high (R2 ¼ 0.91) intra-scorer reproducibility. Statistical Analysis Statistical analyses were carried out using SPSS version 16.0. Actuarial calculations of locoregional control (LRC), RFS and cancer-specific survival (CSS) were obtained using the KaplaneMeier method. CSS was used rather than overall survival to avoid any bias in comparing younger and healthier HPV-positive patients (less likely to die of secondary causes) with HPV-negative patients. Univariate analysis was compared using the Log-rank (Mantel-Cox) method. The Cox proportional hazard model was used for multivariate analysis. The time to primary re-occurrence was defined from the start of treatment with radiotherapy to reported recurrence.

Results There were no differences in survival in both cohorts 1999e2004 and 2005e2010 (Supplementary Figure 1). The 217 patients included in the study had a median age at presentation of 58.5 years. As expected, most of the tumours (92%) originated from tonsil and tongue base sites. Most of the tumours had high clinical stage (American Joint

632

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

Committee on Cancer; AJCC) disease (64% stage IV) and most were moderately differentiated (53%). The prevalence of p16-positive tumours in this cohort was 61% (Table 1). p16 Status and Clinical Characteristics Table 2 compares the clinical characteristics of p16positive and -negative tumours. p16-positive tumours occurred more frequently in younger patients, with a mean age at treatment of 56 years (95% confidence interval 55e57) versus 64 years (95% confidence interval 63e65) years (P < 0.0001) (Figure 1). In comparison with p16negative tumours, p16-positive tumours were also associated with non-smokers, low alcohol use, low T stage, high N Table 1 The clinical characteristics of the patients Clinical characteristics Gender Male Female Median age (years) Smoking status No Ex Current Alcohol Never Moderate Heavy Site Base of tongue Tonsil Other Tumour stage T1 T2 T3 T4 Nodal stage N0 N1 N2 N3 AJCC stage I II III IV p16 status Positive Negative Tumour differentiation Well Moderate Poorly Treatment Radiotherapy alone Chemotherapy þ radiotherapy

Total no available

Number of patients (%)

217 217

163 (75) 54 (25) 58.5

211

42 (20) 82 (39) 87 (41)

193

12 (6) 107 (55) 74 (39)

211

59 (28) 135 (64) 17 (8)

207

48 78 37 44

(23) (38) (17) (22)

207

57 31 108 11

(27) (15) (53) (5)

207

11 26 38 132

(6) (12) (18) (64)

217

129 (60) 88 (40)

179

19 (11) 94 (53) 66 (36)

207

109 (52) 98 (48)

AJCC, American Joint Committee on Cancer.

stage, high AJCC stage and poor differentiation. p16-positive patients more likely to have received chemoradiotherapy (P ¼ 0.0072). There was no difference in gender (P ¼ 0.51) or tumour site (P ¼ 0.41) when the clinical characteristics of p16-positive and -negative tumours were compared. Univariate Analysis Figure 2 shows KaplaneMeier curves stratified by p16 status (Figure 2A, B) and chemoradiotherapy (Figure 2C, D). On univariate analysis, smoking, pretreatment haemoglobin, T stage, nodal status, AJCC stage, differentiation, chemoradiotherapy and p16 status were significantly associated with LRC, RFS and CSS (P < 0.001). Age (P ¼ 0.004) was significantly associated with CSS alone (Table 3). We also looked at the interaction between p16positive/negative patients and the use of chemoradiotherapy (Figure 2E, F). We saw a trend for better survival among the p16 patients, which was less evident in p16-negative patients. Given that p16-positive patients present with a higher grouping, we evaluated the influence of prognostic group (AJCC prognostic group) on survival and observed a significant relationship in the p16-negative patients (Figure 3A, B). This relationship was not seen in the p16-positive patients (Figure 3C, D). Multivariate Analysis Factors associated with LRC, RFS and CSS on univariate analysis (smoking, pretreatment haemoglobin, T stage, nodal status, AJCC stage, grade and p16 status) were included in the Cox proportional hazards model. P16 was significant on multivariate analysis for all end points (Table 4). Tumour grade was significant for both RFS (P ¼ 0.004) and CSS (P ¼ 0.001). Pretreatment haemoglobin (P ¼ 0.001) was associated with RFS. T stage was a significant independent prognostic factor for LRC (P ¼ 0.007) and CSS (P ¼ 0.001). p16 was the strongest independent prognostic variable identified in this study for LRC (hazard ratio ¼ 3.74; 95% confidence interval 1.76e7.95; P ¼ 0.001), RFS (hazard ratio ¼ 6.15; 95% confidence interval 3.57e10.61; P < 0.0001) and CSS (hazard ratio 4.15; 95% confidence interval 2.43e7.08; P < 0.0001).

Discussion The prevalence of OPSCC arising from tonsil and tongue base sites in young non-smoking adults is reported to be on the rise globally [3,4] and in the UK [7]. To our knowledge, this is the first study in the UK to study a large series of patients over a 10 year period, observing differences in clinicodemographic, survival and prognostic features between p16-positive and -negative tumours. The prevalence rate from this study was 61%. However, as this is a radiotherapy-treated cohort from a single site, the true prevalence in the UK remains largely unknown. In this study, we observed the consistent finding of differing clinicodemographic characteristics between p16-

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638 Table 2 The clinical characteristics in p16-positive and -negative tumours Clinical characteristics

p16 positive (%)

Gender Male 99 (77) Female 30 (23) Primary site Tonsil 84 (67) Tongue base 32 (26) Other sites 9 (7) Alcohol None 12 (11) Moderate 67 (59) Heavy 34 (30) Smoking status Never 36 (29) Ex 54 (43) Current 35 (28) T stage 1 34 (27) 2 53 (42) 3 21 (17) 4 17 (14) N stage 0 29 (23) 1 22 (18) 2 71 (57) 3 2 (2) AJCC prognostic group 1 4 (3) 2 16 (13) 3 27 (22) 4 78 (62) Tumour differentiation Well 8 (8) Moderate 48 (46) Poor 47 (46) Chemoradiotherapy No 58 (46) Yes 68 (54)

p16 negative (%)

P value

64 (73) 24 (27)

0.51

51 (59) 27 (32) 8 (9)

0.41

0 (0) 40 (50) 40 (50)

0.0002

6 (7) 28 (33) 52 (60)

0.0001

14 25 16 27

(17) (30) (19) (33)

0.0027

28 9 37 9

(34) (11) (45) (10)

0.014

7 11 10 54

(9) (13) (12) (66)

0.0036

11 (14) 46 (61) 19 (25)

0.0069

51(63) 30 (37)

0.0072

AJCC, American Joint Committee on Cancer.

positive and -negative tumours. p16-positive tumours occurred in younger patients, a finding that has been consistently reported by others [8,20,21]. The median age of 56 years (p16 positive) reported in this study is similar to < 0.0001 100

Age (years)

80

n=84 n=133

60 40 20

n=217 0

Positive

P16 status

Negative

Fig 1. Box plot showing the relationship between p16 status and age.

633

the finding from the DAHANCA 5 trial [21]. Similarly, when lifestyle choices were examined, patients with p16-positive OPSCC were less likely to be smokers and/or drinkers, a feature that has been previously reported [8,20] and is one of the epidemiological characteristics underlying the aetiological association between HPV positive versus negative OPSCC. There were pathological differences between p16-positive and -negative tumours, with positive tumours presenting with a lower T stage and a higher nodal status and as such a higher AJCC stage. This feature has been reported by others [2,22,23]. In the univariate analysis, smoking status was significantly associated with LRC, RFS and CSS, which has been reported previously [24]. In this study, p16 positivity showed a highly significant association with RFS and CSS in univariate analysis. Moreover, it was an independent prognostic factor in this cohort after multivariate analysis, a feature that has been reported by others [2,9,25]. p16 has been shown to be a strong prognostic biomarker after: radiotherapy alone [26], chemoradiotherapy [24,27e29], primary surgery with or without adjuvant radiotherapy [17,30,31]. In this study, 5 year CSS and RFS rates for all patients were 56 and 58%, respectively, which are similar to the reports from a meta-analysis [10]. Stratifying patients based on p16 status, improved 5 year survival rates from 38 to 75% for CSS and 28 to 78% for RFS, similar improvements have been observed by others [9,24,29,32]. In this study we also evaluated the effects of chemoradiotherapy on survival and observed an improvement from 51 to 70% and 49 to 75% for CSS and RFS, respectively. This may have been influenced by the greater number of p16-positive patients in the chemoradiotherapy arm, as p16 patients are younger with better survival rates irrespective of treatment type. To determine whether chemotherapy benefited both p16-positive and -negative groups equally, we substratified the patients and observed a trend for improved survival in the p16-positive group, which was less obvious in the p16-negative group. Although this is a novel finding, the observations were not statistically significant and, due to the small numbers of patients in each arm, the finding must be interpreted with caution. However, it would be worthwhile repeating this analysis in a larger cohort. The retrospective nature of data collection could limit this study, as could the heterogeneity of radiotherapy treatment over a 10 year period. That said, we present a relatively large cohort of OPSCC patients who underwent radiotherapy with curative intent. What is interesting from the current study is that the prognostic strength of p16 surpasses the traditional prognostic markers such as TNM classification and prognostic group in the multivariate analysis; this again has been reported previously [9,25]. In their study, Fischer et al. [25] evaluated prognosis in 102 patients using p16 IHC. Our larger cohort of 217 patients lends further strength to this observation. Given the paradox of p16-positive tumours presenting with higher stage and grade, but with better prognosis, we also explored the role of current clinical staging in p16-positive and -negative groups. Remarkably, the well-reported prognostic association of AJCC stage in OPSCC [33] was clearly shown in p16-negative tumours

634

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

Survival Analysis

A

B

100

P16 negative

P16 negative P16 Positive

P16 positive

80 60

RFS %

80 60

CSS %

100

40

40 20

20

P <0.0001 HR (95% CI) = 5.59 (3.30 - 9.49)

0

0

2

4

0

6

P <0.0001 HR (95% CI) = 5.45 (3.29 - 9.04)

0

2

Time (years)

C

109

92

67

41

11

1

125

107

85

61

36

7

1

81

50

34

19

8

2

1

86

44

27

19

8

2

1

chemoradiotherapy

100

D

No Yes

60

80

40

60

20 0

20

HR (95%CI) = 1.78 (1.09 - 2.89)

2

4

6

0

Time (years) 108

86

62

41

28

6

1

94

73

64

45

22

7

1

100

No Yes

80

80

0

2

109 98

chemoradiotherapy 100

P = 0.005 HR(95%CI) = 1.976 (1.22 - 3.18)

4

6

Time (years)

F

E

No Yes

40

P = 0.03 0

chemoradiotherapy

100

RFS %

CSS %

6

121

80

70 74

53 62

39 43

25 27

5 8

1 1

chemoradiotherapy No Yes P16+ve P=0.18

P16 +ve P=0.27

60

RFS %

CSS %

4

Time (years)

60 40

40

20

P16 -ve P=0.54

20 0

0 0

2

4

59 52

Time (years) 51 39 20 42 30 24

6 6

1 2

-ve 28 54

14 38

13 24

1 4

1 3

7 15

4 8

0

2

68 58

61 51

Time (years) 50 38 19 41 29 23

8 8

1 3

-ve 30 51

15 32

13 17

1 4

1 3

6

64 56

+ve

P16-ve P=0.52

+ve

4

9 15

5 8

6

Fig 2. KaplaneMeier curves of: (A) and (B) comparing 5 year cancer-specific survival (CSS) and recurrence-free survival (RFS) in p16-positive and -negative patients, respectively; (C) and (D) comparing 5 year CSS and RFS in patients treated with or without chemoradiotherapy, respectively; (E) and (F) comparing CSS and RFS in p16-positive and -negative patients, respectively, treated with chemoradiotherapy.

(P < 0.05), but was not significant in the p16-positive tumours (i.e. 61% of cases). Taken together, this shows not only that p16 is the most significant independent prognostic factor in this cohort, but also that the current prognostic classifications has no benefit in p16-positive tumours. There is currently considerable discussion over the best way to classify HPV-positive and -negative tumours using the myriad of techniques available and p16 is not without

faults. There are patients who are p16 positive, yet HPV DNA polymerase chain reaction and/or HPV DNA in situ hybridisation negative [16,18]. The use of p16 alone could lead to an overestimation of the true prevalence of HPV-positive OPSCC. Our intention was not to report on the prevalence of HPV, but rather to observe the differences in the clinicodemographics, tumour morphology and survival patterns between p16-positive and -negative tumours. Significantly larger studies are required to address whether

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

635

Table 3 The univariate Log-rank analysis Characteristics

Locoregional control Hazard ratio (95% confidence interval)

Gender Male 1 Female 1.39 (0.64e2.62) Age (years) <50 1 50e60 0.85 (0.58e1.12) >60 1.74 (1.25e1.84) Alcohol None 1 Lowemedium 0.66 (0.25e0.94) Heavy 1.15 (0.70e1.2) Smoking Current 1 Ex 0.45 (0.06e1.35) Never 0.09 (0.02e0.90) Pretreatment haemoglobin 12 1 14 0.24 (0.69e 0.86) >14 0.45 (0.19e1.06) T stage 1 2 3 4 Nodal status 0 1 1 0.39 (0.07e2.14) 2 1.39 (0.47e4.09) 3 0.89 (0.09e8.03) AJCC stage 1 1 2 0.1.4 (0.3e7.44) 3 1.13 (0.25e5.12) 4a 2.28 (0.67e7.67) 4bc 3.78 (0.76e18.86) Tumour differentiation Well 1 Moderate 1.43 (0.56e 3.2) Poor 1.39 (0.22e4.3) Chemoradiotherapy Yes 1 No 1.30 (0.56e1.8) p16 status Positive 1 Negative 5.01 (2.44e10.3)

P value

Recurrence-free survival Hazard ratio (95% confidence interval)

0.43

1 1.94 (0.68e2.08)

0.9

1 1.49 (1.057e2.062) 0.83 (0.46e1.25)

0.24

1 1.25 (0.65e1.9) 1.56 (1.09e1.99)

0.014

1 0.70 (0.34e1.41) 0.27 (0.08e0.86)

0.001

1 0.66 (0.30e1.44) 0.91 (0.48e1.70)

<0.001

1 1.18 (0.34e4.02) 2.29 (0.65e8.06) 4.48 (1.35e14.8)

P value

Cancer-specific survival Hazard ratio (95% confidence interval)

P value

0.50

1 1.48 (0.78e0.65)

0.22

0.15

1 1.68 (1.19e2.36) 0.98 (0.54e1.32)

0.004

0.12

1 1.32 (0.61e2.77) 2.38 (1.12e4.93)

0.05

0.01

1 0.45 (0.17e1.17) 0.24 (0.08e0.60)

0.001

<0.0001

1 0.75 (0.33e1.69) 0.96 (0.49e1.88)

<0.0001

<0.001

1 1.43 (0.48e4.26) 2.77 (0.89e8.62) 5.30 (1.84e15.29)

<0.0001

0.007

1 0.70 (0.21e2.33) 1.41 (0.55e3.61) 3.73 (1.18e11.78)

0.003

0.008

1 1.47 1.18 2.31 5.79

0.003

0.015

0.026

1 1.03 (0.33e3.19) 1.67 (0.66e4.27) 4.15 (1.31e13.11)

0.03

1 1.30 1.40 2.46 6.13

0.49

1 2.09 (0.97e4.49) 1.73 (0.78e3.85)

0.019

1 2.61 (1.10e6.19) 2.02 (0.82e4.96)

0.45

1 1.97 (1.22e3.18)

0.005

1 1.78 (1.09e2.89)

0.03

<0.001

1 4.60 (2.79e7.60)

<0.0001

<0.001

(0.31e5.46) (0.39 e5.42) (0.76e7.94) (1.60e12.7)

1 6.07 (3.52e10.46)

(0.39e5.49) (0.34e4.06) (0.82e6.48) (1.78e16.86)

AJCC, American Joint Committee on Cancer.

p16 positivity is a phenotype independent of HPV. However, given the aetiological link between HPV and p16, this seems unlikely. These discrepancies could equally arise through technical rather than phenotypic differences. This is an area of debate and more importantly it is unclear whether tumours that are p16 positive but HPV negative share the same prognosis as p16-positive/HPV-positive patients [17,18]. Due to the frequency of discrepant cases, this would not be apparent in survival analysis, particularly in small

cohorts. This knowledge is essential for the appropriate stratification of patients in prospective studies and larger, well-powered studies are needed. Retrospectively, however, p16 performs better than stage in terms of prognostication. Presently, the National Comprehensive Cancer Network guidelinesÔ Version 2.2011 [34] recommend p16 IHC as a prognostic test. The British Association of Otorhinolaryngology on the other hand advices that in ‘routine clinical practice HPV testing is desirable but not mandatory’

636

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

Fig 3. KaplaneMeier curves of (A) and (B) comparing 5 year cancer-specific survival (CSS) and recurrence-free survival (RFS) of p16-negative patients according to their prognostic group (American Joint Committee on Cancer [AJCC] classification); (C) and (D) comparing 5 year CSS and RFS in p16-positive patients according to prognostic group.

[35]. Given the overwhelming evidence that p16-positive patients consistently do better, regardless of treatment modality, including this cohort, all pathology departments should include p16 IHC in routine pathological assessment

of OPSCC, as a simple, sensitive and robust initial triage of cases. We do not suggest limiting prognostication solely on the basis of p16 staining, rather we advocate its inclusion in any prognostic stratification, as available evidence suggest

Table 4 The multivariate Cox-regression analysis Characteristics Locoregional control P value Rank Recurrence-free P value Hazard ratio (95% survival Hazard ratio confidence interval) (95% confidence interval) Tumour differentiation Well Moderate Poor T stage 1 1 2 1.38 (0.28e6.57) 3 2.28 (0.41e12.6) 4 5.28 (1.19e23.4) Pretreatment haemoglobin 12 14 >14 p16 Positive 1 Negative 3.74 (1.76e7.95) NS, non-significant

1 2.03 (0.93e4.46) 2.01 (0.88e4.56)

Rank Cancer-specific P value survival Hazard ratio (95% confidence interval)

Rank

3

3

0.004

2

NS

0.007

1 0.001

1 0.37 (0.19e0.73) 0.14 (0.06e0.33)

0.001

1 6.15 (3.57e10.61)

<0.0001

1 13.23 (1.75e99.55) 4.52 (0.53e38.19) 1 2.60 (1.02e6.61) 4.58 (1.99e10.53) 3.94 (1.15e13.50)

0.01 2

0.001

2

1

NS

1 4.15 (2.43e7.08)

1 <0.0001

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638

it is the most important marker in survival of OPSCC. It is unclear what could be responsible for the differences in survival between p16-positive and -negative patients. One difference is the persistence of the pRB and p53 pathways in HPV-positive OPSCC and the loss in HPV-negative OPSCC [36e39], which may account for a better response to treatment strategies [16,40]. Furthermore, Stransky et al. [41] report that the mutation rate of HPV-positive tumours is half the rate of HPV-negative tumours [41]. The higher mutation rate in the HPV-negative tumours predisposes them to reoccurrence and resistance to therapy. In conclusion, our study shows that p16 expression is the most significant prognostic factor in OPSCC after radiotherapy. We also show that AJCC stage has no prognostic significance in p16-positive patients and as such p16 IHC should be included in patient risk stratification.

Acknowledgement This work was supported by Cancer Research UK Experimental Cancer Centre funding.

Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.clon.2013.07.003.

References [1] Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189(1):12e19. [2] 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 2000;92(9): 709e720. [3] Shiboski CH, Schmidt BL, Jordan RC. Tongue and tonsil carcinoma: increasing trends in the U.S. population ages 20e44 years. Cancer 2005;103(9):1843e1849. [4] Ramqvist T, Dalianis T. An epidemic of oropharyngeal squamous cell carcinoma (OSCC) due to human papillomavirus (HPV) infection and aspects of treatment and prevention. Anticancer Res 2011;31(5):1515e1519. [5] Ramqvist T, Dalianis T. Oropharyngeal cancer epidemic and human papillomavirus. Emerg Infect Dis 2010;16(11): 1671e1677. [6] Reddy S, Rangaiah J, Addiman S, Wareham D, Wilson P, Sefton A. Epidemiology, antibiotic resistance trends and the cost of enteric fever in East London, 2005e2010. Travel Med Infect Dis 2011;9(4):206e212. [7] Reddy VM, Cundall-Curry D, Bridger MW. Trends in the incidence rates of tonsil and base of tongue cancer in England, 1985e2006. Ann R Coll Surg Engl 2010;92(8):655e659. [8] Gillison ML, D’Souza G, Westra W, et al. Distinct risk factor profiles for human papillomavirus type 16-positive and human papillomavirus type 16-negative head and neck cancers. J Natl Cancer Inst 2008;100(6):407e420. [9] Weinberger PM, Yu Z, Haffty BG, et al. Prognostic significance of p16 protein levels in oropharyngeal squamous cell cancer. Clin Cancer Res 2004;10(17):5684e5691.

637

[10] Ragin CC, Taioli E. Survival of squamous cell carcinoma of the head and neck in relation to human papillomavirus infection: review and meta-analysis. Int J Cancer 2007;121(8): 1813e1820. [11] Pai SI, Westra WH. Molecular pathology of head and neck cancer: implications for diagnosis, prognosis, and treatment. Annu Rev Pathol 2009;4:49e70. [12] Robinson M, Sloan P, Shaw R. Refining the diagnosis of oropharyngeal squamous cell carcinoma using human papillomavirus testing. Oral Oncol 2010;46(7):492e496. [13] van der Riet P, Nawroz H, Hruban RH, et al. Frequent loss of chromosome 9p21-22 early in head and neck cancer progression. Cancer Res 1994;54(5):1156e1158. [14] Reed AL, Califano J, Cairns P, et al. High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation in head and neck squamous cell carcinoma. Cancer Res 1996;56(16):3630e3633. [15] Cairns P, Mao L, Merlo A, et al. Rates of p16 (MTS1) mutations in primary tumors with 9p loss. Science 1994;265(5170): 415e417. [16] Ang KK, Sturgis EM. Human papillomavirus as a marker of the natural history and response to therapy of head and neck squamous cell carcinoma. Semin Radiat Oncol 2012;22(2): 128e142. [17] Lewis Jr JS, Thorstad WL, Chernock RD, et al. p16 positive oropharyngeal squamous cell carcinoma: an entity with a favorable prognosis regardless of tumor HPV status. Am J Surg Pathol 2010;34(8):1088e1096. [18] Perrone F, Gloghini A, Cortelazzi B, Bossi P, Licitra L, Pilotti S. Isolating p16-positive/HPV-negative oropharyngeal cancer: an effort worth making. Am J Surg Pathol 2011;35(5):774e777. author reply 777e778. [19] Singhi AD, Westra WH. Comparison of human papillomavirus in situ hybridization and p16 immunohistochemistry in the detection of human papillomavirus-associated head and neck cancer based on a prospective clinical experience. Cancer 2010;116(9):2166e2173. [20] Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol 2011;29(32):4294e4301. [21] Lassen P, Eriksen JG, Hamilton-Dutoit S, Tramm T, Alsner J, Overgaard J. Effect of HPV-associated p16INK4A expression on response to radiotherapy and survival in squamous cell carcinoma of the head and neck. J Clin Oncol 2009;27(12): 1992e1998. [22] Dayyani F, Etzel CJ, Liu M, Ho CH, Lippman SM, Tsao AS. Metaanalysis of the impact of human papillomavirus (HPV) on cancer risk and overall survival in head and neck squamous cell carcinomas (HNSCC). Head Neck Oncol 2010;2:15. [23] Heath S, Willis V, Allan K, et al. Clinically significant human papilloma virus in squamous cell carcinoma of the head and neck in UK practice. Clin Oncol (R Coll Radiol) 2012;24(1): e18ee23. [24] Ang KK, Harris J, Wheeler R, et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 2010;363(1):24e35. [25] Fischer CA, Kampmann M, Zlobec, et al. p16 expression in oropharyngeal cancer: its impact on staging and prognosis compared with the conventional clinical staging parameters. Ann Oncol 2010;21(10):1961e1966. [26] Lassen P, Eriksen JG, Krogdahl A, et al. The influence of HPVassociated p16-expression on accelerated fractionated radiotherapy in head and neck cancer: evaluation of the randomised DAHANCA 6&7 trial. Radiother Oncol 2011;100(1):49e55. [27] Fakhry C, Westra WH, Li S, et al. Improved survival of patients with human papillomavirus-positive head and neck

638

[28]

[29]

[30]

[31]

[32]

[33]

[34]

K.K. Oguejiofor et al. / Clinical Oncology 25 (2013) 630e638 squamous cell carcinoma in a prospective clinical trial. J Natl Cancer Inst 2008;100(4):261e269. Posner MR, Lorch JH, Goloubeva O, et al. Survival and human papillomavirus in oropharynx cancer in TAX 324: a subset analysis from an international phase III trial. Ann Oncol 2011;22(5):1071e1077. Rischin D, Young RJ, Fisher R, et al. Prognostic significance of p16INK4A and human papillomavirus in patients with oropharyngeal cancer treated on TROG 02.02 phase III trial. J Clin Oncol 2010;28(27):4142e4148. Ukpo OC, Flanagan JJ, Ma XJ, Luo Y, Thorstad WL, Lewis Jr JS. High-risk human papillomavirus E6/E7 mRNA detection by a novel in situ hybridization assay strongly correlates with p16 expression and patient outcomes in oropharyngeal squamous cell carcinoma. Am J Surg Pathol 2011;35(9):1343e1350. Granata R, Miceli R, Orlandi E, et al. Tumor stage, human papillomavirus and smoking status affect the survival of patients with oropharyngeal cancer: an Italian validation study. Ann Oncol 2012;23(7):1832e1837. 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 2007;120(8): 1731e1738. Mehanna H, West CM, Nutting C, Paleri V. Head and neck cancerePart 2: Treatment and prognostic factors. Br Med J 2010;341:c4690. Zhang B, Tang PZ. [Introduction to NCCN clinical practice guidelines in head and neck cancer (2009 Chinese version)].

[35] [36]

[37]

[38]

[39]

[40]

[41]

Zhonghua Er Bi Yan Hou Tou Jing Wai Ke Za Zhi 2009;44(9):707e709. Roland NJ. Head and neck cancer: multidisciplinary management guidelines, 4th ed. London: BAHNO; 2011. Hafkamp HC, Speel EJ, Haesevoets A, et al. A subset of head and neck squamous cell carcinomas exhibits integration of HPV 16/18 DNA and overexpression of p16INK4A and p53 in the absence of mutations in p53 exons 5-8. Int J Cancer 2003;107(3):394e400. 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):736e747. Kong CS, Narasimhan B, Cao H, et al. The relationship between human papillomavirus status and other molecular prognostic markers in head and neck squamous cell carcinomas. Int J Radiat Oncol Biol Phys 2009;74(2):553e561. 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 2006; 24(36):5630e5636. Olthof NC, Straetmans JM, Snoeck R, et al. Next-generation treatment strategies for human papillomavirus-related head and neck squamous cell carcinoma: where do we go? Rev Med Virol 2012;22(2):88e105. Stransky N, Egloff AM, Tward AD, et al. The mutational landscape of head and neck squamous cell carcinoma. Science 2011;333(6046):1157e1160.