The Laryngoscope C 2014 The American Laryngological, V
Rhinological and Otological Society, Inc.
Squamous Cell Carcinoma of the Head and Neck Outside the Oropharynx Is Rarely Human Papillomavirus Related Navdeep S. Upile, BmedSci; Richard J. Shaw, MD, FRCS (OMFS); Terry M. Jones, MD, FRCS (ORL-HNS); Paul Goodyear, FRCS (ORL-HNS); Triantafillos Liloglou, PhD; Janet M. Risk, PhD; Mark T. Boyd, PhD; Jon Sheard, FRCPath; Phil Sloan, FRCPath; Max Robinson, FRCPath; Andrew G. Schache, PhD, FRCS (OMFS) Objectives/Hypothesis: The incidence of human papillomavirus (HPV)-driven disease beyond the oropharynx varies greatly in the reported literature. Study Design: Case series. Methods: Two hundred twenty-one samples were strictly classified to the subsites of oral cavity, larynx, or hypopharynx at the time of primary surgery. Formalin-fixed paraffin-embedded samples were subjected to a validated, tiered, diagnostic algorithm of p16 immunohistochemistry, high-risk HPV in situ hybridization, and quantitative polymerase chain reaction for HPV E6 DNA. An additional 60 oropharyngeal cases acted as an internal biological control. Results: An incidence of 4% of HPV-driven cases was observed across the subsites outside the oropharynx compared to 70% of tumors confined within it. Conclusions: This is the first reporting of a broad range of nonoropharyngeal HPV rates using this validated diagnostic algorithm. It remains unclear whether patients with HPV-driven disease originating outside the oropharynx enjoy the same survival advantage apparent in those patients with oropharyngeal squamous cell carcinomas. Key Words: Oral, laryngeal, hypopharyngeal, human papillomavirus, squamous cell carcinoma, incidence. Level of Evidence: 4 Laryngoscope, 124:2739–2744, 2014
INTRODUCTION The evidence that human papillomavirus (HPV) is etiologically linked to squamous cell carcinoma (SCC) of the oropharynx is consistent, with published metaanalyses converging on a proportion of approximately 60% of cases being HPV-mediated.1–3 These findings are of immediate clinical relevance. Oropharyngeal SCC is increasing rapidly in incidence, yet the HPV-related cases carry a superior prognosis.4 Importantly, their differing molecular pathogenesis will direct contrasting prevention strategies and may offer alternative thera-
From the Mersey Head and Neck Oncology Research Group (N.S.U., Department of Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, Cancer Research Centre, University of Liverpool, Liverpool; University Hospital Aintree (N.S.U., R.J.S., T.M.J., P.G., J.S., A.G.S.), Liverpool; and the School of Dental Sciences (P.S., M.R.), Newcastle University, Newcastle Upon Tyne, United Kingdom. Editor’s Note: This Manuscript was accepted for publication June 16, 2014. A.G.S. was the recipient of a Wellcome Trust–FDS Research Training Fellowship. N.S.U. was the recipient of a Royal College of Surgeons Research Fellowship. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Navdeep Upile, BmedSci, North West Cancer Research Centre, Department of Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, University of Liverpool, 200 London Road, Liverpool, UK L3 9TA. E-mail: [email protected] R.J.S., T.M.J., P.G., L.T., J.M.R., M.T.B., A.G.S.),
DOI: 10.1002/lary.24828
Laryngoscope 124: December 2014
peutic avenues, or in the case of HPV-positive disease, potentially de-escalation of therapeutic intensity. The role of HPV in other nonoropharyngeal, head and neck SCC subsites has generated substantially greater controversy. In the oral cavity, the most commonly involved nonoropharyngeal head and neck squamous cell carcinoma (HNSCC) site, 16.8% of 4,195 reported cases have been reported as positive5; however, within these data there is a wide distribution including several reports detailing an HPV-positive rate 50%.8,9 In the larynx, 23.6% of 1,712 reported cases are HPV positive, with a similarly wide range reported.5 In the head and neck field, risk factors such as smoking and alcohol are evidently shared across all sites, and there might appear to be a case for presuming the same for exposure to high-risk HPV. Seen in this light, a 15% to 25% rate of HPV appears etiologically and epidemiologically plausible. It has, however, been shown in more recent and methodologically sound studies that in the case of oral cavity SCC, the percentage of cases is considerably lower, typically around 5%.6,10 It is not surprising that these widely contrasting data have generated considerable confusion. The disagreement between previously published data may reflect true differences between cohort demographics and sampled populations, or the era of tissue collection. However, it is also conceivable that this reflects inconsistency within the HPV testing protocols applied and potential misallocation of tumors to other head and neck subsites, in particular misclassified oropharyngeal Upile et al.: HNSCC Outside the Oropharynx and HPV
2739
tumors. Although the latter is simply a matter for improved tissue bank data, there has been progress in the understanding of what might represent sufficient laboratory data to define HPV etiology in HNSCC. Although there have been proponents for rigor in testing protocol,11–13 much of the published data reflect results from rather less-specific and sometimes inappropriately sensitive techniques for detection of viral DNA that may be prone to polymerase chain reaction (PCR) contamination or other errors. Our previous work has been in validation and comparison between available tests and test combinations14,15 in oropharyngeal SCC. In this work we present data on a previously untested cohort of SCC in the oral cavity, larynx, hypopharynx as well as oropharynx with a uniform HPV testing protocol. We aim to clarify the contribution of HPV in nonoropharynx SCC in a geographically defined population served by a single treatment center.
MATERIALS AND METHODS A retrospective analysis of cases of HNSCC with ethical approval (Ref:07/MRE08/47/South Sefton Research Ethics Committee, EC.47.01–6; North West 5 Research Ethics Committee, EC.09.H1010.5), originating from the period covering 1997 to 2010 (larynx and hypopharynx cases between 97-08 and oral and oropharynx 03-10) was undertaken using tissue microarrays (TMAs). All cases were strictly classified by head and neck subsite (oral, laryngeal, hypopharyngeal, or oropharyngeal) at the time of tissue procurement. For clarity, cases included in this cohort are separate and distinct from those reported upon in recent diagnostic and prognostic analyses.14,15 Patient demographics and tumor characteristics were collated from tissue bank records. Details of tobacco smoking and alcohol consumption were unavailable for inclusion.
TMA Construction Tumor-rich areas of the donor blocks were identified by an experienced pathologist. Hematoxylin and eosin (H&E)-stained slides were examined microscopically, and epithelium-derived tumor areas were circumscribed. A Beecher manual tissue microarrayer MTA-1 (Beecher Instruments, Inc., Sun Prairie, WI) was used to construct TMAs. Donor tissue blocks were at least 4-mm thick to ensure adequate core length. Triplicate tumor cores and matched normal 0.6-mm cores were sited on the recipient blocks. After construction, a cut section of the TMA was again stained with H&E to ensure accurate tumor sampling and transfer.
cases where fresh-frozen tissue samples corresponding to the formalin-fixed, paraffin-embedded (FFPE) block were available, DNA extraction was undertaken using the AllPrep DNA/RNA Mini Kit (Qiagen).
HPV Detection p16 immunohistochemistry. p16 immunohistochemistry (IHC) was carried out using a proprietary kit (CINtec Histology; Roche mtm Laboratories AG, Heidelberg, Germany) on a Benchmark Autostainer (Ventana Medical Systems Inc., Oro Valley, AZ). A tonsil SCC with high p16 expression was used as a positive control. The primary antibody was omitted from negative controls. A binary scoring system was adopted (positive vs. negative). p16 IHC was scored as positive if there was strong and diffuse nuclear and cytoplasmic staining present in >70% of the malignant cells.13,16,17 All other staining patterns were scored as negative. High-risk HPV in situ hybridization. High-risk HPV in situ hybridization (HR-HPV ISH) was carried out using proprietary reagents (Inform HPV III Family 16 Probe (B); Ventana Medical Systems Inc.) on a Benchmark Autostainer (Ventana Medical Systems Inc.). The Inform HPV III Family 16 Probe (B) detects high-risk genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 66. FFPE samples of three cell lines were used as control samples; CaSki cells (HPV16 positive; 600 copies per cell), HeLa cells (HPV18 positive; 10–50 copies per cell), and C-33A (HPV-negative cell line) (Ventana Medical Systems Inc.). A binary scoring system was used (positive vs. negative). The HRHPV ISH test was scored as positive if there was a blue reaction product that colocalized with the nuclei of malignant cells as previously described.16 HPV16 E6 DNA quantitative PCR. HPV16 E6 DNA quantitative PCR (qPCR) was undertaken as previously described.14 Briefly, primers and a FAM_MGB-labelled TaqMan probe, specific to the E6 region of HPV16, were synthesized. Multiplex real-time PCR reactions with the above primer/probe pair were undertaken in conjunction with the commercially available primers and a VIC-TAMRA–labelled probe for the single copy gene RNaseP (Taqman RNase P Control Reagents; Life Technologies, Carlsbad, CA). HPV 16-positive CaSki cell lines and SiHa were used as internal positive controls. Normal bronchial epithelial lung cell line DNA was used as a negative control. Real-time PCR reactions were undertaken in duplicate for all samples on an Applied Biosystems 7500 FAST system (Life Technologies). Assuming an HPV16-driven tumor is composed of a dominant clonal population of cells, we scored as positive those samples with one or more E6 gene copy/diploid genome. A sample was only deemed positive if the threshold was met in both of the duplicate runs.
p16 IHC and HR-HPV ISH Analysis DNA Extraction The blocks from which the TMA was constructed were used for DNA extraction using the QIAamp DNA FFPE Tissue Kit (Qiagen, Venlo, the Netherlands). Optimization of the manufacturer’s protocol (available from Qiagen) was undertaken to maximize DNA yields. This involved extending the manufacturers recommended 1 hour of incubation with proteinase K to 2 hours, employing RNase A and increasing the incubation time of the final buffer with the DNA-containing membrane from 5 minutes to 30 minutes. DNA yields from either protocol were quantified by spectrophotometry using the Nano-Drop 1000 (Thermo-Fisher Scientific, Waltham, MA). In a small number of
Laryngoscope 124: December 2014
2740
The p16 IHC and HR-HPV ISH tests were examined independently by two pathologists (M.R., P.S.). The p16 IHC and HRHPV ISH tests were assessed separately alongside corresponding H&E-stained sections. The scores were collated, and discordant scores were resolved at a meeting between the pathologists to establish a consensus score for each core. The study coordinator compiled the data for the individual cores to formulate the final classification for each case. Cases that had positive results for both p16 IHC and HRHPV ISH were assigned as HPV positive. If only HR-HPV ISH alone was positive, this was scored as negative. The rationale of this is that the chromogenic in situ hybridization only highlights the presence of the HPV DNA, but not viral transcription.
Upile et al.: HNSCC Outside the Oropharynx and HPV
TABLE I. Demographic, Subsite, and Staging Data. Oropharynx
Age, yr Median
53
Mean
55
Gender Male
43
Female Total Subsites of primary tumor Tonsil
25
Base of tongue Soft palate
18 4
Posterior pharyngeal wall
2
Tonsil and base of tongue Tonsil and soft palate
4 5
Base of tongue and soft palate Fig. 1. Tiered diagnostic algorithm used for all samples. CISH 5 chromogenic in situ hybridization; HPV 5 human papillomavirus; HR 5 high risk; IHC 5 immunohistochemical; qPCR 5 quantitative polymerase chain reaction.
Total pT classification
13
2 3
33 12 0
Without evidence of downstream effects, such as disruption of the cell cycle, HPV chromogenic in situ hybridization positivity alone can be viewed as a bystander, without exerting its oncogenic potential. However, if the core was positive for p16 alone, these samples warranted further investigation using HPV16 E6 DNA qPCR. This tiered diagnostic algorithm (Fig. 1) confers greater sensitivity and specificity than the individual tests alone, and a combination of the tests highlights not only the presence but also the downstream effects of HPV integration.
Total pN classification
RESULTS
AJCC/UICC stage 1
58 (96.7%)
0
8
1 2a
7 15
2b
27
2c 3
2 0
Total
Laryngoscope 124: December 2014
1 59 (98.3%)
1
4
Two hundred eighty-one cases of HNSCC were available for TMA construction and were analyzed using the HPV diagnostic testing algorithm (Fig. 1). The case mix, stratified by subsite, and associated demographic data are depicted in Table I. The proportion of tumors demonstrating positive results by the HPV testing algorithm are detailed in Table II, including the biological control of the oropharynx sites. Subsites beyond the oropharynx demonstrated a substantially lower proportion of HPV cases by comparison to those within the oropharynx. The proportion deemed HPV positive was 4.1% overall, and the differences between oral (4%), larynx (3.2%), and hypopharynx (7.1%) were nonsignificant. Tumors from the oropharyngeal control group demonstrated an HPV-positive rate of 70%. Table III provides case specifics for each of the nonoropharyngeal HPV-positive tumors. HPV positive oral cavity tumors (n 5 4) were exclusively confined to the floor of mouth subsite, whereas those in the larynx (n 5 3) were found in the supraglottis (n 5 1) and transglottic (n 5 2) originating above in one instance and below the glottis in the second instance. Two hypopharyngeal SCCs were shown to be HPV positive.
16 59 (98.3%)
59 (98.3%) 3
2
3
3 4a
9 44
Total
59 (98.3%)
Oral cavity
Age, yr Median Mean Gender Male Female Total
58 57.1 72 28 100 (98%)
Subsite of primary tumor Floor of mouth Buccal
28 5
Alveolar (mandible/maxilla)
15
Oral tongue (anterior two-thirds) Retromolar trigone
43 9
Upile et al.: HNSCC Outside the Oropharynx and HPV
2741
TABLE I. (Continued)
TABLE I. (Continued)
Oral cavity
Other oral cavity Total pT classification
Larynx
2 102 (100%)
3
1
Total AJCC/UICC stage
91(100%)
1
6
1
5
2 3
43 6
2 3
14 26
36
4a
4 Total pN classification
91 (89%)
0
36
1 2a
11 1
2b
20
2c 3
8 0
Total AJCC/UICC stage 1
76 (75%) 6
2
15
3 4a
10 48
Total
68 (67%)
Larynx
Age, yr Median Mean
63 64
Gender Male Female Total Subsite of primary tumor Supraglottis
80 11 91 (100%) 24
Glottis
14
Subglottis Supraglottis and glottis
6 14
Glottis and subglottis
12
All three subsites Total
21 91 (100%)
pT classification 1 2
5 18
3
34
4 Total
34 91 (100%)
pN classification 0 1
55 10
2a
2
2b 2c
12 11
Laryngoscope 124: December 2014
2742
46
Total
91 (100%)
Hypopharynx
Age, yr Median Mean
61 61
Gender Male Female Total Subsite of primary tumor Hypopharynx Total pT classification 1
22 6 28 (100%) 28 28 (100%) 6
2
8
3 4
9 5
Total pN classification 0
28 (100%) 6
1
4
2a 2b
2 13
2c
3
3 Total
0 28 (100%)
AJCC/UICC stage 1 2
4 2
3
3
4a Total
19 28 (100%)
AJCC/UICC 5 American Joint Committee on Cancer/International Union Against Cancer.
DISCUSSION Our data show that the proportion of SCCs attributable to HPV in oral, laryngeal, and hypopharyngeal sites is substantially lower than for oropharynx. The use of a strict site classification, a substantial cohort size, and a rigorous validated testing protocol ensures that these data can be seen as robust. Upile et al.: HNSCC Outside the Oropharynx and HPV
TABLE II. Summary of Testing Results. Test, n (%) Site
No.
p16 IHC1
p16 IHC1 and HR HPV ISH 1
DNA qPCR1(in p161 and ISH 2 cases)
HPV1
Nonoropharynx sites Oral
102
8 (8%)
4(4%)
0
9 (4.1%) 4 (4%)
91
6 (6.6%)
3 (3.2%)
0
3 (3.2%)
Hypopharynx Oropharynx
Larynx
28 60
3 (10.7%) 46 (76.7%)
2 (7.1%) 36 (60%)
0 6
2 (7.1%) 42 (70%)
Total
281
63 (22.4%)
45 (16%)
6/18 (33.3%)
51 (18.1%)
HPV 5 human papillomavirus; HR 5 high risk; IHC 5 immunohistochemical; ISH 5 in situ hybridization; qPCR 5 quantitative polymerase chain reaction.
Through the comparison of both oropharyngeal and nonoropharyngeal tumors within the same cohort, the 70% positive rate in the oropharynx can be seen as a biological positive control to further emphasize the small fraction of HPV positivity (4%) within other sites. Without doubt, many biomarker and molecular studies suffer from elements of bias in that research is only performed on the subset of the tissue with consent and when available in the tissue bank. In our research group, it has been more common for patients receiving surgery as a primary modality to be approached for consent to tissue banking than for those having nonsurgical approaches or palliative treatment. In oropharyngeal SCC, this is perhaps unlikely to unduly bias the results, as surgery is the mainstay of treatment in our center.18
TABLE III. Tumor and Patient Demographics for Human Papillomavirus-Positive Nonoropharyngeal Tumors. Subsite
Oral cavity 1 2
59-year-old female, pT3N3b SCC (stage 4a), floor of mouth 48-year-old female, pT2N2c SCC (stage 4a), floor of mouth
3
71-year-old male, pT2N0 SCC (stage 2), floor of mouth
4
72-year-old male, pT4N0 SCC (stage 4a), floor of mouth
Larynx 5
52-year-old male, pT2N2a SCC (stage 4a), supraglottic
6
49-year-old male, pT2N0 SCC (stage 2), supraglottis and glottic
7
70-year-old male, pT2N0 SCC (stage 2), glottic and subglottis
Hypopharynx 8 9
74-year-old male, pT4N2a SCC (stage 4a), hypopharynx 60-year-old female, pT3N1 SCC (stage 3), hypopharynx
Laryngoscope 124: December 2014
In the larynx and hypopharynx, there may be a more significant bias. However, we cannot conclude that this would substantially affect the HPV rate; from a pragmatic viewpoint the characteristics of surgically treated larynx cancer appear not to skew the chance of being HPV linked. In selecting this cohort, specific care has been taken to select cases where site has been determined prior to tissue banking as a prospective collection, and we have not relied on retrospectively analyzed pathology reports for these data. The oropharynx cases selected for this study are an entirely different cohort from our previously published work14 with no overlap. Given the trend toward increasing HPV- positive rates in oropharyngeal squamous cell carcinoma (OPSCC) within Western countries, the HPV-positive rate within the current series (70%) has risen by comparison to rates detailed from the same geographic population from that previous study (57%, 1988–2009). This confirms that the testing carried out on predominantly fixed tissue in this series is comparable to our previous diagnostic data that were established on snap-frozen tissue.14 It is readily acknowledged that the primary site of tumors arising adjacent to the oropharynx, in particular the tonsil, can be difficult to classify by site and therefore introduce potential for misclassification. It is, however, a strength of the tissue procurement system utilized that all attempts have been made at the time of diagnosis and resection to differentiate one from the other. Reassuringly, nonoropharynx HPV-positive cases identified are not simply in sites adjacent to the oropharynx, such as the retromolar triangle, but are instead originating in anatomically discrete sites such as the floor of the mouth. The HPV testing diagnostic algorithm is robust, yet the inclusion of DNA qPCR for HPV16 alone may underestimate positive cases in the ultimate tier of analysis. In previous work, we have found, albeit only in rare instances, that sub-types other than HPV16 are detectable. We would, however, defend this protocol on the basis that frequency with which subtypes other than HPV16 occurs in isolation is
Rhinological and Otological Society, Inc.
Squamous Cell Carcinoma of the Head and Neck Outside the Oropharynx Is Rarely Human Papillomavirus Related Navdeep S. Upile, BmedSci; Richard J. Shaw, MD, FRCS (OMFS); Terry M. Jones, MD, FRCS (ORL-HNS); Paul Goodyear, FRCS (ORL-HNS); Triantafillos Liloglou, PhD; Janet M. Risk, PhD; Mark T. Boyd, PhD; Jon Sheard, FRCPath; Phil Sloan, FRCPath; Max Robinson, FRCPath; Andrew G. Schache, PhD, FRCS (OMFS) Objectives/Hypothesis: The incidence of human papillomavirus (HPV)-driven disease beyond the oropharynx varies greatly in the reported literature. Study Design: Case series. Methods: Two hundred twenty-one samples were strictly classified to the subsites of oral cavity, larynx, or hypopharynx at the time of primary surgery. Formalin-fixed paraffin-embedded samples were subjected to a validated, tiered, diagnostic algorithm of p16 immunohistochemistry, high-risk HPV in situ hybridization, and quantitative polymerase chain reaction for HPV E6 DNA. An additional 60 oropharyngeal cases acted as an internal biological control. Results: An incidence of 4% of HPV-driven cases was observed across the subsites outside the oropharynx compared to 70% of tumors confined within it. Conclusions: This is the first reporting of a broad range of nonoropharyngeal HPV rates using this validated diagnostic algorithm. It remains unclear whether patients with HPV-driven disease originating outside the oropharynx enjoy the same survival advantage apparent in those patients with oropharyngeal squamous cell carcinomas. Key Words: Oral, laryngeal, hypopharyngeal, human papillomavirus, squamous cell carcinoma, incidence. Level of Evidence: 4 Laryngoscope, 124:2739–2744, 2014
INTRODUCTION The evidence that human papillomavirus (HPV) is etiologically linked to squamous cell carcinoma (SCC) of the oropharynx is consistent, with published metaanalyses converging on a proportion of approximately 60% of cases being HPV-mediated.1–3 These findings are of immediate clinical relevance. Oropharyngeal SCC is increasing rapidly in incidence, yet the HPV-related cases carry a superior prognosis.4 Importantly, their differing molecular pathogenesis will direct contrasting prevention strategies and may offer alternative thera-
From the Mersey Head and Neck Oncology Research Group (N.S.U., Department of Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, Cancer Research Centre, University of Liverpool, Liverpool; University Hospital Aintree (N.S.U., R.J.S., T.M.J., P.G., J.S., A.G.S.), Liverpool; and the School of Dental Sciences (P.S., M.R.), Newcastle University, Newcastle Upon Tyne, United Kingdom. Editor’s Note: This Manuscript was accepted for publication June 16, 2014. A.G.S. was the recipient of a Wellcome Trust–FDS Research Training Fellowship. N.S.U. was the recipient of a Royal College of Surgeons Research Fellowship. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Navdeep Upile, BmedSci, North West Cancer Research Centre, Department of Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, University of Liverpool, 200 London Road, Liverpool, UK L3 9TA. E-mail: [email protected] R.J.S., T.M.J., P.G., L.T., J.M.R., M.T.B., A.G.S.),
DOI: 10.1002/lary.24828
Laryngoscope 124: December 2014
peutic avenues, or in the case of HPV-positive disease, potentially de-escalation of therapeutic intensity. The role of HPV in other nonoropharyngeal, head and neck SCC subsites has generated substantially greater controversy. In the oral cavity, the most commonly involved nonoropharyngeal head and neck squamous cell carcinoma (HNSCC) site, 16.8% of 4,195 reported cases have been reported as positive5; however, within these data there is a wide distribution including several reports detailing an HPV-positive rate 50%.8,9 In the larynx, 23.6% of 1,712 reported cases are HPV positive, with a similarly wide range reported.5 In the head and neck field, risk factors such as smoking and alcohol are evidently shared across all sites, and there might appear to be a case for presuming the same for exposure to high-risk HPV. Seen in this light, a 15% to 25% rate of HPV appears etiologically and epidemiologically plausible. It has, however, been shown in more recent and methodologically sound studies that in the case of oral cavity SCC, the percentage of cases is considerably lower, typically around 5%.6,10 It is not surprising that these widely contrasting data have generated considerable confusion. The disagreement between previously published data may reflect true differences between cohort demographics and sampled populations, or the era of tissue collection. However, it is also conceivable that this reflects inconsistency within the HPV testing protocols applied and potential misallocation of tumors to other head and neck subsites, in particular misclassified oropharyngeal Upile et al.: HNSCC Outside the Oropharynx and HPV
2739
tumors. Although the latter is simply a matter for improved tissue bank data, there has been progress in the understanding of what might represent sufficient laboratory data to define HPV etiology in HNSCC. Although there have been proponents for rigor in testing protocol,11–13 much of the published data reflect results from rather less-specific and sometimes inappropriately sensitive techniques for detection of viral DNA that may be prone to polymerase chain reaction (PCR) contamination or other errors. Our previous work has been in validation and comparison between available tests and test combinations14,15 in oropharyngeal SCC. In this work we present data on a previously untested cohort of SCC in the oral cavity, larynx, hypopharynx as well as oropharynx with a uniform HPV testing protocol. We aim to clarify the contribution of HPV in nonoropharynx SCC in a geographically defined population served by a single treatment center.
MATERIALS AND METHODS A retrospective analysis of cases of HNSCC with ethical approval (Ref:07/MRE08/47/South Sefton Research Ethics Committee, EC.47.01–6; North West 5 Research Ethics Committee, EC.09.H1010.5), originating from the period covering 1997 to 2010 (larynx and hypopharynx cases between 97-08 and oral and oropharynx 03-10) was undertaken using tissue microarrays (TMAs). All cases were strictly classified by head and neck subsite (oral, laryngeal, hypopharyngeal, or oropharyngeal) at the time of tissue procurement. For clarity, cases included in this cohort are separate and distinct from those reported upon in recent diagnostic and prognostic analyses.14,15 Patient demographics and tumor characteristics were collated from tissue bank records. Details of tobacco smoking and alcohol consumption were unavailable for inclusion.
TMA Construction Tumor-rich areas of the donor blocks were identified by an experienced pathologist. Hematoxylin and eosin (H&E)-stained slides were examined microscopically, and epithelium-derived tumor areas were circumscribed. A Beecher manual tissue microarrayer MTA-1 (Beecher Instruments, Inc., Sun Prairie, WI) was used to construct TMAs. Donor tissue blocks were at least 4-mm thick to ensure adequate core length. Triplicate tumor cores and matched normal 0.6-mm cores were sited on the recipient blocks. After construction, a cut section of the TMA was again stained with H&E to ensure accurate tumor sampling and transfer.
cases where fresh-frozen tissue samples corresponding to the formalin-fixed, paraffin-embedded (FFPE) block were available, DNA extraction was undertaken using the AllPrep DNA/RNA Mini Kit (Qiagen).
HPV Detection p16 immunohistochemistry. p16 immunohistochemistry (IHC) was carried out using a proprietary kit (CINtec Histology; Roche mtm Laboratories AG, Heidelberg, Germany) on a Benchmark Autostainer (Ventana Medical Systems Inc., Oro Valley, AZ). A tonsil SCC with high p16 expression was used as a positive control. The primary antibody was omitted from negative controls. A binary scoring system was adopted (positive vs. negative). p16 IHC was scored as positive if there was strong and diffuse nuclear and cytoplasmic staining present in >70% of the malignant cells.13,16,17 All other staining patterns were scored as negative. High-risk HPV in situ hybridization. High-risk HPV in situ hybridization (HR-HPV ISH) was carried out using proprietary reagents (Inform HPV III Family 16 Probe (B); Ventana Medical Systems Inc.) on a Benchmark Autostainer (Ventana Medical Systems Inc.). The Inform HPV III Family 16 Probe (B) detects high-risk genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 66. FFPE samples of three cell lines were used as control samples; CaSki cells (HPV16 positive; 600 copies per cell), HeLa cells (HPV18 positive; 10–50 copies per cell), and C-33A (HPV-negative cell line) (Ventana Medical Systems Inc.). A binary scoring system was used (positive vs. negative). The HRHPV ISH test was scored as positive if there was a blue reaction product that colocalized with the nuclei of malignant cells as previously described.16 HPV16 E6 DNA quantitative PCR. HPV16 E6 DNA quantitative PCR (qPCR) was undertaken as previously described.14 Briefly, primers and a FAM_MGB-labelled TaqMan probe, specific to the E6 region of HPV16, were synthesized. Multiplex real-time PCR reactions with the above primer/probe pair were undertaken in conjunction with the commercially available primers and a VIC-TAMRA–labelled probe for the single copy gene RNaseP (Taqman RNase P Control Reagents; Life Technologies, Carlsbad, CA). HPV 16-positive CaSki cell lines and SiHa were used as internal positive controls. Normal bronchial epithelial lung cell line DNA was used as a negative control. Real-time PCR reactions were undertaken in duplicate for all samples on an Applied Biosystems 7500 FAST system (Life Technologies). Assuming an HPV16-driven tumor is composed of a dominant clonal population of cells, we scored as positive those samples with one or more E6 gene copy/diploid genome. A sample was only deemed positive if the threshold was met in both of the duplicate runs.
p16 IHC and HR-HPV ISH Analysis DNA Extraction The blocks from which the TMA was constructed were used for DNA extraction using the QIAamp DNA FFPE Tissue Kit (Qiagen, Venlo, the Netherlands). Optimization of the manufacturer’s protocol (available from Qiagen) was undertaken to maximize DNA yields. This involved extending the manufacturers recommended 1 hour of incubation with proteinase K to 2 hours, employing RNase A and increasing the incubation time of the final buffer with the DNA-containing membrane from 5 minutes to 30 minutes. DNA yields from either protocol were quantified by spectrophotometry using the Nano-Drop 1000 (Thermo-Fisher Scientific, Waltham, MA). In a small number of
Laryngoscope 124: December 2014
2740
The p16 IHC and HR-HPV ISH tests were examined independently by two pathologists (M.R., P.S.). The p16 IHC and HRHPV ISH tests were assessed separately alongside corresponding H&E-stained sections. The scores were collated, and discordant scores were resolved at a meeting between the pathologists to establish a consensus score for each core. The study coordinator compiled the data for the individual cores to formulate the final classification for each case. Cases that had positive results for both p16 IHC and HRHPV ISH were assigned as HPV positive. If only HR-HPV ISH alone was positive, this was scored as negative. The rationale of this is that the chromogenic in situ hybridization only highlights the presence of the HPV DNA, but not viral transcription.
Upile et al.: HNSCC Outside the Oropharynx and HPV
TABLE I. Demographic, Subsite, and Staging Data. Oropharynx
Age, yr Median
53
Mean
55
Gender Male
43
Female Total Subsites of primary tumor Tonsil
25
Base of tongue Soft palate
18 4
Posterior pharyngeal wall
2
Tonsil and base of tongue Tonsil and soft palate
4 5
Base of tongue and soft palate Fig. 1. Tiered diagnostic algorithm used for all samples. CISH 5 chromogenic in situ hybridization; HPV 5 human papillomavirus; HR 5 high risk; IHC 5 immunohistochemical; qPCR 5 quantitative polymerase chain reaction.
Total pT classification
13
2 3
33 12 0
Without evidence of downstream effects, such as disruption of the cell cycle, HPV chromogenic in situ hybridization positivity alone can be viewed as a bystander, without exerting its oncogenic potential. However, if the core was positive for p16 alone, these samples warranted further investigation using HPV16 E6 DNA qPCR. This tiered diagnostic algorithm (Fig. 1) confers greater sensitivity and specificity than the individual tests alone, and a combination of the tests highlights not only the presence but also the downstream effects of HPV integration.
Total pN classification
RESULTS
AJCC/UICC stage 1
58 (96.7%)
0
8
1 2a
7 15
2b
27
2c 3
2 0
Total
Laryngoscope 124: December 2014
1 59 (98.3%)
1
4
Two hundred eighty-one cases of HNSCC were available for TMA construction and were analyzed using the HPV diagnostic testing algorithm (Fig. 1). The case mix, stratified by subsite, and associated demographic data are depicted in Table I. The proportion of tumors demonstrating positive results by the HPV testing algorithm are detailed in Table II, including the biological control of the oropharynx sites. Subsites beyond the oropharynx demonstrated a substantially lower proportion of HPV cases by comparison to those within the oropharynx. The proportion deemed HPV positive was 4.1% overall, and the differences between oral (4%), larynx (3.2%), and hypopharynx (7.1%) were nonsignificant. Tumors from the oropharyngeal control group demonstrated an HPV-positive rate of 70%. Table III provides case specifics for each of the nonoropharyngeal HPV-positive tumors. HPV positive oral cavity tumors (n 5 4) were exclusively confined to the floor of mouth subsite, whereas those in the larynx (n 5 3) were found in the supraglottis (n 5 1) and transglottic (n 5 2) originating above in one instance and below the glottis in the second instance. Two hypopharyngeal SCCs were shown to be HPV positive.
16 59 (98.3%)
59 (98.3%) 3
2
3
3 4a
9 44
Total
59 (98.3%)
Oral cavity
Age, yr Median Mean Gender Male Female Total
58 57.1 72 28 100 (98%)
Subsite of primary tumor Floor of mouth Buccal
28 5
Alveolar (mandible/maxilla)
15
Oral tongue (anterior two-thirds) Retromolar trigone
43 9
Upile et al.: HNSCC Outside the Oropharynx and HPV
2741
TABLE I. (Continued)
TABLE I. (Continued)
Oral cavity
Other oral cavity Total pT classification
Larynx
2 102 (100%)
3
1
Total AJCC/UICC stage
91(100%)
1
6
1
5
2 3
43 6
2 3
14 26
36
4a
4 Total pN classification
91 (89%)
0
36
1 2a
11 1
2b
20
2c 3
8 0
Total AJCC/UICC stage 1
76 (75%) 6
2
15
3 4a
10 48
Total
68 (67%)
Larynx
Age, yr Median Mean
63 64
Gender Male Female Total Subsite of primary tumor Supraglottis
80 11 91 (100%) 24
Glottis
14
Subglottis Supraglottis and glottis
6 14
Glottis and subglottis
12
All three subsites Total
21 91 (100%)
pT classification 1 2
5 18
3
34
4 Total
34 91 (100%)
pN classification 0 1
55 10
2a
2
2b 2c
12 11
Laryngoscope 124: December 2014
2742
46
Total
91 (100%)
Hypopharynx
Age, yr Median Mean
61 61
Gender Male Female Total Subsite of primary tumor Hypopharynx Total pT classification 1
22 6 28 (100%) 28 28 (100%) 6
2
8
3 4
9 5
Total pN classification 0
28 (100%) 6
1
4
2a 2b
2 13
2c
3
3 Total
0 28 (100%)
AJCC/UICC stage 1 2
4 2
3
3
4a Total
19 28 (100%)
AJCC/UICC 5 American Joint Committee on Cancer/International Union Against Cancer.
DISCUSSION Our data show that the proportion of SCCs attributable to HPV in oral, laryngeal, and hypopharyngeal sites is substantially lower than for oropharynx. The use of a strict site classification, a substantial cohort size, and a rigorous validated testing protocol ensures that these data can be seen as robust. Upile et al.: HNSCC Outside the Oropharynx and HPV
TABLE II. Summary of Testing Results. Test, n (%) Site
No.
p16 IHC1
p16 IHC1 and HR HPV ISH 1
DNA qPCR1(in p161 and ISH 2 cases)
HPV1
Nonoropharynx sites Oral
102
8 (8%)
4(4%)
0
9 (4.1%) 4 (4%)
91
6 (6.6%)
3 (3.2%)
0
3 (3.2%)
Hypopharynx Oropharynx
Larynx
28 60
3 (10.7%) 46 (76.7%)
2 (7.1%) 36 (60%)
0 6
2 (7.1%) 42 (70%)
Total
281
63 (22.4%)
45 (16%)
6/18 (33.3%)
51 (18.1%)
HPV 5 human papillomavirus; HR 5 high risk; IHC 5 immunohistochemical; ISH 5 in situ hybridization; qPCR 5 quantitative polymerase chain reaction.
Through the comparison of both oropharyngeal and nonoropharyngeal tumors within the same cohort, the 70% positive rate in the oropharynx can be seen as a biological positive control to further emphasize the small fraction of HPV positivity (4%) within other sites. Without doubt, many biomarker and molecular studies suffer from elements of bias in that research is only performed on the subset of the tissue with consent and when available in the tissue bank. In our research group, it has been more common for patients receiving surgery as a primary modality to be approached for consent to tissue banking than for those having nonsurgical approaches or palliative treatment. In oropharyngeal SCC, this is perhaps unlikely to unduly bias the results, as surgery is the mainstay of treatment in our center.18
TABLE III. Tumor and Patient Demographics for Human Papillomavirus-Positive Nonoropharyngeal Tumors. Subsite
Oral cavity 1 2
59-year-old female, pT3N3b SCC (stage 4a), floor of mouth 48-year-old female, pT2N2c SCC (stage 4a), floor of mouth
3
71-year-old male, pT2N0 SCC (stage 2), floor of mouth
4
72-year-old male, pT4N0 SCC (stage 4a), floor of mouth
Larynx 5
52-year-old male, pT2N2a SCC (stage 4a), supraglottic
6
49-year-old male, pT2N0 SCC (stage 2), supraglottis and glottic
7
70-year-old male, pT2N0 SCC (stage 2), glottic and subglottis
Hypopharynx 8 9
74-year-old male, pT4N2a SCC (stage 4a), hypopharynx 60-year-old female, pT3N1 SCC (stage 3), hypopharynx
Laryngoscope 124: December 2014
In the larynx and hypopharynx, there may be a more significant bias. However, we cannot conclude that this would substantially affect the HPV rate; from a pragmatic viewpoint the characteristics of surgically treated larynx cancer appear not to skew the chance of being HPV linked. In selecting this cohort, specific care has been taken to select cases where site has been determined prior to tissue banking as a prospective collection, and we have not relied on retrospectively analyzed pathology reports for these data. The oropharynx cases selected for this study are an entirely different cohort from our previously published work14 with no overlap. Given the trend toward increasing HPV- positive rates in oropharyngeal squamous cell carcinoma (OPSCC) within Western countries, the HPV-positive rate within the current series (70%) has risen by comparison to rates detailed from the same geographic population from that previous study (57%, 1988–2009). This confirms that the testing carried out on predominantly fixed tissue in this series is comparable to our previous diagnostic data that were established on snap-frozen tissue.14 It is readily acknowledged that the primary site of tumors arising adjacent to the oropharynx, in particular the tonsil, can be difficult to classify by site and therefore introduce potential for misclassification. It is, however, a strength of the tissue procurement system utilized that all attempts have been made at the time of diagnosis and resection to differentiate one from the other. Reassuringly, nonoropharynx HPV-positive cases identified are not simply in sites adjacent to the oropharynx, such as the retromolar triangle, but are instead originating in anatomically discrete sites such as the floor of the mouth. The HPV testing diagnostic algorithm is robust, yet the inclusion of DNA qPCR for HPV16 alone may underestimate positive cases in the ultimate tier of analysis. In previous work, we have found, albeit only in rare instances, that sub-types other than HPV16 are detectable. We would, however, defend this protocol on the basis that frequency with which subtypes other than HPV16 occurs in isolation is
