Human Papillomavirus Testing in Head and Neck Carcinomas: Guideline From the College of American Pathologists Open Access

This evidence-based guideline was developed following the standards endorsed by the National Academy of Medicine, formerly the Institute of Medicine.²⁷  A detailed description of the methods and a systematic review (including the quality assessment and complete analysis of the evidence) used to create this guideline can be found in the supplemental digital content at www.archivesofpathology.org in the May 2018 table of contents.

The CAP convened an EP consisting of members with expertise in head and neck and molecular pathology and surgical, medical, and radiation oncology to develop the guideline. In addition, a research methodologist consultant served on the EP for the systematic review of the evidence. An advisory panel consisting of 2 patient advocates, 4 pathologists, 1 medical oncologist/molecular epidemiologist, 1 radiation oncologist, and 1 methodologist assisted the EP. The following organizations provided official panel representation: the American Academy of Otolaryngology—Head and Neck Surgery Foundation, the American Society of Clinical Oncology (ASCO), and the American Society of Cytopathology.

In addition, the guideline was submitted to ASCO's Head and Neck Guideline Advisory Group and ASCO's Clinical Practice Guideline Committee for review of the final manuscript. No suggestions for revisions were proposed, and it was agreed that the guideline should be considered for endorsement by ASCO.

In accordance with the CAP conflict of interest policy (in effect April 2010), members of the EP disclosed all financial interests from 12 months prior to appointment throughout the development of this guideline. Individuals were instructed to disclose any relationship that could be interpreted as constituting an actual, potential, or apparent conflict. Complete disclosures of the EP members are listed in the Appendix. Disclosures of interest judged by the oversight group to be conflicts are as follows: R.R.S, research grants, National Institutes of Health (Bethesda, Maryland); W.H.W., consultancy, Merck & Co, (Kenilworth, New Jersey). The majority of EP members (9 of 11) were assessed as having no relevant conflicts of interest. The CAP provided funding for the administration of the project; no industry funds were used in the development of the guideline. All panel members volunteered their time and were not compensated for their involvement, except for the contracted methodologist. Please see the supplemental digital content for full details on the conflict of interest policy.

The scope of the panel was to develop evidence-based recommendations for the various methodologies and applications of HR-HPV testing in head and neck carcinomas. The key questions are listed as follows:

1. Should patients with newly diagnosed OPSCC, nonoropharyngeal SCC (non-OPSCC), oropharyngeal non-SCC, nonoropharyngeal non-SCC, and cervical nodal metastatic carcinomas of unknown and/or known primary be routinely tested for HR-HPV?

2. Do relevant clinical outcomes of specific tests or testing algorithms for HR-HPV differ based on items such as specimen size, type and length of tissue fixation, or the criteria/definition for a positive p16 immunohistochemistry (IHC) or in situ hybridization (ISH) test?

3. For patients with OPSCC, non-OPSCC, and cervical nodal metastatic SCC, what is the optimal method of reporting HPV test results to best inform patients and clinicians about the clinical significance of the results (including considerations about uncertainty)?

4. Should patients with recurrent/persistent OPSCC, non-OPSCC, and cervical nodal metastatic SCC be routinely tested for HR-HPV?

5. Should patients with locally and/or regionally recurrent OPSCC, non-OPSCC, and cervical nodal metastatic SCC be routinely tested for HR-HPV?

6. Should patients with distant disease be tested for HR-HPV?

Refer to the supplemental digital content for all of the subquestions under these main key questions.

A comprehensive search for literature was initially performed in MEDLINE using the OvidSP interface on March 3, 2014, encompassing the publication dates of January 1, 1995, to March 3, 2014. A supplemental literature search was completed in PubMed (March 26, 2014) encompassing the publication dates of January 1, 1995, to March 26, 2014. An additional literature search was performed using Scopus (March 29, 2014) to identify relevant articles published between January 1, 1995, and March 29, 2014, in journals not indexed in MEDLINE. The literature search of the electronic databases was conducted in 2 arms: the first combined medical subject headings and keywords to address the concepts head and neck neoplasms, human papillomavirus (HPV), and laboratory testing, and the second combined medical subject headings and keywords for the concepts head and neck neoplasms, human papillomavirus (HPV), and outcomes. The results of both arms of the search were combined and deduplicated. Limits were set for human studies published in English, and a publication filter was applied to exclude lower levels of evidence such as letters, commentaries, editorials, and case reports.

A search for gray (unindexed) literature included a review of guideline and systematic review repository sites (eg, Guidelines International Network, National Guideline Clearinghouse, Cochrane Library, Prospero, Centre for Reviews and Dissemination), and relevant medical organizations' Web sites to identify guidelines, protocols, and standards. A review of meeting abstracts published in the years 2012–2014 from pathology and oncology organizations and EP recommendations completed the systematic literature review. The Ovid search was rerun on July 11, 2016, to identify articles published since March 1, 2014, that provided information that would alter the recommendations in any way.

Detailed information regarding the literature search strategy can be found in the supplemental digital content.

Practice guidelines, consensus documents, systematic reviews, meta-analyses, randomized controlled trials (RCTs), comparative studies, reviews, case-controlled studies, case series, and evaluation studies were eligible for inclusion.

Published studies were selected for full-text review if they met each of the following criteria:

1. Patients with tissue or cytology aspiration material taken from the workup of

• Oropharyngeal primaries

• Cervical nodal metastasis of unknown primary

• Regional or distant metastasis from known or suspected oropharyngeal primary

• Other head and neck sites (eg, sinonasal)

• All carcinomas in the head and neck

1. Human studies

2. Patients of all ages and either sex

3. Studies published in English

4. The study compared, prospectively or retrospectively, laboratory testing methodologies or potential testing algorithms for HPV testing

5. The study addressed one of the key questions

6. The study included measurable data such as the negative predictive value or positive predictive value, if testing methodologies used to determine HPV status, alone and in combination; negative and positive concordance across the platforms; sensitivity and specificity of individual tests; and accuracy in determining HPV status.

Articles were excluded from the systematic review if they were noncomparative or qualitative studies, including editorials, commentaries, or letters; animal studies; full-text articles not available in English; studies that included patients with other tumor types not specified in the inclusion criteria; studies that did not include relevant measurable data; and studies that did not address at least one of the key questions.

Detailed information about the inclusion and exclusion criteria is available in the supplemental digital content.

An assessment of study quality was performed by a research methodologist for all fully published studies meeting inclusion criteria. Studies only available in abstract form did not undergo formal quality assessment. Formal quality assessment involved determining the risk of bias by assessing key indicators, based on study design and methodologic rigor. Refer to the supplemental digital content for the definitions of ratings for strength of evidence (Supplemental Table 1) and for the quality assessment results.

Development of recommendations required that the panel review the identified evidence and make a series of key judgments. Grades for strength of recommendations were developed by the CAP Pathology and Laboratory Quality Center and are described in Table 1.

This guideline will be reviewed every 4 years, or earlier in the event of publication of substantive and high-quality evidence that could potentially alter the original guideline recommendations. If necessary, the entire panel will reconvene to discuss potential changes. When appropriate, the panel will recommend revision of the guideline to the CAP for review and approval.

The CAP developed the Pathology and Laboratory Quality Center as a forum to create and maintain evidence-based practice guidelines and consensus statements. Practice guidelines and consensus statements reflect the best available evidence and expert consensus supported in practice. They are intended to assist physicians and patients in clinical decision making and to identify questions and settings for further research. With the rapid flow of scientific information, new evidence may emerge between the time a practice guideline or consensus statement is developed and when it is published or read. Guidelines and statements are not continually updated and may not reflect the most recent evidence. Guidelines and statements address only the topics specifically identified therein and are not applicable to other interventions, diseases, or stages of diseases. Furthermore, guidelines and statements cannot account for individual variation among patients and cannot be considered inclusive of all proper methods of care or exclusive of other treatments. It is the responsibility of the treating physician or other health care provider, relying on independent experience and knowledge, to determine the best course of treatment for the patient. Accordingly, adherence to any practice guideline or consensus statement is voluntary, with the ultimate determination regarding its application to be made by the physician in light of each patient's individual circumstances and preferences. The CAP makes no warranty, express or implied, regarding guidelines and statements and specifically excludes any warranties of merchantability and fitness for a particular use or purpose. The CAP assumes no responsibility for any injury or damage to persons or property arising out of or related to any use of this statement or for any errors or omissions.

The views expressed in this document do not reflect the official policy of the Department of the Army, the Department of the Navy, the Department of the Air Force, the Department of Defense, or the US government. The identification of specific products or scientific instruments does not constitute endorsement or implied endorsement on the part of the Department of Defense or any component agency.

Of the 2803 unique studies identified in the systematic review, 503 met inclusion criteria and underwent data extraction. One hundred fifty-seven of these studies made up the evidentiary base and informed the guideline statements (recommendations). The vast majority were published, peer-reviewed articles, but 31 studies were published only in abstract form. All 157 underwent data extraction and qualitative analysis. Abstracts included in the 157 studies reported at least partial data. Those that did not report data for any of the outcomes of interest were excluded. Data from abstracts were used only in concert with peer-reviewed data, as they added support for recommendation statements. Abstract data alone were not used in the formulation of recommendations.

The EP met 16 times through Web-based meeting forums from November 22, 2013, through September 21, 2016. Additional work was completed via electronic mail. The EP met in person February 8 and 9, 2014, to formally initiate the project, and again April 9, 2016, to review the evidence to date and draft recommendations.

A public comment period was held from July 18 to August 8, 2016, on the CAP Web site. Fourteen draft recommendations, 2 demographic questions, and 3 questions about feasibility were posted for feedback.

“Agree” and “disagree” responses were captured for every proposed recommendation. In addition, 269 written comments were captured. Seven draft recommendations achieved more than 90% agreement, 6 achieved between 80% and 90% agreement, and 1 received 57% agreement. Each EP member was assigned 1 or 2 draft recommendations for which to review the public comments and to present to the panel for group discussion. After consideration of the comments, 6 draft recommendations were maintained with the original language and 7 were revised. Resolution of all changes was obtained by unanimous consensus of the panel members using nominal group technique (rounds of subsequent teleconference webinars and email discussion). Final EP recommendations were approved by a formal vote. The panel considered laboratory efficiency and feasibility throughout the entire process, although neither cost nor cost-effectiveness analyses were performed. A description of the benefits and harms of implementing the guideline statements is included in the supplemental digital content.

An independent review panel, masked to the EP and vetted through the conflict of interest process, provided final approval on behalf of the CAP Council on Scientific Affairs. In addition, the guideline was submitted to ASCO's Head and Neck Guideline Advisory Group and ASCO's Clinical Practice Guideline Committee for review of the final manuscript. No suggestions for revisions were proposed and it was agreed that the guideline should be considered for endorsement by ASCO. The final recommendations are summarized in Table 2, and an algorithmic approach for the workup of patient specimens is provided in Figure 1.

Pathologists should perform HR-HPV testing on all patients with newly diagnosed OPSCC, including all histologic subtypes. This testing may be performed on the primary tumor or on a regional lymph node metastasis when the clinical findings are consistent with an oropharyngeal primary.

The strength of evidence is convincing to support this guideline statement.

The evidentiary base supporting this recommendation comprised 110 studies, of which 1 was a meta-analysis,²⁸  8 were subgroup analyses from RCTs,^11,29–35  77 were observational studies,^{1,10,36–111}  and 24 were studies only published in abstract form.^112–135  The meta-analysis received a quality score of 7 out of a possible 11 points, and all supporting RCT data had a risk of bias determined to be moderate. The supporting observational studies ranged from low to moderate risk of bias, with the exception of one study that was assessed to have a high risk of bias.¹  This study was a retrospective cohort with retrospective data collection and industry sponsorship. None of the other studies were found to have methodologic flaws that would raise concerns about the findings. Refer to Supplemental Table 2 in the supplemental digital content for the quality assessment results for studies included in the statement 1 evidentiary base.

Breaking from a tradition that has broadly grouped all carcinomas arising from the oral and oropharyngeal subsites as oral cancer, these guidelines maintain a sharp distinction between those carcinomas arising in the oropharynx and those arising in the oral cavity proper. Testing for the presence of HPV must be guided by a familiarity with head and neck anatomy, including those structures that define the oral cavity as separate from the oropharynx (Figure 2). The oral cavity proper comprises the lips, gingiva, retromolar trigone, hard palate, buccal mucosa, mobile tongue, and floor of the mouth, whereas the oropharynx comprises the palatine tonsils, soft palate, base of tongue (posterior to the circumvallate papillae), and lateral and posterior pharyngeal walls. Oropharyngeal tonsillar structures (ie, lingual and palatine tonsils), particular hot spots for HPV-related carcinogenesis, are present in the oropharynx, but not in the oral cavity.

Oropharyngeal SCCs with transcriptionally active HR-HPV represent a unique type of HNSCC. These HPV-positive OPSCCs have risk-factor, demographic, morphologic, molecular, and clinical profiles that stand apart from other HNSCCs.

Human papillomavirus status of a primary or metastatic OPSCC may have diagnostic, staging, and even therapeutic implications. Currently, however, the call for routine HPV testing reflects its standing as a powerful prognostic indicator for patients with OPSCC. The literature overwhelmingly supports the conclusion that HPV status is an important and independent predictor of overall and disease-specific survival for patients with OPSCC. The survival benefit of HPV-positive OPSCC is maintained across nearly all studies, despite significant heterogeneity in patient populations, sample size, methods of HPV detection, tumor stage, tumor treatment, comorbidity, and inclusion of various other prognostic factors in the analysis. In large prospective studies where patient populations with OPSCC are uniformly staged and treated, significant reduction in risk of progression and disease-related death is confirmed for HPV-positive tumors.^{11,12,31,32,35} 

Testing for HR-HPV should be performed on all OPSCCs regardless of histologic type. The morphologic spectrum of HPV-positive OPSCC includes variants with papillary, adenosquamous, lymphoepithelioma-like, sarcomatoid, and basaloid features. Morphologic variation does not seem to influence clinical behavior, and it does not abrogate the need for HPV testing.^136–138 

In the open comment period, of the 168 respondents, 93.45% (n = 157) agreed with the recommendation and 5.36% (n = 9) disagreed. There were 22 written comments. Most of these comments were directed at the method of HPV testing and were, accordingly, taken into consideration in the final drafting of statement 2. Others reflected confusion about the anatomic definition of the oropharynx and its distinction from the oral cavity. To highlight this important distinction, a brief description of the anatomy of the oropharynx and oral cavity has been provided.

For oropharyngeal tissue specimens (ie, noncytology), pathologists should perform HR-HPV testing by surrogate marker p16 IHC. Additional HPV-specific testing may be done at the discretion of the pathologist and/or treating clinician, or in the context of a clinical trial.

The strength of evidence is adequate.

Sixty-seven studies (5 RCTs,^{11,29,32,34,35}  48 observational studies,^* and 14 studies published in abstract form^†) make up the evidentiary base for statement 2. Of these studies, 31 reported on laboratory outcomes of interest (Table 3) and 51 on clinical outcomes (Table 4). The risk of bias assessment of the majority of the included studies ranged from low to moderate, with only 2 studies assessed as high^1,68  because of industry funding. Although the vast majority of included studies were retrospective cohorts, and the inherent limitations of retrospective designs should be taken into consideration, the collection of data did occur prospectively in all but 2 studies.^1,38  None of the other studies were found to have methodologic flaws that would call into question the study findings. Refer to Supplemental Table 3 in the supplemental digital content for the quality assessment results for all studies included in the statement 2 evidentiary base.

Because the literature strongly supports that HR-HPV status is independently prognostic in OPSCC and that it should be routinely assessed, there is a need for consistency in clinical practice. Which test or combination of HR-HPV tests to perform is one of the more controversial issues in head and neck pathology. There are numerous HPV-specific tests, as well as the surrogate markers p16 IHC and hematoxylin-eosin morphology to consider. Although it is ideal to have a reference or standard criterion test, the current literature does not clearly support one. The test should be the one that best stratifies patient survival outcomes while also being practical and inexpensive. p16 IHC is the test the EP considers to best fit that role. p16 is markedly overexpressed in tumor cells with transcriptionally active HR-HPV because the viral E7 oncoprotein destabilizes pRb, functionally removing suppression of p16 expression and allowing tumor cells with high p16 levels to bypass pRb-dependent cell cycle arrest. The result is marked overexpression of p16, making it an excellent surrogate marker of viral infection in the correct context.^139,140  Based on abundant literature on p16 IHC as an independent predictor of improved patient prognosis in OPSCC,^‡ and on its widespread availability, ease and reproducibility of interpretation,⁵³  and excellent performance on small specimen samples such as small biopsies and tissue microarray punches,^11,89,142  the EP recommends that p16 testing be performed. Many consider the detection of HR-HPV E6 and E7 messenger RNA (mRNA) by ISH as the gold standard.^89,143,144  Although this is an excellent test, and perhaps even the ideal test from a purely scientific perspective, it isn't widely available for clinical use, is much more expensive than p16, and is more technically challenging to perform, and the data do not show statistically better performance than p16 IHC alone in OPSCC. Because p16 is only a surrogate marker for HR-HPV, and its overexpression is not always associated with the presence of HR-HPV,^53,75,81,89  as practice changes and HPV-specific tests such as RNA ISH become more widely available clinically, the latter may become the recommended test in the future.

For studies analyzing p16 IHC alone as a prognostic marker, the majority found it to be a marker of favorable outcome in multivariate analysis with univariate hazard ratios for death between 0.2 and 0.5^§ for overall, disease-free, and/or disease-specific survival compared with p16-negative patients. Several prospective and randomized controlled studies showed p16 IHC to be strongly prognostic alone as well.^11,12,32,145  In many of the studies, data extraction and summarization were complicated by p16 results, rather than being analyzed in isolation, being combined with results of HPV-specific test(s) for analysis and data reporting. Correlation rates between p16 IHC and HPV-specific tests were generally high, and best for HPV mRNA tests such as reverse transcriptase polymerase chain reaction and ISH.^‖

In addition to the aforementioned studies, some additional studies not included in our systematic review support this recommendation. Sedghizadeh et al¹⁴¹  performed a systematic review and meta-analysis of studies examining p16 IHC expression and prognosis in OPSCC. Among the 18 studies finally included, they found that p16 status by IHC was prognostic for all survival metrics with hazard ratios between 0.3 and 0.4. Some newer studies compared HR-HPV RNA ISH with p16 IHC and showed high correlation between these tests (for high-incidence US study populations) in OPSCC patients.¹⁴⁶  In 3 large RNA ISH-based studies, correlation rates were 92%, 96%, and 100.0%, respectively.^89,143,144  Rooper et al¹⁴⁷  showed that almost all patients in their practice who were p16 positive but HPV deoxyribonucleic acid (DNA) ISH negative were actually positive for HR-HPV by mRNA using ISH, demonstrating that DNA ISH lacks sensitivity and that the correlation between p16 overexpression and presence of HPV mRNA is indeed high.

In the open comment period, of the 160 respondents, 89.44% (143) agreed with the recommendation, and 6.88% (11) disagreed. There were 32 written comments, the majority of which suggested that HPV-specific testing needs to be performed because p16 IHC is not truly a surrogate marker of HR-HPV. Additional comments pointed out that there are patients with p16-positive and HPV-negative tumors and vice versa. Although the latter 2 points are true, these reflect the concept that p16 IHC is being done specifically to ascertain if a patient's tumor is related to transcriptionally active HPV, rather than regarding it as a prognostic biomarker. All of the many HPV-specific tests that are available for use on tissue sections, such as DNA polymerase chain reaction, DNA ISH, RNA reverse transcriptase polymerase chain reaction, and RNA ISH, are independently prognostic.^{11,12,29,70,141}  There is no shortage of data. However, none of these tests are statistically significantly better than p16 IHC alone.^{53,81,89,92,148}  Some studies, particularly in lower–HPV-incidence regions (selected regions of Europe, for example), have demonstrated that p16-positive, HPV mRNA– and HPV DNA–negative patients have a poorer prognosis than those patients positive for both tests (although interestingly perhaps still better than for p16 and HPV mRNA/DNA double-negative patients).^75,77  Thus, p16 IHC alone may not be the right approach in these other geographic regions. To allow for discretion, the EP has provided the caveat that “HPV-specific testing may be done at the discretion of the pathologist and/or treating clinician, or in the context of a clinical trial.”

Doing both p16 IHC and HPV-specific testing will undoubtedly result in some patients with discrepant test results. There is not currently strong evidence for what to do in these situations, although it is probably good practice at this time not to label patients with discrepant p16- and HPV-specific test results as being in the prognostically favorable category of HPV positive. Particular caution should be exercised for laboratories currently using or considering DNA ISH, because it has been shown to lack sensitivity for HR-HPV and may lead to the above situation (of test discrepancy) more frequently than HPV-specific tests with higher sensitivity.^53,81,89,149 

Comments also dealt with method validation and proficiency testing for p16 IHC. Although the literature has little data comparing p16 IHC antibody clones and methods, most of the studies, despite differing methodologies and clones, show strong, independent, and statistically significant performance of p16 IHC alone as a prognostic marker.^11,32,81,141  Interestingly, most of the polled respondents from a CAP practice patterns survey and other surveys use the E6H4 antibody, and most of the studies in the literature also have used it, consistently showing it to have very good performance.^{11,53,81,84,89}  There are not sufficient data to recommend one antibody, platform, or set of test conditions over another. However, as p16 IHC becomes part of routine clinical practice in OPSCC, test validation and method comparison studies will be critical, and proficiency testing and benchmarking will likely become available. Laboratories should choose tests and model their technical performance after those large studies that validated p16 IHC testing, and must validate their p16 IHC performance in accordance with the IHC validation guideline previously published by the CAP.¹⁵⁰ 

Refer to Tables 3 and 4 for the summary of data for laboratory and clinical outcomes for OPSCC tested with p16 IHC.

Pathologists should not routinely perform HR-HPV testing on patients with nonsquamous carcinomas of the oropharynx.

The strength of evidence is insufficient.

The vast majority of primary oropharyngeal carcinomas are SCCs derived from the epithelium lining the surface of the oropharynx and the tonsillar crypts, but a subset are carcinomas of minor salivary gland origin. Even less commonly, high-grade neuroendocrine (large and small cell) carcinomas arise within the oropharynx, sometimes in association with an HPV-positive OPSCC.

Although HR-HPV may play an etiologic role in some oropharyngeal high-grade neuroendocrine carcinomas, the tumors tend to be clinically aggressive, regardless of HPV status.^151–153  In effect, HPV status does not appear to be a reliable marker for separating aggressive and nonaggressive tumors when it comes to high-grade neuroendocrine carcinomas of the oropharynx. For carcinomas of minor salivary gland origin, there is currently insufficient evidence to support an etiologic role of HPV in these tumors, or to validate the practice of HPV-testing them for prognostic purposes.^154–156  Almost all tested tumors have lacked transcriptionally active HR-HPV.^154,157,158  Accordingly, routine HPV testing is not recommended for nonsquamous carcinomas of the oropharynx, including minor salivary gland carcinomas and high-grade neuroendocrine carcinomas. On the other hand, the presence of glandular differentiation by itself should not be taken as an exclusion criterion for HPV testing. Oropharyngeal SCCs can sometimes exhibit glandular formations as a minor or predominant tumor component, and these adenosquamous carcinomas should undergo routine HPV testing as with other variant forms of OPSCC (see statement 1). Case reports of pure adenocarcinomas with transcriptionally active HR-HPV have been described in the oropharynx (similar to the uterine cervix), but they are so few that no recommendation for testing can be provided.^159,160 

Importantly, conventional squamous differentiation, such as surface dysplasia and keratinization, is not highly developed in most HPV-positive OPSCCs. Instead, HPV-positive OPSCCs are typically nonkeratinizing and show varying degrees of basaloid differentiation.^43,161  Human papillomavirus testing should not be suspended in an OPSCC because it lacks keratinization or exhibits basaloid features, as long as it is proven, with IHC if necessary, to be SCC and not a neuroendocrine or other nonsquamous poorly differentiated carcinoma.

In the open comment period, of the 158 respondents, 88.61% (n = 140) agreed with the recommendation, and 8.86% (n = 14) disagreed. There were 14 written comments. Most of these comments acknowledged the rarity of nonsquamous carcinomas in the oropharynx and encouraged continued HPV testing of these tumors in the research (not clinical) setting. Other comments expressed concerns that statement 3 would inappropriately promote nontesting of some OPSCCs because squamous differentiation is often not readily apparent in those that are HPV positive.

Pathologists should not routinely perform HR-HPV testing on patients with nonoropharyngeal primary tumors of the head and neck.

The strength of evidence is adequate.

This recommendation is supported by 1 subgroup analysis of 3 RCTs and 28 observational studies that met the inclusion criteria for our systematic review.^{66,158,162–187}  Two of these studies were available only in abstract form and, as such, did not undergo formal quality assessment.¹⁸⁶  The evaluable 27 studies were deemed to have varying risk of bias: 5 were deemed low risk of bias,^{162,164,169,176,183}  8 low to moderate,^¶ 10 moderate,^# and 4 high.^{66,158,175,179}  The high risk of bias for the 4 studies was due to funding from industry and missing information on the other assessment criteria. Refer to Supplemental Table 4 for the quality assessment results for statement 4 studies.

There is confusion among many pathologists and treating physicians regarding the appropriateness of HPV testing in nonoropharyngeal head and neck carcinomas. Because of the considerable attention that HPV-positive OPSCC has received, it is understandable that pathologists and treating physicians may want to generalize that experience to carcinomas arising outside the oropharynx, but the EP did not find evidence to support this practice. Routine HPV testing for nonoropharyngeal head and neck carcinomas is not indicated because there is no proven prognostic or therapeutic difference based on its presence or absence (either by any of the various HPV-specific tests or by the surrogate marker p16). If HPV testing were to yield a positive result in a nonoropharyngeal carcinoma, it might mislead treating physicians and patients as to the origin and likely biologic behavior of the carcinoma. This does not mean that there is no potential biological and clinical significance for transcriptionally active HR-HPV in nonoropharyngeal head and neck carcinomas (particularly in specific anatomic subsites such as the nasopharynx or sinonasal tract); it simply means that the clinical significance and ramifications are not established at this time.

Although the EP does not recommend routinely testing nonoropharyngeal carcinomas for HPV, it does recognize that there are occasional situations where it may be indicated. For example, if the anatomic site of tumor origin is not provided, is ambiguous, and/or includes both an oropharyngeal and a nonoropharyngeal site (eg, for large tumors), then HPV testing may be appropriate. As another example, for a patient who had a prior HPV-positive OPSCC, HPV testing in a new non-OPSCC may be appropriate to understand the relationship between the 2 carcinomas (ie, recurrence versus new primary). In these settings, when HPV testing is performed, p16 IHC alone is insufficient because of its suboptimal positive predictive value in nonoropharyngeal sites. p16 IHC can be used to screen a tumor using the same criteria as in the oropharynx. If it is negative, then one can conclude that the tumor is not related to transcriptionally active HR-HPV. If it is positive, however, HPV-specific testing must be performed by one of the available platforms (see algorithm).

The systematic review uncovered 16 studies that investigated HPV testing in nonoropharyngeal head and neck carcinomas (Table 5). These studies were heterogeneous in anatomic subsites evaluated (eg, larynx, oral cavity, sinonasal tract, and other) and in the HPV testing methods used. The studies found that the prevalence of HPV-positive carcinomas, when considering all tests for HR-HPV, is generally low in these nonoropharyngeal sites, ranging from 5.9% to 58.3%.^** When more specific, RNA-based methods for HPV detection were used, or p16 was combined with HPV-specific testing in order to establish the presence of transcriptionally active HR-HPV, the rates were 2.7% to 5.9%.^175,189  Importantly, when the interpretation criteria were reported and appropriate, the positive predictive value of p16 IHC in nonoropharyngeal carcinomas was low, ranging from 22% to 50%, because of the very low overall rates of transcriptionally active HR-HPV in these tumors.^{166,173,175,183,189} 

The systematic review identified 28 studies^†† that investigated the clinical outcomes of nonoropharyngeal HPV-positive carcinomas (Table 6). Once again, these studies were highly variable in anatomic subsites examined and HPV testing methods used and were also heterogeneous in their reported outcomes. Only 7 studies reported a statistically significant difference in overall, disease-free, and/or progression-free survival between HPV-positive and HPV-negative groups, including 5 that found that the patients with HPV-positive carcinomas had better survival^{162,172,184,185}  and 2 that actually found that the HPV-positive group had worse survival.^168,187,190  One additional study found that the HPV-positive group had significantly lower rates of recurrence.¹⁶⁹ 

In addition to the aforementioned studies, systematic reviews, and meta-analyses, Li et al¹⁹²  and Syrjanen and Syrjanen¹⁹⁶  pooled data from numerous heterogeneous studies and found that the overall rates of HPV DNA in laryngeal and sinonasal carcinomas were 28% and 27%, respectively. Other studies using RNA-based HPV detection methods on oral cavity and laryngeal carcinomas reported HPV prevalences of 1.3% and 2.3%, respectively.^197,198 

In the open comment period, there were 154 respondents, of whom 79.7% (n = 123) agreed with the recommendation and 12.3% (n = 19) disagreed. There were 28 written comments, including some that expressed disagreement because they believed this statement included anogenital sites. The recommendation was revised to specify that it was applicable only to the head and neck. Some respondents disagreed with the recommendation because of their experience or reports of occasional HPV-positive tumors arising in other nonoropharyngeal head and neck sites (nasopharynx, oral cavity, hypopharynx, sinonasal tract, and conjunctiva/lacrimal sac were all mentioned). Finally, a few respondents felt the language should be changed to allow testing in cases for which the precise anatomic location was not given.

Pathologists should routinely perform HR-HPV testing on patients with metastatic SCC of unknown primary in a cervical upper– or mid–jugular chain lymph node. An explanatory note on the significance of a positive HPV result is recommended.

The strength of evidence is adequate.

This statement is supported by 4 observational studies that met the inclusion criteria for the systematic review.^199–202  Risk of bias assessments was deemed low in 2 studies^200,202  and low to moderate in 2.^199,201  None of these studies were found to have methodologic flaws that would raise concerns about the findings. Refer to Supplemental Table 5 in the supplemental digital content for the quality assessment results of included studies for statement 5.

Unknown primary is defined as any metastasis for which the primary site has not been clinically identified at the time the biopsy of the metastasis is performed. In this setting, HR-HPV testing may aid in determining the most likely primary site. Hence, HR-HPV status is important for patient management as it informs the clinical team where to search for the primary, or limits the likely area of primary if a definitive lesion is not identified. There are inadequate data at this time to support HR-HPV as a prognostic marker in this setting.

In patients who have an unknown primary SCC in the neck, it is common clinical practice to search for the primary site by performing a thorough endoscopic examination and directed biopsies of likely sources of disease, such as the base of tongue, tonsils, and nasopharynx, as well as tonsillectomy if a primary is not identified on biopsy. The majority of primary tumors can be identified with this approach. Factors that favor an oropharyngeal primary site include cervical (upper or mid jugular chain, synonymous with levels II and III) metastasis location and HR-HPV positivity.^203,204  Thus, HR-HPV status is important because positivity strongly favors oropharyngeal origin and can limit treatment fields, even if a definitive primary is not identified. In addition, more than 70% (to more than 90% in some studies) of SCCs that initially present as an unknown primary can ultimately be confirmed as oropharyngeal in origin after a thorough search.^{203,205–208}  Repeat HR-HPV testing of the primary tumor, once biopsied or resected, as a prognostic marker is not necessary because the metastasis would have already been tested in this setting.

Although the systematic review revealed limited data, it suggests better outcomes for patients with metastatic HPV-positive compared with HPV-negative SCC in metastases of unknown primary. Refer to Tables 7 and 8 for a summary of laboratory data and clinical outcomes for metastatic SCCs of unknown primary. However, most studies had small patient numbers and/or lacked statistical significance, and thus additional evidence is needed before HR-HPV status can be considered a reliable prognostic marker in SCCs of unknown primary.^{199–202,209–211} 

It is important that HR-HPV testing is routinely performed only in metastases of unknown primary located in the appropriate (cervical, specifically upper and mid jugular chain) lymph node groups. The upper and mid jugular chain includes level II and III lymph node groups. The pretest probability of a positive HR-HPV result is high in this location because HPV-positive oropharyngeal SCCs almost always metastasize here, and the primary tumor is often clinically occult.^212,213  Up to one-third of HPV-positive OPSCCs present as an unknown primary (compared with 5%–10% of all head and neck cancers).^212,214  Thus, HR-HPV testing should be performed in all unknown primary cervical upper– or mid–jugular chain metastases.

In contrast, HR-HPV testing should not be routinely performed on metastases of unknown primary outside of the cervical upper– and mid–jugular chain region. Non–upper- and mid-jugular chain lymph node groups include level I (submandibular/submental), level IV (lower jugular), level V (posterior triangle), level VI (pretracheal/prelaryngeal/Delphian) and level VII (superior mediastinal), as well as parotid and supraclavicular nodes. The probability of an HPV-positive metastasis of unknown primary in the above locations is extremely low in the absence of concurrent upper/mid–jugular involvement, and therefore routine HR-HPV testing is not indicated. However, HR-HPV testing may be performed on a non–upper/mid-jugular chain metastasis when clinical suspicion for an HPV-positive SCC is high (“nonroutine” HR-HPV testing).

It is recognized that the specific lymph node group may not always be known to the pathologist examining the metastasis. Some may simply be labeled “right neck” or “left neck” by the ordering physician. Review of the medical record (clinical notes and/or radiology reports) or directly contacting the ordering physician will likely provide sufficient information to identify the involved lymph node group in most cases. There may be rare cases that cannot be localized even after attempts to obtain clinical information. In such cases, HR-HPV testing should be performed.

An explanatory note in the pathology report describing the clinical significance of a positive HR-HPV result is recommended, specifying that HPV-positive SCC metastases most likely originate from the oropharyngeal tonsils and/or base of tongue but rarely may be from other sites (ie, nasopharynx).

In the open comment period, there were 149 respondents; 85.23% (n = 127) agreed, and 11.41% (n = 17) disagreed. There were 22 written comments, including several comments questioning the method of HR-HPV testing, which is the subject of statement 6. Several comments were either in support of or against an explanatory note.

For tissue specimens (ie, noncytology) from patients presenting with metastatic SCC of unknown primary in a cervical upper– or mid–jugular chain lymph node, pathologists should perform p16 IHC.

Note: Additional HPV-specific testing on p16-positive cases should be performed for tumors located outside of level II or III (nonroutine testing) in the neck and/or for tumors with keratinizing morphology.

Strength of evidence is insufficient.

p16 IHC is a very sensitive surrogate marker for HR-HPV that also maintains high positive predictive value when the pretest probability of an HPV-positive SCC is high.^44,146,147  However, the positive predictive value decreases when the pretest probability is low because p16 overexpression can occur by mechanisms other than HR-HPV signaling. For example, as many as 20% to 30% of aggressive head and neck cutaneous SCCs overexpress p16 unrelated to HR-HPV.^215,216 

An algorithmic approach for HR-HPV testing, shown in Figure 1, efficiently triages SCC metastases of unknown primary and reduces unnecessary testing. Following the algorithm, p16 IHC alone is sufficient to determine HR-HPV tumor status when the metastasis is located in one of the upper– or mid–jugular chain (level II and III) lymph node groups and the tumor morphology is nonkeratinizing.^9,213,217  The pretest probability of an HR-HPV–positive SCC is very high in this setting (and therefore p16 IHC is an excellent test). High-risk HPV–specific testing is required to confirm a positive p16 IHC test result only when the tumor morphology is keratinizing and/or the metastasis is located outside of the upper or mid jugular chain (the latter would apply to the setting of nonroutine HPV testing). Confirmatory HR-HPV–specific testing should also be performed if the specific involved lymph node group cannot be determined, regardless of tumor morphology. If the p16 IHC is negative, no further HR-HPV testing is indicated, and the tumor is considered HPV negative because a p16-negative result essentially excludes a transcriptionally active HR-HPV–positive SCC.^44,146,147 

The ability to recognize keratinizing versus nonkeratinizing tumor morphology is important in order to apply the above HR-HPV testing algorithm. Nonkeratinizing SCC resembles transitional epithelium.^9,217  The tumor cells have oval to spindled nuclei, high nuclear to cytoplasmic ratios, and indistinct cell borders, often forming broad, pushing, ribbonlike nests. Limited keratinization may be present and does not mean that the tumor is keratinizing type (as long as the above described histologic features are present). Mitotic activity is typically brisk, with apoptotic debris and/or necrosis in the background. When in doubt as to whether a particular SCC is keratinizing or nonkeratinizing, as may be the case in small biopsy material, HR-HPV–specific testing is recommended. However, one should attempt to histologically classify most SCC metastases. In addition, one must consider that the morphology of the tumor may change in posttreatment specimens so that more keratinization may be present.

In the open comment period, there were 145 respondents; 84.14% (n = 122) agreed, and 11.72% (n = 17) disagreed. There were 21 written comments, with most weighing in on the method of HR-HPV testing: several comments favored p16 IHC testing alone in all settings, some were in support of HR-HPV–specific testing in all cases, and others felt the method should be at the discretion of the pathologist. Several pointed out that the location of the lymph node metastasis (needed to apply the HR-HPV testing algorithm) is not always known to the pathologist.

Pathologists should perform HR-HPV testing on head and neck FNA SCC samples from all patients with known oropharyngeal SCC not previously tested for HR-HPV, with suspected oropharyngeal SCC, or with metastatic SCC of unknown primary.

Note: No recommendation is made for or against any specific testing methodology for HR-HPV testing in FNA samples. If the result of HR-HPV testing on the FNA sample is negative, testing should be performed on tissue if it becomes available. If pathologists use cytology samples for p16 IHC testing, they should validate the criteria (ie, cutoff) for a positive result.

The strength of evidence is adequate.

This statement is supported by 16 studies^{26,201,210,218–230}  that met the inclusion criteria for our systematic review. Five of these studies^{210,224–227}  were reported only in abstract form and did not undergo quality assurance. Of the 11 studies assessed, all were retrospective cohort designs, with the exception of 1 cross-sectional study.²²²  All included studies had prospective data collection, 2 reported blinding of outcome assessors,^222,230  and only 1 reported funding from industry.²³⁰  Ultimately, 4 studies^{219,222,229,230}  were deemed to have a low to moderate risk of bias, 6 moderate,^‡‡ and 1 high,²²⁸  although this study did not have methodologic flaws that would raise concerns about its findings. See Supplemental Table 6 in the supplemental digital content for the quality assessment results for all studies included in the statement 7 evidentiary base.

Among patients with HPV-positive OPSCC, approximately 85% develop lymph node metastases, and approximately 50% of these patients are first diagnosed based upon an enlarged cervical lymph node.^{11,53,55,231–233}  Fine-needle aspiration is frequently used to sample and diagnose metastatic head and neck carcinomas in cervical lymph nodes. Because of the marked tendency for HPV-positive HNSCC to metastasize to cervical lymph nodes, FNA plays a very important diagnostic role in the initial detection of these cancers.^234–237  In some cases, cytologic material may be the only tumor specimen available for diagnostic workup, and in a subset of cases, a primary site of origin will not be identified even after exhaustive clinical and radiologic evaluation. The aspirated material obtained by FNA can be tested by any of a variety of methods for HR-HPV and used to classify the metastatic HNSCC as HPV positive or HPV negative, thus indicating an oropharynx origin.

The systematic review identified a limited number of studies that used samples obtained by FNA for analysis of HR-HPV status of metastatic HNSCC. The range of cytologic studies includes HPV determination using formalin-fixed, paraffin-embedded FNA material in a cell block; liquid-based specimens (Surepath, Becton, Dickinson and Company, Franklin Lakes, New Jersey, and ThinPrep, Hologic, Marlborough, Massachusetts); and scrapes from air-dried or alcohol-fixed smears. Testing methodologies evaluated included p16 IHC,^{222,229,230,238}  DNA ISH,²³⁹  cobas HPV test (Roche Molecular Systems, Inc, Pleasanton, California),^26,228  Cervista HPV HR and Cervista HPV16/18 assays (Hologic),^220,240,241  and Hybrid-Capture 2 assay (Qiagen, Germantown, Maryland).^219,224,242  The literature supports the use of FNA as a valid method for obtaining material for HR-HPV testing.^218,221,223  Sensitivities and specificities of HPV assays for detecting HR-HPV in FNA samples are reported to be greater than 90%; however, there are limited data about the accuracy of any one particular HPV testing method.^§§

A particular difficulty in the assessment of HR-HPV in FNA samples pertains to the interpretation of p16 immunoreactivity in cell block specimens. Although use of the standard criterion of more than 70% positive cells is accepted when applied to tissue biopsy, 3 recent studies suggest that thresholds as low as 10% to 15% for the percentage of positive cells may be valid for cell blocks.^23,229,230  Whichever method is selected to assess the HR-HPV status of an FNA sample, individual laboratory validation is required. Because of the limited data available pertaining to HR-HPV testing in FNA specimens at the current time, HR-HPV testing is recommended on any subsequent tissue specimens that may become available if HR-HPV testing was negative in the FNA specimen.

In the open comment period there were 142 respondents; 84.51% (n = 120) agreed, and 13.38% (n = 19) disagreed. There were 16 written comments, including a number that suggested that either p16 IHC or liquid-based testing methodologies be used. Others commented that confirmatory testing should be performed for FNA samples testing positive by p16 IHC, particularly in cases where the patient is known to have a nonoropharyngeal HNSCC.

Refer to Tables 9 and 10 for the summaries of laboratory data and clinical outcomes for studies where FNAs were used.

Pathologists should report p16 IHC positivity as a surrogate for HR-HPV in tissue specimens (ie, noncytology) when there is at least 70% nuclear and cytoplasmic expression with at least moderate to strong intensity.

The strength of evidence is insufficient.

p16 is a tumor suppressor protein that inhibits CDK4 and CDK6–dependent/cyclin D–mediated phosphorylation of RB required for cell proliferation.²⁴³  Overexpression of p16 is an established surrogate marker for transcriptionally active HR-HPV infection.^{139,140,243,244} 

The criteria for p16 IHC have been established in multiple retrospective and prospective studies that validated the association of p16 immunopositivity with a more favorable prognosis in OPSCCs compared with p16-negative carcinomas.^11,35,81,245  It is important to note that staining must be both nuclear and cytoplasmic to be considered positive. Definitions for what percentage of positive cells is necessary have varied substantially; however, some of the largest and prospective studies, such as Ang et al,¹¹  have supported a stringent cutoff of 70% to 75%. However, in high-incidence countries such as the United States, lesser staining cutoffs may function similarly. With these criteria, the sensitivity of p16 IHC for transcriptionally active HR-HPV approaches 100%. The specificity of p16 IHC in the oropharynx is lower (∼85%–95%) for transcriptionally active HR-HPV, in part because of p16 expression unrelated to HPV.^53,81,246  The interrater agreement among pathologists for p16 IHC interpretation is excellent (κ = 0.95–0.98).⁵³ 

Rare tissue specimens may exhibit an equivocal pattern of p16 staining that fails to meet the recommended threshold for positivity. For instance, cases can exhibit more than 50% and less than 70% moderate to strong nuclear and cytoplasmic staining or diffuse low-intensity nuclear and cytoplasmic staining. Limited evidence suggests a subset of these cases may have transcriptionally active HR-HPV.^53,247  In these situations, an HPV-specific test can be performed at the discretion of the pathologist.

The specificity of p16 IHC outside of the oropharynx is not as well characterized but is lower, in part because of entities that can mimic p16/HPV–positive metastatic OPSCC. For instance, p16 IHC is positive (using the above high expression cutoff) in as many as 20% to 30% of cutaneous head and neck SCCs, which are unrelated to HPV and have no association with clinical outcomes.²¹⁶  Approximately 40% of lymphoepithelial cysts are p16 immunopositive within the epithelial lining, although the staining is typically patchy and involves less than 50% of the epithelium.²⁴⁸ 

In the open comment period there were 140 respondents; 90.00% (n = 126) agreed and 4.29% (n = 6) disagreed with statement 8. There were 22 written comments, including comments that the data supporting the 70% threshold should be referenced. Several comments suggested there should be a description of equivocal patterns of p16 staining that do not meet the threshold for positivity. There was a comment that p16, as a surrogate marker for HPV, can be positive in non–HPV-related tumors. Lastly, there were comments that p16 IHC should be standardized to include the antibody clone used.

Pathologists should not routinely perform low-risk HPV testing on patients with head and neck carcinomas.

The strength of evidence is insufficient.

There is persistent confusion about whether low-risk HPV types (6 and 11) should be tested in HNSCCs, as indicated by published studies from reference laboratories that indicate that this testing is frequently requested.²⁴⁹  Low-risk types of HPV are biologically distinct from high-risk types, largely because of different binding and signaling properties of their respective E6 and E7 proteins.^250–253  Although low-risk HPV types are an established etiologic agent in benign squamous papillomas and warts of various sites, they do not play a significant role in the development of HPV-positive OPSCC.²⁵³ 

Because there is little (if any) benefit of identifying low-risk HPV types in the head and neck, the EP determined that there is no role for routine low-risk HPV in this context. Although the systematic review did not specifically address low-risk HPV types, the expert consensus opinion is that low-risk testing should not be routinely performed.

In the open comment period, there were 140 respondents; 95.71% (n = 134) agreed, and 4.29% (n = 6) disagreed. There were 7 written comments, 2 of which focused on a belief that low-risk HPV can lead to dysplasia. The published literature does not support this association. One comment suggested that low-risk HPV testing should be done in patients with HIV; however, there is no evidence that low-risk HPV results in carcinoma in HIV-positive patients.

Pathologists should not repeat HPV testing on patients with locally recurrent, regionally recurrent, or persistent tumor if primary tumor HR-HPV status has already been established. If initial HR-HPV status was never assessed or results are unknown, testing is recommended. Testing for HPV may be performed on a case-by-case basis for diagnostic purposes if there is uncertainty regarding whether the tumor in question is a recurrence or a new primary SCC.

The strength of evidence is insufficient.

High-risk HPV status established on the primary tumor is the basis for its prognostic value. Recurrences are readily testable and have been demonstrated to show the same HR-HPV status.²⁵⁴  As such, there is no documented value of repeating testing for HR-HPV on locoregionally recurrent or persistent HNSCC. However, because of this consistency of phenotype, HR-HPV testing on a recurrence when the status of the primary OPSCC is unknown would accurately reflect the HPV status and is thus recommended. Particularly with delayed recurrences, a logical clinical question that may arise is the possibility of a new primary tumor. Such scenarios require correlation with clinical and morphologic features, and HPV status may be informative in separating a recurrence from a new primary SCC.

Our systematic review yielded limited data addressing the level of concordance between the HPV status of the primary tumor and the corresponding recurrence. In one study,²⁵⁴  16 locoregional recurrences and 21 metastases were tested along with their untreated primaries. Thirty-six of 37 cases (97%) demonstrated a concordant HPV status. For the 1 discordant case, a second primary was not entirely excluded. However, the study also suggests that technical variability in testing may affect findings in a recurrent site.

In the open comment period there were 94 respondents; 92.47% (n = 86) agreed, and 7.53% (n = 7) disagreed. There were 8 comments, most notably the suggestion to indicate that HPV testing may help separate true recurrence from a separate primary tumor. Concern about technical differences among laboratories affecting repeat testing was raised, as well as a concern about treatment effect.

Pathologists should not routinely perform HR-HPV testing on patients with distant metastases if primary tumor HR-HPV status has been established. Testing for HPV may be performed on a case-by-case basis for diagnostic purposes if there is uncertainty regarding whether the tumor in question is a metastasis or a new primary SCC.

The strength of evidence is insufficient.

High-risk HPV status established on the primary tumor is the basis for its prognostic value. Distant metastases are readily testable and limited data show that they retain the same HR-HPV status, including p16 overexpression.^137,254,255  As such, there is no documented value of repeating testing on a metastatic tumor. However, because of this consistency of phenotype, HR-HPV testing on a metastasis when the status of the primary is unknown would accurately reflect the HPV status of the primary HNSCC and is thus recommended. In some cases, however, there is the possibility that the metastasis represents a new primary SCC. Such scenarios require correlation with clinical and morphologic features, and HPV status may be informative in separating a distant metastasis from a new primary SCC.

Our systematic review yielded limited data addressing the level of concordance between the HPV status of primary tumor and corresponding distant metastasis. Collectively, concordance is noted in about 44 of 45 (97.7%) tested paired primary tumors and metastases in the literature.^137,254,255  In one study,¹³⁷  20 of 20 tested metastases demonstrated concordant HPV status as compared with their primaries. In another study,²⁵⁴  16 locoregional recurrences and 21 distant metastases were tested along with their untreated primaries. Thirty-six of 37 cases (97%) demonstrated concordant HPV status. For the 1 discordant case, a second primary was not entirely excluded. One small series of 4 patients outside of our systematic review confirmed concordance in HPV status between 4 primary and metastatic SCCs.²⁵⁵ 

However, as with recurrences, technical variability in testing may affect findings in a metastatic site,²⁵⁴  and the findings are insufficient to recommend a specific testing algorithm at recurrent sites. Particularly with possible lung metastases, an approach that includes HPV-specific testing for p16-positive tumors should be considered, because of p16 expression in a subset of lung SCCs not associated with HR-HPV.

In the open comment period there were 90 respondents; 95.56% (n = 86) agreed, and 4.44% (n = 4) disagreed. There were 7 comments, most notably echoing the statement that HPV testing may help separate true recurrence from a separate primary tumor. Concern about p16 IHC testing as a stand-alone test in this context was raised, as well as concern about the effect of tumor heterogeneity.

Pathologists should report primary OPSCCs that test positive for HR-HPV or its surrogate marker p16 as HPV positive and/or p16 positive.

The strength of evidence is insufficient.

Oropharyngeal SCC with transcriptionally active HR-HPV is a distinct subtype of head and neck cancer. Because HPV defines this subtype, the HR-HPV status (by HPV-specific and/or surrogate marker p16 testing) should be included in the pathologic diagnosis. In the literature, a number of different terms have been used to describe the HPV status of these tumors, including HPV positive, HPV related, HPV driven, HPV mediated, and p16 positive.^{12,55,148,256}  The above expert consensus opinion is consistent with the terminology used in contemporary classifications of OPSCCs.

The term HPV positive refers to OPSCCs with detectable virus by HPV-specific methods in a tumor that is already established to be p16 positive, referring thus to transcriptionally active HPV. It can also be used to describe tumors that are just positive for p16 by IHC. Although p16 IHC is a surrogate marker for HR-HPV, a positive result in the appropriate clinical and pathologic context is sufficient to classify a tumor as HPV positive for risk stratification.⁵⁵  If the term p16 positive is used in clinical reporting on its own, a comment should be added that describes the strong relationship between p16 immunopositivity and HPV in the respective setting. For tumors that are positive for p16 by IHC with or without accompanying HPV-specific testing, both HPV positive and p16 positive can be used for reporting (ie, HPV-positive or p16-positive OPSCC).

The World Health Organization (WHO) also recommends the term SCC, HPV positive for patients who are either p16 positive (when it is the only test performed) or p16 plus HPV-specific test positive. The new Union for International Cancer Control²⁵⁷  and AJCC staging systems¹⁴  prefer a hybrid term, HPV-mediated (p16⁺) oropharyngeal cancer. The panel considers these synonymous.

The above terminology should be used at the time of diagnosis, if possible. Delays in the reporting of HPV status should be avoided, as the HPV status of OPSCCs defines the tumor subtype, predicts prognosis, and may affect therapeutic decisions. Given the widespread availability and rapid turnaround time of p16 IHC, the HPV status of OPSCC should be rapidly available in the majority of cases.²⁴³ 

In the open comment period there were 136 respondents; 92.65% (n = 126) agreed, and 5.15% (n = 7) disagreed. There were 22 written comments. There were several comments that HPV positive and p16 positive are not equivalent, in part because of a small percentage of p16-positive tumors that are unrelated to HPV. Others commented that reporting terminology needs to explicitly describe the HPV status, given that some clinicians may not understand the relationship between p16 IHC and HPV status. There was a suggestion to avoid the term HPV positive because of the stigma of sexually transmitted diseases. Lastly, there was a suggestion to use the terminology approved by the WHO Classification of Head and Neck Tumours,²⁵⁸  which this guideline agrees with, in accordance with the new 4th edition.

Pathologists should not provide a tumor grade or differentiation status for HPV-positive/p16-positive OPSCC.

The strength of evidence is insufficient.

Tumor grade is a measurement of differentiation, that is, how closely a tumor resembles the normal tissue from which it presumably arises. Tumor grade generally correlates with biologic behavior, so that well-differentiated tumors typically behave less aggressively than poorly differentiated tumors. For SCCs, highly keratinizing tumors with keratin pearl formation and small nuclei are considered well differentiated. Most HPV-positive OPSCCs have a characteristic morphologic appearance. As they are usually nonkeratinizing, with high nuclear to cytoplasmic ratio sand hyperchromatic nuclei, and are arranged in lobules and sheets, they have often been classified as poorly differentiated or high-grade carcinomas. However, these classifiers were developed in head and neck SCC in general and not specifically for HPV-positive OPSCC. In these tumors, this morphology does not predict poor outcomes, but rather is paradoxically associated with a better prognosis in a majority of cases because it predicts HPV positivity.^2,55,90 

Human papillomavirus–positive/p16-positive OPSCCs arise from the tonsillar crypts rather than the surface epithelium. The tonsillar crypts are lined by a specialized reticulated epithelium with an associated lymphocytic infiltrate. Deep in the normal crypts, this epithelium has basaloid cytologic features, absent keratinization, high nuclear to cytoplasmic ratio, and permeating lymphocytes. Westra¹⁶¹  has proposed that HPV-positive oropharyngeal SCCs might best be considered as well-differentiated carcinomas given their resemblance to the nonneoplastic reticulated crypt lining epithelium.

In the open comment period there were 136 respondents; 57.35% (n = 78) agreed, and 36.03% (n = 49) disagreed, making this the recommendation with the lowest agreement rate. There were 26 written comments. Eight comments were acknowledgment of the respondents' lack of expertise to answer the question. Thirteen respondents stated that grading should be performed. Some of these 13 felt that tumor grade and differentiation might assist in staging and future therapy. Some saw “no harm” in providing tumor grade even if it did not correlate with clinical behavior. Some of the respondents confused tumor grade and staging. Several respondents pointed out that the CAP protocol for the examination of specimens from patients with carcinomas of the pharynx²⁵⁹  and the 7th edition of the AJCC staging manual²⁶⁰  require histologic grading on all head and neck cancers. The concerns regarding AJCC 7th edition will be less relevant as the 8th edition of the AJCC staging manual¹⁴  comes into clinical use by early 2018, and familiarity with the 4th edition of the WHO Classification of Head and Neck Tumours²⁵⁸  grows. The latter states that grading is not applicable for HPV-positive OPSCC. In addition, some pointed out the practical complication that one might not know what the p16 status was before signing out a case, and that one would have to provide tumor grade in an addendum if the tumor turned out to be p16 negative.

Pathologists should not alter HR-HPV testing strategy based on patient smoking history.

The strength of evidence is insufficient.

Patients with HPV-positive OPSCC often have improved disease-free and overall survival when compared with those with HPV-negative OPSCC. The strong historical association of OPSCC with tobacco and alcohol use has led to the examination of these and other potential prognostic factors to further risk stratify HPV-positive OPSCC to allow the identification of patients who might benefit from treatment deintensification and subsequent reduction in short- and long-term side effects without compromising overall survival. The initial observation that tobacco use significantly decreased the improved survival seen in HPV-positive OPSCC resulted from a retrospective analysis of Radiation Therapy Oncology Group 0129 data using recursive partition analysis. Although HPV status was the main predictor of survival, tobacco use as measured by pack-years increased the risk of death by 1% per year, independent of OPSCC HPV status.¹¹  The authors concluded that tumor HPV status and tobacco use (>10 pack-years) were robust and independent predictors of survival after chemoradiation therapy. A subsequent retrospective analysis based on the subset of OPSCC patients from Radiation Therapy Oncology Groups 9003 and 0129 confirmed these findings.²⁹  Careful and exhaustive quantitation of tobacco exposure in this analysis demonstrated that the risk of death increased linearly with tobacco smoking as measured by pack-years in HPV-related OPSCC. The increased rate of locoregional failure seen in this cohort suggests a likely direct effect of tobacco use on treatment effectiveness, rather than from competing causes of mortality commonly seen in smokers. Since these initial studies, a number of additional analyses based on recursive partitioning analysis have validated tobacco use as an important variable in treatment response in HPV-positive OPSCC.^70,261  There is also no published evidence that smoking changes the results of any of the HPV-specific tests or p16 IHC. Consequently, the EP does not recommend altering the HR-HPV testing strategy based on smoking history. Rather, tobacco use, as measured by pack-years, is one of several variables, along with HR-HPV status, that the treating physician will use when counseling patients regarding likely treatment outcomes and potentially when selecting therapy in the context of a clinical trial.

In the open comment period there were 134 respondents; 97.76% (n = 131) agreed, and 1.49% (n = 2) disagreed. Of all the recommendations, this one had the highest agreement rate. There were only 8 written comments. Two respondents commented that the smoking history was irrelevant, one respondent expressed concern regarding pathology access to the smoking history, and another expressed concern about the lack of standards in reporting a smoking history. One respondent agreed that HPV and smoking may coexist, and another stressed smoking does not rule out HPV infection. One respondent suggested performing HPV testing in non-OPSCC in the absence of a smoking history and limited keratinization.

The emergence of HPV-positive OPSCC, which is biologically and clinically a unique type of HNSCC, has made it critical that such patients be identified by routine testing for HR-HPV in clinical practice. Knowledge of the HPV tumor status is important for patient prognosis and for the establishment of specific treatments better matched to such tumors. This EP, through a rigorous systematic review, has provided 14 formal recommendations or expert consensus opinions on the nature of HPV testing in various head and neck specimens, scenarios, and settings, with the goal of standardizing what is performed across diverse pathology practice settings. These recommendations will be expected to evolve with future research, literature updates, and reviews in the coming years.

We thank advisory panel members Maura Gillison, MD, PhD; Amy Lynn, MD; Dina Mody, MD; Evan R. Myers, MD, MPH; Cherie Paquette, MD; Michael B. Prystowsky, MD, PhD; Harry Quon, MD; Brian Hill of the Oral Cancer Foundation; and Bert Noojin, JD. We also thank the members of the ASCO Head and Neck Guideline Advisory Group and the ASCO Clinical Practice Guideline Committee, and in particular E. Rosenthal, MD, and A. Loren, MD, for their thoughtful review.